www.elsevier.com/locate/humpath

C-kit expression in renal oncocytomas and chromophoberenal cell carcinomas

Lei Huoa, Jun Sugimurab, Maria S. Tretiakovac, Kurt T. Pattona,Rohit Guptaa, Boris Popova, William B. Laskin MDa, Anjana Yeldandi MDa,Bin Tean Tehb, Ximing J. Yang MD, PhDa,*

aDepartment of Pathology, Northwestern Memorial Hospital, Northwestern University, Feinberg School of Medicine,

Chicago, IL 60611, USAbLaboratory of Cancer Genetics, Van Andel Research Institute, Grand Rapids, MI 49503, USAcDepartment of Pathology, University of Chicago, Chicago, IL 60637, USA

Received 11 May 2004; accepted 3 January 2005

0046-8177/$ – see front matter D 2005

doi:10.1016/j.humpath.2005.01.011

T Corresponding author.

E-mail address: xyang@northwester

Keywords:C-kit;

Renal cell carcinoma;

Oncocytoma;

Angiomyolipoma

Summary C-kit encodes the membrane-bound tyrosine kinase KIT, whose expression has been

identified in several types of human neoplasms. Recently, KIT has been reported to be a marker for

chromophobe renal cell carcinoma (RCC) and renal angiomyolipoma. However, expression of this

molecule has not been adequately studied in other renal tumors, particularly oncocytoma, which may

morphologically resemble chromophobe RCC. In this study, we analyzed c-kit messenger RNA

(mRNA) levels in 17 chromophobe RCCs and 20 renal oncocytomas obtained from complementary

DNA (cDNA) microarrays. Furthermore, comprehensive immunohistochemical analysis of KIT protein

using a monoclonal antibody was performed in 226 renal tumors including chromophobe RCC (n = 40),

oncocytoma (n = 41), clear-cell RCC (n = 40), renal angiomyolipoma (n = 29), and papillary RCC (n =

21) on tissue microarrays (TMAs) and was compared with immunostaining results from

25 chromophobe RCCs and 30 oncocytomas using standard sections. The staining intensity was

semiquantitatively graded on a 3-tier scoring system. All chromophobe RCCs and oncocytomas showed

significant overexpression of c-kit mRNA. The average increase of mRNA compared with normal

kidney tissue was 7.4- fold for chromophobe RCCs and 7.4- fold for oncocytomas. Immunohisto-

chemical expression of KIT was found in most chromophobe RCCs (95% in TMAs and 96% in

conventional sections) and oncocytomas (88% in TMAs and 100% in conventional sections) but was

infrequently observed in renal angiomyolipomas (17%), papillary RCCs (5%), and clear-cell RCCs

(3%). Furthermore, the average KIT immunoreactivity in TMAs was stronger in chromophobe RCC

(1.93) and oncocytoma (2.07) than in other subtypes of renal tumors tested, including angiomyolipomas

(0.17), papillary RCCs (0.05), and clear-cell RCCs (0.03). In conclusion, we found a significant

elevation of c-kit mRNA by cDNA expression microarrays and overexpression of KIT protein by

immunohistochemistry not only in chromophobe RCCs but also in oncocytomas. In contrast,

immunohistochemical expression of KIT was not detected in most other types of renal cell tumors

Human Pathology (2005) 36, 262–268

Published by Elsevier Inc.

n.edu (X.J. Yang).

C-kit in oncocytoma and chromophobe renal cell carcinoma 263

evaluated. The differential expression of c-kit in these renal tumors may have diagnostic and

therapeutic implications.

D 2005 Published by Elsevier Inc.

1. Introduction

Chromophobe renal cell carcinoma (RCC) and renal

oncocytoma are presently considered histogenetically

related neoplasms with a purported origin from the

intercalated cell of the collecting duct [1,2]. Chromophobe

RCC accounts for approximately 5% to 8% of renal cell

neoplasms and has a better prognosis than the other

subtypes of RCC such as clear cell, papillary, and collecting

duct. In contrast, renal oncocytoma is a benign neoplasmwith

an incidence of 5% among renal epithelial tumors [3].

