Intratumoral expression profiling of genes involved in ... - Nature

The Pharmacogenomics Journal (2013) 13, 410 -- 416 & 2013 Macmillan Publishers Limited All rights reserved 1470-269X/13 www.nature.com/tpj

ORIGINAL ARTICLE

Intratumoral expression profiling of genes involved in angiogenesis in colorectal cancer patients treated with chemotherapy plus the VEGFR inhibitor PTK787/ZK 222584 (vatalanib) PM Wilson1, D Yang2, M Azuma3, MM Shi4, KD Danenberg5, D Lebwohl4, A Sherrod1, RD Ladner1, W Zhang3, PV Danenberg6, T Trarbach7, G Folprecht8, G Meinhardt9 and H-J Lenz2 The phase III CONFIRM clinical trials demonstrated that metastatic colorectal cancer patients with elevated serum lactate dehydrogenase (LDH) had improved outcome when the vascular endothelial growth factor receptor (VEGFR) inhibitor PTK/ZK (Vatalanib) was added to FOLFOX4 chemotherapy. We investigated the hypothesis that high intratumoral expression of genes regulated by hypoxia-inducible factor-1 alpha (HIF1a), namely LDHA, glucose transporter-1 (GLUT-1), VEGFA, VEGFR1, and VEGFR2, were predictive of outcome in CONFIRM-1. Tumor tissue was isolated by laser-capture microdissection from 85 CONFIRM-1 tumor specimens; FOLFOX4/placebo n ¼ 42, FOLFOX4/PTK/ZK n ¼ 43. Gene expression was analyzed using quantitative RT-PCR. In univariate analyses, elevated mRNA expression of LDHA, GLUT-1, and VEGFR1 were associated with response to FOLFOX4/PTK/ZK. In univariate and multivariate analyses, elevated LDHA and VEGFR1 mRNA levels were associated with improved progression-free survival in FOLFOX4/PTK/ZK patients. Furthermore, increased HIF1a and VEGFR2 mRNA levels were associated with decreased survival in FOLFOX/placebo patients but not in patients who received FOLFOX4/PTK/ZK. These are the first data suggesting intratumoral mRNA expression of genes involved in angiogenesis/HIF pathway may predict outcome to VEGFR-inhibitors. Biomarkers that assist in directing VEGFR-inhibitors toward patients with an increased likelihood of benefit will improve the cost-effectiveness of these promising agents. The Pharmacogenomics Journal (2013) 13, 410--416; doi:10.1038/tpj.2012.23; published online 5 June 2012 Keywords: angiogenesis; biomarker; colorectal (cancer); hypoxia; vatalanib; VEGFR

INTRODUCTION Vascular endothelial growth factor (VEGF) and its receptors, VEGFR1 and VEFGR2, are intimately involved in the regulation of tumor-associated angiogenesis and have been correlated with poor prognosis in various human malignancies.1,2 Their functions include increasing microvascular permeability, stimulating endothelial cell growth, and promoting angiogenesis and lymphangiogenesis.3--5 As the inhibition of VEGF-induced angiogenesis is highly selective for tumor-associated vessels, therapeutic approaches targeting angiogenesis mediated by VEGF and it’s kinase receptor are expected to be safe and well tolerated in patients with cancer. The anti-VEGF monoclonal antibody bevacizumab has demonstrated improved clinical outcome and a favorable toxicity profile in patients with a variety of human carcinomas.6--8 Hypoxia is a common feature of many solid tumors and has been linked with treatment resistance.9 Hypoxia-inducible factor-1 alpha (HIF1a) is a key transcription factor that activates an

aggressive network of genes involved in glycolytic energy metabolism, angiogenesis, cell survival, and erythropoiesis. Genes regulated by HIF1a include those encoding VEGF and it’s type-1 receptor (VEGFR1), glucose transporter-1 (GLUT-1), and several glycolytic enzymes, such as lactate dehydrogenase 5 (LDHA).10--12 Recent data have suggested that high serum LDH is linked to poor prognosis in non-small cell lung and colorectal cancer, as well as tumor hypoxia, angiogenesis, and over-expression of both VEGF and LDHA.9--15 A direct association between LDH5 up-regulation and HIF1a and HIF2a accumulation in colorectal cancer cells has been reported13 and a subsequent study showed that patients with high serum LDH had significantly increased levels of intratumoral gene expression of VEGF and VEGFR1.16 The accumulating evidence supports the hypothesis that elevated serum LDH is directly associated with up-regulation of genes involved in the HIF1a and tumor-associated angiogenesis pathway.

