Genomic heterogeneity of clear cell renal cell carcinoma Ferronika, Paranita

(1)

Genomic heterogeneity of clear cell renal cell carcinoma Ferronika, Paranita

DOI:

10.33612/diss.101437783

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.

Document Version

Publisher's PDF, also known as Version of record

Publication date:

2019

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):

Ferronika, P. (2019). Genomic heterogeneity of clear cell renal cell carcinoma. Rijksuniversiteit Groningen.

https://doi.org/10.33612/diss.101437783

Copyright

Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

Take-down policy

If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.

Download date: 28-06-2021

(2)

537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika

Processed on: 24-10-2019 Processed on: 24-10-2019 Processed on: 24-10-2019

Processed on: 24-10-2019 PDF page: 1PDF page: 1PDF page: 1PDF page: 1

1 Clear Cell Renal Cell Carcinoma

Paranita Ferronika

(3)

537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika

2 The work presented on this thesis was mainly funded by Nuffic through Netherlands Fellowship Programmes (Grant NFP-PhD.13/119). Printing of this thesis was financially supported by Universitas Gadjah Mada, University of Groningen, and University Medical Center Groningen

Cover design & layout Bianca Pijl, www.pijlldesign.nl Groningen, the Netherlands Shibori pattern design Beni Sulistiono and Neni Lastri Printed by Ipskamp Printing

Enschede, the Netherlands ISBN 978-94-034-2160-5 (print) 978-94-034-2159-9 (digital)

© Copyright: 2019 P. Ferronika, Groningen, the Netherlands

All rights reserved. No part of this thesis may be reproduced, stored in a retrieval system,

or transmitted in any form or by any means, without prior written permission of the

author, or when appropriate, of the publishers of the publications included in this thesis.

(4)

537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika

3 PhD thesis

to obtain the degree of PhD at the University of Groningen

on the authority of the Rector Magnificus prof. C. Wijmenga

and in accordance with the decision by the College of Deans.

This thesis will be defended in public on

Tuesday 26 November 2019 at 11.00 hours

by

Paranita Ferronika

born on 7 November 1983 in Yogyakarta, Indonesia

Clear Cell Renal Cell Carcinoma

(5)

537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika

4 Co-supervisors Dr. K. Kok Dr. H. Westers

Assessment Committee Prof. H. Hollema

Prof. I.J. de Jong

Prof. R.M. Medeiros

(6)

537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika

(7)

537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika

6 Jiacong Wei

(8)

537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika

7

Chapter 1

Chapter 2

Chapter 3

Chapter 4

Chapter 5

Chapter 6

Chapter 7

Introduction

Evaluation of the mutational profile as a prognostic factor in a population-based study of clear cell renal cell carcinoma

Mutational heterogeneity between different regional tumour grades of clear cell renal cell carcinoma

Comprehensive profiling of primary and metastatic ccRCC reveals a high homology of the metastases to a subregion of the primary tumour Copy number alterations assessed at the single-cell level revealed mono- and polyclonal seeding patterns of distant metastasis in a small-cell lung cancer patient

Summary

General discussion and future perspective Nederlandse samenvatting

Acknowledgments

Curriculum vitae

9 21

51

79

111

127

135

151

159

167

(9)

537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika

(10)

537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika

9 Introduction

1

(11)

537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika

(12)

537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika

11 Introduction

1. Clear cell renal cell carcinoma (ccRCC)

1. 1. Clinical and epidemiological aspects of ccRCC

Worldwide, kidney cancer is the 14

^th

most common cancer, accounting for 3% of all sporadic cancers [1]. According to the 2016 World Health Organisation (WHO) classification, kidney cancer is classified into subtypes based on the following characteristics: predominant cytoplasmic and architectural features, anatomic location of tumours, correlation with a specific renal disease background, underlying molecular alterations and familial predisposition syndromes [2].

Renal cell cancer accounts for 90% of all kidney cancer, and clear cell renal cell carcinoma (ccRCC), characterized typically abundant clear cytoplasm due to lipid or glycogen deposition, is its most common subtype (75%) [2, 3]. The worldwide incidence of ccRCC is >337,000 new cases/year [4].

Although there are some hereditary tumour syndromes that feature ccRCC, including von Hippel- Lindau syndrome, most ccRCCs are sporadic.

The ratio between the incidence of localized and metastatic kidney cancer is 2:1 [5]. For patients with localized kidney cancer that is diagnosed early, and treated by total or partial nephrectomy, the prognosis is quite good: the five-year survival rate is up to 90% [6]. However, metastases will still manifest during follow-up in one third of these patients [5]. For patients with metastatic disease, which is treated with systemic therapy either in combination with primary tumour removal (cytoreductive nephrectomy) or not [3], five-year survival decreases to 12% [6].

Prognostic parameters have been defined that predict the clinical outcome of kidney cancer subtypes including ccRCC. The first prognostic assessment is based on the morphological feature: tumour grade. The latest tumour grade assessment is described in the 2016 International Society of Urological Pathology (ISUP)/WHO classification. It classifies the tumour into grade 1-4 based on nucleolar grade and the presence of the extreme form of tumour dedifferentiation [7]. The presence of independent sarcomatoid or rhabdoid dedifferentiation independently decreases overall survival of the kidney cancer patients [8, 9]. Other prognostic parameters used in pathological assessment of kidney cancer are the presence of tumour necrosis, the presence of microvascular invasion, and the tumour subtype itself [10]. These pathological prognostic parameters, in combination with other clinical parameters including tumour stage and size, are used to predict the cancer-specific survival of patients [11].

