International Journal of Colorectal Disease

, 23:1041

Overexpression of FMNL2 is closely related to metastasis of colorectal cancer


  • Xi-Ling Zhu
    • Department of PathologySouthern Medical University
  • Li Liang
    • Department of PathologySouthern Medical University
    • Department of PathologySouthern Medical University
Original Article

DOI: 10.1007/s00384-008-0520-2

Cite this article as:
Zhu, X., Liang, L. & Ding, Y. Int J Colorectal Dis (2008) 23: 1041. doi:10.1007/s00384-008-0520-2


Background and aims

Formin-like 2 (FMNL2) is a member of diaphanous-related formins which can control the actin-dependent processes such as cell motility and invasion. In this study, we investigated the expression of FMNL2 in colorectal cancer (CRC) and its correlation with CRC metastasis.


Paraffin-embedded specimens of CRC (including 75 primary CRC tumors and 45 corresponding metastatic lymph nodes) and normal colorectal mucosa (30 samples) were immunostained with a FMNL2 antibody. Thirty-two paired snap-frozen tumor tissues and adjacent normal colorectal mucosa were subjected to real-time reverse-transcription polymerase chain reaction (RT-PCR). Six CRC cell lines (SW480, SW620, SW480/M5, LoVo, LS174T, and HT29) were assayed for FMNL2 expression by Western blotting and real-time RT-PCR. Their invasive abilities in vitro were determined by Boyden chamber assay.


The immunohistochemical analysis showed FMNL2 expression was considerably higher in CRC tumors and corresponding metastatic lymph nodes than in normal colorectal mucosa (P < 0.05, respectively). Elevated FMNL2 expression was significantly correlated with lymphatic metastasis of CRC (P < 0.05). Real-time RT-PCR analysis confirmed the results obtained by immunohistochemistry. At mRNA and protein levels, FMNL2 expression was substantially upregulated in cell lines derived from CRC metastases (SW620, SW480/M5, and LoVo) compared to ones derived from primary CRC (HT29, LS174T, and SW480; P < 0.05). In vitro cell invasive assay demonstrated that the former three cell lines had higher invasive ability than the latter cell lines.


FMNL2 may play an important role in the metastasis of CRC and may be a useful marker for metastasis of CRC.


Formin-like 2 (FMNL2)Diaphanous-related formins (DRFs)Colorectal cancer (CRC)MetastasisInvasion



formin-like 2


colorectal cancer




diaphanous-related formins


Colorectal cancer (CRC) is one of the most common forms of malignancy, and metastasis is the major cause of mortality in patients with CRC [1, 2]. In our previous study, we analyzed gene expression profiles of three CRC cell lines (SW480, SW620, and SW480/M5) with different metastatic potentials, to screen the metastasis-associated genes of CRC. SW480/M5, established at our laboratory by repeated selection in vivo, was a hepatic metastatic subline of low metastatic SW480. It manifested much higher metastatic potential than the parent line [3]. The high metastatic SW620 was derived from the lymph node metastases of CRC. About 298 unique genes were overexpressed in both SW620 and SW480/M5 compared with SW480. After bioinformatics analysis of those substantially upregulated genes in SW620 and SW480/M5, we hypothesized that formin-like 2 (FMNL2) might be a potential metastasis-associated gene of CRC.

FMNL2 is a member of the diaphanous-related formins (DRFs) that are key regulators of the actin cytoskeleton and act as effectors of Rho family guanosine triphosphatases (GTPases) [4, 5]. Recently, DRFs have emerged as a diverse family of ubiquitous, highly conserved multidomain proteins involved in a growing range of cellular processes such as filopodium formation, cell migration, cytokinesis, cell adhesion, and cell polarity [6, 7], which are frequently deregulated during pathological situations such as tumor cell transformation and metastasis [810]. Currently, some of DRFs, such as mDia1, have been reported to be critically involved in cancer cell progression and invasion [11, 12]. But, till now, the documents and researches about FMNL2 have been less available and its especially biological function in tumor biology has not been illustrated. Here, we evaluated the expression of FMNL2 in CRC cell lines and tissues, then sought to determine its function and significance in metastasis of CRC.

