Thymidylate synthase protein expression in colorectal cancer metastases predicts for clinical outcome to leucovorin-modulated bolus or infusional 5-fluorouracil but not methotrexate-modulated bolus 5-fluorouracil

C. Aschele1,+, D. Debernardis1, R. Bandelloni1, S. Cascinu2, V. Catalano3, P. Giordani3, S. Barni4, D. Turci5, G. Drudi6, S. Lonardi7, L. Gallo1, F. Maley8 and S. Monfardini7

1 Departments of Medical Oncology and Pathology, EO Ospedali Galliera, Genoa; 2 Azienda Ospedaliera di Parma, Parma; 3 Ospedale San Salvatore, Pesaro; 4 Azienda Ospedaliera Treviglio–Caravaggio, Treviglio; 5 Ospedale S Maria delle Croci, Ravenna; 6 Ospedale Infermi, Rimini; 7 Azienda Ospedaliera di Padova, Padua, Italy; 8 Wadsworth Center, New York State Department of Health, Albany, NY, USA

Received 14 January 2002; revised 30 April 2002; accepted 17 May 2002


    Abstract
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
Background:

Different 5-fluorouracil (5-FU) schedules and/or biochemical modulators may result in different mechanisms of cytotoxicity, potentially affecting the correlation between thymidylate synthase (TS) expression and the clinical response to the fluoropyrimidine.

Patients and methods:

TS levels were measured immunohistochemically on archival specimens of colorectal cancer metastases from 124 patients homogeneously treated in a series of clinical trials at our institutions with: (A) leucovorin (LV)-modulated infusional 5-FU (n = 48); (B) LV-modulated bolus 5-FU (n = 41); (C) methotrexate (MTX)-modulated bolus 5-FU (n = 35).

Results:

A statistically significant correlation between TS levels and the clinical response was observed with the regimens involving continuous infusion and/or LV modulation (response rate in patients with low and high TS: 66% versus 24%, P = 0.003, and 50% versus 0%, P = 0.0001, in group A and B, respectively). Conversely, TS levels failed to predict the clinical response within the group of patients treated with MTX-modulated bolus 5-FU (response rate 21% versus 13%, P = 0.50, with low and high TS, respectively). Consistently, the median time to progression/overall survival time in patients with low and high TS were 9 versus 6 months/19 versus 14 months (P = 0.009/0.035, group A), 8 versus 2 months/12 versus 6 months (P = 0.002/0.0006, group B) and 3 versus 2 months/12 versus 13 months (P = 0.14/0.74, group C).

Conclusions:

The correlation between intratumoral TS levels and the clinical response to 5-FU depends strongly on the schedule of administration/biochemical modulators that are used in different 5-FU regimens. These data strengthen the notion that different 5-FU schedules have different mechanisms of cytotoxicity.

Key words: biochemical modulators, colorectal cancer, 5-fluorouracil, response prediction, schedule of administration, thymidylate synthase


    Introduction
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
Thymidylate synthase (TS), a rate limiting enzyme in the DNA synthetic pathway [1], continues to be a major target for colorectal cancer chemotherapy. The TS inhibitor drug 5-fluorouracil (5-FU) has been the only agent showing consistent clinical activity against this malignancy for nearly 40 years [2]. During this time, infusional schedules have been developed [3, 4] and leucovorin (LV) has been combined with the fluoropyrimidine [5] to prolong and potentiate its TS inhibitory activity. These schedules provided the backbone for the combination regimens with irinotecan (CPT-11) or oxaliplatin that are currently used as conventional first-line chemotherapy for advanced colorectal cancer [68]. Recently, pure TS inhibitors were also introduced in the treatment armamentarium for large bowel tumors [9, 10] and oral fluoropyrimidines [11], which provide long-term TS inhibition without requiring infusion pumps and central venous lines, are being approved in many countries for the treatment of this disease.

The role of TS expression as a determinant of fluoropyrimidine cytotoxicity has been established in both preclinical and clinical models. In vitro, high levels of TS activity have been correlated with either acquired or intrinsic resistance to 5-FU or fluorodeoxyuridine [12, 13] and exposure of tumor cells to TS inhibitors has been shown to induce an acute rise in TS levels which may overcome drug-induced inhibition [14]. On clinical grounds, patients with low levels of TS gene or protein expression in their colorectal cancer metastases have shown response rates (RRs) to fluoropyrimidine-based chemotherapy that are 3–10 times higher compared with those obtained in patients with high TS levels in multiple recent studies [1519].

Considering the complex activating pathways and multiple intracellular targets of the fluoropyrimidine [20], this clear-cut correlation is somehow surprising. An improvement in overall response prediction was indeed obtained with the simultaneous assessment of TS and different enzymes involved in 5-FU metabolism [21, 22]. As to the target, we have previously shown that TS inhibition is the dominant site of action with prolonged exposures to the fluoropyrimidine, while incorporation into RNA plays a major role with pulse dosing [23]. Biochemical modulators may also shift the mechanism of action of 5-FU towards a specific cellular target: LV selectively potentiates TS inhibition while methotrexate (MTX) has been shown to enhance 5-FU incorporation into RNA [2426].