1 Department of Pathology, University of Southern California/Norris Comprehensive Cancer Center, Keck School of Medicine, Los Angeles, CA, USA; 2Department of Preventive Medicine, University of Southern California/Norris Comprehensive Cancer Center, Keck School of Medicine, Los Angeles, CA, USA; 3Division of Medical Oncology, University of Southern California/Norris Comprehensive Cancer Center, Keck School of Medicine, Los Angeles, CA, USA; 4Novartis Oncology, East Hanover, NJ, USA; 5Response Genetics, Inc., Los Angeles, CA, USA; 6Department of Molecular Biology and Biochemistry, University of Southern California/Norris Comprehensive Cancer Center, Keck School of Medicine, Los Angeles, CA, USA; 7University of Essen, West German Cancer Center, Essen, Germany; 8University Hospital Dresden, Dresden, Germany; 9Global Clinical Development Oncology Bayer Healthcare Pharmaceuticals, Montville, NJ, USA. Correspondence: H-J Lenz, Division of Medical Oncology, University of Southern California/Norris Comprehensive Cancer Center, Keck School of Medicine, 1441 Eastlake Avenue, Suite 3456, Los Angeles, CA 90033, USA. E-mail: [email protected] Previous Presentation: Portions of these results were presented at the ASCO 2008 Annual Meeting in the Clinical Science Symposium, Biomarkers in Colorectal Cancer Management (Abstract 4002). Received 13 January 2012; revised 6 April 2012; accepted 11 April 2012; published online 5 June 2012

Prognostic factors in patients with mCRC treated with PTK787 PM Wilson et al

411

PTK787/ZK 222584 (PTK/ZK, vatalanib) is a novel oral angiogenesis tyrosine kinase inhibitor (TKI) that is active against all known VEGFR tyrosine kinases and platelet-derived growth factor receptor (PDGFR) tyrosine kinases and therefore offers a novel approach to inhibiting tumor growth facilitated by angiogenesis.17--20 Two randomized, double-blind, placebo-controlled, phase III studies were carried out in metastatic colorectal cancer (mCRC) patients who were receiving first-line (CONFIRM-1) or second-line (CONFIRM-2) chemotherapy with folinic acid (leucovorin), 5-fluorouracil, oxaliplatin (FOLFOX4), and either PTK/ZK or placebo.21,22 Although the primary efficacy objectives of both trials were not met, subgroup analysis on the final data from CONFIRM-1 and CONFIRM-2 demonstrated with statistical significance and striking similarity, that patients with elevated serum LDH levels (more than 1.5 times the upper limit of normal) derived a greater clinical benefit when PTK/ZK was added to a standard FOLFOX4 regimen, compared with FOLFOX4 plus placebo.21,22 The hypothesis proposed herein is that serum LDH may serve as a surrogate predictive marker of an activated tumoral HIF1a and angiogenic gene expression profile with the potential to predict response to VEGFR-inhibitor in combination with chemotherapy. This hypothesis is further supported by a recent study that analyzed LDH5 (the most active LDH isozyme encoded by the LDHA gene) by immunohistochemistry in a balanced subset of patients from the CONFIRM trials and found that tumors with high LDH5 expression had improved response to the FOLFOX4 plus PTK/ZK combination compared with tumors with low LDH5 expression.23 Understanding the molecular mechanisms that govern resistance to VEGFR TKIs will provide critical insight that will assist in the future development of this class of agents and will be paramount in improving cost-effectiveness through the successful selection of patients who may show optimum benefit from VEGFR TKI therapies. Therefore, to investigate the role of genes involved in angiogenesis and HIF-signaling in predicting the efficacy of PTK/ZK in mCRC, we tested the hypothesis that patients with high serum LDH would have increased intratumoral gene expression of VEGFR1, VEGFR2, and additional genes involved in the HIF1a pathway and that such patients would benefit from the addition of PTK/ZK to chemotherapy.

were available for analysis. A pathologist reviewed FFPE tumor blocks for quality and tumor content. Ten sections, each of 1-mm thickness, were obtained from the areas identified with the highest tumor concentration and mounted on uncoated glass slides. Three sections representative of the beginning, middle, and end of the tumor were taken for histological diagnosis and stained with hematoxylin and eosin using the standard method. Before micro-dissection, the sections were deparaffinized in xylene for 10 min, hydrated with 100, 95, and 70% ethanol, and then washed in water for 30 s. Following this, the sections were stained with nuclear fast red (American Master Tech Scientific, Lodi, CA, USA) for 20 s and rinsed in water for 30 s. Samples were dehydrated with 70% ethanol, 95% ethanol, and 100% ethanol for 30 s each, followed by xylene for 10 min, and then the slides were air dried. Laser capture micro-dissection (PALM Microlaser Technologies AG, Munich, Germany) was performed on all tumor samples to ensure that only tumor cells were dissected.25 The dissected areas of tissue were then transferred to a reaction tube containing 400 ml of RNA lysis buffer.

PATIENTS AND METHODS

Statistical analysis

The phase III multinational CONFIRM-1 trial was designed to investigate the first-line therapeutic potential of the VEGFR TKI PTK/ZK in combination with oxaliplatin-based chemotherapy in patients with mCRC. In the CONFIRM-1 trial, 1168 patients were randomized to receive FOLFOX4 with either PTK/ZK or placebo as first line therapy. Progression-free survival (PFS) was the primary endpoint with secondary endpoints of overall safety and overall response rate.21 Of note, in both the CONFIRM trials, patients were stratified according to performance status (PS) and serum LDH levels. PTK/ZK improved the PFS of patients with high baseline serum LDH levels in CONFIRM 1: Hazard ratio for PFS ¼ 0.67; 95% confidence interval (CI): 0.49--0.91; P ¼ 0.009.21 For comparative purposes, similar results were observed for patients with elevated serum LDH in CONFIRM-2, which evaluated the same therapy as second line: Hazard ratio for PFS ¼ 0.63; 95% CI: 0.48--0.83; Po0.001.22 Where applicable, this study was conducted and is reported according to Reporting Recommendations for Tumor Marker Prognostic Studies (REMARK) criteria.24 All patients on the CONFIRM 1 study provided informed consent for the analysis of molecular correlates. All analyses herein were approved by the University of Southern California Institutional Review Board.

