International Journal of Colorectal Disease

, Volume 21, Issue 7, pp 625–631

5-Fluorouracil (5FU) treatment does not influence invasion and metastasis in microsatellite unstable (MSI-H) colorectal cancer


    • Department of Cancer Genetics, Kolling Institute of Medical ResearchRoyal North Shore Hospital and University of Sydney
  • Palaniappan Ramanathan
    • Department of Cancer Genetics, Kolling Institute of Medical ResearchRoyal North Shore Hospital and University of Sydney
  • Anthony Kaufman
    • Department of Anatomical PathologyPalms, Royal North Shore Hospital
  • Bruce G. Robinson
    • Department of Cancer Genetics, Kolling Institute of Medical ResearchRoyal North Shore Hospital and University of Sydney
  • Margaret Schnitzler
    • Department of Cancer Genetics, Kolling Institute of Medical ResearchRoyal North Shore Hospital and University of Sydney
Original Article

DOI: 10.1007/s00384-005-0071-8

Cite this article as:
Warusavitarne, J., Ramanathan, P., Kaufman, A. et al. Int J Colorectal Dis (2006) 21: 625. doi:10.1007/s00384-005-0071-8


Microsatellite instability is a recognised pathway of colorectal carcinogenesis responsible for about 15% of all sporadic colorectal cancers. Recent evidence has suggested that these tumours may not have the same response as microsatellite stable colon cancers to 5-fluorouracil (5FU)-based chemotherapy. The response to 5FU in four microsatellite unstable (MSI-H) cell lines was examined by cell viability assays and invasion assays. Flow cytometry was used to assess the effect of 5FU on MSI-H cell lines. In vivo response to 5FU was assessed by intraperitoneal injection of 5FU or control to 80 nude mice that had received intrasplenic injections of an MSI-H cell line KM12C prior to commencing treatment. There was inhibition of cell growth in MSI-H cell lines when treated with 5FU. There was no difference in invasiveness in the MSI-H cell lines when treated with 5FU. Primary tumours formed in 27 of the untreated and 25 of the 5FU treated mice (p=NS). There was a 36% reduction in splenic weight in those mice treated with 5FU (p<0.03). Metastases formed in 5 of the untreated and 9 of the treated mice (p=0.12). 5FU treatment of MSI-H tumours results in a reduction in growth but does not result in a reduction in invasion or metastasis.


Colorectal cancerMicrosatellite instability (MSI-H)Chemotherapy5-Fluorouracil (5FU)Metastasis





Microsatellite unstable


High-frequency microsatellite unstable


Microsatellite stable


Phosphate-buffered saline






Thymidine synthetase


There are at least two recognised pathways of colorectal carcinogenesis. About 15% of colorectal cancers demonstrate microsatellite instability, in which point mutations at microsatellite repeat loci result in widespread genomic instability [1, 2].

Tumours with high-frequency microsatellite instability (MSI-H) tend to be locally advanced and are less likely to metastasise than microsatellite stable (MSS) colon cancers. In addition, they have a right-sided colonic predeliction and demonstrate a better prognosis stage than that of MSS tumours [3, 4].

Prior to their identification as a separate subset of colon cancer, these tumours have been treated in the same way as MSS carcinomas, but there is now uncertainty regarding the role of cytotoxic chemotherapy in MSI-H colorectal cancer. Two recent retrospective studies have shown that patients with MSI-H tumours had an improved prognosis when treated with 5FU-based chemotherapy [5, 6]. In contrast, Ribic et al. have shown that 5FU treatment in the presence of MSI-H does not offer any benefit to the patient and may even reduce overall survival [7].

As the identification of MSI-H carcinomas as a discrete subset of colorectal cancer is relatively recent and randomisation in multicentre chemotherapy trials has not taken into account the microsatellite status of these tumours, several biases may have been introduced into the above-mentioned studies. Attempting to analyse the microsatellite status in previously randomised data could result in differences in patient characteristics between groups. Differences in tumour stage at presentation can also have a significant effect on the decision to treat, leading to potential biases in tumour selection.