Response prediction based on the assessment of intratumoral TS expression may thus be limited to infusional schedules, in particular with LV modulation, while the clinical activity of bolus regimens may be independent from the level of expression of this enzyme. To test this hypothesis, we have used an immunohistochemical technique suitable for archival material to measure the level of TS protein expression in colorectal cancer metastases from a series of patients homogeneously treated with one of three different 5-FU regimens (bolus or infusion, LV or MTX modulation). These measurements were then retrospectively correlated with patients’ characteristics and a series of parameters of clinical outcome [RR, time to progression (TTP), overall survival (OS)] both on data pooled from all chemotherapy regimens and separately within each treatment group.


    Patients and methods
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
Patients
The study population consisted of 124 patients [68 males and 56 females; median age 62, range 31–81, years; Eastern Cooperative Oncology Group performance status (PS) 0/1/2: 62, 39 and 23 patients, respectively] with metastatic, recurrent or locally advanced unresectable colorectal cancer, homogeneously treated with infusional 5-FU (regimen A), LV-modulated bolus 5-FU (regimen B) or MTX-modulated bolus 5-FU (regimen C) in a series of prospective phase II and III clinical trials at our institutions [2730]. These three regimens were chosen for this study to enable the evaluation of both different 5-FU schedules and different biochemical modulators. In addition, clinical trials testing these regimens had been performed previously at our institutions, allowing us to gather easily high quality clinical data and to retrieve tumor sections from measurable metastatic lesions. Unmodulated bolus 5-FU was not investigated because clinical trials testing bolus 5-FU without biochemical modulators generally closed their accrual before 1990 and its use has been progressively abandoned even in clinical practice during the last 10 years, making it difficult to retrieve tissue samples from metastatic tumor lesions.

Patients treated with regimen A received a 3-week continuous infusion of 5-FU modulated by weekly LV alternated to bi-weekly administrations of bolus 5-FU modulated by MTX (given 24 h earlier). Regimen B consisted of daily administrations of bolus 5-FU modulated by LV repeated for five consecutive days every 4 weeks. The third regimen (C) consisted of bi-weekly courses of bolus 5-FU modulated by MTX (given 24 h earlier). Further treatment details have been previously published [2730]. The cases included in this study represent all the patients from a series of clinical trials of these regimens for whom tumor sections from a measurable metastatic lesion were available. The rates of objective responses according World Health Organization criteria [31], observed in the three groups of patients analyzed in this study (48%, 17% and 17% in group A, B and C, respectively) are similar to those originally reported for the entire patient populations treated with the corresponding 5-FU regimens (48%, 41% and 33% with the infusional regimen [2729], 10% with LV-modulated bolus 5-FU [30] and 14% with MTX-modulated bolus 5-FU [29]).

The main patient and tumor characteristics of each treatment group are reported in Table 1. One-hundred and five (85%) patients had a primary tumor located in the colon and 19 (15%) in the rectum. Fifty-nine patients had metastatic disease limited to the liver, 22 had liver metastases plus other sites of disease and 43 had extra-hepatic disease. At baseline, the median area of the measured tumor lesions was 25 cm2 (range 2–1234 cm2) and the median white blood cell count (WBCC) was 7515 cells/mm3 (range 4200–14 900 cells/mm3). The median baseline serum levels of carcinoembryonic antigen (CEA), lactic dehydrogenase (LDH) and alkaline phosphatase (ALP) were 17.2 ng/ml (range 0.7–17 400 ng/ml), 312 U/l (range 36–3652 U/l) and 219 U/l (range 60–1614 U/l), respectively. Fifteen of 124 patients had previously received adjuvant treatment following resection of the primary tumor (14 chemotherapy, one pelvic radiation therapy).


View this table:
[in this window]
[in a new window]
 
Table 1. Patient and tumor characteristics of each treatment group [regimen A: infusional 5-fluorouracil (5-FU); regimen B: leucovorin-modulated bolus 5-FU; regimen C: methotrexate-modulated bolus 5-FU]
 
Immunohistochemistry
TS protein expression was evaluated immunohistochemically [32], using a rabbit polyclonal antibody to recombinant human TS, as previously described [16]. Archival, formalin-fixed, paraffin-embedded tissue samples from colorectal cancer metastases (n = 99: liver, 79; lung, 12; other, eight), colorectal cancer local recurrences (n = 23; pelvic mass, four; peritoneum, 18; abdominal wall, one) or locally advanced colorectal tumors (n = 2) were used for this study. Ninety-six patients had a single site of disease that included both the lesion analyzed for TS expression and the target lesions that were clinically monitored. Of 41 patients with multiple sites of disease, eight patients with pulmonary and hepatic metastases and five patients with pelvic masses and liver metastases had TS assessed only on liver metastases. Similarly, liver metastases were the only sites of TS assessment in seven patients with hepatic disease and peritoneal implants, bone metastases or nodal metastases. Two patients with liver and hilar node metastases had only the lymph nodes analyzed. TS expression was assessed only on pelvic masses in six patients with concomitant lung disease. The intensity of TS immunostaining was assessed independently by two investigators (D.D. and R.B.) blinded to the clinical data and scored based on a visual grading scale ranging from 0 (undetectable staining) to 4 (very high intensity of staining). Intensity levels 0–2 were then grouped together and considered low expression while levels 3 and 4 were considered high expression. The agreement in TS evaluation between the two observers was greater than 90%. In the seven cases of disagreement, a final score was determined by consensus after re-examination. When heterogeneous levels of TS expression were found within a tumor (in multiple sections from different paraffin-embedded blocks of the same tumor), the level of TS expression of that lesion was defined according to the highest TS score that was recorded.