Because chromophobe RCC and oncocytoma share many

morphological features, their diagnosis sometimes may be a

challenge, especially when a tumor is composed of cells with

abundant eosinophilic cytoplasm. In addition, clear-cell RCC

and papillary RCC may occasionally demonstrate over-

lapping histological features with chromophobe RCC or

oncocytoma. As the prognosis and treatment of these

neoplastic entities differ, a correct diagnosis is essential.

Gene-expression microarray analysis has recently shed

light on the molecular pathogenesis of renal tumors, and it

has also contributed to the identification of new biomarkers

useful for histological diagnosis. For example, differential

expression of glutathione S-transferase a in clear-cell RCC,

a-methylacyl racemase in papillary RCC, carbonic anhy-

drase II in chromophobe RCC, and keratin 19 in transitional

cell carcinoma was initially demonstrated by complementa-

ry DNA (cDNA) microarray studies and later confirmed by

immunohistochemistry [4]. Analysis of gene expression

profiles has shown that chromophobe RCC and oncocytoma

cluster closely together, thereby supporting the notion of a

common histogenetic linkage and distinguishing them from

other renal tumors [4,5]. Thus, a search for novel markers

capable of distinguishing chromophobe RCC from oncocy-

toma has become a relevant diagnostic pursuit.

Expression of KIT, a membrane receptor tyrosine kinase

encoded by c-kit, has been demonstrated in several types of

neoplasms [6-10]. The therapeutic implication of KIT

overexpression in a tumor is the potential efficacious use

of Gleevec (Novartis Pharmaceuticals, East Hanover, NJ),

an anti–tyrosine kinase drug, in treatment because over-

expression of KIT correlates with gain-of- function muta-

tions of the gene [11,12].

Recently, with high-density oligonucleotide arrays,

overexpression of KIT was reported in chromophobe RCC

but not in other types of RCC tested. This result was verified

by immunohistochemistry on only a few examples of

chromophobe and clear-cell RCCs [13], and expression of

KIT in renal oncocytoma was not evaluated. In addition,

recent studies reported KIT immunoreactivity in renal

angiomyolipomas [14], oncocytomas [15-17], and other

types of RCCs [18-20].

Given the diagnostic value and potential therapeutic

implications of c-kit overexpression in tumor cells, we

decided to perform a comprehensive analysis of messen-

ger RNA (mRNA) and protein levels of c-kit on a large

number of renal tumors, particularly oncocytoma, chromo-

phobe RCC, and angiomyolipoma.

2. Materials and methods

2.1. C-kit cDNA expression microarray

cDNA microarrays of kidney tumors were performed as

previously described [4,21,22]. The 37 renal tumors ana-

lyzed by cDNA microarrays included oncocytomas (n = 20)

and chromophobe RCCs (n = 17). Briefly, total RNA was

isolated from renal tumor tissues using Trizol reagent.

Microarrays were performed using gene chips containing

21632 cDNA clones. Polymerase chain reaction–amplified

bacterial colonies were transferred into 384-well plates and

printed onto aminosilane-coated glass slides using a home-

built robotic microarrayer. Fifty micrograms of total RNA

from each tumor and reference were reverse transcribed with

oligo(deoxythymidine) primer and Superscript II (Invitro-

gen, Carlsbad, Calif) in the presence of Cy5-dCTP and Cy3-

dCTP (Amersham Pharmacia Biotech, Peapack, NJ). The

Cy5- and Cy3-labeled cDNA probes were mixed with probe

hybridization solution containing formamide and hybridized

to prewarmed (508C) slides for 20 hours at 508C. Afterhybridization and washing, slides were scanned using

ScanArray Lite (GSI Lumonics, Billerica, Mass) operating

at 532- and 635-nm wavelengths. Images were analyzed with

the GENEPIX PRO 3.0 software (Axon Instruments, Foster

City, Calif). The gene ratios were median centered before

hierarchical clustering by using CLUSTER and were then

visualized by using TREEVIEW (MB Eisen, http://rana.lbl.

gov). The CIT software was used to find genes that were

differentially expressed (using Student t test) between one

histological subtype and the others. To find significant

discriminating genes, 10000 t statistics were calculated by

randomly placing patients’ samples into 2 groups.

