Absence of mutations in cereblon (CRBN) and DNA damage-binding protein 1 (DDB1) genes and significance for IMiD therapy.

Thalidomide and the IMiD immunomodulatory drugs, lenalidomide and pomalidomide, are widely used in the treatment of multiple myeloma (MM), del(5q) myelodysplastic syndromes and other hematologic malignancies, including mantle cell lymphoma. Ito et al.[1] recently identified cereblon as a key target of thalidomide. Subsequent studies confirmed cereblon to be a common target for lenalidomide and pomalidomide, and established its essential role in mediating anticancer and immunomodulatory effects of these drugs.[2, 3] Cereblon is encoded by the CRBN gene on chromosome 3 containing 11 exons, and the fully spliced transcript produces a 51-kDa protein. Cereblon is a component of the cullin ring E3 ubiquitin ligase complex (CRL4CRBN) that also contains DNA damage-binding protein 1 (DDB1), cullin (Cul) 4a and regulator of cullins (Roc) 1.[1] E3 ligases attach ubiquitin moieties to specific substrate proteins in the cell that can mark them for proteasomal degradation. The putative role of cereblon within the E3 ligase complex is that of a substrate receptor.

With the discovery of cereblon, as a target of IMiD therapy, there has been considerable interest in defining whether expression of cereblon protein or the presence of CRBN mutations will impact clinical responses to these drugs.[2, 4, 5, 6, 7] There are limited available data on CRBN gene mutation in the literature. Originally, a nonsense mutation (R419X) of CRBN was described to be associated with autosomal recessive non-syndromic mental retardation.[8] However, the functional link between the mutation in CRBN and the onset of mental retardation has not been demonstrated. Sequencing analyses of CRBN in MM cells from patients identified a truncating mutation (Q99X) and a point mutation (R283K) in 1 of 30 MM patients.[9] In addition, an A/G polymorphism has been identified at −29 nucleotide from the transcriptional start site of the CRBN transcript.[10] So far, mutations in other components of the CRL4 E3 ligase complex (DDB1, Cul4a or Roc1) in MM cell lines or patients have not been described in the limited genome-wide sequencing of MM patients.[11]

Here, we focused on sequencing the exons of CRBN with a goal of identifying missense or nonsense mutations with likely functional and clinical consequence. We analyzed IMiD-sensitive, intrinsically IMiD-resistant, as well as isogenic-sensitive or acquired lenalidomide- and/or pomalidomide-resistant MM cell lines. In addition, 90 MM patient samples, including those from 36 lenalidomide-resistant patients, were evaluated for the presence of CRBN mutations. As shown in Table 1, we found that the vast majority of IMiD-sensitive cell lines harbored the wild-type CRBN gene sequence. Also, none of the three intrinsically resistant cell lines (LP1, RPMI 8226 or JJN3) carried any mutation within the CRBN exons. As described previously, all three cell lines express high levels of cereblon transcript and protein.[12] Thus, the lack of mutation in the CRBN gene strongly suggests the existence of a cereblon-independent mechanism(s) of intrinsic resistance to IMiD drugs in the LP1, RPMI 8226 and JJN3 cell lines. In contrast, a heterozygous CRBN mutation (D249Y) was detected in the lenalidomide-resistant ANBL-6 cell line, while its sensitive parental line did not harbor the mutation. The location of the mutation suggests that it may impact the binding of cereblon to the drug or interacting partner DDB1. However, it is not clear whether the resistant phenotype in the lenalidomide-resistant ANBL-6 cell line is a direct result of the CRBN mutation alone and therefore it requires further evaluation. One copy of CRBN gene was shown to be deleted in the MM1S and MM1S.R MM cell lines.[2] However, in our sequencing of the CRBN gene in these cell lines, we did not find any missense mutation or single-nucleotide variations (SNVs). SNVs, as opposed to missense or nonsense mutations, are synonymous substitutions of nucleotides that do not change the amino acid at a given codon. In all, we found two SNVs in the KMS-12-BM (rs17027638) and OPM-2 cell lines, respectively. The SNV in OPM-2 has not been described in the public database. Among the isogenic sensitive and resistant pairs of cell lines, no new mutation or polymorphic changes were detected.