Statistics
Analyses were conducted on data pooled from all chemotherapy regimens and separately within each treatment group. The association between TS expression and dichotomized clinical variables, including response to treatment, was analyzed using the chi-square or Fisher’s exact test as indicated by the data. The baseline serum levels of CEA, LDH, ALP and WBCC were compared between patients with low and high levels of TS expression using the Student’s t-test for unpaired samples. Progression-free and OS curves were constructed using the Kaplan–Meier method and differences were assessed by the log-rank test.


    Results
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
Sixty-four of 124 patients (52%) showed high levels of TS expression, as defined by a TS score equal to 3 and 4 in 47 and 17 cases, respectively. Among the patients with a low level of TS expression (60 of 124, 48%), 20 had a negative staining for TS, 18 displayed a staining score of 1 and 22 had a score of 2.

The mean and median levels of immunoreactivity (TS score) were similar among the three patient groups as shown in Table 2. Consistently, the proportion of patients with high levels of TS expression was 44%, 66% and 46% in groups A, B and C, respectively.


View this table:
[in this window]
[in a new window]
 
Table 2. Levels of thymidylate synthase (TS) protein expression in each treatment group [regimen A: infusional 5-fluorouracil (5-FU); regimen B: leucovorin-modulated bolus 5-FU; regimen C: methotrexate-modulated bolus 5-FU]
 
No significant differences in the level of TS expression were found with respect to age, sex, PS, previous adjuvant chemotherapy, baseline WBCC and serum CEA and LDH levels (Table 3). As previously observed [16], TS expression seemed to be related to initial tumor bulk. Fifty-six percent of the patients with a measured tumor area greater than the median value for the study had a high TS level compared with 34% of the patients with an initial tumor area of less than this value (P = 0.04). High levels of TS expression were also associated with baseline ALP levels greater than the median value for the study (P = 0.01, Table 3). The proportion of cases with high TS levels observed in the group with a colon primary was twice that in patients with a primary rectal tumor (56% versus 26%, P = 0.002). In contrast, only a marginal association was found between the level of TS expression and the dominant site of metastatic disease, with pelvic recurrences showing higher TS levels compared with other sites (Table 3).


View this table:
[in this window]
[in a new window]
 
Table 3. Patient and tumor characteristics according to the level of thymidylate synthase (TS) protein expression
 
Overall, 36 of 124 patients (29%) achieved an objective response with a strong inverse correlation between intratumor TS levels and treatment outcome (RR 48% versus 10%, in patients with low and high TS, respectively; P <0.00001). When separate analyses were performed within each treatment group (Figure 1), this correlation was evident only for patients treated with regimen A {RR in patients with low and high TS: 66% [95% confidence interval (CI) 79% to 53%] versus 24% [95% CI 36% to 12%]; P = 0.003} and regimen B [RR: 50% (95% CI 65% to 35%) in patients with low TS versus 0% in patients with high TS; P <0.0001]. All the complete responses (CRs) (n = 7) observed with these regimens were in patients with low TS levels while 23 of 28 progressions were observed in patients with high TS. Conversely, intratumor TS levels failed to predict for the clinical response in the group of patients treated with MTX-modulated bolus 5-FU [RR: 21% (95% CI 39% to 3%) versus 12% (95% CI 28% to 0%); P = 0.50]. Among 19 treatment failures observed with this regimen, nine were in patients with low TS and 10 in the group with high TS levels. The only CR was observed in the group with high TS.



View larger version (13K):
[in this window]
[in a new window]
 
Figure 1. Response to different 5-fluorouracil (5-FU) regimens according to the level of thymidylate synthase (TS) protein expression in tumor metastases. TS levels were determined immunohistochemically with a rabbit polyclonal antibody to recombinant human TS on paraffin sections from colorectal cancer metastases as described in Materials and methods. Columns indicate the rates of objective responses [complete response (CR) plus partial response (PR)] achieved in patients with low (shaded columns) and high (open columns) TS levels. Bars represent 95% confidence intervals. SD, stable disease; P, progression. The association of low TS levels with a higher response rate is significant for patients treated with regimens A (66%; P = 0.003; n = 48) and B (50%; P <0.0001; n = 41) but not for those receiving methotrexate-modulated bolus 5-FU (regimen C; P = 0.50; n = 35).

