Whole-genome approach implicates CD44 in cellular resistance to carboplatin

Abstract

Carboplatin is a chemotherapeutic agent used in the management of many cancers, yet treatment is limited by resistance and toxicities. To achieve a better understanding of the genetic contribution to carboplatin resistance or toxicities, lymphoblastoid cell lines from 34 large Centre d'Etude du Polymorphisme Humain pedigrees were utilised to evaluate interindividual variation in carboplatin cytotoxicity. Significant heritability, ranging from 0.17-0.36 (p = 1 x 10(-7) to 9 x 10(-4)), was found for cell growth inhibition following 72-hour treatment at each carboplatin concentration (10, 20, 40 and 80 microM) and IC(50) (concentration for 50 per cent cell growth inhibition). Linkage analysis revealed 11 regions with logarithm of odds (LOD) scores greater than 1.5. The highest LOD score on chromosome 11 (LOD = 3.36, p = 4.2 x 10(-5)) encompasses 65 genes within the 1 LOD confidence interval for the carboplatin IC 50 . We further analysed the IC(50) phenotype with a linkage-directed association analysis using 71 unrelated HapMap and Perlegen cell lines and identified 18 single nucleotide polymorphisms within eight genes that were significantly associated with the carboplatin IC(50) (p < 3.6 x 10(-5); false discovery rate <5 per cent). Next, we performed linear regression on the baseline expression and carboplatin IC(50) values of the eight associated genes, which identified the most significant correlation between CD44 expression and IC(50) (r(2)= 0.20; p = 6 x 10(-4)). The quantitative real-time polymerase chain reaction further confirmed a statistically significant difference in CD44 expression levels between carboplatin-resistant and -sensitive cell lines (p = 5.9 x 10(-3)). Knockdown of CD44 expression through small interfering RNA resulted in increased cellular sensitivity to carboplatin (p < 0.01). Our whole-genome approach using molecular experiments identified CD44 as being important in conferring cellular resistance to carboplatin.

Figure 1

Chemical structure of carboplatin.

Figure 2

Box plots and heritability (h²) values for 34 families are shown for 10, 20, 40 and 80 μM carboplatin. Centre d'Etude du Polymorphisme Humain (CEPH) family identification pertaining to the box plot is located on the x-axis. The middle line within the box represents the mean percentage survival of each family, while the box edges and whiskers represent the standard error of the mean and twice the standard error of the mean, respectively.

Figure 3

Linkage analyses for carboplatin-induced cytotoxicity. Non-parametric quantitative trait locus (QTL) linkage analysis was performed using 34 large pedigrees for the 10, 20, 40 and 80 μM concentrations and IC₅₀ phenotype. Each vertical line indicates chromosome boundaries and the horizontal lines indicate logarithm of odds (LOD) scores. The dotted line indicates the LOD > 1.5 regions, which were used for association analyses.

Figure 4

Baseline expression correlation of CD44 with carboplatin IC₅₀. (a) Linear regression of log₂ transformed carboplatin IC₅₀ values on log₂ transformed baseline CD44 expression (from exon expression array) and in 57 unrelated HapMap cell lines. (b) Quantitative real-time polymerase chain reaction analysis of normalised CD44 expression in lymphoblastoid cell lines with the highest (resistant, n = 8) and lowest (sensitive, n = 10) carboplatin IC₅₀ values. Each box represents biological triplicates run in duplicate (*p < 0.01, Student's t-test).

Figure 5

Effect of small interfering RNA (siRNA)-mediated knockdown of CD44 on the sensitivity of cell lines to carboplatin. (a) Protein lysates from GM06985 and GM11881 were isolated 48 hours post-transfection with 600 pmol CD44 siRNA and non-targeting control siRNA. Western blot analysis was performed using antibodies against CD44 and β-actin (used to normalise the CD44 bands). (b) Carboplatin cytotoxicity of 06985 following treatment with CD44 siRNA or non-targeting control (p < 0.01; Student's t-test for comparison of IC₅₀ values) 48 hours after carboplatin exposure. The solid line indicates transfection with non-targeting siRNA and the dotted line indicates transfection with CD44 siRNA. Each line represents an average of triplicate knockdown experiments, and error bars represent the standard error of the mean. (c) Carboplatin cytotoxicity following CD44 siRNA knockdown compared with the non-targeting group for cell line GM11881 (p < 0.05; Student's t-test for comparison of IC₅₀ values). The solid line indicates transfection with non-targeting siRNA and the dotted line indicates transfection with CD44 siRNA. Each line represents an average of triplicate knockdown experiments and error bars represent the standard error of the mean.