Renal medullary carcinomas: histopathologic phenotype associated with diverse genotypes

www.elsevier.com/locate/humpath

Human Pathology (2011) 42, 1979–1988

Original contribution

Renal medullary carcinomas: histopathologic phenotypeassociated with diverse genotypesZoran Gatalica MD, DSc a,⁎, Stan L. Lilleberg PhDb, Federico A. Monzon MDc,Manika Sapru Koul PhDb, Julia A. Bridge MDd, Joseph Knezetic PhDa,Ben Legendre PhDb, Poonam Sharma MDa, Peter A. McCue MDe

aDepartment of Pathology, Creighton University School of Medicine, Omaha, NE 68131, USAbTranslational and Clinical Research, Transgenomic, Inc, Omaha, NE 68164, USAcDepartment of Pathology, The Methodist Hospital and The Methodist Hospital Research Institute, Houston, TX 77030,and Department of Pathology, Weill Cornell Medical College, New York, NY, USAdDepartment of Pathology, University of Nebraska Medical Center, Omaha, NE 68161, USAeDepartment of Pathology, Thomas Jefferson University Hospital, Philadelphia, PA 19107, USA

Received 28 December 2010; accepted 10 February 2011

0d

Keywords:Kidney;Medullary Carcinoma;Oncogenes;Karyotype;Hypoxia-inducedsignaling

Summary Chromosomal abnormalities and gene mutations have become major determinants in theclassification of kidney carcinomas. Most renal medullary carcinomas develop in patients withhereditary sickle cell disease, but sporadic cases unassociated with sickle cell disease have also beendescribed, for which underlying genetic abnormality is unknown. We evaluated 3 patients with renalmedullary carcinoma (1 patient with sickle cell disease and 2 patients without sickle cell disease) forgerm line and somatic mutations in genes commonly involved in pathogenesis of renal carcinomas usingdenaturing high-performance liquid chromatography and direct sequencing. Chromosomal abnormal-ities were studied by the conventional cytogenetic and SNP arrays analysis. Expression of hypoxia-inducible factor 1α was examined using immunohistochemistry. Two new mutations in the gene forfumarate hydratase were identified in 1 case of medullary renal carcinoma without sickle cell disease: agerm line mutation in exon 6 (R233H) and an acquired (somatic) mutation in exon 8 (P374S). Nofumarate hydratase mutations were identified in the other 2 patients. The second sporadic case ofrenal medullary carcinoma harbored double somatic mutations in von Hippel-Lindau gene, and renalmedullary carcinoma in the patient with sickle cell disease showed von Hippel-Lindau gene promotermethylation (epigenetic silencing). No consistent pattern of chromosomal abnormalities was foundbetween 2 cases tested. All 3 cases showed increased hypoxia-inducible factor 1α expression. Medullaryrenal carcinomas from patients with or without sickle cell disease show involvement of genes importantin hypoxia-induced signaling pathways. Generalized cellular hypoxia (in sickle cell disease) orpseudohypoxia (in tumors with fumarate hydratase and von Hippel-Lindau mutations or epigeneticsilencing) may act alone or in concert at the level of medullary tubular epithelium to promotedevelopment of this rare type of renal carcinoma, which could then be genetically reclassified as eitherfumarate hydratase–associated renal carcinomas or high-grade clear cell renal cell carcinomas.© 2011 Elsevier Inc. All rights reserved.

⁎ Corresponding author.E-mail address: zorangatalica@creighton.edu (Z. Gatalica).

046-8177/$ – see front matter © 2011 Elsevier Inc. All rights reserved.oi:10.1016/j.humpath.2011.02.026

1980 Z. Gatalica et al.

1. Introduction American with an SCD and 2 whites without significant

In 1995, a series of 33 highly aggressive renal carcinomasoccurring in young African Americans with sickle celldisease (SCD) or trait and with characteristic histomorpho-logic findings was reported [1]. These tumors were usuallydetected at an advanced stage and showed histologic patterncharacterized by epithelial cells with reticular, adenoid cysticor yolk sac–like growth pattern, necrosis, stromal desmo-plasia, and prominent inflammation. This type of renal can-cer was named renal medullary carcinoma (RMC) and wasconsidered to be a specific phenotype associated with SCD[1,2]. Since then, multiple cases of medullary carcinomasoccurring in non–African Americans without hemoglobin-opathy were also reported, including a familial renal canceroccurrence [3,4]. Although medullary renal carcinoma iscurrently listed as a separate histopathologic entity in theWorld Health Organization classification of the tumors of thekidney, some authors consider this tumor a variant of col-lecting duct carcinoma (CDC) arising in patients with hemo-globinopathy [4,5].

Chromosomal and molecular genetics studies have shownthat renal cell carcinoma (RCC) is a heterogeneous diseasewith strong histomorphologic phenotype-genotype correla-tions. Cases of clear cell RCC (ccRCC) bearing mutations invon Hippel-Lindau (VHL) gene (located on chromosome3p26-p25) are typically seen in patients with VHL diseaseand constitutional chromosome 3 translocations as well as insporadic ccRCC. Papillary renal carcinomas arise in patientswith hereditary autosomal dominant germ line mutation inMET oncogene on chromosome 7q31. Somatic mutationsin MET gene as well as duplications of chromosome 7 arecommonly seen in sporadic papillary renal carcinoma.Hybrid oncocytic and clear cell renal tumors characterizeBirt-Hogg-Dube syndrome (BHD) caused by a germ linemutation in the folliculin gene on chromosome 17p11.2.Xp11 translocation RCC is a group of neoplasms character-ized by translocations involving a breakpoint at Xp11.2[6-10]. Hereditary leiomyomatosis with RCC (HLRCC) is anautosomal dominant tumor syndrome caused by a germ linemutation in fumarate hydratase (FH) gene on chromosome1q42. This syndrome has been associated with developmentof unusually aggressive kidney carcinomas morphologicallycharacterized as type II renal papillary carcinomas [11].Studies on chromosomal abnormalities in RMC and CDC aresparse. Chromosomal status of medullary carcinomas waspreviously described using classic cytogenetic methods insporadic cases [2,12-14] and using comparative genomichybridization method in a series of 9 cases [3]. Gene ex-pression profile was studied in 2 cases [15]. A small numberof cases histologically classified as collecting duct (Bellini)carcinomas were also studied using classic cytogeneticmethods [16-20]. No studies on specific renal cancer–causing genes are yet published for RMC.