2.2. Clinical material and tissue microarray

Conventional formalin-fixed, paraffin-embedded sections

of chromophobe RCC (n = 25) and renal oncocytoma

(n = 30) accessioned to the Department of Pathology,

Northwestern Memorial Hospital, from 1996 to 2003 were

retrieved and formed the basis of this study.

Fig. 1 C-kit mRNA expression by cDNA microarrays. (A) Increased c-kit mRNA expression in 17 chromophobe RCCs over normal

kidney control indicated in fold increase. Each case was analyzed with 2 probes (N20798 and N24824) as described in Materials and

methods. (B) Increased c-kit mRNA expression in 20 oncocytomas analyzed in a similar fashion.

Table 1 C-kit mRNA expression measured by cDNA expression

microarrays

Tumor type C-kit mRNA

overexpression

Expression

(fold)

Chromophobe RCC 17/17 7.4 F 3.6

Oncocytoma 20/20 7.4 F 4.2

Expression are presented as mean F SD values (calculated after averaging

the 2 values obtained with the 2 probes for each case).

L. Huo et al.264

A total of 171 cases of primary kidney tumors from

nephrectomy specimens were selected from the Surgical

Pathology archives of Northwestern Memorial Hospital and

University of Chicago Hospitals. All paraffin-embedded

donor tissue blocks were sampled with 1.5-mm punchers

using a Beecher tissue microarray (TMA) instrument

(Beecher Instruments Inc, Sun Prairie, Wis). Four paraffin

TMA blocks containing arrayed core cylinders from 40

chromophobe RCCs, 41 oncocytomas, 40 clear-cell RCCs,

21 papillary RCCs, 29 renal angiomyolipomas, and 10

normal kidneys were subjected to routine hematoxylin and

eosin staining and immunohistochemical staining.

2.3. Immunohistochemistry

Standard tissue sections and TMA sections were sub-

jected to immunohistochemical staining using a monoclonal

antibody specific for c-kit protein (Dako, Carpinteria, Calif)

as previously described [23,24]. Briefly, endogenous per-

oxidase activity was blocked with 3% hydrogen peroxide.

Heat-induced epitope retrieval was carried out in citrate

buffer (10 mmol/L, pH 6) for 15 minutes at 1008C in a

microwave oven. A primary mouse monoclonal anti–c-kit

antibody (Dako) at 1:100 dilution was applied for 1 hour at

room temperature. A subsequent reaction was performed

with biotin-free, horseradish peroxidase enzyme–labeled

polymer of EnVision plus detection system (Dako). A

positive reaction was visualized with diaminobenzidine

solution followed by counterstaining with hematoxylin. A

gastrointestinal stromal tumor (GIST) was used as a positive

control. Appropriate negative controls for the immuno-

staining were prepared by using nonimmune mouse IgG.

C-kit in oncocytoma and chromophobe renal cell carcinoma 265

2.4. Evaluation of immunohistochemistry andstatistical analysis

C-kit immunostaining was semiquantitatively evaluated

for intensity (0, negative; 1, weak; 2, moderate; 3, strong

staining) and patterns of reactivity (membranous and/or

cytoplasmic). Statistical analysis was performed using

Student t test.

