Human De Novo Papillary Renal-Cell Carcinomas in a Kidney Graft: Evidence of Recipient Origin With Adenoma-Carcinoma Sequence

Authors


  • The first two authors contributed equally to this work.

Corresponding author: Jérôme Verine, jerome.verine@sls.aphp.fr

Abstract

Papillary renal-cell carcinoma (pRCC) is unusual for its occurrence in kidneys with chronic dysfunction, for its frequent multifocality and for its common association with papillary adenoma, a benign renal lesion morphologically indistinguishable from pRCC. Concomitant development of papillary adenoma and pRCC in five transplanted kidneys, where donor and recipient characteristics are well established, provided a unique opportunity for molecular studies of de novo pRCC carcinogenesis. We aimed to study this tumor type to determine whether or not the different papillary tumors have the same origin, and whether or not papillary adenomas are precursor lesions of pRCC. We performed XY-FISH in sex-mismatched kidney transplants, and polymorphic microsatellite DNA and high-resolution melting of mitochondrial DNA analyzes in all five patients on laser-microdissected tumor cells, then compared these molecular profiles to donor and recipient profiles. This study (i) identified the recipient origin of de novo papillary adenomas and pRCCs in a kidney transplant, (ii) demonstrated an identical origin for precursor cells of papillary adenomas and pRCCs and (iii) showed additional genetic alterations in pRCCs compared to papillary adenomas. This molecular approach of papillary tumors developed in transplanted kidney identified successive steps in carcinogenesis of human de novo papillary renal-cell carcinoma.

Abbreviations
AI

allelic imbalance

BMDC

bone marrow derived cells

BNIP3L

BCL2/adenovirus E1B 19kDa interacting protein 3-like

CDKN2a

cyclin-dependent kinase inhibitor 2A

CK7

cytokeratin 7

FISH

fluorescent in situ hybridization

HRM

high-resolution melting

HV1/HV2

hypervariable regions 1 and 2

KRT7

keratin 7

LOH

loss of heterozygosity

mtDNA

mitochondrial DNA

PCR

polymerase chain reaction

pRCC

papillary renal-cell carcinoma

PTPRC

protein tyrosine phosphatase receptor type C

RCC

renal-cell carcinoma

STR

short tandem repeat

Introduction

Human papillary renal-cell carcinoma (pRCC), representing 10% of RCC [1], has particular characteristics: its occurrence in kidneys with chronic dysfunction, including transplanted kidneys [2], its frequent multifocality [3] and its frequent association with papillary adenoma, an epithelial lesion considered as benign, and morphologically indistinguishable from pRCC except by its size (<5 mm) [3-5].

The concomitant development of multiple tumors and the papillary adenoma–carcinoma association make pRCC an interesting model to gain insight into the evolutionary process of human cancer, for which two models have been proposed: either cancer stem/progenitor cells drive tumor initiation and progression (cancer stem cell model), or a premalignant or malignant cell population over time acquires multiple mutations, genetic instability and uncontrolled proliferation (clonal evolution model) [6-8].

The occurrence of de novo pRCCs and papillary adenomas in kidney transplants, where donor and recipient characteristics are well established, offers a unique opportunity to study the homing of tumor-initiating cells in the two lesions and to compare their molecular characteristics.

We aimed to study whether or not the different papillary tumors have the same donor/recipient origin, and whether or not papillary adenomas are precursor lesions of pRCC in exceptional situation of kidney transplantation. This study was performed using XY-FISH in sex-mismatched kidney transplants, and polymorphic microsatellite DNA and high-resolution melting (HRM) of hypervariable D-loop of mitochondrial DNA (mtDNA) analyzes on laser-microdissected tumor cells.

Materials and Methods

Patients and tissue samples

From 2000 to 2010, 2158 renal transplant recipients were followed up in Hôpital-Saint-Louis, kidney transplantation department, Paris, France. Five of them who developed concomitant pRCC(s) and papillary adenoma(s) in their kidney transplant were included in this study (Table 1). At median age 41 (range 34–47), they had received a renal graft from a deceased donor (median age 33, range 17–43). After transplantation, all patients received 100–300 mg cyclosporine with 2–10 mg prednisone daily, plus 100 mg azathioprine for patient 1 and 2 mg mycophenolate mofetil for patients 2 and 4. Median time lapse between transplantation and pRCC diagnosis was 12 years (range 9–22), median follow-up was 58 months (range 47–107), and in 2011, the five patients were alive without metastasis.

