Targeting Epidermal Growth Factor Receptor–Related Signaling Pathways in Pancreatic Cancer

Abstract: Pancreatic cancer is aggressive, chemoresistant, and character- ized by complex and poorly understood molecular biology. The epidermal growth factor receptor (EGFR) pathway is frequently activated in pancre- atic cancer; therefore, it is a rational target for new treatments. However, the EGFR tyrosine kinase inhibitor erlotinib is currently the only targeted therapy to demonstrate a very modest survival benefit when added to gemcitabine in the treatment of patients with advanced pancreatic cancer. There is no molecular biomarker to predict the outcome of erlotinib treat- ment, although rash may be predictive of improved survival; EGFR expres- sion does not predict the biologic activity of anti-EGFR drugs in pancreatic cancer, and no EGFR mutations are identified as enabling the selection of patients likely to benefit from treatment. Here, we review clinical studies of EGFR-targeted therapies in combination with conventional cytotoxic regi- mens or multitargeted strategies in advanced pancreatic cancer, as well as research directed at molecules downstream of EGFR as alternatives or ad- juncts to receptor targeting. Limitations of preclinical models, patient se- lection, and trial design, as well as the complex mechanisms underlying resistance to EGFR-targeted agents, are discussed. Future clinical trials must incorporate translational research end points to aid patient selection and circumvent resistance to EGFR inhibitors.

Key Words: biomarkers, erlotinib, epidermal growth factor receptor, molecular targeted therapy, pancreatic cancer

Pancreatic cancer continues to present a significant therapeutic challenge. It is characterized by chemotherapy-refractory dis- ease, and our understanding of its complex molecular biology re- mains limited.1,2 Survival rates in pancreatic cancer have shown little improvement in the last 4 decades.3 Treatment of patients with inoperable disease remains palliative and comprises chemotherapy or chemoradiation therapy.4,5 For metastatic disease, the most common interventions involve gemcitabine, either alone or in combination with the small mole- cule tyrosine kinase inhibitor (TKI) erlotinib, nab-paclitaxel, an albumin-bound form of paclitaxel, or a fluoropyrimidine.4,6–8 A phase 3 study in a highly selected patient population showed that the FOLFIRINOX regimen (infusional 5-fluorouracil, folinic acid, irinotecan, and oxaliplatin) significantly improved overall survival (OS) compared with single-agent gemcitabine, although with in- creased toxicity.2,9 Recently, improved progression-free survival (PFS) and OS results have been reported in a phase 3 study of nab-paclitaxel in combination with gemcitabine versus gemcitabine alone.10 Other promising results have been seen in phase 2 studies of the combination of cationic liposomal embedded paclitaxel with gemcitabine11 and with the addition of the hypoxia-activated pro- drug TH-302 to gemcitabine.12 Substantially more progress in treat- ment options is needed, however, to improve the current poor prognosis of patients with advanced pancreatic cancer.

A better understanding of the underlying molecular alter- ations that drive pancreatic cancer may allow the rational de- velopment of targeted therapies. This article reviews key concepts and clinical findings from agents targeting epidermal growth factor receptor (EGFR)–related signaling pathways in pancreatic cancer.

MATERIALS AND METHODS

A MEDLINE search was conducted covering 2002 to 2012 using the following search terms: pancreatic cancer/carcinoma/ tumor; epidermal growth factor receptor/EGFR; erlotinib/Tarceva/ OSI-774/OSI 774. Search results were restricted to English- and German-language papers. In addition, abstracts from key oncology (American Society for Clinical Oncology, European Society of Medical Oncology) and gastroenterology meetings (American Soci- ety for Clinical Oncology Gastrointestinal Cancers Symposium, World Congress for Gastrointestinal Cancer) were searched using the same search terms.