1.2. Frequent somatic alterations in sporadic ccRCC

Early genomic studies revealed large copy number variations in ccRCC, with loss of one copy of

the short arm of chromosome 3 being the most characteristic/prominent [12-14]. Also, according

to the more recent cancer genome atlas project (TCGA) database, the chromosomal alterations in

ccRCC are dominated by the loss of the short arm of chromosome 3 in close to 100% of the cases

(13)

537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika

12 [15]. Other frequent chromosomal alterations include loss of 14q (47%), 9p (32%), 9q (32%) and gain of 5q (60%) [15, 16]. The loss of chr9 or 14q has been found to correlate with a higher tumour grade, the presence of metastasis, and worsened cancer-specific and recurrence-free survivals [17, 18].

As far back as 1994, biallelic inactivation [14] had been used to identify VHL as a major tumour suppressor gene in several cases of sporadic ccRCC. With several hundred tumours now analysed with high- throughput sequencing, the Catalogue of Somatic Mutations in Cancer (COSMIC) and the TCGA now give a clear overview of the mutation frequencies and significant altered pathways in ccRCC [15, 19]. With a mutation frequency of 40-50%, VHL is still the most frequently mutated gene observed in ccRCC. VHL has an important role in the regulation of the two crucial hypoxia inducible factors, HIF-1α and HIF-2α. The list of top-six mutated genes in ccRCC is completed by PBRM1 (38%), SETD2 (12%), BAP1 (9.5%), MTOR (8%), and KDM5C (5%) [15], which all, except MTOR, are known as chromatin-modifying tumour suppressor genes [20-23]. Remarkably, the four most frequently mutated genes –VHL, PBRM1, SETD2 and BAP1 – are all located on the short arm of chromosome 3. Thus, a single event, e.g. loss of one copy of the short arm of chromosome 3, results in hemizygosity for four tumour suppressor genes. In addition, TP53 is among the most frequently mutated genes in ccRCC with a mutation frequency of 3% [15], but this is a relatively low mutation frequency compared to other cancers. ccRCC is thus characterized by the alteration of genes involved in epigenetic regulation (e.g. chromatin modification) and cellular metabolism regulation (e.g. cellular oxygen sensing) [24].

Some mutated genes have been correlated with clinical behaviour in ccRCC. Mutations in TP53, together with BAP1, were recently correlated with worsened survival in ccRCC patients [24, 25].

Based on the same studies, SETD2-inactivating mutations were suggested to correlate with a higher recurrence rate. Other genes worth mentioning because of their clinical significance are VHL and MTOR. The availability of targeted therapies aimed at inhibiting the VHL and mTOR signalling pathways makes screening for mutations in these genes, and other genes in these pathways, important [3, 26].

1.3. Tumour evolution and metastasis development in ccRCC

The mutational events that occur in the evolution of sporadic ccRCC have been the topic of many studies. As discussed above, the major event in the development of sporadic ccRCC is 3p loss or 3p uniparental disomy (copy-neutral loss of heterozygosity), which occurs in 94% of the cases [24]. This event, along with inactivation of the remaining allele of VHL by mutation or promoter methylation or by functional inactivation of other genes encoding proteins of the VHL-elongin BC protein complex (i.e. TCEB1, TCEB2, CUL2/5 and RBX1) contributes to 92% of the ccRCC cases.

This biallelic inactivation of VHL disrupts the VHL/HIF pathway, which has an important role in the

regulation of energy metabolism, angiogenesis, cell proliferation and apoptosis [27]. Von Hippel-

Lindau disease is characterized by a germline mutation causing inactivation of one of the VHL

copies [28]. This disease is clinically classified based on the presence of either pheochromocytoma,

ccRCC, or both [29]. The mean age of onset of hereditary ccRCC in Von Hippel-Lindau patients is

(14)

537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika

13 44 years [30], which contrasts with the mean age of onset of 62 years for sporadic ccRCC in the general population [28]. In both hereditary and sporadic ccRCC, additional mutations in other genes might contribute to tumour progression [24, 31]. These mutations might be present in all population of tumour cells or in a tumour subclone, as indicated by a low number of mutant reads frequency [31].

The development of metastatic cells is initiated by clonal expansion of competent cells from the primary tumour [32]. This process might involve multiple advantageous aberrations that are gained in specific subclones at different time points during evolution [33]. The metastatic- competent cells can originate from a major clone of the primary tumour or from rare subclones that arise at a later stage [34]. These metastatic clones can progress further through clonal evolution and develop multiple subclones within the metastatic site [32].

2. Personalized Medicine and Tumour Heterogeneity

Decades of research have, above all, taught us that cancer is complex. Each type of cancer shows variable behaviour among patients with respect to tumour progression, clinical outcome and response to therapy. Many research projects, including TCGA, have reported on the large variety of gene mutations and gene expression profiles in tumours. The ENCODE project [35] added another layer of information by focusing on the role of epigenetic changes. Personalized medicine aims to use the variations in cancer genetics and epigenetics to improve therapy and its outcome for individual patients. Although the number of studies in the field of personalized medicine is growing fast, development of this area faces further challenges. Genomic variation among tumour cells is not only present among patients, but also within an individual patient, a phenomenon now commonly referred to as intratumour heterogeneity. Intratumour heterogeneity was first recognized in the 19

^th

century by Rudolf Virchow and other pathologists who showed variation in the morphologic features within one tumour (percentage of tumour cells with particular tumour grade distributed among tumour nest) [36]. Fluorescence in-situ hybridization–based studies have shown heterogeneity among tumour cell populations of primary, recurrent and metastatic bladder cancer with respect to chromosome numbers [37]. Flow cytometry has revealed variations in the DNA index in primary tumour and lymph node metastases of individual cervical cancer patients [38]. Such intratumour heterogeneity could challenge targeted therapy because the different individual tumour cells in one patient may respond differently to specific drugs. The existence of intratumour heterogeneity was further reinforced by multiregion sampling of ten cases of ccRCC [39, 40]. Subsequently, this intratumour heterogeneity was also observed in other cancers, including breast cancer, prostate cancer, colorectal cancer, non-small cell lung cancer and liver cancer [41-45].