Tumor response in CONFIRM-1 was assessed according to modified RECIST criteria.21 Non-responders were defined as patients with stable or progressive disease. PFS was calculated as the period from the first day of randomization until the first observation of disease progression or death from any cause. If a patient had not progressed or died, PFS was censored at the time of the last follow-up. Overall survival (OS) was calculated as the time from the first day of randomization until death from any cause or until the date of the last follow-up. Gene expression values are expressed as ratios between two absolute measurements: the gene of interest versus the internal reference gene (bactin). To assess the associations between the expression level of each gene and tumor response, PFS or OS, the expression level was categorized into a low and a high value at optimal cutoffs. The maximal w2 method of Miller and Siegmund26 and Halpern27 was used to determine which gene expression value (optimal cut-offs) best segregated patients into poorprognosis and good-prognosis subgroups, in terms of likelihood of response. The cutoff values chosen in analyzing response were applied for analysis of PFS and OS. The analysis was conducted separately by therapy. The Cox proportional hazards regression model was used to evaluate the independent effects of gene expression levels on PFS and OS, when adjusting baseline PS and serum LDH level. Interactions between treatment and expression values were tested by comparing corresponding likelihood ratio statistics between the baseline and nested models that included the multiplicative product terms. No adjustment for multiple comparisons was performed because this study was exploratory in nature

Laser-capture micro-dissections A total of 85 formalin-fixed paraffin-embedded (FFPE) primary colorectal adenocarcinoma samples from patients enrolled in the CONFIRM-1 trial & 2013 Macmillan Publishers Limited

Quantitative reverse transcription polymerase chain reaction Isolation of RNA from FFPE tumor samples was performed according to a proprietary procedure defined by Response Genetics (Los Angeles, CA, USA; United States patent number 6 248 535). Briefly, FFPE tumor samples were deparaffinized followed by homogenization using mechanical and sonic homogenization in RNA lysis buffer. The homogenized samples were heated, RNA recovered by phenol--chloroform extraction and integrity confirmed by spectrophotometry. cDNA was synthesized and gene expression was analyzed using a quantitative real-time reverse transcription PCR method as described previously.16 Relative mRNA levels were expressed as ratios between the target gene and an internal reference gene (b-actin). Quantification of LDHA, GLUT-1, HIF1a, VEGF, VEGFR1, VEGFR2, and b-actin was performed using a TaqMan-based real-time detection method on board an ABI PRISM 7900 Sequence detection System (Applied Biosystems, Foster City, CA, USA). The PCR reaction mixture consisted of the following: 1200 nmol l1 of each primer, 200 nmol l1 of probe, 0.4 U of AmpliTaq Gold Polymerase, 200 nmol l1 each of dATP, dCTP, dGTP, dTTP, 3.5 mmol l1 MgCl2, and 1 TaqMan Buffer A containing a reference dye, added to a final volume of 20 ml (all reagents from PE Applied Biosystems, Foster City, CA, USA). Cycling conditions were 50 1C for 2 min, 95 1C for 10 min, followed by 46 cycles of 95 1C for 15 s, then 60 1C for 1 min. The primer sequences and details of PCR conditions have been published previously.16

The Pharmacogenomics Journal (2013), 410 -- 416


412 and conducted to generate hypotheses for future studies. All reported P-values were two-sided. All analyses were performed using the SAS statistical package version 9.0 (SAS Institute, Cary, NC, USA).

RESULTS Intratumoral gene expression levels of LDHA, GLUT-1, HIF1a, VEGF, VEGFR1, and VEGFR2 A total of 85 FFPE tumor samples from patients enrolled in CONFIRM-1 were available for analysis; 43 who received FOLFOX4 plus PTK/ZK and 42 who received FOLFOX4 plus placebo. There were no statistically significant differences in clinical and demographic characteristics, including serum LDH, sex, PS, age, and gene expression levels and treatment received (Table 1). Baseline characteristics and clinical outcome in patients from whom FFPE tissue was recovered for this study (response rate, 62%; PFS, 9 months; OS, 25 months) were similar to the entire CONFIRM-1 patient population21 and historical outcomes for firstline therapy in similar patient populations.28 Table 1. Baseline characteristics and intratumoral gene expression by study and treatment

Gene expression levels by treatment and response In CONFIRM-1, elevated mRNA levels of LDHA (P ¼ 0.033), GLUT-1 (P ¼ 0.045), and VEGFR1 (P ¼ 0.012) were predictive of a response to treatment in patients who received FOLFOX4 plus PTK/ZK but not in patients treated with FOLFOX4 plus placebo. Elevated expression of VEGFR2 also demonstrated a statistical trend in favor of improved response to FOLFOX4 plus PTK/ZK (P ¼ 0.071). Furthermore, there was a significant interaction between treatments to predict PTK/ZK activity for GLUT-1 (P ¼ 0.046) and VEGFR1 (P ¼ 0.05) with VEGFR2 again demonstrating a trend toward predicting PTK/ZK treatment (P ¼ 0.074) (Table 2). Gene expression levels by treatment and PFS Increased mRNA levels of LDHA (P ¼ 0.004) and VEGFR1 (P ¼ 0.023) were significantly associated with improved PFS in patients treated with FOLFOX4 plus PTK/ZK in CONFIRM-1. (Table 3, Figure 1a and b). All associations remained significant in the Table 2.