We analyzed 3 high-grade renal carcinomas, morpholog-ically diagnosed as RMCs from 3 male patients: 1 African

comorbidities. We focused on the analysis of genes known tobe important in renal carcinogenesis.

1.1. Patients and methods

The study was approved by Creighton University Institu-tional Review Board.

1.1.1. Patients1.1.1.1. Patient 1. A 34-year-old white man (withoutSCD) presented with hematuria of 8-day duration and rightneck supraclavicular mass for 2 months. Abdominal com-puted tomographic (CT) scan revealed an infiltrating tumorof right kidney with a maximal dimension of 8 cm withextension into the renal vein. In addition, enlarged retro-peritoneal and mediastinal lymph nodes were identified.

Biopsy of supraclavicular mass showed metastatic adeno-carcinoma. Right radical nephrectomy was performed,followed by chemotherapy. The patient died of diseaseapproximately 36 months later.1.1.1.2. Patient 2. A 28-year-old white man withoutSCD presented with complaint of right flank pain and rightlower extremity tenderness. Urinalysis showed microscopichematuria. An ultrasound showed a deep vein thrombosis inthe right lower extremity. A CT scan showed a 10-cm rightrenal mass that virtually replaced the entire renal parenchymaand extensive retroperitoneal adenopathy. A right radicalnephrectomy was performed, followed by chemotherapy.The patient died of metastatic carcinoma approximately14 months thereafter.1.1.1.3. Patient 3. A 32-year-old African American manwith SCD presented with complaint of right flank/abdominalpain and gross hematuria. CT scan showed a 7-cm rightkidney mass with extensive retroperitoneal lymphadenopathy.CT also showed bilateral pulmonary nodules and mediastinallymphadenopathy. A radical nephrectomy with subsequentvideoscopic wedge resection of the right middle lobe of thelung was performed. The patient was treated with radiationand 2 cycles of chemotherapy after which he developed arecurrent tumor in the mediastinum and upper lung lobesresulting in a superior vena cava syndrome. The patient diedof disease approximately 3 months after the diagnosis.

2. Methods

2.1. Immunohistochemistry

Immunohistochemical stains were performed on forma-lin-fixed, paraffin-embedded tissue sections using mono-clonal mouse antibodies against epidermal growth factorreceptor (Clone 31G7; Zymed Laboratories, South SanFrancisco, CA), hypoxia-inducible factor (HIF) 1α (HIF1α)(Abcam, Inc, Cambridge, MA), MLH1 (clone 14; ZymedLaboratories), MSH2 (clone FE11; Zymed Laboratories),

1981Renal medullary carcinomas

MSH6 (clone 44; Zymed Laboratories), and PMS2 (mABA16-4; BD Pharmingen, Franklin Lakes, NJ) in automatedstaining procedures.

2.2. Mutation analysis

2.2.1. Tumor DNA extractionDNA extraction was done after pathologic review and

manual macrodissection to remove nontumor tissue. Sampleareas that contained at least 70% tumor cells were used forDNA extraction. Tissue was digested with 0.4 μg proteinaseK per microliter of digestion buffer (500 mmol/L KCl,100 mmol/L Tris-HCl, 15 mmol/L MgCl2, 0.5% Tween 20)at 50°C overnight. A standard protocol was used to extractDNA from the digested samples. DNA was quantitatedusing a Nano-Drop ND-1000 spectrophotometer (Thermo-Scientific, Wilmington, DE).

2.2.2. Polymerase chain reactionPolymerase chain reaction (PCR) and sequencing primers

were designed using the amplicon and primer selectioncomponent of the Web-based software available atwww.mutationdiscovery.com (Transgenomic, Inc, Omaha,NE). DNA was screened for mutations in genes deemedimportant in renal carcinogenesis (VHL, BHD, and MET aswell as FH and SD) as well as other common oncogenes(EGFR, ERBB2, KRAS, NRAS, BRAF). Amplification ofpatient tumor DNAs was carried out in 50 μL reactions using10 to 15 ng of tumor DNA and 1.25U of either HotMaster TaqDNAPolymerase (Eppendorf, Boulder, CO, USA) or HotStartTaq DNA polymerase (Denville Scientific, Metuchen, NJ,USA)with their respective 1X reaction buffers and 0.2mmol/Lof each deoxynucleotide triphosphate and 0.2 μmol/L each offorward and reverse primer. Thermal cycling was accom-plished using MJ Research (Bio-Rad, Hercules, CA, USA)Dyad and Tetrad DNA Engines and a program of 95°C for2 minutes; 10 cycles of touchdown PCR; and then 30 cycles of95°C for 30 seconds, 58°C for 30 seconds, and 68°C for30 seconds, followed by a final 5-minute extension at 68°C.PCR products were heteroduplexed using a program of 95°Cfor 5 minutes, cooling to 85°C at −2.0°C per seconds, coolingto 25°C at −0.2°C per second, and then incubation at 4°C ina thermocycler. PCR products (5 μL) were analyzed by 2%agarose gel electrophoresis in 1X Tris-acetate-EDTA andvisualized with ethidium bromide.