Targeting the EGFR Signaling Pathway in Pancreatic Cancer EGFR-Related Signaling Pathways

Human EGFR type 1 (HER1/EGFR) is a transmembrane glycoprotein belonging to the HER family of receptor tyrosine kinases (also called ErbB receptors).13 Epidermal growth factor receptor is activated by the selective binding of ligands to its extracellular domain, resulting in receptor dimerization and the induction of a conformational change that allows autophos- phorylation of the intracellular domain of the receptor. This in turn triggers the downstream activation of intracellular signal- ing molecules.13–15 EGFR activates 2 major signaling pathways (Fig. 1): the rat sarcoma/recombinant activated factor/mitogen- activated protein kinase (MAPK) extracellular signal–regulated kinase/MAPKs (RAS/RAF/MEK/MAPK) pathway that affects the downstream target MAPK and directs gene transcription, cell cycle progression, and cell proliferation and the phosphatidylinositol-3 kinase (Pl3K)/Akt pathway that affects the downstream activation of the mammalian target of rapamycin (mTOR) and nuclear factor κB (NF-κB) and generates antiapoptotic and prosurvival sig- nals.14,16 In pancreatic carcinoma xenograft studies, inhibition of EGFR signaling has been shown to slow tumor growth and metas- tases and to potentiate the effects of gemcitabine.17,18 Crosstalk between molecules downstream of the HER kinase family and with EGFR-independent pathways allows for the activation of ad- ditional signaling pathways that would explain the challenges of successfully targeting EGFR with a single drug.13

FIGURE 1. A simplified depiction of the major EGFR dimerization-initiated signaling pathways and the mode of action for anti-EGFR mAbs and TKIs. MOA, mechanism of action; P, phosphate.

Five EGFR-directed therapies are currently available in can- cer treatment: the small molecule EGFR TKIs erlotinib and gefi- tinib, the dual EGFR/HER2 TKI lapatinib, and the anti-EGFR monoclonal antibodies (mAbs) cetuximab and panitumumab. Small molecule TKIs inhibit autophosphorylation by competing with adenosine triphosphate for binding to the intracellular cata- lytic domain of EGFR tyrosine kinase, whereas the anti-EGFR mAbs prevent receptor dimerization by binding to the extracellular domain of EGFR in its inactive configuration and by obstructing li- gand binding (Fig. 1).14 At a cellular level, the effects of these agents include the inhibition of cancer cell proliferation, invasion, and metastasis; decreased angiogenic growth factor production and tumor-induced angiogenesis; and the potentiation of cyto- toxic drugs and radiotherapy.14

Several phases 2 and 3 studies have investigated the addition of anti-EGFR targeted therapies to chemotherapy for the treatment of advanced pancreatic cancer (Supplementary Digital Content Table 1, http://links.lww.com/MPA/A390, which summarizes phases 2 and 3 trials investigating anti-EGFR-based regimens).19–34 His- torically, tumor response rates in patients with advanced pancre- atic cancer treated with single-agent gemcitabine have been less than 10%. Tumor progression typically occurs within 4 months, and median OS is less than 6 months.6 A potential survival ben- efit was noted when EGFR-targeted therapies were added to gemcitabine- and other chemotherapy-based regimens in some,19,24,28 but not all, phase 2 studies.20,21,25,29

Phase 3 trials of anti-EGFR treatments have enrolled molec- ularly unselected populations of patients with both unresectable locally advanced and metastatic pancreatic cancer.31,32,34,35 To date, only 1 such phase 3 trial has shown a statistically significant, albeit very modest, survival benefit. In the PA.3 trial, which inves- tigated the addition of erlotinib to gemcitabine, the objective re- sponse rate was similar in patients treated with erlotinib plus gemcitabine (8.6%) and in those treated with placebo plus gemcitabine (8.0%), but OS (hazard ratio [HR], 0.82; 95% confi- dence interval [CI], 0.69–0.999; P = 0.038) and PFS (HR, 0.77; 95% CI, 0.64–0.92; P = 0.004) were significantly prolonged by the addition of erlotinib (Supplementary Digital Content Table 1, http://links.lww.com/MPA/A390).31 Although the differences in OS and PFS were statistically significant, they represented median improvements of 10 and 6 days, respectively.