Recently, it was found that certain morphologic features, such as rhabdoid differentiation, that

can be present in a specific area within a primary ccRCC, may be associated with the mutational

profile of that particular area [46]. In a few reported cases, intratumour heterogeneity has also

been identified in metastatic ccRCC. Metastatic ccRCC itself is a lethal disease that develops in

(15)

537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika

14 about one third of ccRCC cases. Metastasis in ccRCC is thought to be initiated by clonal expansion of rare subclones within the primary tumour. The identification of these subclones and their genomic profile is complex. It is still not clear how one could recognize the primary tumour subclone from which the metastatic tumour evolved. The identification of genomic profiles from distinct areas of the primary and metastatic tumours can describe genomic alterations that are crucial for the development of metastases. This may help to identify the genomic events that turn primary tumour cells into metastasis-prone cells.

A novel way to interrogate intratumour heterogeneity in high resolution is based on the analysis of single cells. Single-cell sequencing on whole-genome amplified single-cell DNA has recently been applied to study intratumoural heterogeneity in primary tumours of several cancers including breast cancer, kidney cancer, bladder cancer, colon cancer, essential thrombocythemia and glioblastoma [41, 47-52]. Single-cell sequencing enables the detection of genomic alterations in minor tumour subclones that may be missed by bulk sequencing due to dilution within a background of other tumour cells and the normal admixture within a tumour.

By applying these different analyses we hope to better understand how intratumour heterogeneity evolves. Identifying the genomic alterations that characterize the major and minor clones, the early and late events, and the primary tumour and metastases might help to stratify patients in a way that allows for more effective therapy. This may positively influence the clinical outcome of each individual patient.

3. The aim and outline of this thesis

Although ccRCC has been frequently studied, survival prognosis is still generally poor for this tumour type. Some genomic alterations, together with other clinical factors, have been shown to correlate with prognosis in ccRCC; however, their potential correlation with survival prognosis varies between different studies [25, 53]. Variation among studies also exists in the mutation frequency of the identified genes. This variation is influenced by factors such as ethnic background, patient selection, tumour cell content and sequencing platform [15, 54]. The influence of patient selection and tumour cell content might reflect the presence of intertumour and intratumour heterogeneity in ccRCC and the impact of these heterogeneities on clinical studies of ccRCC.

With the work described in this thesis we aimed to contribute to the knowledge of mutational

intertumour and intratumour heterogeneity, including that of metastases, and to that of the

associations of these gene variants with clinical features. The ultimate aim is that these data can

help further improve personalised cancer treatment. In a joint Groningen–Maastricht project, we

studied the mutational spectra in 250 ccRCC cases with the goal of correlating these mutational

patterns with patient characteristics and disease outcome. The first preliminary data on this

project are described in chapter 2. In chapter 3, we studied intratumour heterogeneity in primary

ccRCC from Indonesian and Dutch patients. Our objective here was to evaluate whether or not

we could use a targeted-sequencing approach to identify intratumour heterogeneity and to see

if the mutation pattern correlated with the histological tumour grade across tumour regions. In

(16)

537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika

15 chapter 4, we took a detailed look at intratumour heterogeneity and the profiles of a range of metastases and used this to reconstruct the order of mutational events leading to metastases.

Single-cell sequencing is a novel approach to study intratumour heterogeneity at an even more

detailed level. In preparation for such studies in ccRCC, we had the opportunity, presented in

chapter 5, to perform single-cell PCR-free low-coverage whole-genome sequencing in small cell

lung carcinoma, focussing on copy number alterations.

(17)

537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika

16 References

Lam JS, Shvarts O, Leppert JT et al. Renal cell carcinoma 2005: new frontiers in staging, prognostication and targeted molecular therapy. J Urol 2005; 173: 1853-1862.

Moch H, Cubilla AL, Humphrey PA et al. The 2016 WHO Classification of Tumours of the Urinary System and Male Genital Organs-Part A: Renal, Penile, and Testicular Tumours. Eur Urol 2016; 70: 93- 105.

Ljungberg B, Bensalah K, Canfield S et al. EAU guidelines on renal cell carcinoma: 2014 update. Eur Urol 2015; 67: 913-924.

Torre LA, Bray F, Siegel RL et al. Global cancer statistics, 2012. CA Cancer J Clin 2015; 65: 87-108.

Motzer RJ, Jonasch E, Agarwal N et al. Kidney Cancer, Version 2.2017, NCCN Clinical Practice Guidelines in Oncology. J Natl Compr Canc Netw 2017; 15: 804-834.

Atkins MB, Tannir NM. Current and emerging therapies for first-line treatment of metastatic clear cell renal cell carcinoma. Cancer Treat Rev 2018; 70: 127-137.

Dagher J, Delahunt B, Rioux-Leclercq N et al. Clear cell renal cell carcinoma: validation of World Health Organization/International Society of Urological Pathology grading. Histopathology 2017; 71: 918-925.

Zhang BY, Cheville JC, Thompson RH et al. Impact of Rhabdoid Differentiation on Prognosis for Patients with Grade 4 Renal Cell Carcinoma. Eur Urol 2015; 68: 5-7.

Przybycin CG, McKenney JK, Reynolds JP et al. Rhabdoid differentiation is associated with aggressive behavior in renal cell carcinoma: a clinicopathologic analysis of 76 cases with clinical follow- up. Am J Surg Pathol 2014; 38: 1260-1265.

