Page 80 - Read Online

Page 122 De Mattia et al. Cancer Drug Resist 2019;2:116-30 I http://dx.doi.org/10.20517/cdr.2019.04

therapy, demonstrated that the TYMS genotype associated with the lowest protein expression (2R/2R)
was significantly correlated with an increased risk of adverse events. This result was limited by substantial
heterogeneity of the studies included in the meta-analysis. However, the data were confirmed in a sensitivity
analyses performed in order to consider potential confounding factors (i.e., ethnicity, chemotherapeutic
regimen, biological sample used for DNA extraction, and the method of genotyping).

It should be also noticed that the predictive effect size of TYMS markers on toxicity risk, even if significant,
is very small and therefore translates into a limited clinical utility. The second recent systematic meta-
[39]
analysis , including 4,855 CRC Caucasian patients receiving 5-FU or capecitabine both alone or in
combinatorial regimens, evaluated the predictive role on toxicity risk of TYMS 5’UTR and 3’UTR genetic
variants. This study reported that TYMS 5’UTR-2R and TYMS 3’UTR-6bp ins alleles were associated with
an increased risk of severe toxicity after infusional or bolus 5-FU monotherapy. These two markers were
found to be predictive of severe toxicity also in patients treated with capecitabine. A score, based on the
number (0 to 4) of TYMS toxicity risk alleles, was further created and was shown to be a good predictor of
any capecitabine associated toxicity with a weaker evidence of association in patients receiving infusional or
bolus 5-FU monotherapy. When FP combination therapy was considered, none of the TYMS polymorphisms
were associated with toxicity risk, probably due to the reduced FP dosage in these regimens, the presence of
overlapping confounding toxicities, or the non-adequate sample size of the studies performed. Globally, even
if TYMS polymorphisms could be predictive of toxicity in FPs monotherapy, especially when capecitabine
is administered, the relatively weak size effect and the lack of association in the most commonly used
combinatorial regimes, require further evidence to support their use in clinical practice.

This is consistent with the results of an umbrella systematic review concluding that TYMS polymorphisms
had a statistically significant relationship with FPs-induced toxicity but that this relationship was relatively
[55]
weak and without any clinical significance . Therefore, TYMS variants could possibly be incorporated
into a panel of other predictive markers of FPs toxicity and a personalized drug dosing based on TYMS
genotyping could be applied only after further exhaustive investigation of the effect of a pre-emptive dose
reduction on the patient outcome.

However, even if the impact of TYMS variants on FPs-related toxicity risk appears to be minimal, their
effect in determining the FPs-based therapy effectiveness is more fully established. A recent randomized
trial (MAGIC trial) with a control group, has found that patients treated with 5-FU containing regimen and
harboring the TYMS genotype 2R/2R, associated with the low expression phenotype, present a superior OS
respect those with 2R/3R or 3R/3R genotype; this difference was not detected among patients treated with
surgery alone suggesting a specific interaction between TYMS genetic status and treatment. Patients with the
TYMS 2R/2R genotype have showed also a tendency towards an higher rate of response to FP chemotherapy
[56]
compared to those with the 2R/3R or 3R/3R genotype . Accordingly, in another pooled analysis of three
prospective clinical studies the TYMS genotype 3R/3R genotype was associated with an inferior objective
response (ORR) rate to FPs-containing regime; considering also the G to C substitution in the 3R allele,
patients carrying the 3RG/3RG genotype appeared to have the lowest ORR . Interestingly, a phase II study
[49]
further demonstrated that the prospective use of TYMS genotyping could direct the neoadjuvant 5-FU-based
[57]
chemoradiotherapy (CRT) in patient affected by rectal cancer . Particularly, patients carrying the TYMS
2R/2R or 2R/3R genotypes, who were supposed to have a favorable response to 5-FU, were treated with the
standard CRT, while patients with the TYMS 3R/3R or 3R/4R genotypes, who were deemed to not derive a
significant benefit from 5-FU, were treated with an alternative regimen. These findings support the feasibility
of a personalized treatment aiming to improve the therapy effectiveness based on TYMS genotyping and
encourage further evaluation of this genotype-based strategy.

Interesting data has been also published on variants in enolase superfamily member 1 (ENOSF1) gene
encoding a mitochondrial enzyme also known as reverse TS (rTS). A work on 968 Caucasian capecitabine-