Genomic Analysis of Hepatocellular Carcinoma With Active Hepatitis - - PowerPoint PPT Presentation

SLIDE 1

13th ILCA Annual Conference

20 ► 22 September 2019 │Chicago, USA

13th Annual Conference

20 ► 22 September 2019 Chicago, USA

Genomic Analysis of Hepatocellular Carcinoma With Active Hepatitis B Virus Replication

Huat Chye Lim, MD, and John Gordan, MD, PhD Divisions of Hospital Medicine and Hematology/Oncology University of California, San Francisco

SLIDE 2

13th ILCA Annual Conference

20 ► 22 September 2019 │Chicago, USA

Background: HBV and HBV-related HCC

• Small enveloped DNA virus with partially double stranded genome
• Persists in nuclei of infected hepatocytes via episomal cccDNA
• Rarely integrates into host genome
• HBV-related HCC known to be associated with clinical and genomic differences

Introduction

Synopsis HCC with active HBV replication represents a molecularly distinct subset of HCC associated with differences in mutations, gene expression and survival. Objectives for this talk 1. Which genes are differentially mutated in HCC with active HBV replication? 2. Which genes are differentially expressed in HCC with active HBV replication? 3. How does HBV replication status affect survival in HCC?

SLIDE 3

13th ILCA Annual Conference

20 ► 22 September 2019 │Chicago, USA

• We used GATK PathSeq software, which

performs sequential computational subtraction, to measure HBV RNA in HCC tumors

• Tumor RNA-Seq data were obtained from

two databases:

• TCGA – n = 371
• ICGC – n = 68 (from LIRI-JP project)
• We classified tumors as HBV RNA+ if more

than 1 HBV RNA read was detected per million human reads

• We investigated association between HBV

RNA status and nonsynonymous somatic mutations, gene set expression and survival

Methods

All tumor RNA-Seq reads Quality filtering and duplicate removal Low quality and duplicate reads Filtered reads Low quality and duplicate reads Human reads Non- human reads Subtraction of human reads Low quality and duplicate reads Human reads Mapping to known microbes Other microbe reads HBV reads

Reference: Walker MA, et al. Bioinformatics. 2018;34(24):4287-4289.

SLIDE 4

13th ILCA Annual Conference

20 ► 22 September 2019 │Chicago, USA

HBV RNA+ HBV RNA- p Total patients 124 315 N/A Cohort TCGA 100 81% 271 86% N/A ICGC (LIRI-JP) 24 19% 44 14% Gender Male 100 81% 201 64% 0.0006 (*) Female 24 19% 114 36% Age at diagnosis Mean ± SD 54.2 ± 12.0 62.9 ± 13.1 < 0.0001 (**) Risk factors HBV clinical history 87 70% 44 14% < 0.0001 (*) HCV clinical history 2 2% 79 25% < 0.0001 (*) Alcohol consumption 45 36% 106 34% NS Edmondson grade at diagnosis Grade I 12 10% 47 15% 0.0003 (*) Grade II 50 40% 161 51% Grade III 51 41% 83 26% Grade IV 8 6% 4 1% Unknown 3 2% 20 6% Pathologic stage at diagnosis Stage I 54 44% 129 41% NS Stage II 28 23% 84 27% Stage III 38 31% 73 23% Stage IV 3 2% 10 3% Unknown 1 1% 19 6% Vascular invasion Present 37 30% 105 33% NS Absent 65 52% 176 56% Unknown 22 18% 34 11%

• p values from χ2 (*) and Mann-

Whitney (**) tests

• Stage determined using either

AJCC (for TCGA) or LCSGJ (for ICGC LIRI-JP) criteria

• HBV RNA+ status was

associated with:

• Male gender
• Younger age
• Higher grade
• HBV history
• No HCV history
• There was no

association with stage, vascular invasion or alcohol consumption

Results: HBV RNA Status is Associated With Differences in Clinical Characteristics

SLIDE 5

13th ILCA Annual Conference

20 ► 22 September 2019 │Chicago, USA

Results: HBV RNA Status is Associated With Differential Gene Mutation Rates

• Figure shows the 94 genes where

nonsynonymous mutation rate depended significantly on HBV RNA status

• Most (82/94) were preferentially

mutated in HBV RNA+ tumors

• Bolded genes are among the 30

significantly-mutated genes identified in 2017 TCGA HCC Cell paper

• TP53 and CDKN2A were more

frequently mutated in HBV RNA+

• BAP1 was more frequently

mutated in HBV RNA-

• TP53 had substantially higher

mutation rates than other genes

• Circle size is inversely proportional to log(p)
• Bolded genes are TCGA top 30 SMGs