 
Similar results were obtained analyzing the level of TS expression by responsiveness to chemotherapy within each treatment group (Figure 2). A significantly lower TS immunoreactivity was observed among the patients achieving an objective response, both in the group treated with infusional 5-FU and in the group who received LV-modulated bolus 5-FU (mean/median TS score in responders versus non-responders: 1.3/1 versus 2.5/3, P = 0.01, for regimen A and 0.7/0 versus 2.9/3, P = 0.000004, for regimen B; Student’s t-test for unpaired samples). In contrast, patients treated with MTX-modulated bolus 5-FU displayed similar levels of immunoreactivity between responders and non-responders (mean/median TS score: 1.8/2 versus 2.3/2, respectively; P = 0.50; Student’s t-test for unpaired samples).



View larger version (11K):
[in this window]
[in a new window]
 
Figure 2. Thymidylate synthase (TS) immunoreactivity in metastatic tumor samples from patients achieving an objective response (filled columns) or failing to respond (open columns) to three different 5-fluorouracil (5-FU) regimens (regimen A: infusional 5-FU, n = 48; regimen B: leucovorin-modulated bolus 5-FU, n = 41; regimen C: methotrexate-modulated bolus 5-FU, n = 35). Columns indicate the mean level of TS immunoreactivity as determined with a rabbit polyclonal antibody to recombinant human TS on paraffin sections from colorectal cancer metastases as described in Materials and methods. Bars = standard deviation. The difference in immunoreactivity between responders and non-responders is significant for patients treated with regimens A and B but not for regimen C (P = 0.01, P = 0.000004 and P = 0.50, respectively; Student’s t-test for unpaired samples).

 
At the time of writing, all patients on this study were off-treatment, 119 of 124 had died and 115 had progressed. The median follow-up time is 2.6 years. Figure 3 shows the relative influence of intratumoral TS levels on TTP for patients treated with different 5-FU regimens. Patients with low TS treated with regimens A and B had a significantly longer TTP compared with patients with high TS levels (median values: 9.4 versus 6.0 months, P = 0.009 and 8.0 versus 2.0 months, P = 0.002, respectively; Figures 3A and 3B). However, for patients treated with regimen C the median TTP was 3.0 months (low TS) versus 2.0 months (high TS) (P = 0.14; Figure 3C). The impact of TS level on survival following chemotherapy is analyzed in Figure 4. Patients with low TS treated with regimen A had a median survival time of 18.5 months compared with 14.3 months for patients with high TS levels (P = 0.02; Figure 4A). Similarly, for patients treated with regimen B low levels of TS expression were associated with a significantly longer median survival time (12.5 versus 6.0 months; P = 0.0006; Figure 4B). In contrast, Figure 4C shows that patients who received MTX-modulated bolus 5-FU had similar median survival times which were independent of the level of TS expression (11.7 versus 12.5 months; P = 0.74).



View larger version (16K):
[in this window]
[in a new window]
 
Figure 3. Kaplan–Meier progression-free curves according to intratumor thymidylate synthase (TS) content of patients with advanced colorectal cancer treated with three different 5-fluorouracil (5-FU) regimens: (A) infusional 5-FU; (B) leucovorin-modulated bolus 5-FU; (C) methotrexate-modulated bolus 5-FU. The difference between the two curves is statistically significant for regimens A and B (P = 0.009 and P = 0.002, respectively; log-rank text) but not for regimen C (P = 0.14; log-rank text). TTP, time to progression.

 


View larger version (16K):
[in this window]
[in a new window]
 
Figure 4. Kaplan–Meier survival curves according to intratumor thymidylate synthase (TS) content of patients with advanced colorectal cancer treated with three different 5-fluorouracil (5-FU) regimens: (A) infusional 5-FU; (B) leucovorin-modulated bolus 5-FU; (C) methotrexate-modulated bolus 5-FU. For regimens A and B, the advantage in survival time is significant for patients with tumors displaying a low TS immunoreactivity (P = 0.02 and P = 0.0006, respectively; log-rank text). For regimen C, survival time was not statistically different between patients with low and high TS levels (P = 0.74; log-rank text). OS, overall survival.

 

    Discussion
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
The level of intratumoral TS expression represents the most promising predictor of response to fluoropyrimidine-based chemotherapy in advanced colorectal cancer. Five recent studies consistently demonstrated an inverse correlation between the level of expression of this enzyme measured in colorectal cancer metastases and the clinical response to either 5-FU or 5-FU deoxyribonucleoside [1519]. In these studies, the predictive role of TS was strengthened by the independent predictive value demonstrated in a multivariated analysis [16], the longer survival shown by patients with low TS levels in two out of five cases [15, 16] and the consistent results obtained by independent groups using different techniques to quantify TS expression [15, 16].