2.2.3. Endonuclease scanningHeteroduplexed PCR samples were combined with 15 U

of Surveyor Nuclease and 1 μL of Surveyor Enhancer(Transgenomic, Omaha, NE, USA) and incubated at 42°C for20 minutes. Digestions were terminated by addition of 2 μLstop solution (0.5 mol/L EDTA [pH 8.0]) and analyzed on aWAVE HS System equipped with DNASep HT column(Transgenomic). Twelve microliters of each digestionmixture was injected on a double-stranded sizing multiplefragment application gradient using a 50°C oven temperature.

Detection was 260 nm for UV and 495 nm excitation/537 nmemission for fluorescence. A 100-base pair DNA ladder (NewEngland Biolabs, Ipswich, MA) was run as a size marker.

Instrument control, data acquisition, and data analysiswere done using WAVE Navigator software. Positive andnegative controls were included with each plate of PCRproducts to monitor endonuclease cleavage efficiency.

2.2.4. Gene sequencingExcess PCR primers were removed from 10 μL of PCR

product using the AMPure PCR Purification system(Agencourt Bioscience, Corp, Danvers, MA). Purifiedproduct was eluted in 30 μL of deionized water. Reactionchemistry using BigDye version 3.1 (Applied Biosystems,Carlsbad, CA) and cycle sequencing were adapted from themanufacturer's recommendations. Cycle sequencing prod-ucts were purified using CleanSEQ reagents (AgencourtBioscience). Purified sequencing products were eluted in40 μL of 0.01 μmol/L EDTA. Samples were then run on anABI 3100 Genetic Analyzer (Carlsbad, CA, USA), and thesequence electropherograms were analyzed using Gene CodeSequencer software (Ann Arbor, MI).

2.3. Cytogenetic analysis

Culturing, harvesting, and preparation of slides were per-formed as previously described [21]. Metaphase cells werebanded with Giemsa trypsin, and the karyotypes were ex-pressed according to the International System for HumanCytogenetic Nomenclature (ISCN 2009).

2.4. Virtual karyotyping using SNP array assay

Samples were processed with a formalin-fixed, paraffin-embedded tissue–optimized protocol based on the GeneChipMapping 250K Nsp Assay Kit (Affymetrix, Santa Clara,CA) as described previously [22]. The samples were hybri-dized on GeneChip Mapping 250K Nsp arrays (Affymetrix)for 16 hours at 48°C in a GeneChip 450 hybridization oven(Affymetrix) at 60 rpm. The arrays were washed and stainedaccording to the manufacturer's genotyping protocol. Probeintensity data acquired from the Affymetrix GeneChipOperating System v4.0 were analyzed using AffymetrixGeneChip Genotyping Analysis Software (GTYPE) 4.1. Lossof heterozygosity and copy number estimates were obtainedusing a publicly available analysis package: Copy NumberAnalyzer for Affymetrix GeneChip arrays (CNAG 3.0) asdescribed before [22].

2.5. Microsatellite instability

Microsatellite analysis was performed through PCRfragment size analysis using ABI 3100 Avant system(Applied Biosystems). Two different sets of 5 DNA micro-satellite markers were used, reflecting the availability of the

1982 Z. Gatalica et al.

commercial kits at the time of the analysis. For case 1,BAT25, BAT26, D5S346, D17S250, and D2S123 markerswere used (Roche Diagnostics GmbH, Mannheim, Germany).For case 2, we used BAT 25, BAT 26, NR-21, NR-24, andMONO-27 (Promega Corporation, Madison, WI).

3. Results

3.1. Histopathologic findings

All cases were diagnosed as RMCs based on light micro-scopic findings. Each tumor showed epithelial cells arrangedin large islands amid desmoplastic stroma. Some areasshowed tubule formation and large cystic spaces (Fig. 1A, B,and C). Cytologically, the cells had abundant cytoplasm thatvaried from clear to granular to eosinophilic with well-defined cytoplasmic membrane. Nuclei were large, vesicular,and moderately pleomorphic with prominent nucleoli.Frequent intracytoplasmic lumina are identified. Each casewas reviewed by 1 or more experts in renal pathology (see“Acknowledgment”), and all agreed with the diagnoses ofRMC in every case.

3.2. Mutation analysis

Germ line mutations in hemoglobin B gene were foundonly in the patient with SCD (patient 3) and consisted ofE6V and D73N double mutation, consistent with patient'sknown SCD.

A novel germ line mutation in FH gene (R190H) wasidentified in case 1, a white man without SCD. In this case,the second somatic mutation in FH gene was found in thetumor (Fig. 1G). The 2 other cases showed no FH mutation.

Double somatic mutation within VHL was found in thetumor of the second white male patient without SCD (Fig. 1H).

The tumor in the patient with SCD exhibited promotermethylation (inactivation) of the VHL gene (Fig. 1I).

Analysis of genes (EGFR, HER-2, MET) encodingreceptor tyrosine kinases (RTKs) showed double somaticmutation of EGFR (L858R and T790M) in the case withgerm line FH mutation (case 1; Fig. 2A). Two other casesshowed no mutations in RTKs.

Results of mutation analysis are summarized in Table 1.