Intratumoral gene expression levels by treatment and

response

CONFIRM-1 Patients who provided FFPE samples (N ¼ 85) FOLFOX4+PTK/ZK (n ¼ 43) Baseline characteristics Sex, no. (%) Male/female Median age, years (range) PS, no. (%)a 0 12 LDH, no. (%)b Low/high

FOLFOX4+placebo (n ¼ 42)

31 (72)/12 (28) 63 (28--79)

25 (60)/17 (40) 62 (48--82)

23 (53) 20 (47)

26 (62) 16 (38)

35 (81)/8 (19)

31 (74)/11 (26)

Intratumoral gene expression LDHA No. Median (range) 1.11 GLUT-1 No. Median (range) 2.11 HIF1a No. Median (range) 1.62 VEGF No. Median (range) 6.21 VEGFR1 No. Median (range) 6.28 VEGFR2 No. Median (range) 1.91

42 (0.07--2.54)

41 1.09 (0.14--3.63)

42 (0.26--23.86)

41 2.49 (0.36--38.59)

42 (0.59--3.68)

41 1.62 (0.31--4.98)

42 (3.05--21.90)

41 5.79 (1.29--15.90)

41 (0.49--23.25)

36 5.59 (0.56--28.06)

42 (0.02--8.16)

41 1.64 (0.62--11.63)

Abbreviations: FFPE, formalin fixed paraffin embedded (metastatic colorectal carcinoma sample); GLUT-1, glucose transporter-1; HIF1a, hypoxia-inducible factor-1 alpha; LDHA, lactate dehydrogenase A; VEGF (R1) (R2), vascular endothelial growth factor (type-1 receptor) (type-2 receptor). a PS, World Health Organization performance status: 0 ¼ fully active and able to perform without restriction; 1 ¼ restricted physically strenuous activity; and 2 ¼ capable of self-care but unable to work. b LDH, lactate dehydrogenase: low p1.5 times upper limit of normal (ULN); high 41.5 ULN. There was no significant difference in baseline characteristics by study and treatment: all P-values 40.05 (not shown).


PTK/ZK

No. LDHA p0.36 40.36 P-valuea Pinteraction GLUT-1 p1.50 41.50 P-valuea Pinteraction HIF1a p1.84 41.84 P-valuea Pinteraction VEGF p4.16 44.16 P-valuea Pinteraction VEGFR1 p3.78 43.78 P-valuea Pinteraction VEGFR2 p1.29 41.29 P-valuea Pinteraction

6 36

Placebo

Response yes/no (%)

No.

0/6 (0%) 22/13 (63%) 0.033

2 39

Response yes/no (%)

1/1 (50%) 27/12 (69%) 0.54

N/Ab

11 31

2/9 (18%) 20/10 (67%) 0.045

13 28 1.00

9/4 (69%) 19/9 (68%)

0.046

26 16

17/9 (65%) 5/10 (33%) 0.36

27 14

19/8 (70%) 9/5 (64%) 0.73

7 34

4/3 (57%) 24/10 (71%) 0.66

10 26

6/4 (60%) 17/9 (65%) 1.00

12 29

8/4 (67%) 20/9 (69%) 1.00

0.28

8 34

2/6 (25%) 20/13 (61%) 0.46 0.44

10 31

1/9 (10%) 20/10 (67%) 0.012 0.050

8 34

1/7 (12%) 21/12 (64%) 0.071 0.074

Abbreviations: GLUT-1, glucose transporter-1; HIF1a, hypoxia-inducible factor-1 alpha; HR, hazard ratio; LDHA, lactate dehydrogenase A; VEGF (R1) (R2), vascular endothelial growth factor (type-1 receptor) (type-2 receptor). a Based on the Fisher exact test following 2,000 bootstrap-like simulations to adjust for selection of optimal cut-off point for response to FOLFOX4+PTK/ZK. b P-value for LDHA--treatment interaction could not be calculated owing to the low numbers in the placebo group with low LDHA expression. Statistically significant P-values are highlighted in bold.

& 2013 Macmillan Publishers Limited


413 Table 3.

Intratumoral gene expression levels and PFS by treatment PTK/ZK

LDHA p0.36 40.36 P-valuea Pinteraction GLUT-1 p1.50 41.50 P-valuea Pinteraction HIF1a p1.84 41.84 P-valuea Pinteraction VEGF p4.16 44.16 P-valuea Pinteraction VEGFR1 p3.78 43.78 P-valuea Pinteraction VEGFR2 p1.29 41.29 P-valuea Pinteraction

Placebo

Median PFS

HR (95%CI)

Median PFS

HR (95%CI)

7.5 (1.9, 9.3+) 10.7 (7.6, 11.3)

1 (Reference) 0.29 (0.10, 0.87)

3.5 (3.5, 18.4) 9.2 (7.6, 12.7)

1 (Reference) 0.99 (0.23, 4.23)

0.004

9.3 (7.4, 11.3) 9.4 (7.6, 11.3)