Table 1. Clinical characteristics of five kidney transplant recipients who developed concomitant pRCC and papillary adenoma in their transplanted kidney
PatientSex D/RAge at transplantation D/R(y)Underlying renal diseaseImmunosuppressionTime-lapse between transplantation and pRCC diagnosis(y)Number of papillary hyperplasiaNumber of papillary adenomasNumber of pRCCpRCC size(mm)Fuhrman grade and TNM stagingTreatmentFollow-up (months)
  1. D = donor; R = recipient; CSA = cyclosporin A; Aza = azathioprine; MMF = mycophenolate mofetil; NSS = nephron-sparing surgery; NTxE = transplantectomy.
2F/M22/40IgA nephropathyCSA, prednisone, MMF1901142G2pT1bNoMoNSS61
3F/F33/47GlomerulonephritisCSA, prednisone1211116G2pT1aNoMoNSS65
4F/M41/45GlomerulonephritisCSA, prednisone, MMF2201140G2pT1aNoMoNSS82
5M/F43/34Interstitial nephritisCSA, prednisone106227,10G2pT1aNoMoNTxE54

All surgical specimens, taken for diagnostic purposes, had been formalin fixed and paraffin embedded. Two pathologists (J.V., A.J.) reviewed slides according to the 2004 World-Health-Organization classification [1]. On samples remaining after diagnosis had been established, 5 μm-thick following sister sections were cut from each block for XY-FISH analyses. DNA was extracted from donor and recipient peripheral blood leukocytes.

The study was approved by Hôpital-Saint-Louis institutional review board, and informed consent was obtained in accordance with the Declaration of Helsinki.

Combined XY-FISH and immunostaining

FISH analyses were performed for papillary tumors from patients who (i) received a sex-mismatched kidney transplant (patients 2, 4, 5, Table 2), (ii) received a sex-matched kidney transplant (patients 1, 3), (iii) not undergone kidney transplantation (patients 6, 7). Monoclonal mouse antihuman cytokeratin-7-antibody (CK7, clone OV-TL-12/30, 1/50, Dako), an epithelial cell cytoplasmic protein strongly expressed by papillary renal tumor cells, was used as primary antibody. After DAB staining, sections were HCl treated, sodium chloride/sodium citrate buffer washed, proteinase K digested, and formaldehyde fixed. CEP-X/Y-DNA probes (Vysis/Abbott) were denatured (90°C/10 min) prior to hybridization (42°C/16 h) on Thermoydrite device (Abbott). Hybridized slides were mounted with a DAPI-containing medium (Vector/CliniSciences).

Table 2. Combined XY-FISH and immunostaining in five patients with concomitant papillary adenoma and pRCC, and in two patients with pRCC not in the kidney transplantation setting
    XY FISH
    CK7 positive tumor cells
Sex  
 DonorRecipientRenal tumorXYXXXYTotalX cells %Y cells %XX cells %XY cells %XX or XY corrected*
  • *Tumor samples from two patients with pRCC not in the kidney transplantation setting (patients 6 and 7) were examined. To determine the efficiency of detecting sex chromosome in pRCC tumor cells, a combined cytokeratin-7/XY-FISH protocol was applied. Tumor sections from the two patients were analyzed and percent detection averaged ((68.6+70.3):2=69.45). The normalization factor was derived by dividing 100% by the average percent (100:69.45=1.44). Here the normalization factor was 1.44 for XX and XY cell detections.
Patients with renal tumor after sex-mismatched kidney transplant
2FMPapillary adenoma39073011234.8065.2093.9
   Papillary carcinoma37068010535.3064.7093.3
4FMPapillary adenoma3606209836.7063.3091.1
   Papillary carcinoma3005908933.7066.3095.4
5MFPapillary adenoma 135066010134.7065.3094.1
   Papillary adenoma 238083012131.4068.6098.8
   Papillary carcinoma 143072011537.4062.6090.2
   Papillary carcinoma 23206109334.4065.6094.5
Patient with renal tumor after sex-matched kidney transplantation
1MMPapillary carcinoma18120679718.612.3069.199.5
3FFPapillary carcinoma46097014332.2067.8097,6
Patients with pRCC not in the kidney transplantation setting (positive controls)
6 MPapillary carcinoma162108111813.617.8068.6 
7 FPapillary carcinoma33078011129.7070.30 

Tissue sections of the same thickness were analyzed by two pathologists (A.J., J.V.) on a motorized Z-axis-Olympus BX61 microscope, using bright and epi-fluorescent light. Microscope pictures obtained through UPlanFI100x/1.3NA were captured with ColorViewIII camera using Olympus-SIS software.