The present study confirms the predictive value of TS expression in a large series of 124 patients but clearly indicates that the results are limited to specific 5-FU regimens. Patients with low TS levels had a significantly higher probability of response to infusional and/or LV-modulated 5-FU (regimens A and B) with longer TTP and OS times compared with patients with high TS. In contrast, intratumoral TS levels failed to predict clinical outcome in a group of patients treated with MTX-modulated bolus 5-FU (regimen C). Although the RR to this regimen was slightly higher in patients with low compared with those with high TS, this difference was not statistically significant, with substantial overlapping of the CIs, and was substantially smaller compared with that observed with the other regimens. Furthermore, TS level segregated almost absolutely both complete responders and progressors following treatment with regimens A and B while for regimen C treatment failures were equally distributed in the groups with low and high TS and the only CR was observed in the latter cohort. Consistently, for patients treated with MTX-modulated bolus 5-FU, TTP and OS following chemotherapy were found to be independent of the level of TS expression. Finally, the mean level of TS expression was similar between responders and non-responders in the group of patients treated by MTX-modulated bolus 5-FU while TS immunoreactivity was significantly lower in responders compared with non-responders treated with the other two regimens.

Since this was a retrospective study, the distribution of patients among the three treatment cohorts was not controlled and patients and tumor characteristics may not have been evenly balanced between the three treatment groups. However, all of the patients had been previously accrued in a series of prospective clinical trials with similar eligibility requirements including the need for measurable disease. Indeed, the main patient and tumor characteristics were fairly balanced among the three groups except for a higher proportion of patients with intra-abdominal recurrences in group B and a higher percentage of patients who had received previous adjuvant chemotherapy in group C. In particular, the median age ranged from 60 to 64 years, the median PS from 0 to 1 and the proportion of patients with liver-limited disease from 34% to 58%. In addition, group C patients had the same median PS, the same relative proportion of patients with a primary rectal tumor and a similar proportion of patients with liver metastases as compared with patients treated with regimen A. It is therefore unlikely that the different dependence on TS expression observed for the activity of the three regimens investigated in this study depends on the distribution of patient characteristics in the three treatment groups.

These data indicate that TS inhibition does not represent a major target for MTX-modulated bolus 5-FU. This is consistent with preclinical data showing that the mechanism of action of 5-FU depends on the schedule of administration. We have previously demonstrated that TS inhibition is the main mechanism of action with prolonged exposures to low doses of the fluoropyrimidine, while incorporation into RNA becomes the main site of action with pulse-dosing [23, 33, 34]. In addition, while LV specifically potentiates the TS inhibitory activity of 5-FU by stabilizing the lethal ternary complex that consists of fluorodeoxyuridylate, 5,10-methylenetetrahydrofolate and the enzyme itself [24], pretreatment with MTX has been shown to enhance the incorporation of the fluoropyrimidine into RNA by increasing phosphoribosylpyrophosphate pools and, therefore, the activation of 5-FU to fluorodeoxyuridine triphosphate (FUTP) [25, 26]. A recent clinical observation showing a significant decrease in TS and ß-actin mRNA following the administration of bolus 5-FU lends support to this contention [35]. Our findings contrast with the results of a recent study that showed a correlation between intratumoral TS levels and the clinical response to a regimen of MTX-modulated bolus 5-FU [36]. However, in that study TS expression was assessed on primary colorectal cancer rather than metastatic samples. This may be a major difference as we have previously shown that TS levels in primary tumors are neither related to TS expression in the corresponding metastases nor predictive of their response to palliative chemotherapy [37]. In addition, in the study by Paradiso et al. [36], patients with negative TS immunostaining were compared with patients with any grade of positive immunoreactivity while in the present study patients with negative and low immunoreactivity were more conventionally grouped and compared with patients with high levels of immunostaining. Finally, even in that study the predictive value of TS expression for the clinical response to MTX-modulated 5-FU was small with a borderline statistical significance and a lack of correlation with TTP and OS. This latter point implies a weak potential to predict tumor response as it has been demonstrated that objective responses are associated with increased survival in this disease [38].

Surprisingly, LV-modulated bolus 5-FU proved to be the regimen where the difference in RR between patients with low and high TS was more marked. In particular, the prediction of high TS levels for resistance was absolute in that all the patients with high TS failed to achieve an objective response. Based on previous preclinical data showing that bolus 5-FU acts mainly via incorporation into RNA [23], TS levels were not anticipated to predict the clinical response to this regimen. This unexpected pattern of prediction may have multiple explanations. LV modulation may shift the mechanism of action of bolus 5-FU from incorporation into RNA to TS inhibition. Stabilization of the ternary complex may in fact result in prolonged retention of inhibition even after a short-term exposure to the drug [39]. The significant increase in the level of clinical activity with a doubling of the rates of objective tumor responses obtained with the addition of LV to bolus 5-FU along with the changes observed in the spectrum of toxicity [34] lend support to the notion that LV-modulated and unmodulated bolus 5-FU may have distinct biochemical targets. However, based on the present data it is impossible to determine whether or not TS levels also affect the clinical activity of unmodulated bolus 5-FU and this aspect may be difficult to investigate even in future studies. The use of bolus 5-FU without biochemical modulators has in fact been progressively abandoned in the last 10 years and metastatic tumor samples from patients treated with this regimen may be difficult to retrieve. An additional explanation for the strong TS dependence of this regimen is that bolus 5-FU is administered for five consecutive days with a possible further increase in the duration of TS inhibition.

