Cereblon expression in multiple myeloma: not ready for prime time – Response to Lodé et al

Abstract

We have read with interest the comments provided by Lodé et al (2013) regarding our findings of a correlation of cereblon (CRBN) expression and response in multiple myeloma (MM) patients treated with lenalidomide and dexamethasone (Heintel et al, 2013). We agree that the prime time has not yet arrived for the routine use of CRBN expression as treatment guidance. Lodé et al (2013) raise three important issues that should be evaluated before drawing final conclusions and we wish to add the following: In relation to the first point raised, we suggest that CRBN should be analysed in both the CD138+ and CD138− fraction of MM bone marrow samples in future studies. It is well recognized that at least some of the effect of immunomodulatory drugs (IMiDs) is mediated through the microenvironment by induction of proliferation of T-cells and other effector cells, of expression of co-stimulatory molecules and modification of cytokine signalling (Görgün et al, 2010), resulting in an anti-myeloma immune response. Lopez-Girona et al (2012) showed that CRBN not only mediates the anti-proliferative effects of lenalidomide and pomalidomide in myeloma cells, but also lenalidomide- and pomalidomide -induced cytokine production in T-cells. In our paper we provided data showing that, although median CRBN expression is lower in the CD138− than CD138+ fraction, expression levels of CRBN vary in the MM bone marrow environment and can be considerably higher than in normal bone marrow (Heintel et al, 2013). Thus we speculate that variable levels of CRBN in immune cells in the bone marrow might have contributed to the association with response in our study. We fully agree with the second point raised, that CRBN physiology needs to be studied in detail and the function of the different isoforms elucidated. Lenalidomide treatment was shown to induce 2-fold expression changes in approximately 1200 genes, and the vast majority of these changes seem to be mediated via CRBN (Zhu et al, 2011). Lodé et al (2013) provide important data showing the frequent occurrence of alternative transcripts, accounting in some patients for >50% of total CRBN in some patients. It has been reported that CRBN downregulation mediates resistance to IMiD-based treatment (Zhu et al, 2011; Heintel et al, 2013). As preferential expression of transcripts lacking a functional IMiD binding domain would be anticipated to have a similar effect, we propose studying the expression of alternative transcripts of CRBN with regard to primary resistance and acquisition of resistance to IMiD based treatment. Gandhi et al (2012) reported on an antibody (CRBN65) capable of discriminating the different isoforms by Western blot, and such a tool should be helpful in these studies. In contrast, truncating mutations in CRBN seem to be a rare mechanism of acquisition of resistance (Egan et al, 2013). Once the correlation between CRBN expression and clinical response can be corroborated, antibodies binding to CRBN or its isoforms might provide the basis for developing a reliable routine test for CRBN expression. Developing a flow cytometric technique for detection might provide further insight into intra-tumour heterogeneity of CRBN expression. It should be noted that three other groups have already reported a correlation between CRBN expression and outcome under IMiD therapy, one using tissue microarrays and fluorescence immunohistochemistry to quantify CRBN expression (Klimowicz et al, 2012), the two other groups using gene expression profiling (Broyl et al, 2013; Schuster et al, 2012). The in-silico analysis performed by Lodé et al (2013) provides important insights with regard to CRBN expression in the different molecular MM subgroups. While median CRBN expression is higher in the hyperdiploid subgroup, there seems to be great variability of expression within all subgroups. In our preliminary analysis we could not find a correlation between CRBN expression and fluorescent in situ hybridization-defined high-risk or standard risk disease. We did not study chromosome 3 trisomy in our series of patients but agree that cytogenetic data should be evaluated in relation to CRBN expression. We think that the outcome defined by response is less biased by a possible overweight of hyperdiploid disease, as the main impact of the molecular signature is on survival parameters, not initial response to therapy. A very rapid and deep response to therapy, with regard to serum free-light chain levels, might even identify a subgroup of patients with inferior survival. It should be noted that the groups who have reported a correlation between CRBN expression and outcome with IMiD-based treatment defined by survival have included a bortezomib-treated control group, where this association was lacking (Broyl et al, 2013; Klimowicz et al, 2012), thus making a bias related to a possible prognostic significance of CRBN expression itself less likely.