3.3. Microsatellite instability

No evidence ofmicrosatellite instability (MSI)was found in2 cases tested (cases 1 and 2). In immunohistochemical assays,normal (nuclear) expression of all 4 major DNA mismatchrepair proteins was found in all 3 cases (data not shown).

3.4. Immunohistochemistry

Nuclear accumulation of HIF1α was detected in each of3 cases (Fig. 1D, E, and F), with intensity and distribution

similar to ccRCC-positive controls (not shown). Strong (3+)membranous expression (Fig. 2B) of EGFR was seen incase 1 (tumor with 2 somatic mutations in EGFR) but notin the other 2 cases.

3.5. Cytogenetic analysis

Conventional cytogenetic analysis of the tumor frompatient 1 revealed presence of a complex hyperdiploid clonecharacterized by gains of chromosomes 7, 8, 10, and 11;losses of chromosomes 9 and 13; presence of an isochro-mosome composed of 8q; additional unknown material on10p and 22q; a deletion of 7q; and a dicentric chromosomecomposed of 13q and 21q. Eleven cells were normal male.The karyotype nomenclature is listed in Table 2 for com-parison with other published cytogenetics results in medul-lary renal carcinoma and CDC.

3.6. Virtual karyotyping

Virtual karyotyping from the patient 1 tumor revealed thefollowing chromosomal changes: +2p, +3q, +7, −8p, +8q,−9q, +11, −13q, +20, −22, which is in partial agreement withthe conventional cytogenetics result (see above). Althoughsome discordance is seen between the results of theconventional and virtual karyotypes, these can be explainedby the presence of unidentified material in the conventionalkaryotype that explains gains not detected by cytogenetics,such as 2p and 3q gains, as has been described before [23].

Virtual karyotyping from the patient 2 tumor wassignificantly distinct from patient 1 because it revealed onlyfocal chromosomal gains and losses: +7(q36.3), −9(p21.1-21.3), +10(q26.3), −15(q14-q21.2), +15(q21.3), +21(q22.2-22.3). The largest region of change was a 15.7-Mb region ofloss in chromosome 15. The only overlap seen between the 2cases was the region of gain in chromosome 7 (7q36.3).

4. Discussion

Acceptance of a specific tumor type is greatly aided byconsistent genotype-phenotype correlation. RMC was orig-inally thought to be a specific pathologic phenotype asso-ciated with SCD genotype (it was postulated that βhemoglobin gene mutation had a pathogenetic role intumor development). However, cases of RMC were alsoreported in white patients without hemoglobinopathies [3,4].

Recent data from molecular and cytogenetic studies haveprovided better insight into genetic characterization and,consequently, classification of renal tumors. Specific geneticand characteristic chromosomal changes have been corre-lated with major subtypes of RCC [8,9]. Classification ofrenal tumors based upon cytogenetic or molecular abnor-malities improves the diagnostic accuracy in challengingcases [24]. Furthermore, the molecular differences could beexploited for targeted therapy.

Fig. 1 A to C, Histologic characteristics of renomedullary carcinomas arising in 3 patients (A, patient 1; B, patient 2; and C, patient 3).RMCs are cytologically high grade (large cells with pleomorphic nuclei and prominent nucleoli) organized in islands (sometimes resemblingadenoid cystic carcinomas) separated by dense desmoplasia and characteristically with abundant inflammatory cell infiltrate. D to F, HIF1αexpression (immunohistochemical stain with DAB chromogen, brown). All 3 cases showed expression of HIF1α. G to H, Molecular geneticanalysis of RMC. G, Surveyor endonuclease digestion of PCR product of FH exons showing presence of mutations both in tumor and normaltissue (germ line). H, Surveyor endonuclease digestion of PCR product showing somatic mutations in VHL gene. I, WAVE DHPLC analysisof PCR product representing the VHL promoter region showing methylated VHL promoter (arrow).

1983Renal medullary carcinomas

Fig. 1 (continued )

1984 Z. Gatalica et al.

Because of its low incidence, RMC has not been wellcharacterized from a cytogenetic point of view. Chromosom-al cytogenetic findings of 4 previously published cases (Table2) are compared with the results from 2 of our cases. One ofour cases showed multiple large chromosomal abnormalitiesinvolving whole chromosomes or chromosomal arms, as seenin many distal nephron tumors. Cytogenetic analysis revealedthe presence of a hyperdiploid clone characterized by gain ofchromosomes 7, 8, 10, and 11 and loss of 9 and 13. Array-based karyotyping revealed additional gains in chromosomes2 and 20. Our second case showed only submicroscopicchromosomal gains/losses in the virtual karyotype. Ourfindings are only partially similar to 5 cases of CDC and1 case of RMC described in the literature showingmonosomyof 9 and 13 [2,17,20]. Review of the literature revealed2 cases of CDC that showed trisomy of chromosomes 7 and8 similar to our case [17,18]. Trisomy 7 is shared by manytumor phenotypes affecting different regions of urinary tract,ranging from papillary renal tumors to bladder cancers. Anextra copy of chromosome 7may confer a proliferative clonaladvantage to the tumor cells because they may express higher

level of epidermal growth factor (EGF) receptors whosecoding gene, the oncogene EGFR, is located on thischromosome. Multiple studies have documented overexpres-sion of this receptor in renal malignancies with developmentand progression of metastatic disease [25]. The only overlapseen between our 2 cases was the region of gain inchromosome 7 (7q36.3). Interestingly, this region has beenreported to be involved in translocations in pediatric acutemyelogenous leukemia [26,27].