Based on preclinical data, the highest degree of correlation was expected for infusional schedules. Conversely, only a three-fold difference in RR was observed in the present study between patients with low and high TS treated with regimen A, with 20% of the patients with high TS levels actually achieving an objective response. This may reflect the hybrid nature of the infusional regimen used in this study. LV-modulated infusional 5-FU was in fact alternated with MTX-modulated bolus 5-FU [28]. As previously discussed, the antitumor activity of MTX-modulated bolus 5-FU has been shown to depend on the incorporation of 5-FU metabolites into RNA and it may be unaffected by TS levels [34]. The 20% RR observed in patients with high TS levels may thus reflect the responsiveness to the bolus part of this hybrid regimen. An absolute correlation between high TS levels and resistance, with a threshold value for TS expression above which no responses were obtained, was in fact observed in a previous study investigating a pure LV-modulated infusional 5-FU regimen [15]. Consistent with this hypothesis, the proportion of patients with high TS achieving an objective response in the present study was similar in the group of patients treated with the hybrid regimen and in the group who received MTX-modulated bolus 5-FU.

These data have important clinical implications. The specificity for infusional and/or LV-modulated 5-FU regimens emphasizes the predictive value of intratumoral TS expression for response to the fluoropyrimidine. If the correlation between TS levels and the clinical response merely reflected the more aggressive biological behavior of tumors with high TS, the same correlation would have been observed with every 5-FU regimen and even with other drugs. In line with our results, TS levels in colorectal cancer metastases failed to predict the clinical response to CPT-11, confirming the high drug-specificity of this correlation [40]. Based on the present data, TS quantification cannot be used to predict the clinical response to MTX-modulated bolus 5-FU. Despite a decreased popularity in advanced disease, this regimen is ‘alive and well’ and a randomized comparison with the Machover regimen involving over 2000 patients has been recently completed in the adjuvant setting [41]. In addition, trimetrexate (TMQ) may soon replace MTX as a modulator of bolus 5-FU resulting in increased activity and lower toxicity [42]. Putative markers of response to this regimen include the level of expression of FUTP-catabolizing enzymes and the assessment of early RNA dysfunction following 5-FU administration [35, 43]. Finally, the present findings support the use of MTX-modulated bolus 5-FU in patients with TS-related resistance to the fluoropyrimidine. In particular, patients failing infusional 5-FU may be treated with combination regimens including MTX- or TMQ-modulated bolus 5-FU. New combinations incorporating CPT-11 or oxaliplatin into these regimens should therefore be developed.

In conclusion, this study demonstrates that the prediction of clinical outcome based on the assessment of intratumor TS expression is limited to specific 5-FU regimens. This study is also the first to confirm on clinical grounds that TS is not the main target for MTX-modulated bolus 5-FU, strengthening the notion that the mechanism of action of the fluoropyrimidine depends on the schedule of administration and the nature of the biochemical modulators. This may be of major clinical importance to optimize the development of combination regimens that integrate 5-FU and new drugs like CPT-11 and oxaliplatin. The importance of the schedule of 5-FU administration in determining the toxicity of these combination regimens is indeed already emerging [44].


    Acknowledgements
 
Supported by grants Associazione Italiana per la Ricerca sul Cancro 2001 and National Cancer Institute/Department of Health and Human Services (NCI/HHS) CA 44355.


    Footnotes
 
+ Correspondence to: Dr C. Aschele, Department of Medical Oncology, EO Ospedali Galliera, Mura delle Cappuccine 14, 16128 Genoa, Italy. Tel: +39-010-5634-502; Fax: +39-010-5634-500; E-mail: carlo.aschele{at}galliera.it Back


    References
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
1. Danenberg PV. Thymidylate synthetase: a target enzyme in cancer chemotherapy. Biochim Biophys Acta 1977; 473: 73–92.[ISI][Medline]

2. Bleiberg H. Colorectal cancer—is there an alternative to 5-FU? Eur J Cancer 1997; 33: 536–541.[Medline]

3. Lokich JJ, Ahlgren JD, Gullo JJ et al. A prospective randomized comparison of continuous infusion fluorouracil with a conventional bolus schedule in metastatic colorectal carcinoma: a Mid-Atlantic Oncology Program study. J Clin Oncol 1989; 7: 425–432.[Abstract]

4. de Gramont A, Bosset JF, Milan C et al. Randomized trial comparing monthly low-dose leucovorin and fluorouracil bolus with bimonthly high-dose leucovorin and fluorouracil bolus plus continuous infusion for advanced colorectal cancer: a French intergroup study. J Clin Oncol 1997; 15: 808–815.[Abstract]

5. Advanced Colorectal Cancer Meta-Analysis Project. Modulation of fluorouracil by leucovorin in patients with advanced colorectal cancer: evidence in terms of response rate. J Clin Oncol 1992; 10: 896–903.[Abstract]

6. Douillard JY, Cunningham D, Roth AD et al. Irinotecan combined with fluorouracil compared with fluorouracil alone as first-line treatment for metastatic colorectal cancer: a multicentre randomised trial. Lancet 2000; 355: 1041–1047.[ISI][Medline]

7. Saltz LB, Cox JV, Blanke C, Rosen LS et al. Irinotecan plus fluorouracil and leucovorin for metastatic colorectal cancer. Irinotecan Study Group. N Engl J Med 2000; 343: 904–914.