3. Results

3.1. C-kit mRNA expression

All 17 chromophobe RCCs and 20 oncocytomas showed

an increased level of c-kit mRNA by cDNA microarray

analysis using 2 v-kit probes N20798 and N24824 (Fig. 1

and Table 1) compared with paired nonneoplastic kidney

Fig. 2 KIT immunohistochemical staining on TMAs. (A) Normal kid

showing strong KIT staining (membranous and cytoplasmic). (C) Onco

angiomyolipoma demonstrating weak KIT expression. (E) Papillary RCC

negative staining. (Original magnification �200.)

tissues. As measured from cDNA microarrays, there is a

2.3- fold variation in c-kit mRNA levels of normal kidney

controls ranging from 0.56 to 1.29 (mean F SD, 1 F0.316). In chromophobe RCCs, c-kit was expressed 1.4- to

14.0-fold over the normal kidney control, with a mean of

7.4-fold increase (SD, 3.6). The levels in oncocytomas

ranged from 1.2- to 14.5-fold, with a mean of 7.4-fold

increase (SD, 4.2). The mRNA levels between the

chromophobe RCCs and oncocytomas were not significant-

ly different (P N .05).

3.2. KIT expression by immunohistochemistry

To confirm the above gene-expression microarray results

demonstrating c-kit overexpression in both chromophobe

RCC and oncocytoma at the mRNA level, immunohisto-

chemical analysis of the KIT protein was performed on

tissue sections with a monoclonal antibody specific for KIT.

ney tubules showing focal KIT staining. (B) Chromophobe RCC

cytoma showing strong KIT cytoplasmic staining. (D) A case of

showing negative to minimal staining. (F) Clear-cell RCC showing

Fig. 3 KIT immunohistochemical staining on conventional sections. (A) Chromophobe RCC (hematoxylin-eosin, original magnification

�200). (B) Chromophobe RCC showing combined membranous and cytoplasmic KIT staining. (C) Oncocytoma (hematoxylin-eosin,

original magnification �200). (D) Oncocytoma showing cytoplasmic KIT staining.

L. Huo et al.266

On TMAs, nonneoplastic renal tubules showed focal

positive staining (Fig. 2A), whereas 38 (95%) of 40

chromophobe RCCs and 36 (88%) of 41 oncocytomas

demonstrated strong KIT expression (Fig. 2B and C). In

contrast, only 5 (17%) of 29 angiomyolipomas (Fig. 2D),

1 (5%) of 21 papillary RCCs (Fig. 2E), and 1 (3%) of 40

clear-cell RCCs (Fig. 2F) showed KIT expression. The

average immunostaining intensity was significantly higher

in oncocytomas (2.07) and chromophobe RCCs (1.93) than

in other renal tumors (0.17 for angiomyolipomas, 0.05 for

papillary RCCs, and 0.03 for clear-cell RCCs). In conven-

tional sections, KIT expression was detected in 24 (96%) of

25 chromophobe RCCs (mean intensity, 1.76) (Fig. 3A and

B) with 13 demonstrating predominantly membranous

Table 2 KIT expression by immunohistochemistry in renal

tumors

Tumor type Total no.

of cases

Positivity Average

intensity

Oncocytoma

TMA 41 36 (88) 2.07

Conventional 30 30 (100) 2.58

Chromophobe RCC

TMA 40 38 (95) 1.93

Conventional 25 24 (96) 1.76

Angiomyolipoma 29 5 (17) 0.17

Papillary RCC 21 1 (5) 0.05

Clear-cell RCC 40 1 (3) 0.03

Positivity values are presented as n (%).

staining (mean intensity, 1.8), 4 showing predominantly

cytoplasmic staining (mean intensity, 2.3), and 7 displaying

combined cytoplasmic and membranous staining (mean

intensity, 1.6). All 30 (100%) oncocytomas (mean intensity,

2.58) (Fig. 3C and D) showed strong, predominantly

cytoplasmic staining. The immunohistochemistry results

are summarized in Table 2.

4. Discussion

KIT, a 145-kd membrane-bound tyrosine kinase receptor

for stem cell factor, is encoded by the protooncogene c-kit

[10], which is a homologue of v-kit. KIT activation plays an

important role in the normal development of certain

hematopoietic cells, including mast cells, gonadal germ

cells, melanocytes, and interstitial cells of Cajal of the

gastrointestinal tract, and it is overexpressed in some

neoplasms derived from these cell types [6-10].