Materials and methods

Cell culture

The human CRC cell lines LoVo, LS174T, HT29, SW480, and SW620 were obtained from American Type Culture Collection. SW480/M5 was established at our laboratory. LoVo and LS174T cell lines were cultured in Roswell Park Memorial Institute (RPMI)-1640 medium (GIBCO BRL/Life Technologies, Gaithersburg, MD, USA), while the other cell lines were cultured in Dulbecco’s modified Eagle’s medium(GIBCO BRL/Life Technologies, Gaithersburg, MD, USA) with 100 IU/ml penicillin, 100 μg/ml streptomycin, and 10% heat-inactivated fetal bovine serum in a humidified 5% CO2 atmosphere at 37°C

Patients and tissue specimens

This study was conducted on a total of 150 paraffin-embedded samples collected retrospectively from archival material stored in the Department of Pathology at Southern Hospital (Guangzhou, China). Samples included 30 normal colorectal mucosa, 30 CRC tissues without lymph node metastases, 45 CRC tissues with lymph node metastases, and 45 corresponding lymph node metastases.

The fresh CRC tissues and the corresponding normal tissues were collected from 32 patients who underwent CRC resection at the Department of Surgery of the Southern Hospital between 2005 and 2007 and were different from those used for immunohistochemical analysis. These samples were collected immediately after resection, snap-frozen in liquid nitrogen, and then stored at −80°C until needed. Tumor samples were taken from vital areas of histopathologically confirmed cancers. As for corresponding normal tissue, we used adjacent unaffected mucosa, 2–3 cm distal to the resection margin from the same resected colorectal segment. All tissues obtained were reviewed by at least two experienced pathologists and examined for the presence of tumor cells. Pathological diagnosis and classification were made based on the system of the International Union Against Cancer (UICC). The research protocol was approved by the Ethics Committee at Nanfang Hospital, and all patients gave consent for the study.


The immunohistochemistry (IHC) was performed using streptavidin–peroxidase technique (Beijing Zhong Shan-Golden Bridge Biological Technology Co., Ltd., China). Briefly, after sections were deparaffinized and rehydrated, the antigens were repaired by high-pressure cooking using citrate buffer (pH 6.0). Endogenous peroxidases were blocked by incubating with 0.3% H2O2 for 15 min at room temperature. The sections were incubated for 30 min at 37°C with normal nonimmune serum and then incubated overnight at 4°C with anti-FMNL2 monoclonal antibody (Abnova Corporation, Taiwan; 1:50). After washing three times with phosphate-buffered saline (PBS), the sections were treated with biotin-conjugated second antibody before adding streptavidin–peroxidase. For color reaction, diaminobenzidine was used. If the IHC signals were present, the cytoplasm was stained brown. For negative controls, the antibody was replaced by PBS. The immunohistochemically stained tissue sections were scored separately by two pathologists blinded to the clinical parameters. The scoring approach in the assessment of FMNL2 immunostaining was used according to the protocol [13] which is relatively simple and reproducible. The staining intensity was scored as 0 (negative), 1 (weak), 2 (medium), and 3 (strong). Extent of staining was scored as 0 (0%), 1 (1% to 25%), 2 (26% to 50%), 3 (51% to 75%), and 4 (76% to 100%), according to the percentages of the positive staining areas in relation to the whole cancer area or entire section for the normal samples. The sum of the intensity and extent score was used as the final staining score. Tumors having a final staining score of ≥3 were considered to be positive. The staining of FMNL2 was assessed as follows: (−) means a final staining score of <3; (+) a final staining score of 3; (+ +) a final staining score of 4; and (+ + +) a final staining score of ≥5.

Total RNA extraction, complementary DNA synthesis

TRIzol reagent (Invitrogen Corporation) was used to extract total RNA from approximately 50–100-mg tissue or 107 cells. Genomic DNA was removed by DNase digestion, and purification of total RNA was made through RNeasy Mini Kit columns (Qiagen, Valencia, CA, USA) according to the manufacturer’s instructions. The complementary DNA (cDNA) was synthesized by oligo-dT-primed reverse transcription from 2 μg of total RNA using an access RT system (Promega Corporation, USA) according to manufacturer’s guide.