0.99

N/Ab 1 (Reference) 0.76 (0.36, 1.60)

9.4 (7.3, 14.9) 9.2 (5.7, 13.1)

0.42

1 (Reference) 0.82 (0.39, 1.69) 0.54

0.91 9.4 (7.5, 11.3) 9.4 (7.4, 11.3)

1 (Reference) 1.03 (0.52, 2.03)

7.6 (7.3, 14.9) 9.3 (5 7, 9 4)

0.94

1 (Reference) 1.28 (0.61, 2.66) 0.46

0.53 7.5 (3.5, 9.4) 11.3 (7.6, 11.3)

1 (Reference) 0.61 (0.27, 1.39)

7.6 (7.6, 18.4) 9.2 (7.3, 9.4)

0.18

1 (Reference) 1.59 (0.60, 4.17) 0.29

0.091 7.5 (3.5, 9.3) 10.7 (7.6, 11.3)

1 (Reference) 0.45 (0.20, 1.02)

7.6 (5.7, 13.1) 7.6 (5.7, 9.4)

0.023

1 (Reference) 1.41 (0.62, 3.18) 0.33

0.024 9.3 (7.5, 11.3) 9.4 (7.6, 11.3)

1 (Reference) 0.72 (0.30, 1.69)

13.1 (7.6, 18.4) 7.6 (5.7, 9.4)

0.39

1 (Reference) 1.79 (0.77, 4.19) 0.12

0.10

Abbreviations: GLUT-1, glucose transporter-1; HR, hazard ratio; HIF1a, hypoxia-inducible factor-1 alpha; LDHA, lactate dehydrogenase A; PFS, progression-free survival; VEGF (R1) (R2), vascular endothelial growth factor (type-1 receptor) (type-2 receptor). a Based on log-rank test. b P-value for LDHA--treatment interaction could not be calculated owing to the low numbers in the placebo group with low LDHA expression. Statistically significant P-values are highlighted in bold.

multivariate Cox model when adjusting for baseline PS and LDH level (P ¼ 0.013 for LDHA and P ¼ 0.036 for VEGFR1; Table 4). Gene expression levels and OS by treatment No significant association between gene expression levels and OS was demonstrated in patients treated with FOLFOX4 plus PTK/ZK (Table 5). However, elevated HIF1a and VEGFR2 expression were significantly associated with decreased OS in patients receiving FOLFOX4 plus placebo in CONFIRM-1 (P ¼ 0.026 and P ¼ 0.003, respectively; Table 5). Patients who received FOLFOX4 plus PTK/ZK showed no difference in OS between high and low VEGFR2 and HIF1a expression. The associations between HIF1a and VEGFR2 expression and OS remained significant after adjusting for baseline PS and LDH level (P ¼ 0.008 for HIF1a and P ¼ 0.007 for VEGFR2; Table 4).

DISCUSSION The CONFIRM trials ended with the conclusion that elevated serum LDH was a negative prognostic factor but that these patients appeared to derive significant benefit from inclusion of VEGFR-targeted agent alongside chemotherapy.21,22 Hypoxia& 2013 Macmillan Publishers Limited

triggered or oncogene-activated HIF pathways are common events in human malignancies and are linked to aggressive tumor biology, resistance to chemotherapy, and poor prognosis.9,14,15 Activation of the VEGF/VEGFR axis triggers multiple signaling pathways that result in the promotion of endothelial cell survival, mitogenesis, migration, differentiation, changes to vascular permeability, and mobilization of endothelial progenitor cells.3--5 Over-expression of VEGF mRNA and protein is associated with tumor size, progression, and poor prognosis in a variety of malignancies. Despite the overall poor prognosis of these tumors, there is a significant rationale suggesting that these tumors may in fact show greater benefit from therapies that counter tumor angiogenesis driven by hypoxia and the VEGF pathway. The data presented herein supports this hypothesis and suggests that intratumoral expression profiling of genes involved in the HIF1a pathway may be predictive and prognostic in patients with mCRC treated with FOLFOX4 in combination with PTK/ZK. These data are consistent with our previous results16 and with the hypothesis that patients with increased serum LDH have significant up-regulation of VEGF and VEGFR1 gene expression in the tumor and may benefit from VEGFR TKI therapy. A recent study analyzed serum LDH and tumoral expression of LDH5, the The Pharmacogenomics Journal (2013), 410 -- 416


414 Table 4. Multivariable analysis of intratumoral gene expression levels and PFS or OS PTK/ZK, PFS HR (95%CI) LDHA p0.36 40.36 P-valuea

1 (Reference) 0.24 (0.08, 0.73) 0.013

VEGFR1 p3.78 43.78 P-valuea

1 (Reference) 0.39 (0.16, 0.94) 0.036 Placebo, OS HR (95%CI)

VEGFR1 ≤ 3.78 VEGFR1 > 3.78

HIF1a p1.84 41.84 P-valuea

1 (Reference) 3.47 (1.38, 8.73) 0.008

VEGFR2 p1.29 41.29 P-valuea

1 (Reference) 7.78 (1.77, 34.18) 0.007

Abbreviations: CI, confidence interval; HIF1a, hypoxia-inducible factor-1 alpha; HR, hazard ratio; LDHA, lactate dehydrogenase A; OS, overall survival; PFS, progression-free survival; VEGF (R1) (R2), vascular endothelial growth factor (type-1 receptor) (type-2 receptor). a Based on Wald test within the multivariable Cox proportional hazards regression model adjusting baseline performance status and LDH level.