Many in vitro studies have shown limited sensitivity of MSI-H colorectal cancers to 5FU [810]. It has been suggested that the lack of the DNA mismatch repair system, a feature of MSI-H tumours, leads to a halt in the G2 cell cycle arrest and apoptosis mediated by the mismatch repair system [11]. In addition, MSI-H tumours with a p53 mutation tend to be more responsive to 5FU treatment [12, 13]. Since most MSI-H tumours do not demonstrate mutations in p53, this could be viewed as a mechanism by which these tumours demonstrate resistance to 5FU. However, other studies have shown that the efficacy of 5FU in MSI-H tumours may be p53 independent [14].

Thus, there is conflicting in vitro and clinical evidence regarding the sensitivity of MSI-H colorectal tumours to 5FU. This is of particular importance as patients with MSI-H tumours who receive 5FU treatment may have a worse prognosis if treated [7]. In addition, the efficacy of 5FU as a palliative therapy in patients with metastatic MSI-H colorectal cancer has not been evaluated.

The aims of this study were to examine the in vitro and in vivo efficacy of 5FU in the treatment of a series of MSI-H and MSS cell lines and to assess the efficacy of 5FU treatment on metastasis of MSI-H colon cancer.

Materials and methods

We used the MSI-H cell lines KM12C (kind gift of Dr. Isiah Fiddler, MD, Anderson Cancer Centre, Houston, TX), HCT116, LOVO and LIM1215. These MSI-H cell lines have all been previously shown to have no mutations in p53 [12, 15, 16].

The cells were grown to subconfluent state in RPMI supplemented with 10% fetal calf serum and incubated in 5% CO2–95% air at 37°C. When cells reached 75% confluence, they were harvested, resuspended and counted. Experiments were conducted within 6 weeks following thawing of the cells from frozen stock.

5FU was handled according to the manufacturers’ instructions. For in vitro experiments, a concentration of 50 μM was used. All in vitro experiments were performed in triplicate and repeated. For in vivo experiments, the 5FU was diluted in phosphate-buffered saline (PBS) to a concentration of 1 mg/ml, then administered to Balb/c nude mice at a dose of 30 mg/kg.

Cell viability assays

Cells at a concentration of 1×104 per milliliter were treated with media containing 5FU or control media for 6 days. The Celltitre Glo Luminescent Cell Viability Assay (Promega Corporation, Madison, WI, USA) was used to assess cell survival.

Flow cytometry

MSI-H cells were treated with media containing 5FU or control media for 24 h. Bromodeoxyuridine (BrDU) was added for 1 h, and cells were harvested, fixed and treated with anti-BrDU Fluorescein (FITC) antibody and analysed in a flow cytometry. Previous cell cycle studies in our laboratory have determined the methodological error (coefficient of variance) for each of the cell cycle phases. These values were taken into account when assessing the significance of results obtained from this assay [17].

Invasion assays

Growth-factor-reduced Matrigel (BD Biosciences) was used to assess the invasive capability of the cells. Cells at a concentration of 1×105 per milliliter in media containing 5FU or control media were placed on the Matrigel. After 24 h, the Matrigel was removed and cells attached to the transwell chamber were stained and counted.

In vivo studies

A previously described, mouse metastatic model was used in this study [18, 19]. A single MSI-H cell line KM12C was chosen as representative and is a validated method in a comparative study [20]. Approval for animal studies was obtained from the Joint Royal North Shore/University of Technology of Sydney Animal Care and Ethics Committee. Balb/c nude mice were used for all in vivo experiments. Mice were aged 6–7 weeks at the time of commencement of experiments, were housed in sterile barrier cages and were fed autoclaved chow and water.

Under 2% isofluorane in a laminar flow hood, a 0.5-cm incision was placed in the left flank of each mouse. The muscle was divided and the spleen was exteriorised. The MSI-H cell line KM12C was used for this study. Cells which were previously suspended at a concentration of 1×105/100 μl sterile PBS were injected into the spleen using a 30-gauge needle. The muscle and skin were sutured, and the mice were then recovered and placed within the barrier cages. Ten mice were placed in each barrier cage. The mice were then monitored at regular intervals and randomised to receive 5FU 30 mg/kg or equivalent normal saline by intraperitoneal injection. Injections were performed on days 5, 7, 12, 14, 21, 28 and 42.

Forty mice served as controls and 41 received intraperitoneal injections of 5FU. Mice were randomly allocated to the two groups in groups of 10, and the investigator was blinded to the status of the treatment received until all mice were euthanased and tumours were counted and measured.