Table 1

Summary of mutations and SNVs in CRBN and DDB1 in MM cell lines and patients

IMiD-sensitive MM cell lines	CRBN nucleotide position and change	CRBN amino-acid position and change	DDB1 nucleotide position and change	DDB1 amino-acid position and change
H929 parental, U266, EJM, SKMM2	None		None
OPM-2	G1209A silent heterozygous	T403T	None
KMS-12-BM parental	T735C silent homozygous	Y245Y	None
Intrinsic IMiD-resistant MM cell lines
LP1	None		None
RPMI 8226	None		None
JJN3	None		None
Paired sensitive and acquired resistant cell lines
H929 sensitive control	None		None
H929 4 resistant clones	None		None
KMS-12-BM sensitive control	T735C silent homozygous	Y245Y	None
KMS-12-BM LEN-resistant	T735C silent homozygous	Y245Y	None
KMS-12-BM POM-resistant	T735C silent homozygous	Y245Y	None
MM1S sensitive	None		None
MM1S/R10R LEN-resistant	None		None
ANBL-6 sensitive	None		A909T heterozygous	E303D
	None		C153T silent heterozygous	P51P (rs2230356)
ANBL-6 LEN-resistant	G745T heterozygous	D249Y	A909T heterozygous	E303D
	None		C153T silent heterozygous	P51P (rs2230356)
Patient samples (90)
Newly diagnosed (24)	T735C silent homozygous, C219T silent heterozygous	2 × Y245Y (rs17027638), 1 × H73H	C153T silent heterozygous, C939T silent heterozygous	2 × P51P (rs2230356), 1 × C313C (rs150106100)
RRMM (30)	T735C silent homozygous	2 × Y245Y (rs17027638)	C153T silent heterozygous	4 × P51P (rs2230356)
LEN-resistant RRMM (36)	None		None

Abbreviations: CRBN, cereblon; DDB1, DNA damage-binding protein 1; IMiD, immunomodulatory drug; LEN, lenalidomide; MM, multiple myeloma; POM, pomalidomide; RRMM, relapsed/refractory multiple myeloma; SNV, single-nucleotide variation.

We next sequenced the CRBN gene from 90 MM patients. All samples collected in this study followed institutional procedures for ethical guidelines and informed consent for the analysis. Samples used for the sequencing analysis were either CD138+ cells isolated from bone marrow aspirates (n=36) or bone marrow mononuclear cells (BMMC; n=54) with a plasma cell content of at least 20%. The patients were further divided according to disease status: newly diagnosed (n=24); relapsed/refractory (n=30, 19 of whom were heavily pretreated with multiple regimens that included lenalidomide); or relapsed and/or refractory with lenalidomide resistance (n=36; clinical criteria for resistance were progression within two cycles or relapse within 6 months of completing the last line of an IMiD-based regimen). In the 24 newly diagnosed patients, two SNVs were detected, and in the 66 relapsed and refractory patients two SNVs were also detected: one of the SNVs was identified in four cases at position 735 (T>C; Y245Y) that was seen previously in the cell lines and the other SNV was found once at 219 (C>T; H73H). Both SNVs have been previously described in public databases (http://www.ncbi.nlm.nih.gov/projects/SNP/). Strikingly, we did not find any mutation in the CRBN gene in any of the lenalidomide-resistant patient samples.

As DDB1 interaction with cereblon is critical for the E3 ligase function, the DDB1 gene may potentially harbor mutations of clinical significance. So, we extended our sequencing analysis to DDB1 to search for the presence of mutations in MM cell lines and patients. Of more than 20 cell lines tested, only a single DDB1 heterozygous mutation (E303D) was identified in the ANBL-6 parental cell line (Table 1). There is no known functional consequence of this mutation, as ANBL-6 is sensitive to lenalidomide. In the 54 BMMC patient samples tested from newly diagnosed and relapsed and refractory patients, two different SNVs were detected: One of the SNVs occurred six times at nucleotide position 153 (C>T; P51P) and the other once at nucleotide position 939 (C>T; C313C). Both SNVs have been described in the dbSNP database (http://www.ncbi.nlm.nih.gov/projects/SNP/).

There are a number of implications of the sequencing analyses presented here. Although the number of cell lines and patients tested is limited, our findings suggest that incidence of CRBN and DDB1 gene mutations in MM cell lines and patients are rare and will have a limited impact on defining resistance to IMiD therapy. In the three intrinsically IMiD-resistant cell lines that clearly express detectable levels of cereblon, the absence of CRBN and DDB1 mutations suggest that potential cereblon-independent mechanisms of resistance exist. We did, however, detect a mutation in the lenalidomide-resistant ANBL-6 cell line. Thus, the development of lenalidomide resistance in the ANBL-6 cell line may be linked to a CRBN mutation and provides the proof of principle that resistance to IMiD exposure could arise by a mutation within the CRBN gene. Finally, unlike for ATP-dependent kinase inhibitors, where there is precedence for development of drug resistance due to selection of mutant forms of the drug targets, this does not appear to be a common mechanism for IMiD therapy.

It has been shown by us and others that cereblon levels (measured by gene expression or protein assays) are lower compared with pretreatment levels when cell lines or patients become resistant to lenalidomide or pomalidomide.[2, 13] It is likely that epigenetic, transcriptional and/or posttranscriptional mechanisms—rather than CRBN mutation—are predominantly involved in the development of CRBN-dependent mechanisms of resistance to IMiD-based therapy. Therapeutic intervention to overcome IMiD resistance may have to address the consequences arising from these alternative mechanisms.