Unlike previously published cases of CDC and RMC thatreported monosomy of chromosomes 8, 10, and 11[2,12,16,20], our case 1 showed gain of these chromosomes.Cytogenetically, loss of chromosome 1 material has beenconsistently found in distal nephron tumors including CDCs[16,27-29]. Abnormality of chromosome 1 is also seen 2 casesof RMC [2,12]. In a series of 9 RMCs studied by comparativegenomic hybridization, no chromosomal change was observedin 8 cases; and a loss of chromosome 22was seen in 1 case [4].From these studies, it is evident that no single chromosomalnumerical or structural abnormality was consistently observedin either collecting duct or medullary carcinomas.

Fig. 2 A, DNA sequencing after WAVE fragment collection demonstrating EGFR exon 21 L858R mutation. B, Immunohistochemicaldetection of strong (3+) EGFR expression in case 1. A, DNA sequencing after WAVE fragment collection demonstrating EGFR exon 21L858R mutation. B, Immunohistochemical detection of strong (3+) EGFR expression in case 1.

1985Renal medullary carcinomas

Molecular genetic studies on RMC are rare. One reportdescribed t(9;22) and t(10;16) in a single patient with RMC;BCR-ABL rearrangement was confirmed by fluorescencein situ hybridization (FISH) [12], but this was not found in3 patients in a different report [30]. Similarly, we found noevidence of this translocation in our case 2 (using FISH notshown). Recently, the anaplastic lymphoma kinase wasfound rearranged in 1 case of an unclassified RCC arising ina young African American with sickle cell trait but not in2 cases of RMC [14]. We found no evidence of anaplasticlymphoma kinase rearrangement in cases 1 and 2 using FISHmethod (not shown). Although histologically completelydistinct from Wilms tumor, expression profiling of RMCrevealed close clustering with Wilms tumor and urothelialcarcinoma [15]. The young age of the patients and the citedmolecular genetic similarity to urothelial carcinoma, whichcould be seen as a part of Lynch syndrome [31], prompted usto investigate MSI in our series of cases. No evidence of MSIand of the loss of major mismatch repair proteins (MLH-1,MSH-2, and MSH-6) was found. Although MSI has not beenspecifically investigated in patients with RMCs, 1 previousstudy showed no evidence of MSI in renal carcinomasarising in young individuals [32].

Molecular genetic studies of renal neoplasms have iden-tified a limited number of genes characteristically associated

Table 1 Summary of molecular genetic analysis in renomedullary ca

RMC Association Hemoglobingene mutation

Kidney cancer ge

Germ line mutatio

Case 1 HLRCC Negative FH: R233H

Case 2 Nonsyndromic(sporadic)

Negative Negative

Case 3 SCD HBB: E6V;HBB: D73N

Negative

with defined histopathologic phenotypes [9]. Somatic andgerm line mutations and promoter hypermethylation in VHLgene (located on chromosome 3p26-p25) have beenassociated with clear cell renal carcinomas [33]. Activatingmutations in MET were associated with papillary carcinoma[34] and inactivating mutations of BHD with chromophobecarcinoma [35]. Besides papillary type II carcinomas [36],rare cases of renal tumors resembling medullary carcinoma/CDC were described in patients with HLRCC syndrome.HLRCC syndrome is a variant of an autosomal dominantsyndrome of multiple cutaneous and uterine leiomyomatosiscaused by germ line heterozygous loss-of-function mutationsin the FH gene. The prevalence of renal cancer in multiplecutaneous and uterine leiomyomatosis lies between 1 (2%) in46 families who were not radiologically screened and 2 (6%)in 32 families who were radiologically screened [11]. Ourfinding of a germ line mutation in FH gene in patient1 should be considered diagnostic of HLRCC; however, itwas not associated with additional signs and symptoms ofhereditary leiomyomatosis.

FH plays an essential role in the mitochondrial tricarbox-ylic acid or Krebs cycle, by catalyzing the conversion offumarate to malate. The primary function of the tricarboxylicacid cycle is the oxidation of pyruvate, supplied by theglycolytic pathway, for energy production. The abrogation of

rcinomas arising in patients with SCD and without SCD

nes RTK mutation

n Somatic mutation Methylation

FH: P374S Negative EGFR, L858Rand T790M

VHL: E37K, L89H Negative Negative

Negative VHL Negative

Table 2 Comparison of published genetic findings in renal collecting duct (CDC or Bellini duct carcinoma) and RMCs

Authors Cases and Diagnosis Genetic changes Reference no.

Gatalica et al 1 (CG) 47,XY,+del(7)(q32q36),+i(8)(q10),−9,+add(10)(p13),+11,−13,dic(13;21)(p13;p13),add(22)(q13) Present study1 (VK) +2p,+3q,+7,−8p,+8q,−9q,+11,-13q,+20,−222 (VK) +7(q36.3),−9(p21.1-21.3),+10(q26.3),−15(q14-q21.2),+15(q21.3),+21(q22.2-22.3)RMC

Avery et al 1 RMC 72,XXY,del(1)(p32),t(1;?)(q21;?),del(2)q(21q31),+del(2)(q21q31),t(3;8)(p21;q24),+del(5)(q13)x2,+6,+del(7)(q32)x2,−8,−9,t(9;?)(p13;?)x2,−11,+14,−15,−16,t(16;?)(q22;?),−17,+18,t(19;?)(q13;?),−21,−22,+4mar

[2]

Stahlschmidt et al 1 RMC 84-94,XX,+X,−Y,−Y,+I(1)(q10),−7,−8,+I(8)(q10)x2,t(9;22)(q34;q11)x2,t(10;16)(q22;q22)x2,i(17)(q10),add(22)(q11)x2[cp5]

