Oncotarget, May, Vol.4, No 5
www.impactjournals.com/oncotarget/
P53 Mutations in Advanced Cancers: Clinical Characteristics, Outcomes, and Correlation between Progression-Free Survival and Bevacizumab-Containing Therapy Rabih Said1,*, David S. Hong1,*, Carla L. Warneke2, J. Jack Lee2, Jennifer J. Wheler1, Filip Janku1, Aung Naing1, Gerald S. Falchook1, Siqing Fu1, Sarina Piha-Paul1, Apostolia M. Tsimberidou1, Razelle Kurzrock3 1
Department of Investigational Cancer Therapeutics (Phase I Clinical Trials Program), The University of Texas MD Anderson Cancer Center, Houston, TX 2
Department of Biostatistics, The University of Texas MD Anderson Cancer Center, Houston, TX
3
UCSD Moores Cancer Center
*
Denotes equal contribution
Correspondence to: David Hong, email:
[email protected] Keywords: P53 mutations, PTEN loss, bevacizumab, Li-Fraumeni syndrome Received: April 10, 2013
Accepted: May 1, 2013
Published: May 1, 2013
This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
ABSTRACT: Background: Mutations in the p53 gene are amongst the most frequent aberrations seen in human cancer. Our objective was to characterize the clinical characteristics associated with p53 mutation in patients with advanced cancer. Methods: We retrospectively reviewed and analyzed the clinical features and response to standard systemic therapy of 145 patients with documented tumor p53 mutational status (mutant-type [mtp53] vs. wild-type [wtp53]) referred to the Clinical Center for Targeted Therapy. Results: Sixty-six (45.5%) patients had mtp53. Mutations in p53 occurred more frequently in older patients (p= 0.015) and in Caucasians (p=0.024). The incidence of liver metastases was 69.2% vs. 43%, p=0.002 in mtp53 and wtp53, respectively. PTEN loss by immunohistochemistry was found more frequently in mtp53-bearing tumors compared to wtp53 (33.3% vs. 10%, p=0.007). The best progression-free survival (PFS) on standard systemic therapy was significantly longer with bevacizumab-containing regimens as compared to non-bevacizumab containing regimen in patients with mtp53 (median 11.0 [95% CI 5.9-16.0], n=22 vs. 4.0 months [95% CI 3.6-5.7], n=35, p2 years Median number of phase 1 1 (0 – 4) therapies (range) Median number of prior cancer 3 (0 – 12) therapies Median number of metastases 3 (0 – 9) (range)** Characteristics
mtp53, N=66 (%) 56.1 (22.2-72.6) 29 (43.9%)
wtp53, N=79 * P-value (%) 51.0 (14.5-75.3) 0.015 16 (20.3%) 0.004
34 (47.2) 32 (43.8)
38 (52.8) 41 (56.2)
0.740
58 (50.4) 8 (26.7)
57 (49.6) 22 (73.3)
0.024
1 (33.3) 20 (64.5) 3 (60.0) 5 (41.7) 6 (75.0) 3 (23.1) 7 (53.9) 3 (100.0) 0 (0.0) 8 (34.8) 5 (55.6) 3 (33.3) 2 (15.4)
2 (66.7) 11 (35.5) 2 (40.0) 7 (58.3) 2 (25.0) 10 (76.9) 6 (46.1) 0 (0.00) 3 (100.0) 15 (65.2) 4 (44.4) 6 (66.7) 11 (84.6)
8 (12.1) 45 (69.2) 49(75.4) 20 (30.8) 22 (33.9) 51 (78.5) 10 (15.4) 24 (36.9) 7 (10.8) 9 (28.1) 4 (6.2)
9 (11.4) 34 (43.0) 54(68.4) 14(17.7) 29 (36.7) 59 (74.7) 20 (25.3) 24 (30.4) 8 (10.1) 10 (25.0) 12 (15.2)
1.000 0.002 0.458 0.078 0.730 0.694 0.156 0.478 1.000 0.794 0.112
7.6 (0 – 361.7)
9.9 (0 – 235.2)
0.536
15 (22.7)
17 (21.5)
1.000
1 (0 – 4)
1 (0 – 4)
1.000
3.5 (0 – 10)
3 ( (0 – 12)
0.080
4 (0 – 9)
3 ( 0 – 8)
0.080
*P-values are from Fisher’s exact test or Kruskal-Wallis test, as appropriate. ** Patients may have multiple sites of metastasis and some patients may have unavailable data on their sites of metastasis. The P-value in each row is computed for testing the association of each metastasis site and p53 mutation.