Data from the PA.3 trial,31 the AIO (Abeitsgemeinschaft Internistische Onkologie) trial,33 and a phase 2 trial of bevacizu- mab plus gemcitabine/erlotinib36 show that patients who develop rash with erlotinib have improved OS compared with those with- out rash. However, it has not been determined if rash is truly pre- dictive of erlotinib efficacy or if merely the disease has a better prognosis through patient-related factors.
Preclinical studies suggest additive effects when anti-EGFR agents are combined with inhibitors of the vascular endothelial growth factor (VEGF).37,38 In phase 2 studies of treatment-naive advanced pancreatic cancer, combining gemcitabine-based che- motherapy with bevacizumab (an antibody directed against VEGF) and anti-EGFR therapy (cetuximab or erlotinib) showed modest benefit,39,40 but these were not universally considered sufficient to warrant phase 3 evaluation.41 A phase 3 study of patients with metastatic disease found that adding bevacizumab to gemcitabine plus erlotinib did not extend OS, but there was a statistically signif- icant gain in PFS in the overall population and some evidence of survival benefit in patients with more aggressive disease.33 The 4-drug regimen of gemcitabine/capecitabine plus bevacizumab/ erlotinib achieved a 23% confirmed radiologic response rate and an OS of 12.8 months in a phase 2 trial that enrolled patients with a World Health Organization performance status of 2 or less.42,43 Treatment with erlotinib and bevacizumab was relatively ineffec- tive in patients with gemcitabine-refractory metastatic pancreatic cancer in another phase 2 study.44

EGFR and Ligand Expression

Among patients with non–small cell lung cancer (NSCLC), activating EGFR mutations are key predictors of response to EGFR TKI therapy.45–47 In contrast, EGFR expression, EGFR gene amplification, EGFR intron 1 polymorphism, or the pres- ence of the EGFR exon 12 R497K point mutation do not appear to be useful in predicting the survival benefit of these agents in pa- tients with advanced pancreatic cancer.32,48–50 Amphiregulin, an EGFR ligand present in pancreatic cancer cells, has been impli- cated in the development of pancreatic cancer.51–54 An explor- atory analysis conducted within the BioMarker Identification trial (MARK) recently identified amphiregulin as a potential bio- marker of response to EGFR-targeted therapy in patients with ad- vanced pancreatic cancer.55

KRAS Mutation

Mutations in the Kirsten rat sarcoma viral oncogene homolog (KRAS) proto-oncogene that result in constitutive activation of the downstream effector of EGFR are associated with resistance to anti-EGFR therapies in patients with advanced colorectal cancer.56,57 Support for KRAS wild-type status as a predictor of response to EGFR-targeted therapy comes from a recent retrospective anal- ysis of 136 patients for whom KRAS wild-type status was associ- ated with a significant survival benefit among those treated with erlotinib (OS, 9.7 vs 5.2 months; P = 0.002), but not among pa- tients treated without erlotinib (OS, 7.0 vs 7.0 months; P = 0.121).58 KRAS mutation status was neither prognostic for OS nor predictive of therapeutic response to erlotinib therapy in the phase 3 PA.3 study.48 The retrospective analysis was hampered by the small number of analyzed tumor samples (117 of 569). Similar findings were reported from a small phase 2 trial of erlotinib-based therapy in patients with advanced pancreatic can- cer in which tumor samples were prospectively collected.28 In the phase 2 MARK trial,49 patients with unresectable locally ad- vanced or metastatic pancreatic cancer who had failed or were in- eligible for first-line chemotherapy were randomized to receive erlotinib 150 mg/d (n = 104) or placebo (n = 103) until disease progression. Tumor biopsies were collected prospectively, and the primary end point was identification of biomarkers predictive of PFS benefit with erlotinib. Progression-free survival did not differ significantly by KRAS mutation status, and KRAS mutation
status did not predict response to erlotinib.