Recently, Yamazaki et al [13] reported KIT overexpres-

sion in chromophobe RCCs but not conventional RCCs and

therefore proposed KIT as a marker for chromophobe RCC.

Their conclusions were based on a limited number of cases

and renal oncocytoma, a benign renal cell neoplasm that

shows not only a morphological overlap but also a genetic

relationship to chromophobe RCC [4,5], which was not

included in their study. Because chromophobe RCC and

oncocytomas are morphologically and genetically related

[23], a biomarker should be tested in both tumors before it

can be declared bspecific.Q The RON oncogene product was

C-kit in oncocytoma and chromophobe renal cell carcinoma 267

initially touted as boncocytoma specific.Q However, we haveshown convincing evidence that RON immunoreactivity in

both oncocytomas and chromophobe RCC [24].

Petit et al [15] and Pan et al [17] reported KIT

immunoreactivity in most chromophobe RCCs and renal

oncocytomas but not in other renal tumors. Other studies

reporting KIT immunoreactivity in renal angiomyolipomas

and papillary RCC [14,18,20] raised the question of

possible false-positive staining. Herein, we demonstrate

elevated levels of c-kit mRNA in all 17 chromophobe

RCCs and 20 oncocytomas evaluated by cDNA microarray

analysis, thus reducing the possibility of immunohisto-

chemical cross reactivity with other proteins and validating

the elevated c-kit expression in both oncocytomas chro-

mophobe RCC. In addition, the immunohistochemical

study of KIT expression in a large number of renal

neoplasms confirms our cDNA microarray results by

demonstrating that most chromophobe RCCs and oncocy-

tomas express KIT; in contrast, angiomyolipoma, papillary,

and clear-cell RCC show lower KIT expression. The results

of our study compare favorably with those by Petit et al

and Pan et al (Table 3).

We do not have an explanation for the discrepancy

between the KIT immunoreactivity in all papillary RCCs

reported by Lin et al [20] and negative or weak KIT staining

in other 4 studies including ours. With our staining condition,

a small percentage of papillary RCC and angiomyolipomas

displayed faint or weak (intensity score, 1-2) KIT immunos-

taining (Fig. 3D and E). In fact, in the study by Miliaras et al

[18], the KIT immunoreactivity in other renal tumors was also

weak, scored as 2 of 4, which was consistent with our finding.

Therefore, a small percent of papillary RCC (5%-26% Table

3) and angiomyolipomas (17%-21%, Table 3) may be

positive for KIT; however, their weaker KITstaining intensity

is different from the strong staining observed in chromophobe

RCC and oncocytomas.

Angiomyolipoma, a member of the perivascular epithe-

lioid cell tumor family [25], is an essentially benign

neoplasm composed of thick-walled blood vessels, smooth

muscle, and adipose tissue in varying proportions. Makhlouf

et al [14] reported KIT immunoexpression in all 21

angiomyolipomas they tested, including 15 examples from

Table 3 KIT immunoreactivities reported

Authors Ref CC Pap

Yamazaki et al [13]

Makhlouf et al [14]

Petit et al [15] 0/29 0/10

Pan et al [17] 0/256 0/25

Miliaras et al [18] 2/13 2/7

Lin et al [20] 0/28 18/18

Subtotal 2/326 (1%) 20/60 (33%)

This study 1/40 (3%) 1/21 (5%)

Total 2/366 (0.5%) 21/81 (26%)

CC indicates clear-cell RCC; Pap, papillary RCC; Chr, chromophobe RCC; Onc,

the liver and 6 tumors from the kidney. The authors

suggested that angiomyolipoma be included in the differ-

ential diagnosis of KIT-positive tumors because some

angiomyolipomas share morphological features with GIST

and melanoma, which are 2 other KIT-expressing tumors.