Delahunt B, Cheville JC, Martignoni G et al. The International Society of Urological Pathology (ISUP) grading system for renal cell carcinoma and other prognostic parameters. Am J Surg Pathol 2013; 37:

1490-1504.

Frank I, Blute ML, Cheville JC et al. An outcome prediction model for patients with clear cell renal cell carcinoma treated with radical nephrectomy based on tumor stage, size, grade and necrosis: the SSIGN score. J Urol 2002; 168: 2395-2400.

van den Berg A, Hulsbeek MF, de Jong D et al. Major role for a 3p21 region and lack of involvement of the t(3;8) breakpoint region in the development of renal cell carcinoma suggested by loss of heterozygosity analysis. Genes Chromosomes Cancer 1996; 15: 64-72.

Foster K, Crossey PA, Cairns P et al. Molecular genetic investigation of sporadic renal cell carcinoma:

analysis of allele loss on chromosomes 3p, 5q, 11p, 17 and 22. Br J Cancer 1994; 69: 230-234.

Gnarra JR, Tory K, Weng Y et al. Mutations of the VHL tumour suppressor gene in renal carcinoma. Nat Genet 1994; 7: 85-90.

Network CGAR. Comprehensive molecular characterization of clear cell renal cell carcinoma. Nature 2013; 499: 43-49.

Moore LE, Jaeger E, Nickerson ML et al. Genomic copy number alterations in clear cell renal carcinoma:

associations with case characteristics and mechanisms of VHL gene inactivation. Oncogenesis 2012; 1: e14.

Arai E, Ushijima S, Tsuda H et al. Genetic clustering of clear cell renal cell carcinoma based on array- comparative genomic hybridization: its association with DNA methylation alteration and patient outcome. Clin Cancer Res 2008; 14: 5531-5539.

La Rochelle J, Klatte T, Dastane A et al. Chromosome 9p deletions identify an aggressive phenotype of clear cell renal cell carcinoma. Cancer 2010; 116: 4696-4702.

Forbes SA, Beare D, Boutselakis H et al. COSMIC: somatic cancer genetics at high-resolution. Nucleic Acids Res 2017; 45: D777-D783.

Duns G, Hofstra RM, Sietzema JG et al. Targeted exome sequencing in clear cell renal cell carcinoma tumors suggests aberrant chromatin regulation as a crucial step in ccRCC development. Hum Mutat 2012; 33: 1059-1062.

1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

(18)

537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika

17

Pena-Llopis S, Vega-Rubin-de-Celis S, Liao A et al. BAP1 loss defines a new class of renal cell carcinoma.

Nat Genet 2012; 44: 751-759.

Varela I, Tarpey P, Raine K et al. Exome sequencing identifies frequent mutation of the SWI/SNF complex gene PBRM1 in renal carcinoma. Nature 2011; 469: 539-542.

Dalgliesh GL, Furge K, Greenman C et al. Systematic sequencing of renal carcinoma reveals inactivation of histone modifying genes. Nature 2010; 463: 360-363.

Sato Y, Yoshizato T, Shiraishi Y et al. Integrated molecular analysis of clear-cell renal cell carcinoma. Nat Genet 2013; 45: 860-867.

Manley BJ, Zabor EC, Casuscelli J et al. Integration of Recurrent Somatic Mutations with Clinical Outcomes:

A Pooled Analysis of 1049 Patients with Clear Cell Renal Cell Carcinoma. Eur Urol Focus 2017; 3: 421-427.

Koshkin VS, Rini BI. Emerging therapeutics in refractory renal cell carcinoma. Expert Opin Pharmacother 2016; 17: 1225-1232.

Linehan WM, Srinivasan R, Schmidt LS. The genetic basis of kidney cancer: a metabolic disease. Nat Rev Urol 2010; 7: 277-285.

Nordstrom-O’Brien M, van der Luijt RB, van Rooijen E et al. Genetic analysis of von Hippel- Lindau disease.

Hum Mutat 2010; 31: 521-537.

Zbar B, Kishida T, Chen F et al. Germline mutations in the Von Hippel-Lindau disease (VHL) gene in families from North America, Europe, and Japan. Hum Mutat 1996; 8: 348-357.

Maher ER, Yates JR, Harries R et al. Clinical features and natural history of von Hippel-Lindau disease. Q J Med 1990; 77: 1151-1163.

Brugarolas J. Molecular genetics of clear-cell renal cell carcinoma. J Clin Oncol 2014; 32: 1968- 1976.

Caldas C. Cancer sequencing unravels clonal evolution. Nat Biotechnol 2012; 30: 408-410.

Neelakantan D, Drasin DJ, Ford HL. Intratumoral heterogeneity: Clonal cooperation in epithelial- to-mesenchymal transition and metastasis. Cell Adh Migr 2015; 9: 265-276.

Huang Y, Gao S, Wu S et al. Multilayered molecular profiling supported the monoclonal origin of metastatic renal cell carcinoma. Int J Cancer 2014; 135: 78-87.

Roadmap Epigenomics C, Kundaje A, Meuleman W et al. Integrative analysis of 111 reference human epigenomes. Nature 2015; 518: 317-330.

Navin NE, Hicks J. Tracing the tumor lineage. Mol Oncol 2010; 4: 267-283.

Sauter G, Moch H, Gasser TC et al. Heterogeneity of chromosome 17 and erbB-2 gene copy number in primary and metastatic bladder cancer. Cytometry 1995; 21: 40-46.

Nguyen HN, Sevin BU, Averette HE et al. Evidence of tumor heterogeneity in cervical cancers and lymph node metastases as determined by flow cytometry. Cancer 1993; 71: 2543-2550.

Gerlinger M, Horswell S, Larkin J et al. Genomic architecture and evolution of clear cell renal cell carcinomas defined by multiregion sequencing. Nat Genet 2014; 46: 225-233.

