Orthotopic implantation mouse model and cDNA microarray analysis indicates several genes potentially involved in lymph node metastasis of colorectal cancer

Abstract

In colorectal cancer (CRC) patients, metastasis to the regional lymph node (LN) is an important first step in the dissemination of cancers. To identify the genes possibly involved in LN metastasis of CRC, we analyzed LN metastases in an orthotopic implantation mouse model with 22 CRC cell lines using Matrigel, an extracellular matrix protein derived from mice sarcoma, and combined the data with gene expression profiles of cDNA microarray of those cell lines. With this implantation analysis, the incidence of LN metastasis was 60% in 228 orthotopically implanted mice and varied from 100% to 0% among the cell lines. KM12c and Clone A showed LN metastasis in all orthotopically implanted mice, but DLD-1, HCT-8, and SW948 did not show LN metastases at all. In contrast, the incidence of liver and lung metastasis in 22 CRC cell lines was 13% and 1%, respectively. Combining those data with cDNA microarray in vitro, we isolated 636 genes that were differentially expressed depending on the incidence of LN metastasis. Among those genes, the expression level of ring finger protein 125 (RNF125), previously known as an E3 ubiquitin ligase in T cell activation, was significantly different between primary tumors in Stage III CRC patients with LN metastasis and Stage II patients without LN metastasis. In conclusion, the orthotopic implantation mice model with Matrigel was useful, and we isolated candidate genes such as RNF125 that possibly play an important role in LN metastasis of CRC cells.

Figure 1

Orthotopic implantation of colorectal cancer (CRC) cells using Matrigel (a), and their macroscopic (b, c, e, f) and histological (d, g) characteristics. (a) The cecal wall bulges slightly after 100 µL injection of Matrigel solution containing 2.0 × 10⁶ CRC cells (white arrows). (b) Eight weeks after implantation of the LoVo cell line. The local tumor growth was recognized at the cecum. The mesenteric LN was not swollen. (c) Eight weeks after implantation of Clone A. The local tumor and a swollen LN were recognized. (d) Pathohistology of LN metastasis shown in (c) (hematoxylin–eosin [HE] staining, 400×). (e) Liver metastasis after implantation of HCT‐15 cells. (f) Liver translocated outside the body. The number of suspicious metastatic nodules was counted and the tissue was prepared with standard HE staining to confirm metastatic cancer cells. (g) Pathohistology of liver metastasis as shown in (f) (HE, 400×). Black arrows in (d) and (g) indicate invasive fronts of cancer cells. Ce, cecum; F, lymphoid follicle; I, ileum; L, liver; LM, liver metastasis; LN, lymph node; M, lymph node or liver metastasis; Sp, spleen; T, orthotopically implanted tumor onto cecal walls; TC, transverse colon.

Figure 2

cDNA microarray and reverse transcription–polymerase chain reaction (RT‐PCR) analyses with colorectal cancer cell lines. (a) cDNA microarray using the KM12c cell line. (b) RT‐PCR with cell lines. 1–6, high lymph node (LN) metastasis group; 7–11, low LN metastasis group; N.C., negative control. Internal control is glyceraldehyde‐3‐phosphate dehydrogenase (GAPDH) mRNA.

Figure 3

Semiquantitative reverse transcription–polymerase chain reaction analysis of RNF125 and CRIP1 mRNA expression levels in cell lines (a) and clinical samples (b) obtained from stage II and stage III patients. The relative expression levels of RNF125 or CRIP1 were defined by mRNA expression levels of ring finger protein 125 (RNF125) or cysteine‐rich intestinal protein 1 (CRIP1)/mRNA expression level of glyceraldehyde‐3‐phosphate dehydrogenase (GAPDH) ratio, represented as arbitrary units (AU). All data are the mean ± SD of values analyzed in triplicate using cDNA of cell lines or duplicate using cDNA of clinical samples.