A Novel Splice Site Variant of hOCT-1 and Response to Imatinib.

The prognosis for patients with chronic myeloid leukaemia (CML) has vastly improved through the use of tyrosine kinase inhibitors (TKIs), such as imatinib. However, a proportion of patients will not respond, or will lose their initial response, and the biological mechanisms underlying this heterogeneity are poorly understood. Human organic cation transporter-1 (hOCT-1 or SLC22A1), the main transporter for imatinib, has been proposed as one determinant. This study set out to assess the prognostic value of hOCT-1 expression and the presence of polymorphisms in the hOCT-1 gene.

hOCT-1 mRNA levels in 153 diagnostic whole blood samples from two different patient cohorts (one from a trial population, and the other from patients treated only at our institution) were measured by RT-qPCR (normalised against two control genes). hOCT-1 exon 7 DNA (and cDNA transcripts) were sequenced in 156 patients, and four cells lines, to identify insertions/deletions or single nucleotide polymorphisms (SNPs). Fragment length analysis using gene-scanner technology was also used to identify patients with insertions/deletions and to correlate genotypes with cDNA transcript lengths. Time to each endpoint (remission/imatinib failure) was compared according to the level of expression for each gene, or according to genotype, using Kaplan-Meier analyses and the log-rank test.

No significant differences in outcomes were found when comparing patients with high or low hOCT-1 expression (whether defined using the median or other cut-offs). The 408V>M (g.1222G>A) SNP in hOCT-1 exon 7 was found in 102/156 patients (22 homozygotes and 80 heterozygotes) and was associated in all cases with an eight base-pair insertion at the exon-intron boundary (rs113569197). This insertion was found to create an alternative splice site, leading to the transcription of an additional RNA/cDNA transcript in these patients, the sequence of which contains a premature stop codon soon after the splice site. Additionally, M420del was found in 52 patients (three homozygotes and 49 heterozygotes) and was not found in alleles containing the eight base pair insertion. The six possible combinations of these three alleles (N=no insertion/deletion, 8⁺=8bp insertion and 3⁻=3bp deletion) were found to give rise to five possible combinations of RNA transcript lengths, since patients with 8⁺8⁺ produce identical transcript lengths to those with 8⁺N (i.e. the normal transcript and one with an additional 8bp and premature stop).

In the trial cohort (n=109), significant differences in time to 10% (p=0.01), 1% (p=0.0002) and 0.1% (p=0.0003) molecular responses (by the international scale) and time to imatinib failure (p=0.02) were seen when patients with 8⁺8⁺/8⁺N were compared to those with the remaining four genotypes (NN/N3⁻/3⁻8⁺/3⁻3⁻). However, this association was not replicated in the internal cohort, which was smaller (n=47) and more heterogeneous in terms of baseline characteristics and management.

These findings may explain discrepancies in the results of previous studies that have examined the association between hOCT-1 expression and outcome, since a number used primer/probe sets that would be affected by the presence/absence of these polymorphisms. Our results suggest that while hOCT-1 expression is not a determinant of response, alterations in splice sites or amino acid sequence due to insertions/deletions may be. Further work is required to clarify the impact of these polymorphisms on hOCT-1 protein levels, function and responses to imatinib.

White:Novartis Oncology: Honoraria, Research Funding; BMS: Research Funding; CSL: Research Funding. Marin:Novartis: Research Funding; BMS: Research Funding. Apperley:Novartis, Bristol Myers-Squibb, and ARIAD: Honoraria, Research Funding. Goldman:Novartis, Bristol Myers-Squibb, and Amgen: Honoraria.