Months since randomization to FOLFOX4 plus PTK/ZK

Figure 1. (a) Kaplan--Meier curves showing lactate dehydrogenase A (LDHA) gene expression levels and progression-free survival (PFS) in patients treated with FOLFOX4 plus PTK/ZK in CONFIRM-1. (b) Kaplan--Meier curves showing vascular endothelial growth factor type1 receptor (VEGFR1) gene expression levels and PFS in patients treated with FOLFOX4 plus PTK/ZK in CONFIRM-1.

LDH isozyme encoded by the LDHA gene, in a balanced subset of patients from the same CONFIRM trial populations. The authors reported that LDH5 was a significant negative prognostic marker only in patients who did not receive PTK/ZK and the addition of this angiogenesis inhibitor countered the adverse prognostic effect observed in patients with elevated LDH5.23 It is plausible that overexpression of LDH5 is a surrogate marker for the presence of aggressive HIF-activated tumor biology that includes increased angiogenesis mediated by VEGF and it’s receptors that is attenuated by PTK/ZK. We previously confirmed a direct link between elevated expression of HIF1a pathway genes, such as LDHA, VEGF, VEGFR1, VEGFR2, and GLUT-1 in patients with high serum LDH.16 In the present study, we expanded upon these initial observations to confirm a direct link between increased expression of these genes and response to PTK/ZK. In this study, univariate analysis of response rates to chemotherapy showed that high expression of LDHA, GLUT-1, and VEGFR1 were predictive of response following treatment with FOLFOX4 plus PTK/ZK but not in patients treated with FOLFOX4 plus placebo. In multivariate analysis, the association for LDHA and The Pharmacogenomics Journal (2013), 410 -- 416

VEGFR1 and PFS remained significant, suggesting that tumors with an HIF1a-induced pathway benefit the most from VEGFR-TKI when given in combination with chemotherapy. In terms of OS, there was no significant difference in survival benefit for patients with high vs low expression of VEGF, VEGFR1, or VEGFR2 when treated with FOLFOX4 plus PTK/ZK, whereas patients with high expression of VEGFR2 and HIF1a treated with FOLFOX4 plus placebo had a significantly shorter OS in multivariate analysis. This was a highly statistically significant interaction, suggesting a strong negative prognostic influence for these genes that was countered by the addition of PTK/ZK. Response (or disease stabilization) to an antiangiogenic inhibitor and/or chemotherapy is transient and is typically followed by a distinct change in tumor biology leading to the development of resistance and subsequent disease progression in the majority of patients. The importance of the genes that regulate treatment outcome are likely to change during this process and may switch from those involved in a predictive role (i.e., those genes that determine response to therapy) to those whose prognostic influence is likely to dictate the subsequent course of the disease and OS as resistance develops. It is possible that therapeutic intervention with PTK/ZK in patients with elevated HIF1a and VEGFR2 resulted in a shift in tumor biology to alternate angiogenic pathways that facilitated continued disease progression and ultimately impacted OS. This is supported by the results of CONFIRM-1 where a statistically significant improvement in PFS but no OS benefit was observed in patients with elevated LDH.21 The current study is a retrospective subset analysis but represents a focused and concerted research effort based on the current biological understanding to generate hypotheses for future testing in clinical trials. As with all retrospective pilot studies, this analysis has potential limitations. First, our findings & 2013 Macmillan Publishers Limited


415 Table 5.

Intratumoral gene expression levels and OS by treatment PTK/ZK

LDHA p0.36 40.36 P-valuea Pinteraction GLUT-1 p1.50 41.50 P-valuea Pinteraction HIF1a p1.84 41.84 P-valuea Pinteraction VEGF p4.16 44.16 P-valuea Pinteraction VEGFR1 p3.78 43.78 P-valuea Pinteraction VEGFR2 p1.29 41.29 P-valuea Pinteraction

Placebo

Median OS

HR (95%CI)

Median OS

HR (95%CI)

23.3 (23.3, 30.8) 26.6 (21.4, 33.2)

1 (Reference) 0.70 (0.28, 1.73)

3.5 (3.5, 38.0+) 25.5 (19.4, 46.6+)

1 (Reference) 0.95 (0.13, 7.10)

0.42

26.6 (22.7, 33.2) 27.9 (21.4, 42.7+)

0.96

N/Ab 1 (Reference) 0.61 (0.29, 1.30)

24.3 (17.2, 46.6+) 33.9 (20.1, 45.0+)

0.18

1 (Reference) 0.92 (0.39, 2.16) 0.84

0.58 26.1 (22.1, 27.9) 30.8 (21.1, 35.8)

1 (Reference) 0.93 (0.45, 1.94)

45.0+ (24.8, 45.0+) 17.2 (14.3, 37.1)

0.84

1 (Reference) 2.39 (1.06, 5.38) 0.026

0.086 23.3 (21.4, 31.2) 26.6 (22.1, 42.7+)

1 (Reference) 0.59 (0.26, 1.34)

15.9 (11.6, 38.8+) 25.5 (20.1, 46.6+)