Real-time PCR

Real-time polymerase chain reaction (PCR) was performed using Mx3000P real-time PCR System (Stratagene Corporation) and SYBR PremixEx TaqTM (Takara Biotechnology Dalian Co., Ltd.), using the following thermal cycling profile: 95°C for 5 min, followed by 45 cycles of amplification (95°C for 40 s, 58°C for 40 s, 72°C for 40 s), followed by dissociation curve analysis to validate the amplification of a single product. Each reaction consisted of 10 μl 2 × ExTaq, 0.4 μl Dye II, 1 μl cDNA template, 0.4 μl each primer set (10 μM), and 7.8 μl nuclease-free water. Experiments were performed in triplicate in the same reaction. The quality of isolated RNA, its precise amount (based on absorbance) added in reverse transcription, as well as efficiency of reverse transcription were difficult to assess. Therefore, human β-actin gene was amplified as an internal control. Gene-specific primers were as follows: forward primer for FMNL2 5′-TAA TCA GCA TTA GCA TTT CTG AGG-3′ and reverse primer 5′-AGG AGA GTA AGG CCA GGT TCC-3′ (145 bp); forward primer for β-actin 5′-CTC CAT CCT GGC CTC GCT GT-3′ and reverse primer 5′-GCT GTC ACC TTC ACC GTT CC -3′ (268 bp). Relative quantity of target transcripts FMNL2 in each sample was expressed as N-fold differences relative to control, or 1×sample, and according to the equation: \(N = {\text{2}} - \left( {\Delta {\text{Ctsample}} - \Delta {\text{Ctcontrol}}} \right)\). ΔCt values of samples and control were determined by subtracting average Ct value of target transcripts FMNL2 from the average Ct value of β-actin gene [14]. Here, the corresponding normal tissue and SW480 became “control.” Each sample was tested three times and all other quantities were expressed as an n-fold difference relative to the corresponding normal tissue or SW480.

Western blotting

Cells were washed twice with cold phosphate-buffered saline and lysed on ice in radioimmunoprecipitation assay buffer (1× PBS, 1% NP40, 0.1% sodium dodecyl sulfate (SDS), 5 mM ethylenediaminetetraacetic acid, 0.5% sodium deoxycholate, 1 mM sodium orthovanadate) with protease inhibitors. The protein lysates were resolved in 10% SDS-polyacrylamide gel, electrotransferred to polyvinylidene fluoride membranes (Immobilon P, Millipore Corporation, USA), and blocked in 5% nonfat dry milk in Tris-buffered saline (100 mM NaCl, 50 mM Tris, 0.1% Tween-20, pH 7.5). Membranes were immunoblotted with anti-FMNL2 monoclonal antibody (Abnova corporation, Taiwan; 1:500) or anti-β-actin monoclonal antibody (Santa Cruz Biotechnology, Inc.; 1:1,000), overnight at 4°C, followed by their respective secondary antibodies conjugated to horseradish peroxidase. The signals were detected by enhanced chemiluminescence (Pierce Biotechnology, Inc.).

In vitro invasion assay

In vitro cell invasive assay was performed based on the principle of the Boyden chamber assay [15] with modifications. The top and bottom of the cell invasion chamber (Corning company) were separated by a polycarbonate filter with 8-μm pore size, over which a thin layer of ECMatrix was dried. Serum-free medium was added to the top chamber to rehydrate the ECMatrix layer for 2 h at room temperature. A single cell suspension of 10,000 tumor cells in serum-free medium was added to the top chamber, while 500 μl 10% fetal bovine serum–RPMI-1640 was added into the bottom chamber as a chemoattractant. The cells that had migrated through the membrane and had stuck to the lower surface of the membrane were fixed with methanol and stained with hematoxylin. Migration was determined by the count of the cells with a microscope at ×200 magnification. Five random visual fields were counted for each well, and the average was determined.

Statistical analysis

Statistical analysis was carried out using SPSS 13.0 for Windows. Results were presented as mean ± standard deviation (SD). Statistical differences of IHC were calculated using Kruskal–Wallis and Mann–Whitney test. The independent samples T test was used for real-time reverse-transcription (RT)-PCR. Statistical differences between groups were evaluated by one-way analysis of variance (ANOVA). Least-significant difference (LSD test) was used for multiple comparisons. All P values for multiple testing were corrected by Bonferroni correction. P < 0.05 was considered statistically significant.