[12]

Chatelain et al 1 RMC Tetraploidy without structural abnormalities [13]Debelenko et al 1 RMC 46,XX,inv(2)(p23q11.2) [14]Gregori Romeroet al

1 91-91,XXY,−Y,+12,+12,−15,−16,−18,+mar [17]2 53,XY,+2,t(2;7)(p22;q11),der(2)t(2;7)(p22;q11),+3,+r(3),add(5)(p15),+7,−8,+12,+17,+20,−213 44-47,X,−Y,+9,+16,−21All CDC

Cavazzana et al 1 CDC 53,XY,del(1)(p34),+iso(1q),+iso(5p),+4,+7,+8,−14del(16)(q22),+17,−18,+20,+20,−22 [18]Antonelli et al 1 40,X,−Y,add(1)(p11),r(2)(p25q37),−6,der(13)t(13;15)(q24;q15),−15,−18,add(19)(p11),−21,−22 [20]

2 37,X,−Y,der(1)t(1 ;13)(q10 ;q10),−3,−6,−11,−13,−14,−18,der(19)t(1;19)(q12;q13),−21,−223 40,−X,−Y,del(1)(p32),−10,−18,−20,−224 33,X,−X,−1,t(2;12)(q33;q34),−3,−4,−6,−9,−11,−13,−14,−15,der(16)t(1;16)(q21;q11),−17,−21,−225 41,X,−Y,−1,der(3)add(3)(p26),+7,+7,−8,−9,der(11)(q23),add(11)(p25),−13,−14,−14,−206 45,−X,del(1)(p21),der(1)t(1;14)(q10;p10),der(1)t(1;?3)(p10;q10), del(7)(q32),der(9)del(9)(q21),add(12)(q24),−14,del(15)

(q21),add(17)(q25),add(19)(q13),+20,+21,−22All CDC

Füzesi et al 1 37,XX,−1,−4,−6,−9,−13,−14,−15,−18,−22 [16]2 37,XX,−1,−4,−6,−9,−13,−14,−15,−18,−223 39,XY,−1,−6,−8,−11,−14,−15All CDC

Abbreviations: CG indicates classic cytogenetics; VK, SNP array virtual karyotype.

1986Z.

Gatalicaet

al.

1987Renal medullary carcinomas

FH activity (seen in biallelic germ line mutations in fumaratedeficiency syndrome) has profound cellular consequencesthat are incompatible with normal embryologic develop-ment resulting in severe neurologic impairment and earlydeath [37]. No FH mutations were identified in 2 other casesof RMC in our series; but instead, in both patients' tumors,VHL genes were affected. In the second patient with RMCwithout SCD, there were 2 somatic mutations in VHL,whereas in the third patient (with SCD), there was an epi-genetic silencing of VHL gene promoter. Results of thegenetic testing could justify reclassification of these tumors aseither FH-associated renal carcinomas or high-grade ccRCCs.

Although seemingly disparate pathways, increased cyto-plasmic fumarate and a loss of functional VHL proteinexpression may have similar effects on cellular metabolismand survival. It was recently shown [38] that renal cancers inHLRCC exhibit elevated expression of HIF-1 and HIF-2proteins as well as the protein product of the HIF-regulatedgene Glut-1. Under normal oxygenation, HIF1α is hydrox-ylated and slated for degradation via VHL-dependentubiquitination pathway. However, under hypoxic or pseu-dohypoxic conditions, prolyl hydroxylase enzymes areinhibited; and HIF1α escapes degradation. Stable HIF1αthen translocates to nucleus where it can heterodimerize withHIF1β (ARNT) to form an active HIF transcription complexleading to expression of genes that can enhance survival inhypoxic conditions: glucose metabolism and angiogenesis,or reduced apoptosis (cell death) [39]. This situation issimilar to the effect of alteration of VHL in ccRCCs, wherethe loss of VHL leads to high HIF1α protein levels and up-regulation of HIF-dependent transcripts [40,41].

In SCD, even in the absence of genetic alteration in VHL,there is a generalized hypoxic state with increased steady-state levels of hypoxia-induced serum factors such asangiopoietin, vascular endothelial growth factor (VEGF),erythropoietin, and others [42].

Taken together, our results suggest that, in patients withmedullary carcinomas, there is a common underlying hy-poxic cellular environment favoring activation of HIFpathways. In the patients with SCD, there is a predispositionto a generalized hypoxia (due to the germ line mutations inhemoglobin gene), which in combination with a somaticalteration in the VHL gene leads to activation of HIF cellsignaling (as in case 3, characterized by SCD and somaticinactivation of VHL promoter), all of which can lead to thedevelopment of RMC. Similarly, RMC can develop in non-SCD patients where cellular pseudohypoxia in collecting ductepithelium is caused by either biallelic inactivation of FH (asin case 1) or an absence of functional VHL tumor suppressorprotein (as in case 2 with biallelic inactivation of VHL). Ineach pathologic scenario, HIF1α is not degraded and is able toform a transcriptionally active HIF complex. As a transcrip-tion factor, HIF1α is involved in the regulation of manybiological processes that facilitate both oxygen delivery andadaptation to oxygen deprivation by regulating genes that areinvolved in glucose uptake and energy metabolism, angio-

genesis, erythropoiesis, cell proliferation, and apoptosis [39].Our results corroborate an earlier observation of an increasedVEGF and HIF expression in RMC [4].