0.19

1 (Reference) 0.85 (0.29, 2.49) 0.76

0.60 27.8 (23.3, 33.2) 26.6 (22.1, 42.7+)

1 (Reference) 0.64 (0.30, 1.38)

25.5 (15.4, 38.8+) 24.3 (17.2, 37.1)

0.24

1 (Reference) 1.13 (0.45, 2.86) 0.79

0.32 30.8 (23.3, 35.8) 26.2 (21.4, 31.2)

1 (Reference) 0.99 (0.42, 2.31)

45.0+ 24.3 (15.9, 33.9)

0.98

1 (Reference) 6.67 (1.59, 28.01) 0.003

0.014

Abbreviations: GLUT-1, glucose transporter-1; HR, hazard ratio; HIF1a, hypoxia-inducible factor type-1 alpha; LDHA, lactate dehydrogenase A; OS, overall survival; VEGF (R1) (R2), vascular endothelial growth factor (type-1 receptor) (type-2 receptor). a Based on log-rank test. b P-value for LDHA--treatment interaction could not be calculated owing to the low numbers in the placebo group with low LDHA expression. Statistically significant P-values are highlighted in bold.

are based on a relatively small number of patients who were recruited from a large international study over a period of 4 years; second, owing to the retrospective nature, additional analyses of tumor angiogenesis such as microvessel density or venous/ lymphatic vascular invasion were not possible; and third, only six genes within the angiogenesis pathway were examined. Nonetheless, this type of retrospective study is an ideal platform for testing a novel hypothesis and for generating data that can be confirmed prospectively in future clinical trials evaluating this class of agents. Care was taken to select candidate genes with a documented role in the angiogenesis pathway, which have been found to be associated with prognosis in previous studies at our institution or in published manuscripts. In addition, internal validation analysis was performed to reduce the likelihood of overanalyzing this dataset. Limitations aside, we have identified independent molecular markers associated with clinical outcome in patients treated with FOLFOX4 plus PTK/ZK. Importantly, the significant observations we report are entirely consistent with the underlying tumor biology and the subsequent hypothesis tested. Elevated expression levels of genes involved in angiogenesis and HIF-tumor biology, and in particular VEGFR1, were independently associated & 2013 Macmillan Publishers Limited

with clinical outcome in patients enrolled in CONFIRM-1 treated with PTK/ZK, indicative of a potential role in predicting the efficacy of VEGFR TKI therapy. Recently, additional VEGFR TKI agents have been developed, but they have not been shown to increase the efficacy of FOLFOX4.29 It is plausible that a subgroup of patients who are likely to benefit from a VEGFR-TKI could be identified by molecular predictive markers. However, as with all retrospective analyses, larger independent, prospective, biomarker-embedded clinical trials are needed to confirm and validate these preliminary findings.

CONFLICT OF INTEREST Gerold Meinhardt is employed by and has stock ownership in Bayer Healthcare Pharmaceuticals. Michael M. Shi and David Lebwohl are employed by Novartis. David Lebwohl has stock ownership in Novartis. Heinz-Josef Lenz is a consultant for Novartis and has contracted research with Novartis.

ACKNOWLEDGEMENTS This study was sponsored by Novartis and Bayer Schering Pharma as part of a codevelopment program.



416

REFERENCES 1 Ferrer FA, Miller LJ, Lindquist R, Kowalczyk P, Laudone VP, Albertsen PC et al. Expression of vascular endothelial growth factor receptors in human prostate cancer. Urology 1999; 54: 567--572. 2 Dvorak HF. Vascular permeability factor/vascular endothelial growth factor: a critical cytokine in tumor angiogenesis and a potential target for diagnosis and therapy. J Clin Oncol 2002; 20: 4368--4380. 3 Liotta LA, Steeg PS, Stetler-Stevenson WG. Cancer metastasis and angiogenesis: an imbalance of positive and negative regulation. Cell 1991; 64: 327--336. 4 Senger DR, Van de Water L, Brown LF, Nagy JA, Yeo KT, Yeo TK et al. Vascular permeability factor (VPF, VEGF) in tumor biology. Cancer Metastasis Rev 1993; 12: 303--324. 5 Brown LF, Berse B, Jackman RW, Tognazzi K, Guidi AJ, Dvorak HF et al. Expression of vascular permeability factor (vascular endothelial growth factor) and its receptors in breast cancer. Hum Pathol 1995; 26: 86--91. 6 Hurwitz H, Fehrenbacher L, Novotny W, Cartwright T, Hainsworth J, Heim W et al. Bevacizumab plus irinotecan, fluorouracil, and leucovorin for metastatic colorectal cancer. N Engl J Med 2004; 350: 2335--2342. 7 Miller K, Wang M, Gralow J, Dickler M, Cobleigh M, Perez EA et al. Paclitaxel plus bevacizumab versus paclitaxel alone for metastatic breast cancer. N Engl J Med 2007; 357: 2666--2676. 8 Aghajanian C, Finkler NJ, Rutherford T, Smith DA, Yi J, Parmar H et al. OCEANS: A randomized, double-blinded, placebo-controlled phase III trial of chemotherapy with or without bevacizumab (BEV) in patients with platinum-sensitive recurrent epithelial ovarian (EOC), primary peritoneal (PPC), or fallopian tube cancer (FTC). ASCO Meeting Abstracts 2011; 29(15_suppl): LBA5007. 9 Hockel M, Vaupel P. Tumor hypoxia: definitions and current clinical, biologic, and molecular aspects. J Natl Cancer Inst 2001; 93: 266--276. 10 Semenza GL. HIF-1: mediator of physiological and pathophysiological responses to hypoxia. J Appl Physiol 2000; 88: 1474--1480. 11 Lu H, Forbes RA, Verma A. Hypoxia-inducible factor 1 activation by aerobic glycolysis implicates the Warburg effect in carcinogenesis. J Biol Chem 2002; 277: 23111--23115. 12 Maxwell PH, Pugh CW, Ratcliffe PJ. Activation of the HIF pathway in cancer. Curr Opin Genet Dev 2001; 11: 293--299. 13 Koukourakis MI, Giatromanolaki A, Simopoulos C, Polychronidis A, Sivridis E. Lactate dehydrogenase 5 (LDH5) relates to up-regulated hypoxia inducible factor pathway and metastasis in colorectal cancer. Clin Exp Metastasis 2005; 22: 25--30. 14 Kim JW, Gao P, Dang CV. Effects of hypoxia on tumor metabolism. Cancer Metastasis Rev 2007; 26: 291--298. 15 Harris AL. Hypoxia--a key regulatory factor in tumour growth. Nat Rev Cancer 2002; 2: 38--47. 16 Azuma M, Shi M, Danenberg KD, Gardner H, Barrett C, Jacques CJ et al. Serum lactate dehydrogenase levels and glycolysis significantly correlate with tumor VEGFA and VEGFR expression in metastatic CRC patients. Pharmacogenomics 2007; 8: 1705--1713.


