Table of Contents Table of Contents
Previous Page  643 656 Next Page
Information
Show Menu
Previous Page 643 656 Next Page
Page Background

We evaluated mutation signatures

[16]

to identify the

prominent causes of mutation in UTUC. Three out of

21 signatures were the most prominent (Supplementary

Fig. 2) including the C

>

G transversions and C

>

T

transitions associated with APOBEC enzymes (#4 and

#9), and the C

>

T transitions at CpG islands (#6). APOBEC

signatures were the most prominent, seen in 45% of samples

(14/31).

To characterize the mutational landscape of UTUC we

selected genes that were frequently mutated (normalized

by gene length) plus genes previously identified to be

significantly mutated

[11] ( Fig. 1 )

. One hypermutated

sample was excluded for the purpose of determining the

top mutated genes.

FGFR3

was the most commonly mutated

gene (74% mutation rate in all tumors) and although it was

seen in a higher proportion of low-grade tumors (92%

mutation rate in low-grade tumors) the frequency was not

significantly different to that in high-grade tumors (Fisher’s

exact test,

p

= 0.091). Notably, 60% of high-grade tumors in

this study had

FGFR3

mutations, compared to 35.4% in the

prior report

[11]

. Of the 20

FGFR3

mutations, 15 were S249C,

two were Y373C, and there were single mutations at R248C

and S371C; one patient harbored a

FGFR3-TACC

fusion.

Other frequently altered genes were similar to previous

reports

( Fig. 1

and Supplementary Fig. 3). Novel mutations

were found in

NPHS1

(11%) and

RHOB

(11%).

3.3.

Deregulated p53 signaling and genomic instability of high-

grade UTUC tumors

Comparison of mutational profiles of low- and high-grade

tumors revealed a significantly higher incidence of muta-

tions in components of p53 signaling

( Fig. 2

A) and greater

genomic instability (Supplementary Figs. 4 and 5). Out of

599 gene sets from Biocarta, KEGG, and the National Cancer

Institute pathway interaction database, three were signifi-

cantly altered following correction for multiple testing, all

related to p53 signaling (Supplementary Table 5). Gene sets

were considered altered in a sample if a mutation (copy

number, single-nucleotide variant, or indel mutation)

occurred in one or more genes in the gene set. Interestingly,

10/15 high-grade samples had a mutation in

TP53

,

ATM

, or

ATR

, while only one out of 12 low-grade samples exhibited a

mutation in the same genes

( Fig. 2 A

). Enrichment of

mutations in other genes interacting with p53 suggests

deregulated cell-cycle and apoptotic pathways in the high-

grade tumors (Supplementary Fig. 6).

Consistent with a previous report

[11]

, high-grade UTUC

tumors tended to have greater genomic instability

( Fig. 2 B

).

A deletion region at 9p21.3

( Fig. 2

B and Supplementary Fig.

4) in primarily high-grade samples includes loss of genes

encoding the cell-cycle regulators CDKN2A and CDKN2B

found in this region.

Protein-level differences (measured by RPPA) between

low- and high-grade samples indicated altered p53 signal-

ing, apoptosis, and insulin/growth factor signaling in the

high-grade samples

( Fig. 2

C). High-grade samples had

higher TFRC

[22]

(negatively regulated by TP53) and pro-

apoptotic signals (active caspase-3 and PUMA), and lower

phospho-NDRG1

[23]

(upregulated by TP53).

3.4.

RNA expression subtypes

We used unsupervised consensus clustering of RNAseq

expression data to segregate samples into four molecular

subtypes that were significantly correlated with important

Table 1 – Clinical characteristics of the bi-institution upper tract urothelial carcinoma cohort

Total

BCM

MDACC

p

value

a

Patients (

n

)

31

9

22

Median age at diagnosis, yr (IQR)

74 (68–80)

74 (70–80)

74 (66–80)

0.6

y

Female,

n

(%)

10 (32)

2 (22)

8 (36)

0.7

Race,

n

(%)

1

Black

2 (6.5)

0 (0)

2 (9.1)

Hispanic

2 (6.5)

0 (0)

2 (9.1)

White

27 (87)

9 (100)

18 (82)

Tobacco use,

n

(%)

0.1

Current

4 (13)

1 (11)

3 (14)

Former

16 (52)

7 (78)

9 (41)

None

11 (36)

1 (11)

10 (46)

Pathologic T stage,

n

(%)

0.4

Ta/1

21 (68)

5 (56)

16 (73)

T2/3/4

10 (32)

4 (44)

6 (27)

High pathologic grade,

n

(%)

17 (55)

5 (56)

12 (54)

1

Recurrence,

n

(%)

Any

16 (52)

3 (33)

13 (59)

0.3

Local

4

2

2

Bladder

12

2

10

Distant

4

1

3

Median FU for surviving patients, mo (IQR)

b

20 (14–31)

42 (31–45)

18 (13–22)

NA

Median overall survival, mo (IQR)

18 (14–34)

25 (5.3–43)

18 (14–30)

NA

Deceased (

n

)

13

4

9

NA

BCM = Baylor College of Medicine; MDACC = MD Anderson Cancer Center; IQR = interquartile range; FU = follow-up; NA = not applicable.

a

Significance test: Fisher’s exact test or Kruskal-Wallis rank sum test

y

.

b

FU range for surviving patients, 3–66 mo.

E U R O P E A N U R O L O G Y 7 2 ( 2 0 1 7 ) 6 4 1 – 6 4 9

643