FMNL2 was overexpressed in CRC-metastasis-derived cell lines compared with primary-CRC-derived ones.

To determine that FMNL2 was expressed in CRC cell lines with different metastatic abilities and illustrate the relationship with metastasis, real-time RT-PCR and Western blotting were performed in three primary-CRC-derived cell lines (SW480, LS174T, HT29), two lymph-node-metastases-derived cell lines (SW620, LoVo), and one liver-metastasis-derived cell line (SW480/M5).

Western blot results showed that a specific band of FMNL2 protein (124 kDa) was detected in six CRC cell lines. Higher level of the FMNL2 protein was detected in SW620, LoVo, and SW480/M5 cell lines derived from CRC metastases than three primary-CRC-derived ones (Fig. 1).
Fig. 1

Detection by Western blotting of the FMNL2 protein in six CRC cell lines. Blots were stripped and reprobed with a human β-actin probe to confirm equal loading. In total, 50 μg of total protein was loaded in each lane

For real-time RT-PCR detecting, relative quantities of FMNL2 mRNA in CRC cell lines were expressed as N-fold difference in relation to SW480 and normalized to the β-actin as a reference gene. FMNL2 mRNA levels of SW620, SW480/M5, LoVo, and LS174T cell lines were increased by 18-fold, 14.7-fold, 5.58-fold, and 1.17-fold compared with that of SW480, respectively. While FMNL2 mRNA level of HT29 was decreased by 0.43-fold. These findings were consistent with the results obtained by Western blotting (Fig. 2).
Fig. 2

Expression of FMNL2 mRNA in six CRC cells were measured using real-time RT-PCR. The start amount of cDNA isolated from SW480 was set to one; the relative quantification of FMNL2 mRNA in SW620, SW480/M5, and LoVo were higher than LS174T and HT29. The differences among them were significant by one-way ANOVA analysis

Next, we compared the invasive abilities of all cell lines by Boyden chamber assay. The results showed that SW480/M5 (223.0 ± 10.9), LoVo (193.0 ± 13.9), and SW620(175.6 ± 6.9) had higher invasive abilities than those of SW480 (116.2 ± 12.3), LS174T (85.2 ± 6.9), and HT29 (52.8 ± 9.8, P < 0.05, ANOVA, LSD test; Fig. 3). All the results suggested that FMNL2 may be related with invasion and metastasis of CRC.
Fig. 3

In vitro invasion assay was carried out to compare and quantify the invasiveness of SW480/M5, LoVo, SW620, SW480, LS174T, and HT29. Results were representative of three independent experiments, and bars represented the mean ± SD. The result indicated that LoVo, SW620, and SW480/M5 had a higher invasive ability than SW480, LS174T, and HT29 (P < 0.05, ANOVA, LSD test)

Overexpression of FMNL2 was related to lymphatic metastasis of CRC

To detect the expression of FMNL2 in CRC and analyze the relationship with CRC metastasis, we carried out a comprehensive immunohistochemical analysis which included normal colorectal mucosa, primary CRC tumor, and corresponding lymph node metastases. Analysis of FMNL2 expression in the arrayed 75 primary CRC tumors revealed that 80% of samples showed positive staining for FMNL2, with 4% of samples demonstrating strong(+ + +) intensity of FMNL2 expression and 31% moderate (+ +) intensity (Fig. 4b–c). By contrast, FMNL2 was rarely expressed in normal colorectal mucosa, and only 27% of samples demonstrated weak (+) intensity of FMNL2 expression (Fig. 4a), with none demonstrating strong (+ + +) or moderate (+ +) intensity. This indicated considerable overexpression of the FMNL2 in CRC tumors compared with normal tissues (P = 0.000, Table 1). A significant association was found between the intensity of FMNL2 expression and lymphatic metastasis of CRC. In 45 lymph node metastases, moderate and strong intensity was observed in 47% and 13% of samples, respectively (Fig. 4d–e), suggesting that higher expression of FMNL2 was observed in lymph node metastases than in primary CRC (P < 0.05, Table 1). The primary CRC tumors with lymphatic metastases were more frequently positive for FMNL2 than ones without lymphatic metastases (86% vs. 70%, P = 0.031, Table 2). In all cases, FMNL2 was mostly detected in the cytoplasm. There were no significant correlations between the expressions of FMNL2 and pathologic variables of the patients including patient gender, age, tumor differentiation, depth of invasion, or localization of tumor within the different sections of the colon. In addition, no correlation was observed between FMNL2 expressions and TNM–UICC tumor stages (Table 3).
Fig. 4