For chromosomal analysis, 10 sequential Z-stack images of the same field at 0.5 μm intervals were captured with motorized z-axis system. X and Y signals were counted on nonoverlapping CK7-positive tumor cells. Since tissue sections can pass through nuclei with loss of chromosomic material, the percentage of XX or XY tumor cells is underevaluated by direct counting. To correct this, the observed number of XX or XY tumor cells was multiplied by a “normalization factor” [9] derived from analyses of pRCC tumor cells developed in two nontransplanted patients (positive controls): one male (patient 6) and one female (patient 7) (see Table 2).

Laser microdissection

For PCR-STR and HRM mtDNA studies tumor-enriched DNA was required. Laser microdissection (PALM-Zeiss) was performed on snap-frozen samples. A minimum of 1000 CK7-positive tumor cells were laser microdissected for each sample. DNA was extracted using a Qiagen-kit.

Quantitative PCR

Total RNA was extracted from laser-microdissected CK7-positive tumor cells and reverse transcribed using random primers with SuperScriptTM II Reverse Transcriptase (Invitrogen). qPCR was performed to determine KRT7 (also known as CK7) and PTPRC (protein tyrosine phosphatase receptor type C, also known as CD45, leukocytic marker) gene expression using Hs00559840_m1 (NM_005556.3) and Hs00365634_g1 (NM_002838.3) primers (Applied Biosystems), respectively. Human TBP primer was used as a housekeeping gene (Hs99999910_m1, NM_003194.4). All experiments were performed in triplicate.

Relative quantification was established using, as positive controls, laser-microdissected normal renal tubules for KRT7 mRNA, and peripheral blood mononuclear cell for PTPRC mRNA.

Short-tandem repeat analyses

Eleven highly polymorphic STR sequences were amplified [10]: D3S1597, D3S3611, D5S2095, D6S440, D8S261, D8S1820, D9S162, D17S802, D17S1879, HPRTB and DXS101. We also performed analysis of the homologous amelogenin gene X- and Y-specific alleles [11]. For donor/recipient origin of papillary renal tumors, donor and recipient DNA were first compared. Fluorescent allelic profiles obtained from CK7-positive laser-microdissected tumor cells were then compared with those obtained from donor and recipient cells for informative microsatellite markers. All tests were performed twice.

For a genetic alteration study, PCR products were analyzed [12]. Informative cases were scored as loss of heterozygosity (LOH) when the intensity of signal for one allele in tumor tissue specimens was decreased by > 50% in comparison to allelic signal observed in normal tissue specimens, i.e. donor or recipient DNA. Allelic imbalance (AI) was scored when the signal for one allele was decreased >20% and <50%. Homozygous markers were quoted “not informative”.

PCR-HRM analyses

Primers were designed to flank hypervariable regions 1 and 2 (HV1/HV2) of the noncoding displacement loop of mtDNA from NCBI Reference Sequence NC_012920.1 [13]. They were synthesized by Eurogentec (Table 3).

Table 3. PCR primer sequences for hypervariable regions HVI and HV2
PCRPrimer nameSequenceAmplicon length (bp)
HVII 16105 F5′-TGCCAGCCACCATGAATA-3′243
 I 16348 R5′-GACTGTAATGTGCTATGTACGGTAAA-3′ 
HV2II 45 F5′-ATGCATTTGGTATTTTCGTCTG-3′242
 II 287 R5′-TTGTTATGATGTCTGTGTGGAAAG-3′ 

PCR was carried out on LightCycler480 (Roche) on 20 μL total volume containing 5 μL of genomic DNA (20 ng), 15 μL of LightCycler 480 High Resolution Melting Master 1X (Roche) and 0.4 μM of each forward and reverse primer. PCR included initial denaturing step (94°C/10 min), followed by 50 denaturing cycles (94°C/2 s), annealing (57°C/6 s) and extension (72°C/6 s). After PCR, a postamplification melting curve program was initiated by heating (94°C/10 s), cooling (55°C/10 s) and increasing temperature by 0.1°C/ s to reach 95°C. Each PCR run included a no-template control, and each sample was run in triplicate. Postamplification fluorescent melting curves were analyzed using LC480 Gene Scanning software v1.2.9 (Roche).