MED12 DNAH6 CNOT2 CHD5 CDKN2A NRXN1 CCNL2, FOXG1, KAT6A, NBEA NBEA SPTBN2 CC2D2A, EGFLAM, FBXO42, NCOA2, OR5D14 TP53 BAP1 SVEP1 SSPO BRD7 SYNE2 ZNF208 CSDE1, EXO1 MAPK9 and 36 other genes BPTF CFAP47 CDH4 and 18 other genes TDRD5 FBN1 BRCA2, CHD9, DOPEY2, TNXB, TRIP12 COL4A5, ENGASE, ZNF135

• 0,01

0,00 0,01 0,02 0,03 0,04 0,05 0,06 0,07 0,08

• 0,01

0,01 0,03 0,05 0,07 0,09 0,11 HBV RNA- Mutation Rate HBV RNA+ Mutation Rate 0.35 Preferentially mutated in HBV RNA+ tumors Preferentially mutated in HBV RNA- tumors

Reference: Cancer Genome Atlas Research Network. Cell. 2017;169(7):1327-1341.e23.

SLIDE 6

13th ILCA Annual Conference

20 ► 22 September 2019 │Chicago, USA

Results: HBV RNA+ Status is Associated With Differential Gene Set Expression

• Figure shows GSEA enrichment map of Gene

Ontology gene sets with FDR < 10% in TCGA dataset (n = 531 of 4,464)

• All such gene sets were enriched only in HBV

RNA+ tumors

• Enriched gene sets included:
• Multiple DNA damage repair pathways
• Genes upregulated by MYC and mTORC1
• Genes upregulated in “proliferative” HCC:

▪ Boyault subclass G1-G3 HCC ▪ Hoshida subclass S2 HCC ▪ Lee subclass A HCC ▪ Chiang “proliferation” subclass HCC

• We then evaluated association between HBV

RNA status and several measures of genomic instability

RNA processing and splicing Methylation and chromatin modification Cell cycle regulation Mitotic spindle regulation DNA damage repair Chromatin modification DNA replication Translation RNA processing Nuclear transport Nucleases Transcription Enrichment in HBV RNA+ 0.0001 FDR 0.1 Circle size is proportional to gene set size

Genes upregulated in FDR q Boyault subclass G1-G3 HCC 0.003 Hoshida subclass S2 HCC 0.011 Lee subclass A HCC 0.012 Chiang “proliferation” subclass HCC 0.015

References: Boyault S, et al. Hepatology. 2007;45(1):42-52. Hoshida Y, et al. Cancer Res. 2009;69(18):7385-92. Lee JS, et al. Hepatology. 2004;40(3):667-76. Chiang DY, et al. Cancer Res. 2008;68(16):6779-88.

SLIDE 7

13th ILCA Annual Conference

20 ► 22 September 2019 │Chicago, USA

• HRD score for TCGA dataset was

calculated as the sum of three independent HRD measures:

• Large-scale state transitions
• Loss of heterozygosity
• Telomeric allelic imbalance
• HBV RNA+ status was associated

with increased HRD score (22.19 for RNA+ vs. 15.97 for RNA-, p = 1e-6)

• There was no association with tumor

mutational burden (TMB)

Results: HBV RNA+ Status is Associated With Increased Homologous Recombination Deficiency (HRD) Score

• Error bars show mean ± SD
• p value from Mann-Whitney test

HBV RNA+ HBV RNA- 20 40 60

HRD score

p = 1e-6

Reference: Knijnenburg TA, et al. Cell Rep. 2018;23(1):239-254.e6.

SLIDE 8

13th ILCA Annual Conference

20 ► 22 September 2019 │Chicago, USA Reference: Cancer Genome Atlas Research Network. Cell. 2017;169(7):1327-1341.e23.

HBV RNA+ HBV RNA- All Gene Cox Coefficient p Cox Coefficient p Cox Coefficient p TP53 0.0420 0.9148 0.5393 0.0284 0.4334 0.0356 CTNNB1