Immunohistochemical analysis of FMNL2 expression in CRC and adjacent normal colorectal mucosa. The sections were stained with DAB, counterstained with Meyer’s hematoxylin, and cover-slipped for microscopic observation. Brown areas were judged to be positive. a The normal colorectal mucosa negligibly stained for FMNL2. b Definite FMNL2 staining in the colorectal tumor cell cytoplasm of the CRC without metastasis. c Intense FMNL2 expression in the CRC with metastasis. d, e the lymph node metastases intensely stained for FMNL2. (ad original magnification ×200, e original magnification×100, d enlarged drawing of e)

Table 1

Expression of FMNL2 in normal colorectal mucosa, primary CRC tumors, and lymph node metastases

Sample (number)

Expression level, number (percentage)


+ +

+ + +

P value

Normal mucosa (30)

22 (73)

8 (27)




Primary CRC tumors (75)

15 (20)

34 (45)

23 (31)

3 (4)


Lymph node metastases (45)

9 (20)

7 (16)

22 (49)

7 (15)


aKruskal–Wallis test, P < 0.001

bMann–Whitney test, normal mucosa vs. primary CRC tumors (after Bonferroni correction, P < 0.025 was considered statistically significant)

cMann–Whitney test, primary CRC tumors vs. lymph node metastases

Table 2

The relationship between FMNL2 expression and metastasis of CRC

Sample (number)

Expression level, number (percentage)




P value

CRC without metastasis (30)

9 (30)

14 (47)

7 (23)



CRC with metastasis (45)

6 (13)

20 (44)

16 (36)

3 (6)

aMann–Whitney test

Table 3

Expression of FMNL2 in CRC and its correlation with pathoclinical characteristics

Clinical data (total number)

Expression level, number (percentage)






Male (41)

7 (17)

18 (44)

14 (34)

2 (5)


Female (34)

8 (24)

16 (47)

9 (26)

1 (2)


<60 (24)

4 (16)

10 (42)

9 (38)

1 (4)


≥60 (51)

11 (22)

24 (47)

14 (27)

2 (4)

Degree of differentiation of the primary tumors

Well (25)

5 (20)

11 (44)

8 (32)

1 (4)


Moderate (27)

6 (22)

12 (44)

7 (26)

2 (7)

Poor (23)

4 (17)

11 (48)

8 (35)


Depth of invasion

Muscularis extrema (44)

9 (20)

20 (45)

13 (30)

2 (5)


Serosa (31)

6 (19)

14 (45)

10 (32)

1 (3)

Sites of the cancer

Ascending colon (19)

3 (16)

10 (52)

6 (32)



Descending colon (20)

4 (20)

7 (35)

7 (35)

2 (10)

Sigmoid colon (17)

4 (24)

8 (47)

5 (29)


Rectum (19)

4 (21)

9 (47)

5 (26)

1 (5)

UICC stages

I (18)






II (12)






III (45)

6 (13)

20 (44)

16 (36)

3 (6)


aMann–Whitney test

bKruskal–Wallis test

To assess the molecular basis for FMNL2 overexpression, real-time RT-PCR was performed on 32 paired snap-frozen normal tissues and CRC tumors. It would be advantageous to analyze such paired samples, to clarify the degree of upregulation in individual cases. Relative quantities of FMNL2 mRNA were expressed as N-fold difference in relation to the corresponding normal tissue and normalized to the β-actin as a reference gene. FMNL2 mRNA expression was higher in primary CRC tumors with lymphatic metastasis than in ones without lymphatic metastasis (P = 0.009), supporting the immunohistochemical experiments. No significant correlation was found between FMNL2 mRNA expressions with depth of invasion or tumor differentiation. The results indicated that increased expression of FMNL2 was attributable to FMNL2 upregulation at the transcriptional level. Totally, FMNL2 expression was closely associated with lymph node metastasis of CRC.