There were no adverse outcomes related to the intrasplenic injection or intraperitoneal injections of 5FU.

Eight weeks after the injection of cells, the mice were euthanased by overdosing on isofluorane, followed by removal of the liver and spleen. Tumours were counted and measured, and liver and spleen weights were recorded. All tumours were examined microscopically by a gastrointestinal histopathologist who was blinded to the treatment allocation.


Statistical analysis was performed with the Students’ t test for all in vitro studies. The Fisher’s exact test was used to analyse the data from the in vivo study on metastasis. A p value of less than 0.05 was considered significant.


Cell viability

There was a variable response to 5FU treatment in MSI-H cell lines, as shown in Fig. 1.
Fig. 1

Box plots for cell count as shown in relative light luminescence units (RLU) for control and 5FU-treated MSI-H cell lines. Circles represent outlying values

Flow cytometry

Cells treated with 5FU were arrested at S phase as shown by a significant increase in the percentage of cells in S phase that does not take up BrDU (Table 1 and Fig. 2).
Table 1

Flow cytometry data for selected MSI-H cell lines


G1 (%)

S (dividing) (%)

G2/M (%)

S (non-dividing) (%)
















HCT116 + 5FU










LIM1215 + 5FU




Fig. 2

Flow cytometry analysis for HCT116 (a) and HCT116 treated with 5FU (b) showing an increase in the non-dividing S phase when treated with 5FU

Invasion assays

Figure 3 summarises the difference in mean cell count per high-power field in treated and untreated cells. There was no significant difference in the rate of invasion in MSI-H colon cancers when treated with 5FU.
Fig. 3

Box plots depicting cell count per high-power field for control and 5FU-treated MSI-lines as seen in Matrigel invasion assay

In vivo model

Table 2 summarises the results. Of the 41 mice receiving 5FU, 25 (61%) developed primary splenic tumours and 9 (37.5%) of these developed liver metastases. Twenty-seven (65%) of the control mice developed primary tumours and 5 (18.5%) of these developed liver metastases. The primary tumour and liver metastasis rate for the control group was similar to that published in previous studies using the same cell line [19, 21].
Table 2

Summarised in vivo data

Cell type

KM12C, n=40

KM12C with 5FU treatment, n=41



27 (67%)

25 (61%)


Primary and metastasis

5 (19%)

9 (36%)


Spleen weight (g)




Liver weight (g)




The spleen weight was used as an indication of the volume of the tumour. In mice that did not develop tumours, there was no variation in the spleen weight despite treatment. In the control group, the average spleen weight was 2.3 g, and in the 5FU-treated group, the average weight was 1.5 g. This 36% reduction in the tumour bulk was significant (p<0.03). This reduction is similar to the reduction in cell survival demonstrated in our in vitro studies and is an indication that the drug delivery was appropriate.

These findings demonstrate that there was no significant difference in the primary tumour formation rate, but a reduction in tumour volume was achieved by 5FU treatment. The metastatic rate was not influenced by 5FU treatment.


In this study, we have shown that 5FU treatment reduces cell proliferation in MSI-H cell lines but does not reduce invasion or metastasis in MSI-H cells. In the in vivo study, there was some reduction in tumour size but no reduction in the rate of tumour formation in response to 5FU. Flow cytometry data confirmed that 5FU had an effect on the cycling of the cells by causing S phase arrest. However, the cell viability data showed that most MSI-H cells had limited response to 5FU. This could be due to a lack of induction of apoptosis as a result of the abnormal DNA mismatch repair system.

No previous studies have examined the effect of 5FU treatment on metastasis in MSI-H CRC. To assess this, we used both in vitro invasion assays and a previously described nude mouse metastatic model [1820]. In vitro and in vivo data in this study revealed that there was no decrease in invasion or the metastatic rate related to 5FU treatment in MSI-H cell lines. This is a significant finding given that the mainstay of treatment for metastatic colorectal cancer is 5FU-based chemotherapy.

MSI-H tumours have been shown to have a better prognosis and lower metastatic rate than MSS colonic cancers. The main treatment in colorectal cancer is surgery, but in the presence of lymph node metastasis, 5FU-based treatment leads to a 5–10% benefit in overall survival [22]. However, this evidence is based on studies that did not take into account the molecular basis of the tumour, and thus, no randomisation occurred according to microsatellite status. Several more recent studies have attempted to consider microsatellite status and the effect of chemotherapy with conflicting results [5, 7, 23]. The effect of 5FU on invasion and metastasis of MSI-H tumours has not been previously assessed.