However, our study demonstrated that most (83%) of renal

angiomyolipomas showed little or no expression of KIT

when compared with the strong KIT immunoreactivity in

GIST (positive control), renal oncocytoma, and chromo-

phobe RCC. The discrepancy in these results may be

attributed to the use of different dilutions of antibodies that

react with different epitopes in the molecule, different

antigen- retrieval techniques used, or differences in inter-

pretation of staining patterns and intensity of reaction.

The recent use of TMAs has resulted in more efficient

immunohistochemical phenotyping of tissue. However,

technical issues regarding TMA testing of renal tumors

have not been adequately addressed. In this study, we found

that there is no difference of KIT immunoreactivity in

chromophobe RCC between TMA (95%) and conventional

sections (96%). However, TMA slightly underestimated

KIT expression in oncocytomas (88%) compared with

results from standard tissue sections (100%). This difference

is most likely caused by sampling nonimmunoreactive areas

of the tumor. Therefore, when TMA sections are used for

immunohistochemistry analysis, it is important to keep this

issue in mind.

KIT immunoexpression correlates with c-kit gain-

of-function mutations and thus makes it a potential target

for Gleevec, a competitive antagonist of the ATP binding

site within the tyrosine kinase domain of the molecule

[26,27]. The efficacy of Gleevec in the treatment of meta-

static and inoperable GISTs and related tumors has recently

been documented [11,12]. Surgical resection, the treatment

of choice for chromophobe RCC and oncocytomas, may not

be suitable for certain patients because of potential clinical

complications. Gleevec may provide an option for patients

with contraindications to nephrectomy, provided that the

radiological or pathological diagnosis can be established

preoperatively and the tumor shows c-kit mutations.

In conclusion, using a combination of cDNA expression

microarrays and immunohistochemistry on TMAs and

Chr Onc AML

4/4

6/6

22/25 10/14 0

24/29 5/7 0/23

4/7 7/7 2/4

20/20

74/85 (87%) 22/28 (78%) 8/33 (24%)

62/65 (95%) 66/71 (93%) 5/29 (17%)

136/150 (91%) 88/98 (90%) 13/62 (21%)

renal oncocytoma; AML, renal angiomyolipoma.

L. Huo et al.268

standard tissue sections, we confirmed KIT as a good

biomarker for chromophobe RCC and oncocytoma because

it exhibited a high degree of sensitivity and specificity.

Although KIT does not distinguish chromophobe RCCs

from benign renal oncocytomas, KIT immunoreactivity may

be useful in separating chromophobe RCC from potential

mimics featuring cells with abundant eosinophilic cyto-

plasm. Overexpression of KIT in defined subtypes of renal

tumors indicates a potential therapeutic role for tyrosine

kinase inhibitors in managing patients with these tumors.

References

[1] Bonsib SM, Bray C, Timmerman TG. Renal chromophobe cell

carcinoma: limitations of paraffin-embedded tissue. Ultrastruct Pathol

1993;17:529-36.

[2] Chao DH, Zisman A, Pantuck AJ, et al. Changing concepts in the

management of renal oncocytoma. Urology 2002;59:635 -42.

[3] Lieber MM. Renal oncocytoma. Urol Clin North Am 1993;20:355-9.

[4] Takahashi M, Yang XJ, Sugimura J, et al. Molecular subclassification

of kidney tumors and the discovery of new diagnostic markers.

Oncogene 2003;22:6810-8.

[5] Young AN, Amin MB, Moreno CS, et al. Expression profiling of renal

epithelial neoplasms: a method for tumor classification and discovery

of diagnostic molecular markers. Am J Pathol 2001;158:1639 -51.

[6] Russell ES. Hereditary anemias of the mouse: a review for geneticists.

Adv Genet 1979;20:357 -459.

[7] Isozaki K, Hirota S, Nakama A. Disturbed intestinal movement, bile

reflux to the stomach, and deficiency of c-kit–expressing cells in Ws/

Ws mutant rates. Gastroenterology 1995;109:456-64.

[8] Kitamura Y, Go S. Decreased production of mast cells in S1/S1d

anemic mice. Blood 1979;53:492 -7.