The development of metastasis is a major cause of death in many human cancers. The metastasis of tumor cells is a very complex process and requires cancer cells to survive and proliferate outside their tissue of origin. Recently, growing evidence indicates that the actin-based processes, such as the cell polarity, cell migration, and filopodium formation, may also be important in advanced tumors and may correlate well with their invasion into adjacent tissues and the formation of metastases [9, 10]. And DRFs have emerged as potent regulators of actin dynamics in vitro and in cells [16, 17]. FMNL2 belongs to DRFs, containing a series of domains and functional motifs [18]. A unique and highly conserved C-terminal formin homology 2 (FH2) binds actin filament barbed ends and moves processively as these barbed ends elongate or depolymerize. The FH1 domain influences FH2 domain function through binding to the actin monomer-binding protein and profilin. The FH1 and FH2 domains are flanked by an array of regulatory domains, and FMNL2 is regulated by autoinhibition through interaction between the diaphanous inhibitory domain and diaphanous autoregulatory domain and activated by Rho GTPase binding to GTPase-binding domains [1921]. By bioinformatics analysis, people found that FMNL2 was expressed in several normal tissues and in diffuse-type gastric cancer, breast cancer, colorectal cancer, chondrosarcoma, melanoma, and glioblastoma [18]. But the molecular function of FMNL2 has not been reported.

To the best of our knowledge, this was the first time that FMNL2 had been characterized in CRC. In order to validate the relationship of FMNL2 with CRC metastasis, we investigated the expression of FMNL2 in six CRC cell lines with different metastatic potential. At both mRNA and protein levels, the expression of FMNL2 was increased in CRC-metastases-derived cell lines relative to primary-CRC-derived ones. It is well known that not all cells of an identical tumor have similar metastatic potential and only a few of them are highly metastatic[8]. The CRC cell lines derived from metastatic CRC are highly metastatic and may possess some special properties that can facilitate metastasis. Boyden chamber assay also validated that CRC-metastasis-derived cell lines had higher invasive abilities in vitro. FMNL2 could facilitate the metastasis of CRC cell lines.

Then, we evaluated FMNL2 expression by immunohistochemistry in 150 paraffin-embedded samples, followed by quantitative analysis of FMNL2 mRNA expression using real-time RT-PCR in 32 paired fresh samples (Fig. 5). The results of IHC showed that FMNL2 expression was significantly higher in primary CRC and lymph node metastases than in normal colorectal tissue, and the lymph node metastases tended to express FMNL2 more than the primary CRC. The expression of FMNL2 was upregulation in primary CRC tumors with lymph node metastases compared with ones without lymph node metastases. FMNL2 mRNA expression was higher in the lymph node metastasis-positive primary CRC tumors than in the negative ones (P = 0.009), supporting the immunohistochemical experiments. However, we did not correlate the FMNL2 expression with clinical prognosis. It would be an interesting focus for future studies. The above results suggested that overexpression of FMNL2 could enhance motor ability of tumor cells and facilitate them to invade into lymphatics or vasculature and therefore being a prerequisite for development of local lymph node and distant metastases.
Fig. 5

Quantitative mRNA analysis of FMNL2 by real-time RT-PCR. mRNA was isolated from fresh CRC tissues and the corresponding normal tissues (n = 32). The start amount of cDNA isolated from the corresponding normal tissues was set to one

In conclusion, our studies provided a new clue for illustrating the aggressive behavior of CRC cells. The special molecular mechanisms of FMNL2 in CRC metastasis awaits further studies to be elucidated. FMNL2 may be a useful marker for predicting tumor metastasis and a therapeutic target for the treatment of CRC patients with metastasis.


This work was supported by the National Natural Science Foundation of China Nos. 30400206, 30370649, 30670967, and 30770977.

Copyright information

© Springer-Verlag 2008