Immunohistochemistry

Automated indirect-immunoperoxidase method was performed (Discovery/Roche) using monoclonal mouse antihuman p16INK4a antibody (clone JC8, 1/50, Biocare Medical) and polyclonal rabbit anti-BNIP3L (BCL2/adenovirus E1B 19kDa interacting protein 3-like) antibody (1/50, Abcam), as primary antibodies. The absence of primary antibody and irrelevant antibody of same isotype was used as controls. Two pathologists (J.V., A.J.) performed analyses.

Results

Papillary renal-cell carcinomas and associated papillary adenomas

In five kidney-transplanted patients with concomitant pRCC(s) and adenoma(s) here studied, all pRCCs were type 1 pRCCs with a median size of 18 mm (range 7–42) (Table 1). Indirect evidence of de novo origin of these pRCCs was provided by the absence of renal mass and atypical cyst in native kidneys at diagnosis and during follow-up, by evolution without metastasis in the course of close clinical follow-up and imaging (median 58 months, range 47–107), by a long time lapse (median 12 years, range 9–22) between kidney transplantation and pRCC diagnosis, and by a low Fuhrman-nuclear-grade (G2) and TNM staging (pT1a/pT1bN0M0).

Four patients had one pRCC and one papillary adenoma distant from each other. Patient 5 had two pRCCs and two papillary adenomas on same surgical sample, together with six papillary hyperplasias (size <1 mm). There was no contact between two concomitant pRCCs, 6 cm apart, nor between pRCCs and papillary adenomas. Papillary hyperplasias, papillary adenomas and pRCCs all had similar degree of cellular differentiation (Figure 1A).

Figure 1.

In situ analyses of concomitant papillary hyperplasia, papillary adenoma and papillary renal-cell carcinoma in the grafted kidney of a female patient with kidney transplant from a male donor (patient 5). (A) Papillary hyperplasia and papillary adenoma, two lesions considered as benign, and papillary renal cell carcinoma have the same morphological aspect. At higher magnification (small panels), the cellular differentiation is the same. The only morphological feature available to differentiate them is the size of lesions: less than 1 mm for papillary hyperplasia, 1–5 mm for papillary adenoma, and over 5 mm for papillary renal-cell carcinoma. (B) Combined CK7 immunostaining with FISH analysis for X chromosome (green signal) and Y chromosome (red signal) of papillary adenoma 1. CK7+ tumor cells have XX genotype on overlay, as shown at higher magnification (small panel), consistent with recipient origin. For papillary carcinoma 1, CK7+ tumor cells have XX genotype on overlay, as shown at higher magnification (small panel), consistent with a recipient origin. Scale bars: 100 μm (A); 20 μm (B).

In follow-up of kidney-transplant recipients, one episode of acute humoral rejection was observed at day 15 in patient 3, and recurrent episodes of acute graft pyelonephritis in patient 5.

Sex chromosome genotype of de novo pRCCs

The clinical situation of sex-mismatched kidney transplant enabled us to study chimerism of de novo pRCC samples using XY-FISH combined with cytokeratin-7 immunostaining on same tissue section.

Among three sex-mismatched patients studied, patient 5, a female recipient of male kidney, had XX genotype in 62.6% and X genotype in 37.4% of 115 CK7-positive pRCC1 tumor cells studied (Table 2 and Figure 1B), and XX genotype in 65.6% and X genotype in 34.4% of 93 CK7-positive pRCC2 tumor cells studied. No CK7-positive tumor cell from these two pRCCs had XY or Y genotype whereas XY genotype was noted in endothelial cells and XX genotype in inflammatory cells inside tumors. As Y chromosome loss in men is an early karyotypic change in papillary subtype of pRCC [1], we checked that CK7-positive tumor cells with X genotype were not XY donor-derived tumor cells having lost Y. Therefore, we analyzed CK7-positive pRCC tumor cells from one male and one female who had received a sex-matched kidney transplant (patients 1 and 3) and one male and one female who had not undergone kidney transplantation (patients 6 and 7 as positive controls) (Table 2). After application of a “normalization factor” (1.44) (note above Table 2) derived from these positive controls, proportions of XX genotype in two concomitant pRCCs of patient 5 were 62.6%  × 1.44 = 90.2% (pRCC1) and 65.6%  × 1.44 = 94.5% (pRCC2), respectively. These data were consistent with a XX genotype for all tumor cells of these two concomitant de novo type 1 pRCCs.