the cell cycle, controls DNA repair mechanisms and activates apoptotic pathways (5). In addition, p53 protein plays a role in regulating angiogenesis at least in part www.impactjournals.com/oncotarget
through direct binding to the hypoxia-induced factor-α (HIF-α) subunit, leading to HIF-α destruction (6). Under physiologic conditions, intracellular levels of p53 protein 706
Oncotarget 2013; 4: 705-714
Table 2: p53 mutational status and co-existing molecular aberrations mtp53, N=66
wtp53, N=79
* P-value
Molecular Aberrations
No. of Patients No. of Patients No. of patients No. of Patients Tested with aberration(%) Tested with aberration(%)
PTEN loss
51
17 (33.3)
50
5 (10.0)
0.007
BRAF mutation
44
1 (2.3)
56
3 (5.4)
0.629
PIK3CA mutation
56
3 (5.4)
66
9 (13.6)
0.221
KRAS mutation
48
7 (14.6)
56
9 (16.1)
1.000
EGFR mutation
40
1 (2.5)
46
0
0.465
c-KIT mutation
36
0
46
0
NA
PTEN mutation
19
3 (15.8)
23
0
0.084
NRAS mutation
29
4 (13.8)
33
1 (3.0)
0.176
GNAQ mutation 20
0
25
2 (8.0)
0.495
MET mutation
1 (4.4)
25
0
0.479
23
*P-values are from Fisher’s exact test
RESULTS
are maintained at low levels by a complex network of proteins that include murine double minute 2 (Mdm2) (7). Targeting aberrant p53 has proven challenging. Mdm2 inhibitor molecules have demonstrated preclinical promise in a wide variety of tumors with wild-type p53 (810). Importantly, MK-1775, a potent and selective small molecule inhibitor of Wee-1 kinase (a tyrosine kinase that phosphorylates and inactivates CDC2 and is involved in G2 checkpoint signaling) selectively sensitizes tumors to DNA damaging agents, probably because p53 is a key regulator in the G1 checkpoint and p53-deficient tumors rely only on the G2 checkpoint after DNA damage (11). It is currently being studied in phase II trials of ovarian cancer with mtp53 (NCT01164995 and NCT01357161) (http://Clinicaltrials.gov). Overall, however, there is a paucity of molecules targeting p53 mutations, and, because these mutations are found in over 50% of cancer, identifying ways to counteract them is important. The clinical correlates of p53 mutations malignancies have not been fully delineated. Preclinical data have shown that p53 mutation accelerates cancer progression and increases tumor invasiveness and metastasis, which is not, however, the entire picture (12, 13). Here, we studied p53 mutational status in patients with advanced malignancies referred to the Phase I Clinical Trials Program in The University of Texas MD Anderson Cancer Center. Our objectives were to identify clinical, prognostic and predictive characteristics associated with p53 mutational status in advanced solid tumors.
www.impactjournals.com/oncotarget
Patient characteristics The clinical and demographic characteristics of our patient population are summarized in Table 1. Overall, starting in May 2010, the p53 mutational status of tumors was identified in 145 patients. The median age of the patients at diagnosis was 53.3 years (range, 14.5 to 75.3 years). Women comprised 50.3% (n=73) of the study population. Most patients were Caucasian (n = 115, 79.3%) and the remaining 30 patients were African American (n=15, 10.3%), non-white Hispanic (n= 11, 7.6%), or Asian (n= 4, 2.8%). The type of primary cancer varied among patients. The most common histologic subtypes were colorectal carcinoma (n=31, 21.4%), sarcoma (n=23, 15.9%), ovarian carcinoma (n=13, 9.0%) and melanoma (n=13, 9.0%). The histologic subtypes of patients reflected the pattern of referrals to the Phase I Clinic. The median time from diagnosis to metastasis/ recurrence was 8.9 months (range 0 – 361.7, months). Forty-eight (33.1%) patients had metastatic disease at diagnosis. The most common metastatic sites were the lymphatic system (n=110, 76.4%), lungs (n=103, 71.5%) and liver (n=79, 54.9%). The number of metastatic sites, as reported by the last available imaging studies, ranged from 0 to 9 (median 3). The total number of prior standard systemic therapies before referral to the Phase I Clinical Trials Program ranged from 0 to 9 (median 3) and the total 707
Oncotarget 2013; 4: 705-714
number of phase I clinical trials that patients enrolled on ranged from 0 to 4 (median 1).
tumors was reviewed in patients’ electronic medical records. Three of 66 (4.5%) patients with mtp53 met the criteria for Li-Fraumeni-like syndrome (14, 15). Among those three patients, two had sarcoma (ages at diagnosis 27 and 44) with multiple family members (first, 2nd and 3rd degree) having breast cancer, brain tumors, lung cancer and other malignancies. One patient with ovarian cancer (negative BRCA1/2 mutation) diagnosed at age 47 had multiple family members with breast cancer at a young age (42 and 48 years old), leukemia (20 years old), brain tumor and other malignancies.