8. de Gramont A, Figer A, Seymour M et al. Leucovorin and fluorouracil with or without oxaliplatin as first-line treatment in advanced colorectal cancer. J Clin Oncol 2000; 18: 2938–2947.[Abstract/Free Full Text]

9. Rustum YM, Harstrick A, Cao S et al. Thymidylate synthase inhibitors in cancer therapy: direct and indirect inhibitors. J Clin Oncol 1997; 15: 389–400.[Abstract]

10. Zalcberg JR, Cunningham D, Van Custem E et al. ZD1694: a novel thymidylate synthase inhibitor with substantial activity in the treatment of patients with advanced colorectal cancer. Tomudex Colorectal Study Group. J Clin Oncol 1996; 14: 716–721.[Abstract]

11. Hoff PM, Royce M, Medgyesy D et al. Oral fluoropyrimidines. Semin Oncol 1999; 26: 640–646.[ISI][Medline]

12. Washtein WL. Increased levels of thymidylate synthase in cells exposed to 5-fluorouracil. Mol Pharmacol 1984; 25: 171–177.[Abstract]

13. Berger SH, Jenh CH, Johnson LF et al. Thymidylate synthase overproduction and gene amplification in fluorodeoxyuridine-resistant human cells. Mol Pharmacol 1985; 28: 461–467.[Abstract]

14. Chu E, Drake JC, Koeller DM et al. Induction of thymidylate synthase associated with multidrug resistance in human breast and colon cancer cell lines. Mol Pharmacol 1991; 39: 136–143.[Abstract]

15. Leichman CG, Lenz HJ, Leichman L et al. Quantitation of intratumoral thymidylate synthase expression predicts for disseminated colorectal cancer response and resistance to protracted-infusion fluorouracil and weekly leucovorin. J Clin Oncol 1997; 15: 3223–3229.[Abstract]

16. Aschele C, Debernardis D, Casazza S et al. Immunohistochemical quantitation of thymidylate synthase predicts for clinical outcome in patients with advanced colorectal cancer treated by alternated bolus and infusional 5-fluorouracil. J Clin Oncol 1999; 17: 1760–1770.[Abstract/Free Full Text]

17. Davies MM, Johnston PG, Kaur S et al. Colorectal liver metastasis thymidylate synthase staining correlates with response to hepatic arterial floxuridine. Clin Cancer Res 1999; 5: 325–328.[Abstract/Free Full Text]

18. Kornmann M, Link KH, Lenz HJ et al. Thymidylate synthase is a predictor for response and resistance in hepatic artery infusion chemotherapy. Cancer Lett 1997; 118: 29–35.[ISI][Medline]

19. Cascinu S, Aschele C, Barni S et al. Thymidylate synthase protein expression in advanced colon cancer: correlation with the site of metastasis and the clinical response to leucovorin-modulated bolus 5-fluorouracil. Clin Cancer Res 1999; 5: 1996–1999.[Abstract/Free Full Text]

20. Pinedo HM, Peters GF. Fluorouracil: biochemistry and pharmacology. J Clin Oncol 1988; 6: 1653–1664.[Abstract]

21. Salonga D, Danenberg KD, Johnson M et al. Colorectal tumors responding to 5-fluorouracil have low gene expression levels of dihydropyrimidine dehydrogenase, thymidylate synthase, and thymidine phosphorylase. Clin Cancer Res 2000; 6: 1322–1327.[Abstract/Free Full Text]

22. Metzger R, Danenberg K, Leichman CG et al. High basal level gene expression of thymidine phosphorylase (platelet-derived endothelial cell growth factor) in colorectal tumors is associated with nonresponse to 5-fluorouracil. Clin Cancer Res 2000; 4: 2371–2376.[Abstract]

23. Aschele C, Sobrero A, Faderan MA et al. Novel mechanism(s) of resistance to 5-fluorouracil in human colon cancer (HCT-8) sublines following exposure to two different clinically relevant dose schedules. Cancer Res 1992; 52: 1855–1864.[Abstract]

24. Moran RG, Scanlon KL. Schedule-dependent enhancement of the cytotoxicity of fluoropyrimidines to human carcinoma cells in the presence of folinic acid. Cancer Res 1991; 51: 4618–4623.[Abstract]