In two other sex-mismatched transplant patients, two male recipients with a female graft (patients 2 and 4), pRCCs had no XY genotype of the recipient, i.e. the absence of Y signal, and 93.3% (patient 2) and 95.4% (patient 4) of XX tumor cells after normalization (Table 2).

Donor/recipient origin of de novo pRCCs

Two hypotheses could explain XX genotype of two pRCCs in patient 5, i.e. a recipient origin or a donor origin with copy number alteration of sex chromosomes (Y chromosome deletion and partial or complete duplication of X chromosome, including the Xp11.1-q11.1 pericentromeric region recognized by CEP X/Y DNA probe used for FISH). STR-PCR analyses of laser-microdissected pRCC cells were performed (Figure 2A). A prior quality control of laser-microdissected pRCC tumor cells using qPCR analyses showed that laser-microdissected cells were epithelial cells (KRT7-positive), and that there was no contamination by leucocytes (PTPRC-positive) from the recipient (Figure 2B).

Figure 2.

Similar recipient origin of the two papillary adenomas and the two pRCCs in patient 5. (A) Papillary adenoma and pRCC CK7-positive cells were laser microdissected from 7 μm-thick tissue sections. (B) qPCR analyses showed that laser microdissected cells were cytokeratin (KRT7) positive and that there was no contamination by leucocytes (PTPRC also known as CD45) from the recipient. Comparison of the DNA profiles of the laser-microdissected cells from papillary adenoma and papillary carcinoma cells, as well as from normal donor cells and normal recipient cells were analyzed. (C) Analysis of the homologous amelogenin gene X- and Y-specific alleles showed that only the X-specific product (106 bp) was amplified in the four papillary tumors, i.e. two papillary adenomas and two pRCCs, as in the recipient, contrary to the donor where the X- and Y-specific products (106 bp/112 bp) were detected. D, STR sequence analysis at HPRTB, a microsatellite marker on X chromosome, showed two identical alleles for the two papillary adenomas, the two pRCC and the recipient profiles, i.e. 278 bp/290 bp, with additional allelic imbalance (arrowhead) in three of these tumors. E, STR sequence analysis at D3S3611 locus showed that the two papillary adenoma cells, the two papillary carcinoma cells, and the recipient cells are identical and heterozygous for this locus (106 bp and 120 bp), whereas the donor DNA was heterozygous but different (115 bp and 121 bp). Furthermore an allelic imbalance (arrowhead) was noted in the two pRCCs. Scale bars: 200 μm.

A prior quality control of laser-microdissected pRCC tumor cells using qPCR analyses showed that laser-microdissected cells were KRT7-positive and PTPRC-negative, thus confirming that laser-microdissected cells were epithelial cells (KRT7-positive), and that there was no contamination by leucocytes (PTPRC-negative) from the recipient (Figure 2B).

We first examined homologous amelogenin gene X- and Y-specific alleles in the donor, the recipient, and the two pRCC DNA samples. This study showed that only X-specific product (106 bp) had been amplified in the two pRCCs and in the recipient, contrary to the donor where X- and Y-specific products (106 bp/112 bp) were detected (Figure 2C). We also analyzed two highly polymorphic microsatellite markers located in X chromosome, i.e. HPRTB (Xq26), and DXS101 (Xq22.1), located outside the human pseudo-autosomal regions [14]. These two markers showed two identical alleles for the two pRCC and the recipient profiles, i.e. 278 bp/290 bp for HPRTB (Figure 2D) and 209 bp/212 bp for DXS101. Gender genotype of two pRCC tumor cells were XX, suggesting a recipient origin.

To get further inside this donor/recipient analysis, we studied nine highly polymorphic microsatellite dinucleotide markers located on autosomes. Five of them, i.e. D3S1597, D3S3611, D5S2095, D9S162, D17S1879, were discriminant between donor and recipient (patient 5). The pRCC1 and pRCC2 STR-PCR profiles of these five markers were identical with the recipient profiles (as shown for D3S3611 in Figure 2E), whereas none was similar to the donor profiles.

Altogether these results demonstrated the recipient origin of these two concomitant de novo pRCCs.

In the two other patients with sex-mismatched kidney transplant (patients 2 and 4) and in the two patients with sex-matched kidney transplant (patients 1 and 3), STR-PCR analyses showed the donor origin of all pRCCs.