Evaluation of mutations in genes encoding tyrosinereceptor kinases in our cases revealed double EGFR somaticmutations in 1 non-SCD case. Such coexisting 2 somaticEGFR mutations were also previously identified in non–small cell lung cancers and were recently associated withvariable response to a selective tyrosine kinase inhibitorGefitinib. Lack of RTK gene mutations and expression of theEGFR in the other 2 cases make the use of RTK inhibitorsless likely to be successfully used in patients with RMC. Onthe other hand, rather consistent HIF up-regulation offers apromise of therapy targeted at angiogenesis pathways, suchas VEGF [43].

In conclusion, results of our study indicate that RMC isnot a single chromosomally or genetically defined type ofrenal cancer. Rather, it appears that this high-grade, morpho-logically defined aggressive carcinoma arises from thecollecting duct system in conditions favoring cellularhypoxia (such as SCD) or mutations (germ line or somatic),which affect hypoxia-sensing pathway (pseudohypoxia)leading to deregulation of HIF1α signaling, a commonlyrecognized carcinogenetic pathway.

Acknowledgments

The cases were reviewed by (late) Dr C. J. Davis, MD(AFIP, Washington, DC); Dr John Eble, MD (IndianaUniversity, Indianapolis, IN); and Dr R. O. Petersen, MD,PhD (Thomas Jefferson University, Philadelphia, PA), whoconcurred with the diagnoses of medullary carcinomas. Theauthors thank Ms Mindee Curtis, HT (Creighton MedicalLaboratories); Ms Lisa Linder Stephenson, MT; and MrMichael Fisher, HT (Creighton University Medical Center),for their excellent technical assistance in immunohistochem-ical staining and Ms Bridget Lacquer (Creighton MedicalLaboratories) for MSI testing and Karla Alvarez, BSc (TheMethodist Hospital Research Institute), for performance ofthe virtual karyotypes.

References

[1] Davis Jr CJ, Mostofi FK, Sesterhenn IA. Renal medullary carcinoma.The seventh sickle cell nephropathy. Am J Surg Pathol 1995;19:1-11.

[2] Avery RA, Harris JE, Davis Jr CJ, Borgaonkar DS, Byrd JC, WeissRB. Renal medullary carcinoma: clinical and therapeutic aspects of anewly described tumor. Cancer 1996;78:128-32.

[3] O'Donnell PH, Jensen A, Posadas EM, et al. Renal medullary-likecarcinoma in an adult without sickle cell hemoglobinopathy. Nat RevUrol 2010;7:110-4.

[4] Swartz MA, Karth J, Schneider DT, Rodriguez R, Beckwith JB,Perlman EJ. Renal medullary carcinoma: clinical, pathologic,immunohistochemical, and genetic analysis with pathogenetic impli-cations. Urology 2002;60:1083-9.

1988 Z. Gatalica et al.

[5] Kovacs G, Akhtar M, Beckwith BJ, et al. The Heidelberg classificationof renal cell tumours. J Pathol 1997;183:131-3.

[6] Delahunt B, Velickovic M, Grebe SK. Evolving classification of renalcell neoplasia. Expert Rev Anticancer Ther 2001;1:576-84.

[7] Furge KA, Lucas KA, Takahashi M, et al. Robust classification ofrenal cell carcinoma based on gene expression data and predictedcytogenetic profiles. Cancer Res 2004;64:4117-21.

[8] Linehan WM, Walther MM, Zbar B. The genetic basis of cancer ofthe kidney. J Urol 2003;170:2163-72.

[9] Linehan WM, Vasselli J, Srinivasan R, et al. Genetic basis of cancerof the kidney: disease-specific approaches to therapy. Clin Cancer Res2004;10:6282s-9s.

[10] Ross H, Argani P. Xp11 translocation renal cell carcinoma. Pathology2010;42:369-73.

[11] Alam NA, Olpin S, Leigh IM. Fumarate hydratase mutations andpredisposition to cutaneous leiomyomas, uterine leiomyomas and renalcancer. Br J Dermatol 2005;153:11-7.

[12] Stahlschmidt J, Cullinane C, Roberts P, Picton SV. Renal medullarycarcinoma: prolonged remission with chemotherapy, immunohisto-chemical characterization and evidence of bcr/abl rearrangement. MedPediatr Oncol 1999;33:551-7.

[13] Chatelain D, de Pinieux G, Slama J, et al. Renal medullary carcinoma,a new clinico-pathologic entity. Immunohistochemical, ultrastructural,flow cytometric and cytogenetic study of a case. Ann Pathol 1999;19:320-4.

[14] Debelenko LV, Raimondi SC, Daw N, et al. Renal cell carcinoma withnovel VCL-ALK fusion: new representative of ALK-associated tumorspectrum. Mod Pathol 2010;24:430-42, doi:10.1038/modpathol.2010;213.

[15] Yang XJ, Sugimura J, Tretiakova MS, et al. Gene expression profilingof renal medullary carcinoma: potential clinical relevance. Cancer2004;100:976-85.

[16] Fuzesi L, Cober M, Mittermayer C. Collecting duct carcinoma: cyto-genetic characterization. Histopathology 1992;21:155-60.

[17] Gregori Romero MA, Morell Quadreny L, Llombart Bosch A. Cyto-genetic analysis of three primary Bellini duct carcinomas. GenesChromosomes Cancer 1996;15:170-2.

[18] Cavazzana AO, Prayer-Galetti T, Tirabosco R, et al. Bellini ductcarcinoma. A clinical and in vitro study. Eur Urol 1996;30:340-4.

[19] Verdorfer I, Culig Z, Hobisch A, et al. Characterisation of a collectingduct carcinoma by cytogenetic analysis and comparative genomichybridisation. Int J Oncol 1998;13:461-4.