Gerlinger M, Rowan AJ, Horswell S et al. Intratumor heterogeneity and branched evolution revealed by multiregion sequencing. N Engl J Med 2012; 366: 883-892.

Navin N, Kendall J, Troge J et al. Tumour evolution inferred by single-cell sequencing. Nature 2011; 472:

90-94.

Wei L, Wang J, Lampert E et al. Intratumoral and Intertumoral Genomic Heterogeneity of Multifocal Localized Prostate Cancer Impacts Molecular Classifications and Genomic Prognosticators. Eur Urol 2016.

Kim TM, Jung SH, An CH et al. Subclonal Genomic Architectures of Primary and Metastatic Colorectal Cancer Based on Intratumoral Genetic Heterogeneity. Clin Cancer Res 2015; 21: 4461-4472.

de Bruin EC, McGranahan N, Mitter R et al. Spatial and temporal diversity in genomic instability processes defines lung cancer evolution. Science 2014; 346: 251-256.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41

42

43

44

(19)

537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika

18

Xue R, Li R, Guo H et al. Variable Intra-Tumor Genomic Heterogeneity of Multiple Lesions in Patients With Hepatocellular Carcinoma. Gastroenterology 2016; 150: 998-1008.

Singh RR, Murugan P, Patel LR et al. Intratumoral morphologic and molecular heterogeneity of rhabdoid renal cell carcinoma: challenges for personalized therapy. Mod Pathol 2015; 28: 1225-1235.

Patel AP, Tirosh I, Trombetta JJ et al. Single-cell RNA-seq highlights intratumoral heterogeneity in primary glioblastoma. Science 2014; 344: 1396-1401.

Wang Y, Waters J, Leung ML et al. Clonal evolution in breast cancer revealed by single nucleus genome sequencing. Nature 2014; 512: 155-160.

Yu C, Yu J, Yao X et al. Discovery of biclonal origin and a novel oncogene SLC12A5 in colon cancer by single-cell sequencing. Cell Res 2014; 24: 701-712.

Hou Y, Song L, Zhu P et al. Single-cell exome sequencing and monoclonal evolution of a JAK2- negative myeloproliferative neoplasm. Cell 2012; 148: 873-885.

Li Y, Xu X, Song L et al. Single-cell sequencing analysis characterizes common and cell-lineage- specific mutations in a muscle-invasive bladder cancer. Gigascience 2012; 1: 12.

Xu X, Hou Y, Yin X et al. Single-cell exome sequencing reveals single-nucleotide mutation characteristics of a kidney tumor. Cell 2012; 148: 886-895.

Hakimi AA, Mano R, Ciriello G et al. Impact of recurrent copy number alterations and cancer gene mutations on the predictive accuracy of prognostic models in clear cell renal cell carcinoma. J Urol 2014;

192: 24-29.

Scelo G, Riazalhosseini Y, Greger L et al. Variation in genomic landscape of clear cell renal cell carcinoma across Europe. Nat Commun 2014; 5: 5135.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

(20)

537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika

(21)

537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika

(22)

537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika

21 Evaluation of the mutational profile as a prognostic factor in apopulation-based study of clear cell renal cell carcinoma

Paranita Ferronika, Jeroen A A van de Pol, Helga Westers, Martijn M Terpstra, Kim de Lange, Rolf H Sijmons, Leo J Schouten, Klaas Kok

*Both authors contributed equally as first authors

Manuscript in preparation

2

(23)

537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika

22 Abstract Background

Profiling prognosis-related mutations in clear cell renal cell cancer (ccRCC), an often lethal disorder, might help to better stratify patients for treatment options in the future and thereby improve their clinical outcome. However, data on the survival effects associated with specific combinations of genes remains scarce. Therefore, we aimed to investigate the influence of the seven most frequently mutated genes in ccRCC, both as single factors and in combination, on cancer-specific survival (CSS).

Methods

DNA was isolated from formalin-fixed paraffin-embedded tumour blocks for all 366 ccRCC cases identified in the prospective Netherlands Cohort Study on diet and cancer. For 252 cases, DNA quantity and quality was sufficient to perform sequencing using a targeted next generation sequencing panel. One hundred and ten of these cases had complete clinical study data and a sequencing coverage of >20x for at least six of the seven genes studied. In these 110 cases, we tested the association of each mutated gene with CSS. Cox proportional hazards multivariate models were used to estimate hazard ratios (HR), after adjusting for the a priori-selected confounders: age, sex, tumour grade, tumour size, tumour stage and presence of mutated genes other than the one for which CSS was calculated. A two-sided p-value <0.05 was considered statistically significant. Multiple testing correction with the Benjamini-Hochberg method was performed where appropriate.

Results

Mutations in one or more of the seven genes were found in 64 of 110 cases (58%). Combined VHL and PBRM1 mutations were present in 11 (10%) cases, with other combinations being far less frequent. We observed a statistically significant effect on CSS for combined VHL and PBRM1 mutations (HR (95% CI) 0.14 (0.03-0.72), p-value = 0.019). In the analysis of mutated genes as single factors, PBRM1 was associated with a better CSS (HR (95% CI) = 0.29, (0.10-0.81), p-value = 0.018), while the finding for VHL was only borderline significant (HR (95% CI) = 0.47, (0.22-1.00), p-value = 0.051). Overall, we found an increased survival in cases with a mutation compared to cases without a mutation (HR (95% CI) = 0.33, (0.15-0.73), p-value = 0.006). However, all findings for single gene survival analysis became non-significant after correction for multiple testing.

Conclusion

Our study demonstrates a statistically significant association of combined PBRM1 and VHL mutations with a more favourable ccRCC-specific survival. This might be related to the role of PBRM1 and VHL mutations as early events in ccRCC instead of as driver genes for tumour progression.