In the phase 3 AIO trial in advanced pancreatic cancer, tumor samples were evaluable from 173 of the 281 randomized patients, and KRAS mutations were detected in 70% of the samples.33 As erlotinib therapy was used in both treatment arms, it was not pos- sible to evaluate KRAS mutation status as a predictive marker of erlotinib efficacy. Median OS was longer in patients with KRAS wild-type than in those with mutated KRAS (7.9 vs 5.7 months), and in a univariate analysis, KRAS mutation status correlated sig- nificantly with OS (HR, 1.68; 95% CI, 1.17–2.41; P = 0.005), but not with objective response rate or time to treatment failure.33

Collectively, these studies failed to demonstrate a definite role for KRAS mutation status in predicting response to anti- EGFR therapies. Nevertheless, there was some evidence that KRAS status has a prognostic role in advanced pancreatic cancer.

Proteomic Profiles

Serum proteomics represent another potential means of predicting benefit from EGFR-directed therapies and has led to the development of a clinically validated59 test (VeriStrat; Biodesix,Boulder, Colo) to guide treatment with EGFR TKIs in patients with advanced NSCLC. The test is also being investigated for use in pancreatic cancer. The assay uses an algorithm or “proteo- mic signature” to predict survival outcomes of NSCLC patients treated with gefitinib or erlotinib based on the pattern of 8 signa- ture peaks detected in patient serum samples subjected to matrix- assisted laser desorption ionization mass spectroscopic analysis. The algorithm was shown to be predictive of treatment outcomes, not merely prognostic of disease outcomes; thus, it may offer a means of selecting patients likely to benefit from EGFR TKI ther- apy. Furthermore, the VeriStrat proteomic signature has been shown to predict survival in NSCLC patients with wild-type EGFR independent of KRAS mutation status.60 The validation of an analogous proteomic signature predictive of response to EGFR TKIs in patients with advanced pancreatic cancer would be of great interest in this field.

Lessons Learned From Targeting EGFR in Pancreatic Cancer

A number of studies were undertaken to determine the role of targeting EGFR in pancreatic cancer. Unfortunately, after more than a decade of clinical investigation, no progress has been made beyond finding a very modest improvement with the use of erloti- nib. The major reason for this lack of progress is our very limited understanding of the biology of pancreatic cancer, including EGFR signaling and its related pathways. There is no model system that accurately mirrors the clinical course of human pancreatic cancer.1 Cell-based assays and xenograft models, although useful in delin- eating drug mechanisms of action and exploring biomarkers, are unable to capture the complexity of tumor biology, and no single cell line replicates the heterogeneity of pancreatic cancer.61

Resistance Mechanisms to Anti-EGFR Therapy

A key issue has been our poor understanding of resistance mechanisms of pancreatic cancer to traditional chemotherapy and radiotherapy, as well as to targeted therapies. Understanding resistance to anti-EGFR agents is key to developing better thera- pies for pancreatic cancer and to improving patient selection by using appropriate molecular markers that would identify patients who will not respond to such therapies. Of those patients who do achieve a tumor response, most will eventually have disease progression, which suggests acquired resistance.62 Major molecu- lar resistance mechanisms have been identified that can be exploited for the development of newer targeted therapies and/or predictive biomarkers. Other characteristics of pancreatic cancer that make it highly resistant to traditional chemotherapy are also likely to affect its response to EGFR-targeted therapies, including the dense desmoplastic stroma, a hypoxic microenvironment that may impair drug delivery, and the presence of highly tumorigenic stem cells.2,63

Alterations, Redundancies, and Crosstalk in EGFR-Related Pathways

The activation of parallel or downstream pathways that per- mit unregulated tumor cell growth independent from direct EGFR stimulation results in resistance to anti-EGFR therapies.57

