Phosphorylation of p44/42 MAPK (ERK1/2) at Thr202/Tyr204 provides a critical measure of signal transduction linking extracellular stimuli to processes such as cell proliferation, differentiation, motility, and death. Phosphorylated ERK (pERK) immunohistochemistry has been used as a pharmacodynamic biomarker in clinical trials to evaluate the activity of numerous compounds targeting HER2, EGFR, RAF, MEK, ERK and other upstream molecules. Some studies have identified reduction in pERK in response to effective therapy. However, one of the challenges in using phospho-specific antibodies for immunohistochemistry is the concern that cross-reactivity may lead to erroneous conclusions. The following antibodies were tested for acceptable IHC performance: Clone MAPK-YT will only recognize ERK1 or ERK2 when dually phosphorylated at Thr202 and Tyr204 (Erk1) or Thr185 and Tyr187 (Erk2), while the other antibodies will also bind if ERK is phosphorylated only at the Threonine residue. Antigen retrieval and antibody titration studies were performed to optimize immunohistochemical staining performance. Cell lysates from the HEPG2 cell line (untreated, U0126 and sorafenib treated) were prepared for analysis via Western blotting and formalin-fixed, paraffin-embedded cell pellets were prepared for analysis via IHC. U0126 is a MEK inhibitor, while sorafenib inhibits multiple kinases including RAF. Development of the 9101 polyclonal antibody was discontinued due to inadequate sensitivity. The three monoclonal antibodies were used to stain samples of breast, colon and lung cancer using an HRP/DAB-based detection system. The same antibodies were used to develop an assay with similar performance with an AP/Vulcan Red-based detection system, and this assay was used to stain melanoma samples.

Immunohistochemical (IHC) analysis of ERK phosphorylation is commonly performed on pre- and post-treatment cancer biopsies as part of pharmacodynamic biomarker assessments for targeted oncology therapeutics. The current study was performed to assess the specificity and sensitivity of four commercially-available pERK antibodies for IHC analysis, which included three monoclonal antibodies (clones 20G11, MAPK-YT, D13.14.4E) and one polyclonal antibody. To facilitate this analysis, the HEPG2 liver cancer cell line was grown in the presence or absence of inhibitors of the RAF/MEK/ERK pathway (U0126 or sorafenib). Cell lysates were prepared for analysis via Western blotting and formalin-fixed, paraffin-embedded cell pellets were prepared for analysis via IHC. All antibodies produced bands at 44 and 42 kDa, the expected molecular weights of ERK1/2, when used for Western analysis of untreated cells and demonstrated a reduction or ablation of these bands following treatment of cells with inhibitors of the RAF/MEK/ERK pathway. When examined via IHC, the monoclonals demonstrated nuclear and cytoplasmic staining in HEPG2 cells with increased sensitivity relative to the polyclonal antibody. The monoclonals also demonstrated greatly reduced staining in HEPG2 cells treated with inhibitors of the RAF/MEK/ERK pathway. These antibodies were further characterized for reactivity in human tumor specimens representing breast cancer, colon cancer, lung cancer and melanoma. Melanoma tissues were stained with a red chromogen; all others were stained with DAB. Differences in the frequency of expression, percentage of expression, and protein localization were identified between the three monoclonals. In summary, this work has identified commercially-available antibodies that appear to be of suitable specificity and sensitivity for use in IHC assays.

Comparison of Commercially Available Phosphorylated ERK Antibodies for Immunohistochemical Biomarker Analysis

Michael Jeffers1, Lisa M. Dauffenbach2, Carol Peña1, Chetan Lathia1, Ajay Bhargava3, Christopher A. Kerfoot2 1Bayer HealthCare Pharmaceuticals, Montville, NJ, 2Mosaic Laboratories, Lake Forest, CA, 3NorthEast Bioanalytical Laboratories, Hamden, CT

Abstract

Introduction and Methods

Conclusions • All four antibodies recognized pERK1/2 via immunoblotting (correct size protein

that is reduced by U0126 or sorafenib). • The 9101 polyclonal antibody demonstrated lower sensitivity for IHC than the

monoclonal antibodies and was discontinued. • Although the monoclonal antibodies reportedly recognize different (but partially

overlapping) forms of pERK, they exhibited similar results in this head-to-head comparison across tumors using a DAB-based assay. Clone 20G11 demonstrated slightly greater frequency of staining than the other two monoclonal antibodies.