Keywords

clear cell renal cell carcinoma, next generation sequencing, survival, VHL, PBRM1.

(24)

537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika

23 1. Introduction

Kidney cancer is an often lethal disease with a mortality rate of 3 per 100.000 people in Western Europe (available from: https://gco.iarc.fr/today, accessed [20-07-2019]) [1]. Renal cell cancer accounts for approximately 90% of all kidney cancer, and clear cell renal cell carcinoma (ccRCC) is its most common subtype (75%) [2, 3]. Clinical characteristics such as tumour size, tumour stage, tumour grade, and necrosis have been established as prognostic factors for ccRCC [4].

However, the mutational profile could also potentially be a prognostic factor. The generation of mutational profiles of cancer, including of ccRCC, allows for the exploration of this possibility.

High-throughput sequencing databases such as the Catalogue of Somatic Mutations in Cancer (COSMIC) and The Cancer Genome Atlas (TCGA) project already provide a clear overview of the frequently mutated genes in ccRCC [5, 6]. Based on the 512 ccRCC cases in the PanCancer study of TCGA, the top-five mutated genes in ccRCC are: VHL (42%), PBRM1 (38%), SETD2 (12%), BAP1 (9.5%), and MTOR (8%) [6]. Mutations in BAP1, SETD2, KDM5C, and TP53 [7-9] have all been associated with an unfavourable prognosis in ccRCC, although only BAP1 showed significant association across several studies (Supplementary Table S4). Association with a better overall survival has also been reported for mutations in PBRM1 [10].

Although studies have looked at the association between single mutations in ccRCC and prognosis, multiple mutations are typically present in the tumours of single patients and studies relating combinations of mutations to survival are still scarce. Further complicating matters, the largest previous studies used data from different countries with different health care systems and from hospital (sometimes specialised centre)-based series, which suggests that outcomes might be different in more homogenously collected series. Therefore, in the present study, we used a national Dutch-population- based prospective cohort study. We looked not only at the most frequently mutated genes in ccRCC – VHL, PBRM1, SETD2, BAP1, MTOR, KDM5C, and TP53 – as single factors, but also aimed to evaluate the influence of combinations of these genes on cancer-specific survival (CSS). Profiling the prognosis-related mutations in ccRCC might help to better stratify patients for medical management options in the future, and thereby improve clinical outcome.

2. Materials and methods 2.1. Study Population

Our study population was derived from the Netherlands Cohort Study on diet and cancer

(NLCS), a prospective cohort study initiated in 1986 [11]. As described in detail elsewhere, the

NLCS included 120,852 participants aged 55-69 years at baseline [12]. The entire cohort was then

followed for cancer incidence through record linkage with the Netherlands Cancer Registry and

PALGA, the Dutch pathology registry. The completeness of cancer follow-up through record

linkage is estimated to be at least 96% [13].

(25)

537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika

24 2.2. Sample Collection

In total, 608 renal cell carcinoma cases were identified during the 20.3 years of follow-up from 1986 until 2006 (Figure 1) [11]. Among these cases, 568 histologically confirmed renal cell carcinoma cases from 51 pathology laboratories were selected using information from PALGA for the collection of formalin-fixed paraffin embedded (FFPE) tissue blocks. FFPE blocks were successfully retrieved from 454 cases. Tumour type histology was revised by two experienced uropathologists using the WHO-classification of tumours [3]. Of the 454 cases, 366 were of the ccRCC type [14].

DNA was isolated from collected paraffin tissue blocks in two series. Series 1 was composed of

samples from patients diagnosed between 1986 and 1997, who had a follow-up time of up to

11.3 years, using DNA samples collected in 2003 [15]. Series 2 was composed of samples from

patients diagnosed between 1997 and 2006, who had a follow-up time of up to 22.3 years, using

DNA samples collected in 2012 [11]. Follow-up was continued until 31-December- 2009. For

our study, 252 cases were selected from both series based on the availability of sufficient good

quality DNA (fragments >200 bp). The tumour-cell fraction in these samples was estimated by a

uropathologist by visual inspection of H&E-stained tissue sections. In Series 1, this fraction varied

between 20% and 100% (median 95%). For Series 2, the tumour cell fraction was estimated to be

100%, as all tumour blocks were subjected to macrodissection in order to enrich for tumour cells

before DNA analysis. The investigations involving human samples were conducted according to

the Declaration of Helsinki [16], and the study was approved by the Medical Ethical Committee of

the Maastricht University and the University Hospital Maastricht.

(26)

537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika

25 Figure 1. Overview of sample selection of clear cell renal cell carcinoma cases from the Netherlands

cohort study on diet and cancer (NCLS), in which DNA collection was performed in two series, Series 1

(year 2003) and Series 2 (year 2012).

(27)

537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika

26 2.3. Targeted sequencing

All DNA samples were subjected to a targeted sequencing approach based on the Single Primer Enrichment Technology (Ovation

^TM

Custom Target Enrichment System, NuGEN, San Carlos, CA, USA). For this purpose we designed a custom landing probe panel for enrichment of the consensus coding regions of the 32 most frequently mutated genes in ccRCC, including VHL and MTOR [6, 17], supplemented with 10 genes associated with the VHL/HIF pathway and the (PI3K)-AKT-MTOR pathway in ccRCC [18]. The complete list of genes is provided in chapter 3 of this thesis.