25. Benz C, Schoenberg M, Choti M et al. Schedule-dependent cytotoxicity of methotrexate and 5-fluorouracil in human colon and breast tumor cell lines. J Clin Invest 1980; 66: 1162–1165.[ISI][Medline]

26. Cadman E, Heimer R, Benz C. The influence of methotrexate pre-treatment on 5-fluorouracil metabolism in L1210 cells. J Biol Chem 1981; 256: 1695–1704.[Abstract/Free Full Text]

27. Sobrero AF, Aschele C, Guglielmi AP et al. Schedule-selective biochemical modulation of 5-fluorouracil: a phase II study in advanced colorectal cancer. Clin Cancer Res 1995; 1: 955–960.[Abstract]

28. Aschele C, Guglielmi A, Frassineti GL et al. Schedule-selective biochemical modulation of 5-fluorouracil in advanced colorectal cancer: a multicentric phase II study. Br J Cancer 1998; 77: 341–346.[ISI][Medline]

29. Sobrero A, Zaniboni A, Frassineti GL et al. Schedule specific biochemical modulation of 5-fluorouracil in advanced colorectal cancer: a randomized study. Ann Oncol 2000; 11: 1413–1420.[Abstract]

30. Labianca R, Cascinu S, Frontini L et al. High versus low dose levo leucovorin as a modulator of 5-fluorouracil in advanced colorectal cancer: a GISCAD phase III study. Ann Oncol 1997; 8: 169–174.[Abstract]

31. Miller AB, Hoogstraten B, Staquet M et al. Reporting results of cancer treatment. Cancer 1981; 47: 207–214.[ISI][Medline]

32. Hsu SM, Raine L, Fanger H. A comparative study of the peroxidase-antiperoxidase method and an avidin biotin complex method for studying polypeptide hormones with radioimmunoassay antibodies. Am J Clin Pathol 1981; 5: 734–738.

33. Sobrero AF, Aschele C, Guglielmi AP et al. Synergism and lack of cross-resistance between short-term and continuous exposure to fluorouracil in human colon adenocarcinoma cells. J Natl Cancer Inst 1993; 85: 1937–1944.[Abstract]

34. Sobrero AF, Aschele C, Bertino JR. 5-Fluorouracil in colorectal cancer: a tale of two drugs. Implications for biochemical modulation. J Clin Oncol 1997; 15: 368–381.[Abstract]

35. Ehrnrooth E, Sorensen B, Poulsen JH et al. Changes in thymidylate synthase mRNA in blood leukocytes from patients with colorectal cancer after bolus administration of 5-fluorouracil. Acta Oncol 2000; 39: 53–57.[ISI][Medline]

36. Paradiso A, Simone G, Petroni S et al. Thymidylate synthase and p53 primary tumour expression as predictive factors for advanced colorectal cancer patients. Br J Cancer 2000; 82: 560–567.[ISI][Medline]

37. Aschele C, Debernardis D, Tunesi G et al. Thymidylate synthase protein expression in primary colorectal cancer compared to the corresponding distant metastases and relationship with the clinical response to 5-fluorouracil. Clin Cancer Res 2000; 6: 4797–4802.[Abstract/Free Full Text]

38. Buyse M, Thirion P, Carlson RW et al. Relation between tumor response to first-line chemotherapy and survival in advanced colorectal cancer: a meta-analysis. Meta-Analysis Group in Cancer. Lancet 2000; 356: 373–378.[ISI][Medline]

39. Radparvar S, Houghton PJ, Houghton JA. Effect of polyglutamylation of 5,10-methylenetetrahydrofolate on the binding of 5-fluoro-2'-deoxyuridylate to thymidylate synthase purified from a human colon adenocarcinoma xenograft. Biochem Pharmacol 1982; 23: 190–197.

40. Saltz L, Danenberg K, Paty P et al. High thymidylate synthase (TS) expression does not preclude activity of CPT-11 in colorectal cancer (CRC). Proc Am Soc Clin Oncol 1998; 17: 281a (Abstr 1080).

41. Lionetto R. INTACC-02: 5 Fluorouracil (5FU) + 6-S-leucovorin (6-S-LV) + levamisole (LEVA) vs methotrexate (MTX) + 5FU + LEVA: an Italian Intergroup Study of Adjuvant Therapy for Resected Colon Cancer. Proc Am Soc Clin Oncol 1999; 18: 291a (Abstr 118).

42. Punt CJ. Trimetrexate as a biochemical modulator of 5-fluorouracil and leucovorin in colorectal cancer. Semin Oncol 2000; 27: 88–90.[ISI][Medline]

43. Gorlick R, Rode W, Miles JS et al. Purification and cloning of uridine monophosphate kinase (UMPK) and its role in 5-fluorouracil resistance. Proc Am Soc Clin Oncol 1996; 16: 189 (Abstr 403).

44. Sargent DJ, Nieclzwiecki D, O’Connell MJ et al. Recommendation for caution with irinotecan, fluorouracil, and leucovorin for colorectal cancer. N Eng J Med 2001; 345: 144–146.[Free Full Text]