The anti-metabolite 5FU is a prodrug which when metabolised can cause both RNA- and DNA-mediated cytotoxicities. DNA-induced damage is thought to be more important for chemotherapy, and RNA-directed effects for cytotoxicity [24]. DNA-directed cytotoxicity is usually a result of FdUTP incorporation into DNA or to the inhibition of thymidine synthetase (TS), a central enzyme in de novo pyramidine synthesis [9]. The RNA-directed effects of 5FU results in disruption of post-transcriptional modifications of tRNA and inhibition of the processing of rRNA [25].

The tumour suppressor p53 has been shown in vitro to be an important determinant of resistance to 5FU. Loss of p53 function can reduce cellular sensitivity to 5FU [26, 27]. However, in the setting of MSI-H tumours, the presence of p53 mutations has been shown to increase the sensitivity to 5FU treatment [12, 13]. The MSI-H cell lines used in this study had no mutations in p53 [16].

Cells deficient in mismatch repair (MMR) genes are resistant to many DNA damaging agents [11]. The MMR system is not only important in recognising and repairing damaged DNA, but it also has a prominent role in inducing apoptosis when DNA repair is not possible [11]. Thus, in the absence of a functioning MMR system, cells with MMR deficiency do not undergo G2 cell cycle arrest when exposed to chemotherapeutic agents, and therefore, they survive. The DNA damage induced by these agents may not be lethal as a result of the inability of the MMR system to repair the lesion or to induce apoptosis. MMR-deficient cells thus survive despite 5FU treatment and may develop further mutations that may cause resistance to other forms of treatment [11].

MMR-deficient colorectal cancers have been shown to be resistant to treatment with chemotherapeutic agents such as cisplatin and methylating agents [11, 28]. There is conflicting evidence as to the efficacy of 5FU in the setting of MMR deficiency.

Several in vitro studies have examined the efficacy of 5FU in MSI-H cell lines. These studies have shown that the direct effect of 5FU on MSI-H cell lines is modest with some, even showing a growth advantage in the presence of 5FU treatment [8, 10]. Furthermore, restoring MMR function reverses the resistance to 5FU and other agents, indicating that an intact mismatch repair system is essential for the cytotoxic effects of 5FU [8, 29]. Carethers et al. [8] demonstrated that HCT116 cells transfected with chromosome 3 do not demonstrate MSI-H and show an increased response to 5FU when compared to untransfected cells.

In a retrospective study, Hemminki et al. reported the findings of 5FU treatment in 11 MSI-H and 84 MSS Duke’s stage C colorectal cancers. There was a 90% 3-year disease-free survival in MSI-H tumours when compared with 43% 3-year disease-free survival in non-MSI-H tumours at a similar stage. Elsaleh et al. [23] in a descriptive study showed that there was a definite survival advantage when MSI-H tumours were treated with 5FU-based chemotherapy. The largest study of the effects of 5FU-based chemotherapy in MSI-H colorectal cancer demonstrated that patients receiving 5FU-based chemotherapy with MSI-H lesions did not have an improved prognosis when compared with those who did not receive 5FU [7]. This study analysed the MSI status of 570 colorectal tumours in patients randomised to receive 5FU or no chemotherapy and assessed survival based on treatment and MSI status. The overall prognosis in patients with MSI-H lesions who received treatment was worse than those who did not receive treatment. The hazard ratio for death was 2.14 in those with MSI-H tumours who received treatment compared with 0.72 in those with MSS tumours who received treatment. This may be related to the morbidity associated with the 5FU, although the authors urge caution in interpreting these data. As the treatment may not have any added benefit in the presence of metastatic disease, the morbidity of such a treatment must be carefully considered. It may be that the poorer prognosis observed by Ribic et al. [7] may be the effect of treatment-related morbidity in the presence of a non-responding tumour.

The findings of this study are consistent with the clinical observation that 5FU may not be as beneficial in MSI-H colorectal cancers as that in MSS colorectal cancers.

Copyright information

© Springer-Verlag 2006