Donor/recipient origin of papillary adenomas

In patient 5, combined XY-FISH and CK7 immunostaining analyses showed XX genotype in 65.3% (94.1% after normalization) of 101 CK7-positive tumor cells analyzed in papillary adenoma 1, and 68.6% (98.8% after normalization) of 121 CK7-positive tumor cells analyzed in papillary adenoma 2 (Table 2 and Figure 1B). After a control of laser-microdissected papillary adenoma tumor cells using qPCR (Figure 2B), profiles of homologous amelogenin gene X- and Y-specific alleles and seven other discriminant donor/recipient microsatellite markers (D3S1597, D3S3611, D5S2095, D9S162, D17S1879, HPRTB, DXS101) were studied. All these profiles were identical for the papillary adenomas 1 and 2, and the recipient (as shown for amelogenin locus, HPRTB, and D3S3611 in Figures 2C, 2D, and 2E, respectively).

Thus combined XY-FISH and CK7 immunostaining, and STR-PCR analyses, showed that the two papillary adenomas in patient 5 were, like the two concomitant pRCCs, recipient derived.

In four other patients, STR-PCR profiles showed that all four papillary adenomas were donor derived. Combined XY-FISH and CK7 immunostaining analysis confirmed the donor origin of papillary adenomas in patients 2 and 4 with sex-mismatched kidney transplant, i.e. XX genotype in 93.9% (patient 2) and 91.1% (patient 4) after normalization (Table 2).

Thus, molecular analyses performed in all five patients showed that papillary adenomas had an identical donor or recipient origin to pRCCs.

Papillary adenomas versus pRCCs of recipient origin

In patient 5, the HRM mtDNA curves of hypervariable regions 1 and 2 of the two papillary adenomas were very similar to the recipient curves, and distinct from the two concomitant pRCC curves (Figure 3A). Since widespread heterogeneity (heteroplasmy) has been found in mtDNA of normal human cells as well as human cancer cells [15], we further checked our HRM results using Sanger sequencing, and no heteroplasmy was detected.

Figure 3.

Comparison of the two papillary adenomas with the two pRCCs of recipient origin. (A) Using HRM on mitochondrial PCR product, mtDNA profiles obtained from the hypervariable regions 1 (A left) and 2 (A right) were analyzed for the four renal papillary lesion cells (i.e. papillary adenoma 1, papillary adenoma 2, papillary carcinoma 1 and papillary carcinoma 2) and for the normal recipient cells, the donor melting curve being the reference. The mtDNA profiles of papillary adenoma 1 and papillary adenoma 2 cells were identical between them and with the recipient mtDNA profile. Papillary carcinoma 1 and papillary carcinoma 2 melting curves were different from the melting curves common to the two papillary adenomas. (B) At the D9S162 locus, papillary adenoma 1 and papillary adenoma 2 were heterozygous (182 bp and 202 bp) and papillary adenoma 2 showed an allelic imbalance (arrowhead) on one allele (202 bp) by comparison with recipient profile. For the same locus, papillary carcinoma 1 and papillary carcinoma 2 were also heterozygous (182 bp and 202 bp) but both of them showed a loss of heterozygosity on the other allele (182 bp). C, p16INK4a protein, product of CDKN2A gene located close to D9S162, was expressed in the nucleus of papillary adenoma cells. Although also expressed in the nucleus, p16INK4a predominated in the cytoplasm of papillary carcinoma cells. D, D8S1820 microsatellite locus showed an identical heterozygous profile for papillary adenoma 1 and papillary adenoma 2 (106 bp and 114 bp). Papillary carcinomas 1 and 2 also had identical profiles, and both showed a loss of heterozygosity on one allele (106 bp LOH). E, BNIP3L, chosen because the BNIP3L gene is located close to D8S1820, was expressed in the cytoplasm of all papillary adenoma cells whereas in papillary renal carcinoma BNIP3L expression was low and limited to the cytoplasm of few tumor cells. Scale bars: 10 μm.

To further study molecular differences between concomitant papillary adenomas and pRCCs in patient 5, we compared more precisely STR-PCR profiles of highly polymorphic autosomal microsatellite nucleotide markers. Two of them (DS8S261, D17S1879) were common to all four papillary lesions. HPRTB showed an AI for papillary adenoma 2, pRCC1, and pRCC2 (Figure 2D) whereas DXS101 exhibited an AI only for pRCC1. An AI was also noted at D3S3611 in the two pRCCs (Figure 2E). For the four remaining informative microsatellite markers, there was a LOH in papillary carcinoma 1 at D5S2095, in papillary carcinoma 2 at D3S1597, and in two pRCCs at D8S1820, and D9S162. If no LOH was found in papillary adenoma 1 or 2, an AI was detected in papillary adenoma 2 at D9S162 (Figures 3B for D9S162 and 3D for D8S1820).