P53 mutational status and clinical features Of 145 patients, 66 (45.5%) had tumors harboring the p53 mutation. The presence of a p53 mutation was significantly associated with patient age at diagnosis. Cancer diagnosed at an older age was associated with a greater number of p53 mutations than was cancer diagnosed at a younger age. At diagnosis, the median age of patients with p53 mutated tumors was 56.1 years (22.2 -72.6 years) versus 51.0 years (14.5 - 75.3 years) for patients with wtp53 tumors (p =0.0145). In addition, each year’s increase in age at diagnosis was associated with a 4% increase in the odds of having a p53 mutation (OR = 1.04, 95% CI 1.01 to 1.07, p =0.019). The percentage of p53 mutations varied with tumor type, as expected (Table 1). Of note, all three patients with pancreatic cancer were found to have mtp53 tumors, and all three patients with renal cell carcinoma were found to have wtp53. Mutant p53 tumors metastasized more frequently to the liver than did wtp53 bearing tumors (69.2% vs. 43.0%) (p = 0.002). In addition, mtp53 tumors trended toward retroperitoneal metastasis compared to wtp53 tumors (30.8 % vs. 17.7 %, p=0.078). There was no statistically significant difference in the percentage of metastases between mtp53 and wtp53 tumors to the lung, brain, lymph nodes or to any other metastatic sites. Univariate analysis showed that mtp53 tumors occurred more frequently in Caucasian patients (50.4%) compared to non-Caucasian patients (26.7%; p=0.024). However, there was no correlation between p53 mutational status and gender, number of metastatic sites and time from diagnosis to metastasis. The family history of cancer in patients with mtp53
Types of p53 mutations and the co-existing molecular aberrations Five (7.6%) patients were found to have two molecular abnormalities of the p53 gene in two different exons and two (3.0%) patients were found to have two molecular abnormalities of the p53 gene in the same exons. Insertion or deletion (indels) was seen in 7 (10.9%) patients. The most common p53 mutations were seen in exon 5 (n=23, 33.3%), followed by exon 6 (n=14, 20.3%), exon 7 (n=12, 17.4%), exon 4 (n=9, 13.0%), exon 8 (n=9, 13.0%) and exon 9 (n=2, 3.0%). The number of other molecular aberrations tested ranged from 0 to 3 (median 1). Insufficient tissue availability for testing was the main reason for the inability to test for other aberrations. However, some data on PI3KCA, KRAS, BRAF, c-KIT and EGFR mutations, and PTEN status assessed by IHC, was available for more than 50% of patients’ specimens. The number and proportions of all molecular aberrations and their relationships to p53 mutational status are summarized in Table 2. PTEN loss by IHC, but not the other aberrations, was statistically correlated with mtp53 (33%, 17/51 mtp53 vs. 10%, 5/50
Figure 1a: Kaplan Meier curve showing PFS on best standard systemic treatment in patients with mtp53 comparing bevacizumab (n = 22) vs. non-bevacizumab containing regimens (n = 35). Figure 1b: Kaplan Meier curve showing PFS on best standard systemic treatment in wtp53 comparing bevacizumab (n = 7) vs. non-bevacizumab containing regimens (n = 48). www.impactjournals.com/oncotarget
708
Oncotarget 2013; 4: 705-714
Table 3: Multivariate Cox proportional hazards regression models predicting duration of longest PFS Clinical Feature Age* Race (Non-Caucasian vs. Caucasian) Tumor histology group§(Recommended vs. not recommended) Bevacizumabcontaining regimen (Yes vs. no)
mtp53
wtp53
Hazard Ratio 95% CI
P-value Hazard Ratio
95% CI
P-value
0.96
0.94-0.99
0.010
0.99
0.97-1.02
0.651
0.74
0.34-1.62
0.456
1.54
0.84-2.82
0.167
0.56
0.29-1.09
0.086
0.84
0.38-1.84
0.657
0.21
0.09-0.454