• 0.1966

0.6505 0.1321 0.6242 0.0087 0.9680 ALB

• 0.4302

0.5601

• 0.4065

0.2834

• 0.3447

0.3030 AXIN1 0.5324 0.3328 0.3711 0.3730 0.3568 0.2738 BAP1

• 15.3500

0.9979

• 1.9399

0.0081

• 1.7160

0.0183 KEAP1 1.4484 0.0135

• 0.4677

0.5147 0.1392 0.7420 NFE2L2

• 16.2300

0.9980 0.3430 0.4209 0.2735 0.5197 LZTR1

• 17.5800

0.9977 0.0118 0.9908

• 0.6234

0.5365 RB1 1.4337 0.0428 0.3752 0.4280 0.4903 0.1877 PIK3CA

• 17.2800

0.9965 0.1847 0.7583

• 0.2901

0.6274 RPS6KA3

• 16.4300

0.9976 0.0003 0.9996

• 0.1785

0.7270 AZIN1 0.5137 0.6240

• 13.8100

0.9958 0.9450 0.3518 KRAS

• 15.3500

0.9979

• 0.0901

0.9288

• 0.4257

0.6728 IL6ST

• 15.7900

0.9971

• 1.0137

0.1645

• 1.0452

0.1492 RP1L1

• 15.6800

0.9970 0.5156 0.3868 0.4072 0.4892 CDKN2A

• 1.0873

0.3009

• 16.0700

0.9956

• 1.0655

0.2930 EEF1A1 NA NA 0.8659 0.0447 0.9484 0.0276 ARID2 1.2225 0.0462 0.2595 0.5757 0.6644 0.0464 ARID1A 0.5161 0.4150 1.0912 0.0008 0.9356 0.0011 GPATCH4 1.7634 0.0952

• 15.4800

0.9948

• 0.0130

0.9900 ACVR2A

• 0.3718

0.7198 0.4871 0.3541 0.3774 0.4150 APOB

• 1.5385

0.1443 0.6915 0.0164 0.1962 0.4749 CREB3L3 NA NA 5.5943 0.0001 6.0263 0.0000 NRAS

• 15.5300

0.9978

• 14.9700

0.9959

• 14.9500

0.9950 AHCTF1

• 14.6900

0.9971

• 0.0716

0.8900

• 0.0266

0.9587 HIST1H1C

• 16.0800

0.9978

• 0.4665

0.6444

• 0.9352

0.3529

• We used Cox multivariate

regression to evaluate effect of mutation and HBV RNA status

• n overall survival
• Covariates: HBV RNA status,

mutation status, age, sex, grade, stage, cohort

• Table shows Cox coefficients for

the 30 SMGs identified in 2017 TCGA HCC Cell paper

• Mutations associated with

survival difference in HBV RNA-:

• Increased survival: BAP1
• Decreased survival: TP53,

EEF1A1, ARID1A, APOB

• Mutations associated with

survival difference in HBV RNA+:

• Decreased survival: KEAP1,

RB1, ARID2

Results: HBV RNA Status is Associated With Survival Differences

1000 2000 3000 4000 50 100 Time (days) Percent survival BAP1+ BAP1-

BAP1 mutations were associated with increased survival in HBV RNA- patients

p = 0.008

1000 2000 3000 50 100

Time (days) Percent survival KEAP1+ KEAP1-

KEAP1 mutations were associated with decreased survival in HBV RNA+ patients

p = 0.01

BAP1 KEAP1

SLIDE 9

13th ILCA Annual Conference

20 ► 22 September 2019 │Chicago, USA

Summary

1. Which genes are differentially mutated in HCC with active HBV replication?

• These included TP53 and CDKN2A in HBV RNA+ tumors and BAP1 in HBV RNA- tumors
• Most differentially mutated genes were preferentially mutated in HBV RNA+ tumors
• TP53 mutation rates were substantially higher than other differentially mutated genes
• All Gene Ontology gene sets with FDR < 10% were enriched only in HBV RNA+ tumors
• These included multiple DNA damage repair gene sets and multiple gene sets upregulated in

“proliferative” HCC

• HBV RNA+ status was associated with increased homologous recombination deficiency score
• BAP1 mutations were associated with increased survival in HBV RNA- patients
• KEAP1 mutations were associated with decreased survival in HBV RNA+ patients

2. Which genes are differentially expressed in HCC with active HBV replication? 3. How does HBV replication status affect survival in HCC?

SLIDE 10

13th ILCA Annual Conference

20 ► 22 September 2019 │Chicago, USA

Conclusions and Future Directions

Acknowledgements

• HCC with active HBV replication represents a molecularly distinct subset of HCC

associated with differences in mutations, gene expression and survival

• HBV replication is a factor that could impact HCC prognosis and response to targeted therapies
• HBV status is a factor that should be more rigorously defined and stratified in HCC clinical trials
• Limitations of our study include that TCGA and ICGC comprise mainly early stage/resectable

tumors and contain incomplete HBV serologic data

• Questions for future investigation include:
• Is this molecularly distinct subset of HCC amenable to targeted therapy?
• Is there a role for direct measurement of tumor HBV replication in HCC?
• How concordant is tumor HBV RNA status with serum HBV markers (e.g., serologies and viral load)?
• Mentors: John Gordan and Katie Kelley
• For valuable feedback: Gary Chan, Mary Feng, Christina Hwang, David Quigley, Neil Mehta,

Dom Mitchell and Rigney Turnham