To confirm in patient 5 the genetic differences between papillary adenomas and pRCCs at protein level, we performed immunohistochemistry. Since the CDKN2A (cyclin-dependent kinase inhibitor 2A) gene, encoding for p16INK4a protein, is located at 2.5 MB to D9S162 and has been implicated in renal carcinogenesis [16], antihuman p16INK4a antibody was tested. p16INK4a was expressed in nucleus of papillary adenomas 1 and 2. Although expressed in nucleus, p16INK4a predominated in cytoplasm of papillary carcinomas 1 and 2 (Figure 3C). Antihuman BNIP3L antibody was also studied, since the BNIP3L (BCL2/adenovirus E1B 19kDa interacting protein 3-like) gene is located at 1.6 MB to D8S1820 and known to be a tumor suppressor gene [17]. BNIP3L was expressed in cytoplasm of all papillary adenoma cells whereas in pRCCs BNIP3L expression was low and limited to cytoplasm of few tumor cells (Figure 3E).

In four concomitant pRCCs and papillary adenomas of one sex-mismatched kidney-transplant recipient, molecular analyses demonstrated a similar recipient origin, and additional genetic abnormalities in pRCCs when compared to papillary adenomas.

Discussion

The systematic study performed herein on five kidney transplant recipients with concomitant pRCC and papillary adenoma using three independent methods enabled to demonstrate recipient origin of two pRCCs and two papillary adenomas developed in the same transplanted kidney (patient 5).

The recipient origin of de novo pRCCs in a transplanted kidney has not been reported to date. In four previous cases, chimerism studies using XY-FISH and polymorphic microsatellite DNA analyses had shown a donor origin [18-20]. Identification of recipient origin in pRCCs implies migration and homing of recipient cells to transplanted kidney, but the nature of these recipient cells is not characterized. The first hypothesis is homing of metastatic cells from a pRCC developed in one of native kidneys, since native kidneys of patient 5 had not been removed. Against metastatic origin hypothesis of these pRCCs is the absence of renal mass and atypical cyst in native kidneys at diagnosis and during follow-up; an evolution without metastasis in course of close clinical follow-up and imaging (median 58 months, range 47–107); a long time-lapse (median 12 years, range 9–22) between kidney transplantation and pRCC diagnosis; and a low Fuhrman nuclear grade (G2) and TNM staging (pT1a/pT1bN0M0). Finally, other papillary lesions considered as benign and frequently associated with primary pRCCs [4, 21] were concomitantly found on the same surgical piece, i.e. six papillary hyperplasias and two papillary adenomas; the latter were also recipient derived. Altogether, these data do not allow to retain this assumption.

Engraftment of exfoliated normal tubular epithelial cells coming from native kidneys into transplanted kidney via a mechanism of vesico-ureteral reflux might also occur. In kidney transplant recipients as in normal individual, thousands of exfoliated renal tubular epithelial cells are shed into urine each day [22], able to establish viable cultures [23, 24] and form nephrogenic adenomas [25]. Patient 5 had a vesico-ureteral reflux but no residual diuresis before kidney transplantation and no nephrogenic adenoma.

Finally, migration and homing of circulating normal recipient-derived stem/progenitor cells to transplanted kidney realize probably the most likely hypothesis. Populations of progenitor/stem cells resident outside the organ are able to contribute to renewal of different lineages, even in tissue from a separate germ layer [26]. Bone marrow-derived cells (BMDC) have a high degree of plasticity, and are able to differentiate in numerous tissue lineages [26]. Although BMDC seem to act essentially through a paracrine/endocrine mechanism on resident renal tubular epithelial cells [27, 28], they can also participate in renal tubular regeneration by direct incorporation into tubular epithelium [29-31]. Origin of these normal recipient-derived stem/progenitor cells could also be the resident renal stem/progenitor cells of adult kidney, since multipotent stem/progenitor cells have been characterized, particularly in Bowman's capsule, and proximal tubules [32, 33]. Triggering factor for homing of these normal recipient-derived stem/progenitor cells could be tissue injury [34-36], as acute graft pyelonephritis described in this patient.