17 Sitaras NM, Sariban E, Bravo M, Pantazis P, Antoniades HN. Constitutive production of platelet-derived growth factor-like proteins by human prostate carcinoma cell lines. Cancer Res 1988; 48: 1930--1935. 18 Buchdunger E, Zimmermann J, Mett H, Meyer T, Muller M, Regenass U et al. Selective inhibition of the platelet-derived growth factor signal transduction pathway by a protein-tyrosine kinase inhibitor of the 2-phenylaminopyrimidine class. Proc Natl Acad Sci USA 1995; 92: 2558--2562. 19 Carmeliet P, Ferreira V, Breier G, Pollefeyt S, Kieckens L, Gertsenstein M et al. Abnormal blood vessel development and lethality in embryos lacking a single VEGF allele. Nature 1996; 380: 435--439. 20 Wood JM, Bold G, Buchdunger E, Cozens R, Ferrari S, Frei J et al. PTK787/ZK 222584, a novel and potent inhibitor of vascular endothelial growth factor receptor tyrosine kinases, impairs vascular endothelial growth factor-induced responses and tumor growth after oral administration. Cancer Res 2000; 60: 2178-2189. 21 Hecht JR, Trarbach T, Hainsworth JD, Major P, Jager E, Wolff RA et al. Randomized, placebo-controlled, phase III study of first-line oxaliplatin-based chemotherapy plus PTK787/ZK 222584, an oral vascular endothelial growth factor receptor inhibitor, in patients with metastatic colorectal adenocarcinoma. J Clin Oncol 2011; 29: 1997--2003. 22 Van Cutsem E, Bajetta E, Valle J, Kohne CH, Randolph Hecht J, Moore M et al. Randomized, placebo-controlled, phase III study of oxaliplatin, fluorouracil, and leucovorin with or without PTK787/ZK 222584 in patients with previously treated metastatic colorectal adenocarcinoma. J Clin Oncol 2011; 29: 2004--2010. 23 Koukourakis MI, Giatromanolaki A, Sivridis E, Gatter KC, Trarbach T, Folprecht G et al. Prognostic and predictive role of lactate dehydrogenase 5 expression in colorectal cancer patients treated with PTK787/ZK 222584 (Vatalanib) antiangiogenic therapy. Clin Cancer Res 2011; 17: 4892--4900. 24 McShane LM, Altman DG, Sauerbrei W, Taube SE, Gion M, Clark GM. REporting recommendations for tumor MARKer prognostic studies (REMARK). Nat Clin Pract Oncol 2005; 2: 416--422. 25 Bonner RF, Emmert-Buck M, Cole K, Pohida T, Chuaqui R, Goldstein S et al. Laser capture microdissection: molecular analysis of tissue. Science 1997; 278: 1481--1483. 26 Miller R, Siegmund D. Maximally selected chi square statistics. Biometrics 1982; 38: 1011--1016. 27 Halpern J. Maximally selected chi square statistics for small samples. Biometrics 1982; 38: 1017--1023. 28 Giacchetti S, Perpoint B, Zidani R, Le Bail N, Faggiuolo R, Focan C et al. Phase III multicenter randomized trial of oxaliplatin added to chronomodulated fluorouracil-leucovorin as first-line treatment of metastatic colorectal cancer. J Clin Oncol 2000; 18: 136--147. 29 Robertson JD, Botwood NA, Rothenberg ML, Schmoll HJ. Phase III trial of FOLFOX plus bevacizumab or cediranib (AZD2171) as first-line treatment of patients with metastatic colorectal cancer: HORIZON III. Clin Colorectal Cancer 2009; 8: 59--60.

& 2013 Macmillan Publishers Limited