*2.5. Immunohistochemistry*

Mouse tumors were first fixed by formalin and embedded in para ffin. For hematoxylin & eosin (H&E) staining, the section was de-para ffinized and stained by the standard hematoxylin and eosin procedure to visualize tissue structures. For immunostaining, rabbit anti-Erk1/2 (Cell Signaling, Cat. No. 9102) and anti-pErk1/2 (Thermo Fisher, Waltham, MA, USA, Cat. No. 36-8800) antibodies were used. Sections were de-para ffinized using a standard procedure and blocked using 1.9% H2O2 in methanol at room temperature for 10 min. Sections were heated at 100 ◦C for 20 min in the antigen retrieval citra solution (BioGenex, San Ramon, CA) and blocked by the serum-free protein blocker (Dako, Glostrup, Denmark, Cat. No. X0909) for 5 min at room temperature. After incubation with the rabbit anti-Erk1/2 or anti-pErk1/2 antibody diluted at 1:50 overnight at 4 ◦C, biotin-conjugated anti-rabbit IgG (Jackson ImmunoResearch, West Grove, PA, USA, Cat. No. 111-066-144) was applied for 20 min at room temperature, followed by washing and incubation with streptavidin peroxidase (Biogenex, Fremont, CA, USA, Cat. No. HK330-9KT) for 15 min at room temperature. Antibody binding was visualized by the 3,3--Diaminobenzidine (DAB) chromogen system (Dako, Glostrup, Denmark). Subsequently, sections were counterstained by hematoxylin. Immunohistochemistry (IHC) quantification of representative tumor tissue sections was carried out with open source software Fiji ImageJ (NIH, Bethesda, MD, USA) using JPEG files. Mean optical density (OD) was calculated as the log average (maximal intensity/mean intensity) after image processing with color deconvolution and background subtraction.

## *2.6. Statistical Analysis*

The results are presented as a mean value plus or minus the standard deviation. Data were analyzed by GraphPad Prism 5.0. The *p*-values were determined by a Mantel–Cox test. A *p*-value under 0.05 was accepted as statistically significant.