[9] Turner AM, Zsebo KM, Martin F. Nonhaematopoietic tumor cell lines

express stem cell factor and display c-kit receptors. Blood 1992;

80:374-81.

[10] Duffaud F, Blay J. Gastrointestinal stromal tumors: biology and

treatment. Oncology 2003;65:187 -97.

[11] Demetri GD, von Mehren M, Blanke CD, et al. Efficacy and safety of

imatinib mesylate in advanced gastrointestinal stromal tumors. N Engl

J Med 2002;347:472-80.

[12] Verweij J, van Oosterom AT, Blay JY, et al. Imatinib (Gleevec) an

active agent for gastrointestinal stromal tumors (GIST), but not for

l

-

l

l

;

.

f

l

f

f

l

r

f

;

other soft tissue sarcoma (STS) subtypes not characterized for KIT and

PDGF-R expression, results of EORTC phase II studies. J Clin Onco

2002;21:403a.

[13] Yamazaki K, Sakamoto M, Ohta T, et al. Overexpression of KIT in

chromophobe renal cell carcinoma. Oncogene 2003;22:847-52.

[14] Makhlouf HR, Remotti HE, Ishak KG. Expression of KIT (CD117) in

angiomyolipoma. Am J Surg Pathol 2002;26:493 -7.

[15] Petit A, Castillo M, Santos M, et al. KIT expression in chromophobe

renal cell carcinoma. Am J Surg Pathol 2004;28:676-8.

[16] Pan CC, Chen PC, Chiang H. KIT (CD117) is frequently overex

pressed in thymic carcinomas but is absent in thymomas. J Patho

2004;202(3):375 -81.

[17] Pan CC, Chen PCH, Chiang H. Overexpression of KIT (CD117) in

chromophobe renal cell carcinoma and renal oncocytoma. Am J Clin

Pathol 2004;121:878 -83.

[18] Miliaras D, Karasavvidou F, Papanikolaou A, et al. KIT expression in

fetal, normal adult, and neoplastic renal tissues. J Clin Patho

2004;57:463-6.

[19] Castillo M, Petit A, Mellado B, et al. C-kit expression in sarcomatoid

renal cell carcinoma: potential therapy with imatinib. J Urol 2004

171:2176-80.

[20] Lin ZH, Han EM, Lee ES, Kim CW, et al. A distinct expression

pattern and point mutation of c-kit in papillary renal cell carcinomas

Mod Pathol 2004;17:611 -66.

[21] Takahashi M, Yang XJ, Lavery TT, et al. Gene expression profiling o

favorable histology Wilms tumors and its correlation with clinica

features. Cancer Res 2002;62(22):6598-605.

[22] Takahashi M, Sugimura J, Yang XJ, et al. Gene expression profiling o

renal cell carcinoma and its implication in diagnosis, prognosis and

therapeutics. Adv Cancer Res 2003;89:157 -81.

[23] Yang XJ, Sugimura J, Tretiakova M, et al. Gene expression profiling

of renal medullary carcinoma. Cancer 2004;100:976-85.

[24] Patton KT, Tretiakova MS, Yao JL, Papavero V, et al. Expression o

RON proto-oncogene in renal oncocytoma and chromophobe rena

cell carcinoma. Am J Surg Pathol 2004;28:1045-50.

[25] Bonetti F, Pea M, Martignoni G, et al. Clear cell (dsugarT) tumor of the

lung is a lesion strictly related to angiomyolipoma: the concept of a

family of lesions characterized by the presence of the perivascula

epithelioid cells (PEC). Pathology 1994;26:230-6.

[26] Heinrich MC, Rubin BP, Longley BJ, Fletcher JA. Biology and

genetic aspects of gastrointestinal stromal tumors: KIT activation and

cytogenetic alterations. Hum Pathol 2002;33:484-95.

[27] Hirota S, Isozaki K, Moriyama Y, et al. Gain-of-function mutations o

c-kit in human gastrointestinal stromal tumors. Science 1998

279:577-80.