Patients and materials
A total of 625 CRC patients were selected at Shinshu University Hospital, Matsumoto, Japan from 2004 to 2014. PD-CRC was defined as the majority of the tumor being occupied by a PD-CRC component. All 29 PD-CRC cases were selected from the above patients. The clinicopathological features of these cases were evaluated.
Histopathology, immunohistochemical staining, and evaluation
All samples were fixed in 8% formaldehyde and paraffin tumor blocks were made. Tumor blocks of CRC were selected to prepare a tissue microarray (TMA). The most representative region of each CRC sample was selected. Tissue cores were punched out from each block using thin-walled 3-mm stainless steel needles (Azumaya Medical Instruments Inc., Tokyo, Japan), and arrayed on a recipient paraffin block. Serial sections of 4-μm thickness cut from these blocks were stained with hematoxylin and eosin (HE) or immunostained with antibodies against MLH1 (ES05, mouse monoclonal; dilution, 1:50; Agilent Technologies, Santa Clara, CA, USA), PMS2 (EP51, rabbit monoclonal; dilution, 1:40; Agilent Technologies), MSH2 (FE11, mouse monoclonal; dilution, 1:50; Agilent Technologies), MSH6 (EP49, rabbit monoclonal; dilution, 1:50; Agilent Technologies), β-catenin (mouse monoclonal; dilution, 1:500; Becton-Dickinson & Company, Franklin Lakes, NJ, USA), or CD8 (CD8/144B, mouse monoclonal; dilution 1:50; Dako, Copenhagen, Denmark). For antigen retrieval, sections were boiled in 0.05% citraconic anhydride solution pH 7.4 (Immunosaver; Nissin EM, Tokyo, Japan) for 45 min for MLH1, PMS2, MSH2, and MSH6, or microwaved in 0.45% Tris/5 mM EDTA for 25 min for β-catenin and CD8. Detection of MMR proteins was performed using a NovoLink polymer detection system (Leica Microsystems GmbH, Wetzlar, Germany) and that of β-catenin and CD8 was performed using an Envision detection system (Agilent Technologies) according to the manufacturers’ recommendations.
In accordance with a previous report [14], the immunohistochemical staining for MLH1, PMS2, MSH2, and MSH6 was scored as positive when a nuclear staining pattern was observed. In addition, at least 5% of tumor cells in individual tissue cores were required to be stained. Cases of PD-CRC were determined to have MMR protein deficiency when at least one of MLH1, PMS2, MSH2, and MSH6 was negative.
β-Catenin staining was evaluated as previously described [15]. The results were calculated as IHC scores, where IHC score = percentage of nuclear positive cells × staining intensity. Nuclear staining was classified into five grades from 0 to 4. We defined staining intensity as follows: 0, negative; 1, weak; 2, moderate; 3, strong; and 4, very strong. The nuclear β-catenin IHC score ranged from 0 to 400. The number of CD8+ TILs was calculated in the three most infiltrated fields for each case using an intermediate-power field.
LGR5 RNA in situ hybridization
Detection of LGR5 mRNA was performed with an RNAscope® kit (Advanced Cell Diagnostics, Hayward, CA, USA) according to the manufacturer’s instructions using unstained sample tissue slides. Briefly, tissue sections were pretreated by heating and protease was applied prior to hybridization with an LGR5-specific probe. The detailed procedure was described in a previous publication [16]. Brown dots present in the nucleus and/or cytoplasm were recognized as positive staining. LGR5 expression was quantified using a five-level scoring system recommended by the manufacturer (0, no staining; 1, 1–3 dots/cell; 2, 4–10 dots/cell; 3, >10 dots/cell; 4, >15 dots/cell with >10% of dots in clusters). The H-score was calculated as: (% of grade 1 cells × 1) + (% of grade 2 cells × 2) + (% of grade 3 cells × 3) + (% of grade 4 cells × 4). The overall H-score for each patient was calculated based on the H-score per high-power field (400× magnification). Furthermore, any cell with one or more dots was regarded as LGR5-positive.
Statistical analysis
Statistical analysis was performed using JMP version 10 (SAS Institute Japan, Tokyo, Japan). Spearman’s rank correlation coefficient analysis was used to assess correlations. The Wilcoxon rank sum test or chi-square test was also applied to assess statistical significance. A value of P<0.05 was considered significant.