• The IHC assay using a red chromogen for analysis of melanomas demonstrated a higher degree of staining with clone 20G11 than the other monoclonals. This increased sensitivity may be a combination of antibody performance and assay sensitivity.

Results

2012 AACR Abstract 742

Characterized Cell Line Testing

Comparison of Average pERK IHC Result Between Antibodies

pERK in Melanoma

pERK in Breast, Colon and Lung Cancer

Figure 1: Immunohistochemistry staining results of optimized IHC assays applied to HEPG2 cell line treated with various conditions. The monoclonal antibodies MAPK-YT, 20G11 and D13.14.4E demonstrated acceptable sensitivity in untreated HEPG2 cells and a reduction in cells treated with UO0126 or sorafenib. This data was concordant with Western blot data. The polyclonal antibody 9101 did not demonstrate acceptable sensitivity.

Figure 3: All three monoclonal antibodies produced slightly different results in melanoma samples. Assays were titered on a tonsil sample for similar sensitivity with a red chromogen-based detection method. When the assay was used to analyze melanoma samples, the staining intensity within fibroblasts was similar between antibodies, but clone 20G11 demonstrated increased staining of cancer cells.

Breast Cancer C1

Breast Cancer C5

Colon Cancer A4

Colon Cancer C1

Lung Cancer A1

Lung Cancer D4

Melanoma A5

Figure 2: All three monoclonal antibodies produced similar staining in breast, colon and lung cancer samples. Subcellular localization of staining was nuclear and cytoplasmic. Staining was observed to varying degrees in cancer cells, fibroblasts and inflammatory cells.

Figure 4: Breast cancer had the greatest frequency of pERK expression, followed by non-small cell lung cancer, melanoma and colon cancer. Clone 20G11 demonstrated the greatest sensitivity across all cancer types. Results were concordant between the 3 monoclonal antibodies for breast, colon and lung cancer, but diverged in melanoma. The 20G11 clone using the red chromogen-based detection kit had much greater sensitivity in melanoma. 20G11 and D13.14.4E, the two antibodies that recognize both mono and dually-phosphorylated ERK, had the greatest correlation. In melanoma, the correlations between antibody staining results were the lowest.

0

50

100

150

Breast Cancer Colon Cancer Lung Cancer Melanoma

H-S

core

pERK Staining in Cancer

MAPKYT20G11D13.14.4E

MAPK-YT 20G11 D13.14.4E

Melanoma B4

Melanoma C2

Melanoma C6

MAPK-YT 20G11 D13.14.4E MAPK-YT 20G11 D13.14.4E

HEPG2 Untreated

HEPG2 + U0126

HEPG2 + Sorafenib

9101

Unt

reat

ed

U01

26

Sora

feni

b

Total ERK

pERK D13.14.4E

pERK 9101

pERK 20G11

pERK MAPK-YT

Clone Species Vendor Cat# Singly phos Dually phosMAPK-YT Mouse Sigma M8159 No Yes

20G11 Rabbit Cell Signaling 4376 Yes YesD13.14.4E Rabbit Cell Signaling 4370 Yes YesPolyclonal Rabbit Cell Signaling 9101 Yes Yes

Cancer Type Comparison # Analyzed Pearson'sBreast CA MAPKYT vs 20G11 18 0.995Colon CA MAPKYT vs 20G11 24 0.846Lung CA MAPKYT vs 20G11 24 0.987Melanoma MAPKYT vs 20G11 18 0.474

Breast CA 20G11 vs D13.14.4E 18 0.946Colon CA 20G11 vs D13.14.4E 24 0.988Lung CA 20G11 vs D13.14.4E 24 0.969Melanoma 20G11 vs D13.14.4E 18 0.747

Breast CA MAPKYT vs D13.14.4E 18 0.947Colon CA MAPKYT vs D13.14.4E 24 0.910Lung CA MAPKYT vs D13.14.4E 24 0.972Melanoma MAPKYT vs D13.14.4E 18 0.455

Pathology: Eric P. Olsen, M.D.; Immunohistochemistry: Sherif K. Girees, Gela A. Sia Acknowledgements