Postulated mechanisms of resistance include “oncogenic shift,” crosstalk with other receptors or tyrosine kinases, and EGFR mutations. “Oncogenic shift” and receptor crosstalk in- volve the participation of alternative receptor tyrosine kinases, which interact with or even replace EGFR signaling.62,64 Exam- ples are the increased VEGF/VEGF receptor expression or the dysregulation of EGFR internalization and degradation leading to sustained signaling and activation of HER3.62 In addition,kinase crosstalk may allow activation of alternative signaling pathways.12,65–67 EGFR mutations with impact on response to therapy may either interfere with TKI binding and/or increase af- finity for adenosine triphosphate, as shown for the T790M muta- tion in NSCLC cells, or constitutively activate the receptor, as shown for the EGFR variant III (EGFRvIII), which lacks the ligand-binding domain and thus confers resistance to EGFR anti- bodies in head and neck squamous cell carcinoma.68

Dysregulation of the PI3K/Akt/mTOR Signaling

Dysregulation of the PI3K/Akt/mTOR signaling axis is a key feature of pancreatic cancer and arises from KRAS mutation, in- creased expression of EGFR, and loss of phosphatase and tensin ho- molog (PTEN) function.69 Loss of PTEN function enables the uncoupling of the Akt pathway from EGFR, resulting in constitutive Akt activation and aberrant PTEN expression, which has been re- ported in pancreatic cancer cells.70 Activation of this pathway leads to enhanced cell survival, likely contributes to its aggressive nature, and is considered important in the resistance of pancreatic cancer to EGFR TKIs.62 Further evidence that uncoupling the Akt pathway from EGFR underlies EGFR TKI resistance comes from preclinical studies showing that xenografts derived from PTEN-mutant human glioma cells overexpressing EGFRvIII are resistant to treatment with either gefitinib or a pan-PI3K inhibitor alone, but that com- bined treatment effectively blocks further tumor growth.71

Epithelial-to-Mesenchymal Transition

Epithelial-to-mesenchymal transition (EMT) describes the conversion of cancer cells from an epithelial phenotype (adherent cells bound and restricted within an organized tissue) to a mesen- chymal phenotype in which they act as independent fibroblast- like cells with the ability to migrate and to invade the extracellular matrix. Epithelial-to-mesenchymal transition is triggered by a com- plex interplay of extracellular signals and growth factors, including epidermal growth factor, and it underlies the high metastatic poten- tial of pancreatic cancer cells.72,73 Epithelial-to-mesenchymal tran- sition appears to be important in conferring drug resistance not only to conventional chemotherapeutic agents, such as gemcitabine, but also to EGFR-targeted therapies73,74 and as such may offer opportu- nities for new treatment targets, as well as development of a predic- tive biomarker. In vitro studies have shown that gene signatures associated with EMT predict response to erlotinib in NSCLC and pancreatic cancer cell lines and tumors.75,76 Although preliminary clinical evidence may suggest a role for EMT status as a predictor of response to anti-EGFR therapy in lung cancer,75 its relevance to pancreatic cancer remains to be assessed. The Radiation Therapy Oncology Group–led study RTOG-0848 is investigating the role of EMT prospectively in patients with resected pancreatic cancer, who are randomized to gemcitabine with or without erlotinib in a phase 3 trial.

c-MET

c-MET, a receptor tyrosine kinase associated with the ligand hepatocyte growth factor, is expressed in normal and malignant cells.77 Elevated levels of c-MET are present in pancreatic cancer cells, and its stimulation via exogenous hepatocyte growth factor triggers proliferation and movement.78 c-MET may underlie resis- tance to EGFR inhibitors65 and is an emerging molecular target in a number of cancers.