The DNA samples from 252 out of 366 cases were subjected to our sequencing protocol. Aliquots of 500 ng DNA were sheared into 500-bp fragments by Adaptive Focused Acoustics

^TM

- (Covaris, Woburn, MA, USA), and subjected to targeted sequencing using a custom panel kit (Ovation

^TM

Custom Target Enrichment System, NuGEN, San Carlos, CA, USA). The library preparation steps were carried out according to the protocol of the manufacturer. Enriched libraries were sequenced with the Illumina HISEQ 2500

^TM

(Illumina, San Diego, CA, USA) using single-end next generation sequencing with 100 bp reads.

2.4. Sequence data analysis and somatic mutation identification

The sequencing data were processed according to the Genome Analysis Toolkit (GATK) best practice recommendations using a pipeline in which HaplotypeCaller from GATK and FreeBayes were used as variant caller [19, 20]. Called variants were annotated and filtered to identify true somatic mutations, as described previously [21]. For the mutation analysis, 121 cases were included that had a coverage per exon of >20 reads for at least six of the seven selected genes, and an average coverage of >10 reads for the remaining gene. Next, we determined the somatic variant with the highest mutant read frequency (MRF) in each patient. Variants with an MRF ≥50%

of the highest MRF seen for that sample were considered major variants likely to be present in the majority of the tumour cells. To eliminate false positive mutations due to sequencing errors, we excluded the major clone variants that were present in more than four samples. If necessary, we reassigned the somatic variant with the highest MRF and redefined the major clone variants. Only variants with a major clonal mutation and ≥4 alternate reads in at least one of the 121 samples were included in the subsequent analysis. The integrative Genomic Viewer was used to confirm the authenticity of any doubtful somatic mutations [22]

2.5. Clinical characteristic assessment

Tumour size was assessed based on the largest diameter and was categorized into two tier

groups: diameter ≤70 mm and diameter >70 mm [7]. Lymph node involvement was defined by

the presence of metastasis tumour in at least one lymph node. The morphological features of

ccRCC and the tumour grade were assessed by two pathologists [11] using the Fuhrman tumour

grade system [23, 24]. Tumour stage was determined using the Union International Contre le

Cancer (UICC) TNM staging system [25]. Information on the cause of death, RCC-related (ICD-O-

3C64) or other, was obtained from Statistics Netherlands (CBS).

(28)

537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika

27 2.6. Statistical Analysis

Statistical analysis was performed using Stata statistical software: release 15 (StataCorp., 2015, College Station, TX). The association of each mutated gene with CSS was tested. The survival time was measured as the time of first diagnosis to the time of death in years. Cox proportional hazards models were used to estimate hazard ratios (HR) and 95% confidence intervals (CI) [26]. Analyses were done for an age- and sex-adjusted model (first model) and a multivariable-adjusted model including the following a priori- selected confounders: age, sex, tumour grade, tumour size, and tumour stage (second model). Mutated genes were added to the multivariable-adjusted model (second model) and run as the third model. A two- sided p-value <0.05 was considered statistically significant. Multiple testing correction with the Benjamini- Hochberg method [27]

was performed for the third model. False-discovery-adjusted p-values, so called q- values, were considered statistically significant if q <0.05. Sensitivity analyses were done by truncating the follow-up time to 10 and 5 years. The proportional hazards assumption was tested using the scaled Schoenfeld residuals and log-log curves [28]. Kaplan-Meier curves and Wilcoxon tests were used to evaluate the CSS of ccRCC cases with and without each mutated gene. For all cases included in the Cox regression analysis, we also made an inventory of the co-occurrence of mutations involving multiple genes in order to select the most frequent combinations to test for their effect on CSS.

The associations between mutated genes and tumour grade, tumour size, pathologic T stage, lymph node involvement, metastasis, and UICC stage were examined by univariable analysis using Chi-square testing. All tests were done two-sided, and a p-value <0.05 was considered statistically significant. Multiple testing correction with the Benjamini-Hochberg [27] method was done per gene, and a q-value <0.05 was considered statistically significant.

3. Results

3.1. Targeted sequencing

Of all 266 ccRCC samples with sufficient DNA quantity and quality, 252 samples were successfully sequenced using our gene panel. The average coverage per gene varied strongly, from 13 to 57 reads. Our targeted NGS panel covered the consensus coding sequence of 42 genes, and the complete mutational profiles will be published elsewhere. For the current study, in which we set out to analyse the association of mutated genes with clinical characteristics of the patients, we focussed on the set of six genes with a mutation frequency ≥5% based on the PanCancer TCGA database: VHL, PBRM1, SETD2, BAP1, MTOR, and KDM5C (Network, 2013). We also included TP53 because its mutation frequency, as reported in the COSMIC database (https://cancer.sanger.ac.uk/

cosmic/browse/tissue, accessed 23-06-2014), is relatively high (8%) [17] . Out of our complete

panel, 121 cases had an average read depth of at least 20 for six out of the seven genes and a

read depth of at least 10 for the seventh gene. We applied this criterion to minimize the chance of

false-negative results, and thus continued our statistical analyses with this set of 121 cases.

(29)

537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika

28 Overall, 72 of the 121 (59.5%) cases had mutations in at least one of the seven target genes: 45 (37.2%) in VHL, 27 (22.3%) in PBRM1, 14 (11.6%) in SETD2, 12 (9.9%) in KDM5C, 7 (5.8%) in BAP1, 3 (2.5%) in MTOR, and 3 (2.5%) in TP53. The characteristics of the cases and the mutation frequency of the genes included in the analysis are described in Table 1.