[20] Antonelli A, Portesi E, Cozzoli A, et al. The collecting duct carcinomaof the kidney: a cytogenetical study. Eur Urol 2003;43:680-5.

[21] Bridge Jr RS, Bridge JA, Neff JR, Naumann S, Althof P, Bruch LA.Recurrent chromosomal imbalances and structurally abnormal break-points within complex karyotypes of malignant peripheral nerve sheathtumour and malignant triton tumour: a cytogenetic and molecularcytogenetic study. J Clin Pathol 2004;57:1172-8.

[22] Yamamoto G, Nannya Y, Kato M, et al. Highly sensitive method forgenomewide detection of allelic composition in nonpaired, primarytumor specimens by use of affymetrix single-nucleotide-polymor-phism genotyping microarrays. Am J Hum Genet 2007;81:114-26.

[23] Monzon FA, Alvarez K, Gatalica Z, et al. Detection of chromosomalaberrations in renal tumors: a comparative study of conventionalcytogenetics and virtual karyotyping with SNP microarrays. ArchPathol Lab Med 2009;133:1917-22.

[24] Kim HJ, Shen SS, Ayala AG, et al. Virtual-karyotyping with SNPmicroarrays in morphologically challenging renal cell neoplasms: apractical and useful diagnostic modality. Am J Surg Pathol 2009;33:1276-86.

[25] Merseburger AS, Hennenlotter J, Simon P, et al. Membranousexpression and prognostic implications of epidermal growth factorreceptor protein in human renal cell cancer. Anticancer Res 2005:251901-7.

[26] von Bergh AR, van Drunen E, van Wering ER, et al. High incidence oft(7;12)(q36;p13) in infant AML but not in infant ALL, with a dismaloutcome and ectopic expression of HLXB9. Genes ChromosomesCancer 2006;45:731-9.

[27] Steiner G, Cairns P, Polascik TJ, et al. High-density mapping ofchromosomal arm 1q in renal collecting duct carcinoma: region ofminimal deletion at 1q32.1-32.2. Cancer Res 1996;56:5044-6.

[28] Polascik TJ, Cairns P, Epstein JI, et al. Distal nephron renal tumors:microsatellite allelotype. Cancer Res 1996;56:1892-5.

[29] Speicher MR, Schoell B, du Manoir S, et al. Specific loss ofchromosomes 1, 2, 6, 10, 13, 17, and 21 in chromophobe renal cellcarcinomas revealed by comparative genomic hybridization. Am JPathol 1994;145:356-64.

[30] Simpson L, He X, Pins M, et al. Renal medullary carcinoma and ABLgene amplification. J Urol 2005;173:1883-8.

[31] Lynch HT, Taylor RJ, Lynch JF, et al. Multiple primary cancer,including transitional cell carcinoma of the upper uroepithelial tract ina multigeneration HNPCC family: molecular genetic, diagnostic, andmanagement implications. Am J Gastroenterol 2003;98:664-70.

[32] Kanomata N, Eble JN, Halling KC. Microsatellite instability isuncommon in young patients with renal cell carcinoma. Cancer GenetCytogenet 1998;101:123-7.

[33] Nickerson ML, Jaeger E, Shi Y, et al. Improved identification of vonHippel-Lindau gene alterations in clear cell renal tumors. Clin CancerRes 2008;14:4726-34.

[34] Schmidt L, Duh FM, Chen F, et al. Germline and somatic mutations inthe tyrosine kinase domain of the MET proto-oncogene in papillaryrenal carcinomas. Nat Genet 1997;16:68-73.

[35] Gad S, Lefèvre SH, Khoo SK, et al. Mutations in BHD and TP53genes, but not in HNF1beta gene, in a large series of sporadicchromophobe renal cell carcinoma. Br J Cancer 2007;96:336-40.

[36] Tomlinson IP, Alam NA, Rowan AJ, et al. Multiple leiomyomaconsortium: germline mutations in FH predispose to dominantlyinherited uterine fibroids, skin leiomyomata and papillary renal cellcancer. Nat Genet 2002:406-10.

[37] Bourgeron T, Chretien D, Poggi-Bach J, et al. Mutation of thefumarase gene in two siblings with progressive encephalopathy andfumarase deficiency. J Clin Invest 1994;93:2514-8.

[38] Isaacs JS, Jung YJ, Mole DR, et al. HIF overexpression correlateswith biallelic loss of fumarate hydratase in renal cancer: novel roleof fumarate in regulation of HIF stability. Cancer Cell 2005;8:143-53.

[39] Haase VH. The VHL/HIF oxygen-sensing pathway and its relevanceto kidney disease. Kidney Int 2006;69:1302-7.

[40] Iliopoulos O, Levy AP, Jiang C, Kaelin Jr WG, Goldberg MA.Negative regulation of hypoxia-inducible genes by the von Hippel-Lindau protein. Proc Natl Acad Sci U S A 1996;93:10595-9.

[41] Maxwell PH, Wiesener MS, Chang GW, et al. The tumour suppressorprotein VHL targets hypoxia-inducible factors for oxygen-dependentproteolysis. Nature 1999;399:271-5.

[42] Duits AJ, Rodriguez T, Schnog JJ, CURAMA Study Group. Serumlevels of angiogenic factors indicate a pro-angiogenic state in adultswith sickle cell disease. Br J Haematol 2006;134:116-9.

[43] Rini BI, Michaelson MD, Rosenberg JE, et al. Antitumor activity andbiomarker analysis of sunitinib in patients with bevacizumab-refractory metastatic renal cell carcinoma. J Clin Oncol 2008;26:3743-8.