Insulinlike Growth Factor Receptor 1 Signaling

Insulinlike growth factor 1 (IGF-1) and its receptor (IGF-1R) are present at elevated levels in pancreatic tumors.79 Preclinical evidence suggested that signaling through the IGF-1R pathway may be responsible for the resistance of tumors against anti- EGFR drug therapy.80 A recent phase 2 study examining the effi- cacy of simultaneously blocking the EGFR and IGF-1R pathways in first-line metastatic pancreatic cancer found that the addition of the anti–IGF-1R mAb cixutumumab to gemcitabine plus erlotinib did not improve PFS or OS.81 Dalotuzumab (MK-0646), a human- ized IGF-1R mAb, is undergoing early clinical trials in combina- tion with gemcitabine or gemcitabine plus erlotinib in advanced pancreatic cancer patients. Preliminary results from a phase 2 trial of 53 evaluable patients with stage IV previously untreated pancre- atic cancer showed that median PFS was significantly longer in patients treated with dalotuzumab plus gemcitabine (17 weeks; P = 0.0425) than with dalotuzumab in combination with gemci- tabine plus erlotinib (8 weeks) or with gemcitabine plus erlotinib alone (8 weeks).80 In this study, patients experiencing progres- sion in the gemcitabine/erlotinib control arm were crossed over to triplet therapy. A significantly greater proportion of patients treated with dalotuzumab plus gemcitabine survived longer than 40 weeks compared with the other 2 treatment groups; however, OS was not significantly different between the 3 treatment arms. Early research suggests tissue expression of IGF-1 may offer a way to predict response to IGF-1R–directed therapy. An analysis of IGF-1 mRNA expression in archival core biopsies from 50 patients enrolled in the phase 2 dalotuzumab study showed a 96% reduction in the risk of disease progression or death in pa- tients treated with dalotuzumab plus gemcitabine compared with the gemcitabine plus erlotinib at higher IGF-1 expression levels (P = 0.08).82

Cyclooxygenase 2

Cyclooxygenase 2 (COX-2) is an inducible enzyme overex- pressed in pancreatic cancer that modulates angiogenesis and me- tastasis. The complex EGFR and COX-2 pathways interact at multiple levels, and targeting COX-2 pathways using a selective COX-2 inhibitor has been shown to potentiate the growth inhibitory and proapoptotic effects of erlotinib in human pancreatic cancer cell lines with high basal levels of EGFR expression.83 However, the Apricot-P study found no improvement in PFS with the addi- tion of the COX-2 inhibitor apricoxib to combination therapy with gemcitabine plus erlotinib in advanced pancreatic cancer.84

Future Directions

Preclinical Models

An important advance in preclinical models has been the de- velopment of genetically engineered mouse models that better mimic the stromal reactions of human disease, perhaps better sim- ulating the genetic heterogeneity of human cancers.1,61 The use of primary, patient-derived tumor xenografts may better represent the clinical activity of new drugs but is limited by the poor availability of tumor tissues, and an improved understanding of potential dif- ferences between xenografts derived from primary and metastatic sites is needed.1

Novel Drug Administration Strategies: Intermittent High-Dose Therapy

Abnormal signaling through PI3K is a predominant feature of pancreatic cancers, making EGFR, which activates the PI3K/ Akt signaling pathway, an attractive treatment target. However, because signaling through PI3K is critical in the maintenance of normal tissues, it is unlikely that complete target inhibition can be sustained over long periods. Some researchers have suggested that a better therapeutic ratio can be achieved by administering targeted agents intermittently at higher doses,66 which may pro- duce more complete PI3K inhibition in the tumor, while also allowing time for normal tissue recovery. Furthermore, evidence supporting intermittent dosing of an EGFR TKI with chemother- apy comes from preclinical studies in NSCLC cells that show a possible mechanistic interference when chemotherapy and EGFR TKIs are administered concurrently. Continuous treatment with erlotinib induces G1 arrest of tumor cells that may hinder the ef- fects of cell cycle–specific cytotoxic agents, whereas administra- tion of the chemotherapeutic agent prior to erlotinib achieves a synergistic cytotoxic effect.85,86 Similar findings have been re- ported with sequential administration of pemetrexed and erlotinib in human pancreatic cancer cells.87 Phase 1/II clinical studies have investigated sequential or intermittent dosing of EGFR TKIs and chemotherapy in the treatment of NSCLC and other solid tumors, with some promising results reported.88–93 A subsequent phase 3 trial of this regimen with erlotinib in NSCLC reported significant OS and PFS improvement over chemotherapy plus placebo.94 This intermittent dosing strategy is currently being investigated in a phase 2 study (NCT00810719) in which patients with meta- static or recurrent pancreatic cancer will receive 4-week cycles of gemcitabine and erlotinib, the latter administered at 150 mg/d intermittently (ie, on days 2–5, 9–12, and 16–26 of each cycle).