Table 1. Characteristics of the ccRCC cases; the Netherlands Cohort Study on diet and cancer, 1986–

2006

Characteristic Cohort, n = 121

Age at diagnosis, years, median (range) 71 (57-88)

Survival rate, months, median (range) 68 (0-245)

Sex, n (%)

Male 74 (61.2)

Female 47 (38.8)

Tumour diameter, mm, median (range) 60 (10-180)

Tumour diameter, two tier

≤ 70 mm, n/N (%) 77 (63.6)

> 70 mm, n/N (%) 36 (29.8)

NA 8 (6.6)

Tumour laterality

right 60 (49.6)

left 59 (48.8)

bilateral 2 (1.7)

Primary tumour grade, n/N (%)

1 16 (13.2)

2 50 (41.3)

3 34 (28.1)

4 21 (17.4)

Pathologic T stage, n/N (%)

T1 7 (5.8)

T2 69 (57.0)

T3 41 (33.9)

T4 1 (0.8)

NA 1 (0.8)

Pathologic node stage, n (%)

N0 85 (70.2)

N1/N2 8 (6.6)

Nx 26 (21.5)

NA 2 (1.7)

Metastatic stage, n (%)

M0 80 (66.1)

M1 15 (12.4)

Mx 24 (19.8)

NA 2 (1.7)

IUCC pathologic stage, n (%)

I 7 (5.8)

II 62 (51.2)

III 34 (28.1)

IV 16 (13.2)

NA 2 (1.7)

(30)

537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika

29

Genes

MTOR, mutated (%) 3 (2.5)

WT, n (%) 118 (97.5)

VHL, mutated (%) 45 (37.2)

WT, n (%) 76 (62.8)

SETD2, mutated (%) 14 (11.6)

WT, n (%) 107 (88.4)

PBRM1, mutated (%) 27 (22.3)

WT, n (%) 94 (77.7)

BAP1, mutated (%) 7 (5.8)

WT, n (%) 114 (94.2)

TP53, mutated (%) 3 (2.5)

WT, n (%) 118 (97.5)

KDM5C, mutated (%) 12 (9.9)

WT, n (%) 109 (90.1)

3.2. Survival analysis

After exclusion of cases with missing clinical information for the a priori confounders, 110 cases remained for survival analysis. Of these, no mutation was identified in 46 cases. In age- and sex-adjusted Cox regression models (model 1, Table 2), we observed an increased CSS, i.e.

ccRCC-related, for patients with PBRM1 or VHL mutations as compared to patients without these mutations, although this was not statistically significant (HR (95% CI) = 0.40 (0.16-1.03), p-value

= 0.059 and HR (95% CI) = 0.71 (0.35-1.43), p-value = 0.333, respectively). In the multivariable- adjusted results (model 2; Table 2), which included the clinical characteristics of the patients (tumour grade, tumour size, and tumour stage), the CSS increased compared to model 1. PBRM1 was associated with a statistically significant better CSS, while the findings for VHL were borderline significant (HR (95% CI) = 0.29, (0.10-0.81), p-value = 0.018 and HR (95% CI) = 0.47 (0.22-1.00), p-value = 0.051, respectively). After mutual adjustments for other genes included in the analyses, excluding MTOR due to the instability of the estimations (model 3), the effect of PBRM1 on CSS became slightly stronger (HR (95% CI) = 0.22 (0.07-0.70), p-value 0.011), while the association of VHL with CSS attenuated slightly (HR (95% CI) = 0.54 (0.23-1.23), p-value = 0.139). The statistically significant effect of the PBRM1 mutation on CSS became non-significant after multiple testing correction using the Benjamini-Hochberg method (adjusted p-value = 0.066). The association of mutated PBRM1 or mutated VHL to CSS was visualized by the Kaplan-Meier curves (Figure 2A).

Cases with PBRM1 mutations had a higher survival rate than the cases without the mutation,

although the association had only a borderline statistical significance (p-value = 0.055, Wilcoxon

test). The mutations in VHL also seemed beneficial to CSS, although this was not statistically

significant (p-value = 0.155, Wilcoxon test). The survival curves levelled out and converged after

a survival time of approximately 15 years (Figure 2B). The associations between other converged

genes and CSS should be interpreted with caution considering the low number of ccRCC-related

deaths in our study population. Overall, we saw that mutations in most genes (with the exception

of KDM5C) are related to an increased survival, although most of the findings were not statistically

significant and based on a low number of cases (Table 2).

(31)

537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika 537042-L-sub01-bw-Ferronika

30

ab c Model 1 Model 2 Model 3 Mutation Total no. of cases

No. of ccRCC related deaths

Survival time, years HR(95% CI)

p- vaHR(95% CI) p-HR(95% CI) lue value

p- vaq- lue value BAP1 No104397351 Ref. 1 Ref. 1 Ref. Yes6 1 570.35(0.05- 2.57)0.3040.31(0.04- 2.36)0.2570.18(0.02- 1.51)0.1140.342

KD M5C No100377291 Ref. 1 Ref. 1 Ref. Yes103 621.08(0.31- 3.83)0.9011.36(0.37- 4.98)0.6441.15(0.31- 4.24)0.8310.831

M TOR No108407791 Ref. 1 Ref. 1 Ref. Yes2 0 13N/A- N/A- N/A-

PBR

M1 No84355701 Ref. 1 Ref. 1 Ref. Yes265 2220.40(0.16- 1.03)0.0590.29(0.10- 0.81)0.0180.22(0.07- 0.70)0.0110.066

SE TD2 No98377201 Ref. 1 Ref. 1 Ref. Yes123 720.96(0.29- 3.17)0.9400.57(0.14- 2.42)0.4480.69(0.16- 2.93)0.6170.740

TP 53 No107387851 Ref. 1 Ref. 1 Ref. Yes3 2 6 5.55(1.25- 24.62)0.0270.60(0.11- 3.41)0.5670.38(0.06- 2.37)0.3020.453 VHL

No70295011 Ref. 1 Ref. 1 Ref. Yes40112900.71(0.35- 1.43)0.3330.47(0.22- 1.00)0.0510.54(0.23- 1.23)0.1390.278