The exceptional clinical situation of concomitant occurrence of papillary adenomas and pRCCs also raises the question of evolutionary process implicated in development of de novo pRCCs in our patients. Papillary adenoma is a very common renal lesion, estimates to over 40% in a population >70 years old [37] which shares a high degree of morphological similarity with pRCC. Consequently, it has been considered as a potential precursor of pRCC, but considerable differences in incidence of these two lesions have cast doubts on this hypothesis [5]. Moreover, nosological positioning of papillary adenoma in spectrum of renal papillary lesions can also be discussed: is it a benign lesion or a true carcinoma of less than 5 mm? Definition of papillary adenoma's size, always somewhat arbitrary, has considerably evolved over the last four decades, from <3 cm to <5 mm. Herein, in each of our patients, we identified at molecular level similar donor/recipient origin for papillary adenomas and pRCCs. In patient 5 with recipient-derived papillary renal lesions, we firstly performed HRM analysis of mtDNA hypervariable regions 1 and 2, since mtDNA mutations play an important role in development and progression of cancers [38]. HRM mtDNA curves of the two papillary adenomas were very similar to recipient curves, and totally distinct from the two concomitant pRCC curves. Using STR-PCR, we then focused on 9p21 band, which has been reported as frequently altered in renal carcinoma [3, 16]. This analysis showed a LOH at D9S162 microsatellite marker in the two pRCCs whereas an allelic imbalance was noted in one of two papillary adenomas. At protein level, this 9p21 LOH in pRCCs was in accordance with a loss of nuclear p16INK4a, product of the CDKN2A gene, in the two pRCCs, while it was normally present in the two papillary adenomas. As inactivation of the CDKN2A gene locus has been found to correlate with loss of nuclear p16INK4a expression [39], these data confirm that additional p16INK4a alterations occur in pRCCs. We also studied the 8p21.3 locus, a chromosomal region known to harbor tumor suppressor genes, including BNIP3L [17]. In the two pRCCs, molecular analyses detected a LOH at D821820 microsatellite marker that was absent in the two papillary adenomas. This was confirmed at protein level, with loss of cytoplasmic expression of BNIP3L in pRCCs whereas it was present in papillary adenomas.

Thus, comparative molecular analyses of laser-microdissected papillary adenomas and pRCCs enabled us to conclude that papillary adenomas developed in patient 5 have genomic and mitochondrial profiles close to normal recipient DNA suggesting that these two tumors are not true carcinomas. This implies that papillary adenoma-initiating cells cannot be cancer stem cells. Two possibilities can then be discussed for tumor progression of pRCC: either two types of recipient-derived precursor cells homed in grafted kidney, a cancer stem cell for pRCC and a noncancer stem cell for papillary adenoma, these two lesions being then distinct; or tumor progression to pRCC resulted from a multistep process. Since in this context sequential biopsies cannot be performed for ethical reasons, no formal proof of a papillary adenoma–papillary carcinoma sequence can be established in our patient. Nevertheless, cellular and molecular studies we performed in patient 5 showed that papillary adenomas and pRCCs (i) had the same recipient origin, (ii) were located in the same original tumor microenvironment, (iii) had the same cellular aspect, and thus the same degree of cellular differentiation, (iv) had size differences, benign tumors being smaller than malignant ones and (v) had molecular differences, malignant tumor having additional genetic alterations. All these elements are in favor of a multistep molecular process leading to pRCC, via the intermediate benign or premalignant papillary adenoma as in colon cancer [40]. However, recent theorical syntheses do not exclude development of cancer stem cells at different steps of a clonal evolution [7, 8].

In conclusion, we here identified the recipient origin of de novo papillary adenomas and pRCCs in a kidney transplant, demonstrated an identical origin for precursor cells of papillary adenomas and pRCCs, and showed additional genetic alterations in pRCCs compared to papillary adenomas. This molecular approach of tumors developed in transplanted kidney through not interfering with clinical practice, identified successive steps in the carcinogenesis of human de novo papillary renal-cell carcinoma.

Acknowledgments

We owe thanks to S Germain for critical review of manuscript, and I Ferreira and F Boudihel for expert molecular and immunohistochemical assistance. Ms. A Swaine reviewed the English language. This study was supported by grants from Agence Nationale pour la Recherche (ANR), and Cancéropôle-Institut National du Cancer (InCa).

Disclosure

The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.