EGFR Targeting as Part of Multitargeted Therapy in Pancreatic Cancer
Genomic analyses have detected a large number of genetic alterations that affect 12 core signaling pathways in the major- ity of pancreatic cancer cases,15,95 highlighting its complexity and genetic diversity. Strategies to overcome specific resistance to EGFR targeting and improve outcomes have involved the de- velopment of multitargeted therapies.2 Phases 2 and 3 clinical trials in which anti-EGFR agents have been combined with other targeted therapies in pancreatic cancer are summarized in Table 1.36,39–42,44,80,81,84,96–98

Combining TKIs and mAbs Against EGFR

A multitargeting strategy is dual inhibition of the EGFR pathway using both a TKI and a mAb. Dual inhibition using erlo- tinib and panitumumab has been shown to potentiate antitumor activity in preclinical pancreatic cancer models, and a phase 2 study found that the addition of panitumumab to gemcitabine plus erlotinib significantly improved OS, but not PFS in patients with untreated metastatic pancreatic cancer.98

Targeting Molecules Downstream of EGFR as an Alternative to Receptor Targeting
The EGFR activates several intracellular signaling pathways (Fig. 1), including the PI3K/Akt, RAS/RAF/MEK/MAPK, phos- pholipase Cγ1/protein kinase C, and signal transducer and activator of transcription pathways, modulating downstream targets, includ- ing mTOR and NF-κB.12,15 These downstream effectors represent alternative therapeutic targets in pancreatic cancer.

mTOR targeted agents, alone or in combination with EGFR inhibitors, are under investigation in early clinical trials in pancreatic cancer,99,100 as are agents targeting NF-κB, such as curcumin.101,102 Single-agent everolimus showed minimal clinical activity in gemcitabine-refractory metastatic pancreatic cancer,99 and 2 other phase 2 studies reported disappointing results with single-agent temsirolimus and combined everolimus plus erlotinib in patients with advanced pancreatic cancer.99

A phase 3 study in patients with advanced pancreatic cancer found that targeting RAS with a farnesyltransferase inhibitor combined with standard gemcitabine therapy did not achieve any improvement in OS compared with gemcitabine alone.103 Al- ternative strategies for targeting RAS are the use of antisense ther- apy and RNA interference, which have produced some promising results in preclinical and early clinical trials.104,105 Direct targeting of MAP2K, the principal downstream effector of the RAS path- way, with the MAP2K inhibitor CI-1040 in combination with EGFR inhibition with gefitinib also showed promise in preclinical studies.106

CONCLUSIONS

Targeting the EGFR signaling pathway has been clinically studied in patients with pancreatic cancer for more than a decade. To date, the EGFR TKI erlotinib is the only targeted therapy to demonstrate a statistically significant, albeit very modest, survival benefit in pancreatic cancer.31 It is now evident that EGFR expres- sion does not predict the biologic activity of anti-EGFR drugs in pancreatic cancer and, unlike the case of NSCLC, no EGFR mu- tations have been identified to help select patients most likely to benefit from treatment. Also, the role of KRAS mutation, which is very common in pancreatic cancer, still remains to be defined. Obviously, there needs to be a better understanding of biology for the design of treatments that address the complexity of signal- ing pathways in pancreatic cancer and to develop predictive bio- markers for patient selection. The development of predictive biomarkers to identify those patients whose tumors rely on EGFR signaling for their growth is a key issue, as is the rational design of therapeutic strategies, such as multitargeted therapies, to circum- vent resistance to EGFR inhibitors. As a basis, there is a crucial need for molecular analyses of tumor tissue, which have been lacking from most phase 3 trials in advanced pancreatic cancer.107 Prospective tissue banking and repeat biopsies assessing re- sponses during BLU 451 treatment are needed.61