1. ¹UPRES-EA 2128, UFR de Médecine, Caen, France
2. ²Laboratoire d'Hématologie, CHU Côte de Nacre, Caen, France
3. ³Laboratoire de Microbiologie de l'Environnement, EA-956 USC INRA, Caen, France

Correspondence: B Sola, UPRES-EA 2128, UFR de Médecine, CHU Côte de Nacre, 14032 Caen cedex, France. Fax: + 33 2 31474084

TO THE EDITOR

Cyclin D1 controls the early phase of cell cycle, and overexpression of cyclin D1 is associated with the pathogenesis of B-cell disorders such as mantle cell lymphoma (MCL), hairy cell leukemia (HCL) and multiple myeloma (MM). Several molecular mechanisms including the t(11;14)(q13;q32) could account for the activation of the CCND1 gene and the subsequent expression of cyclin D1. The human CCND1 gene encoding cyclin D1 spans 15 kb, includes five exons and four introns, and generates two major mRNAs of 1.7 and 4.4 kb that differ in their 3'-part, the 3'UTR region being absent in the shorter form.¹ Betticher et al² have described alternative forms of cyclin D1 transcript, called [b], of 1.7 and 4.4 kb, that is, of the same size as transcripts called [a] previously reported, generated by the absence of splicing at the exon 4/intron 4 boundary. Predicted cyclin D1 proteins translated from alternative splicing differ only in the last 55 amino acids of the carboxy terminus. Cyclin D1 protein forms [a] and [b] of, respectively, 36 and 31 kDa, possess the cyclin box necessary for cyclin-dependent kinase (CDK) binding and for enzymatic activity of the cyclin D/CDK4/6 complexes. However, the PEST destruction box responsible for the rapid turnover of cyclin D1 protein is present only in the form [a]. In order to study a possible role of [b] transcripts in B-cell pathogenesis, we have analyzed the presence of the two [a] and [b] transcripts by RT-PCR in MM cell lines and then in MCL and HCL patients. As reported in Figure 1a, cyclin D1 form [a] (PCR fragment of 886 bp) and cyclin D1 form [b] (PCR fragment of 615 bp) mRNAs were present in the same cell lines: NCI H929, RPMI 8226, U266 and Karpas 620. Only minute amounts of transcript [b] were found in the RPMI 8226 cell line and detected only under optimal conditions. LP1 and OPM-2 cell lines did not express either cyclin D1 form [a] or [b] mRNAs. Thus, cyclin D1 form [b] transcript could be present in t(11;14)(q13;q32)-bearing B-malignancies (Karpas 620) in good correlation with a previous report,³ and also in cell lines harboring either an insertion of a switch region (U266) or a trisomy of chromosome 11 (NCI H929 and RPMI 8226).⁴ There is no correlation between the presence of form [b] transcript and the mechanism of gene activation, and the two alternative form [a] and [b] transcripts are coexpressed in the studied MM cell lines. The presence of the two alternative forms of cyclin D1 was then analyzed in cyclin D1-expressing hemopathies: six MCL and five HCL and in two cyclin D1-nonexpressing chronic lymphocytic leukemias (CLL). Cyclin D1 form [a] mRNA was present in all MCL samples (6/6, Figure 1b and not shown) and in HCL samples (3/5, Figure 1c). Cyclin D1 form [b] was present in all MCL cells (Figure 1b and not shown) but not in HCL cells (Figure 1c). Cyclin D1 form [a] and [b] transcripts were absent in cells isolated from CLL patients (Figure 1b and not shown). It has been reported that in breast cancer cells, the level of transcript [b] was 20–40 times lower than transcript [a] (Hosokawa et al³ and references therein), and in order to enhance the sensitivity of the form [b] RT-PCR, we used a seminested PCR. As presented in Figure 1d, a unique fragment 225 bp long was present in form [a]-expressing HCL patients. The same fragment was also detected in RPMI 8226, NCI H929 and U266 cell lines previously recorded positive for the [b] form (Figure 1a) and absent in the negative control. The identity of the nested-PCR amplified product from two positive patients and NCI H929 and U266 cell lines was verified by direct sequencing (not shown). Thus, the cyclin D1 forms [a] and [b] are always coexpressed in MCL and HCL tumoral cells, but with various levels according to the pathology and again irrespective of CCND1 gene activation mechanism. Interestingly, we and others have reported previously that, in HCL B-cells, the levels of cyclin D1 mRNA species (and protein) are much lower than in MCL, a situation reflected by the RT-PCR analysis. The use of real-time quantitative PCR technology will allow the exact determination of each cyclin D1 form [a] and [b] in the various pathologies.

Figure 1.

RT-PCR analysis of cyclin D1 forms [a] and [b] in MM cell lines and in patients. Total RNA was extracted from cultures LP1, OPM-2, NCI H929, RPMI 8226, U266 and Karpas 620 cell lines (a) or from peripheral blood cells of patients with MCL (b), HCL (c and d) or CLL (b). RNA was subjected to RT-PCR; PCR products obtained with primers specific for cyclin D1 form [a] and form [b], respectively of 886 and 615 bp length were run on 1.5% agarose gels and visualized after ethidium bromide staining. A non-specific band (white star) of molecular weight larger than 886 bp length is sometimes visible. Negative controls (T-) or positive controls (T+) were run in the same gels. A unique upstream primer D1AS (position 272, complete cds, M64349, Genbank), 5'-CCT ACT TCA AAT GTG TGC AGA AG-3', was used with two different downstream primers D1AAS (form [a], position 1158, complete cds, M64349, Genbank), 5'-CAA GGA GAA TGA AGC TTT CCC TT-3' and D1BAS (form [b], position 20, intron 4, X88930, Genbank), 5'-GGG ACA TCA CCC TCA CTT AC-3'. For the seminested PCR, an upstream primer aD1S 5'-GAA CAA ACA GAT CAT CCG CAA ACA-3' (position 663, M64349) was coupled with the downstream primer D1BAS.

Full figure and legend (177K)

The next step was the study of cyclin D1 forms [a] and [b] proteins. Western blot analysis was performed on proteins purified from the MM cell lines and from MCL patients with a panel of antibodies (Ab) able to discriminate between the two forms of the proteins. The DCS-6 and sc-8396 monoclonal anticyclin D1 Abs raised against the entire recombinant cyclin D1 protein detect the forms [a] and [b], and the polyclonal anti-human cyclin D1 Ab raised against the carboxy-terminus of the protein (sc-718) detects only cyclin D1 form [a]. With both DCS-6 and sc-718 Abs, a unique cyclin D1 form [a] of 36 kDa was detected in all samples in which cyclin D1 mRNA form [a] has been detected previously, that is in NCI H929, RPMI 8226, U266 and Karpas 620 and in MCL samples (data not shown and ref.1). The cyclin D1 form [b] was never detected even for samples expressing high levels of mRNA. In order to ensure a complete separation of the two forms of cyclin D1 and to enhance immunodetection of the form [b], proteins extracted from Karpas 620 cells were separated by 2-D electrophoresis followed by immunoblotting, using the same Abs. The main polypeptide recognized by sc-718 Ab displayed migration characteristics (MW=36 kDa, pI=5.2) corresponding to the theoretical ones of the cyclin D1 form [a] (white arrow, Figure 2a). As expected, this protein was also detected with the DCS-6 and sc-8396 Abs (white arrows, Figure 2b and c). Nonspecific spots (see the left and lower part of Figure 2b) were revealed with DCS-6 Ab. However, no other signals were detected by both DCS-6 and sc-8396 Abs, and none of those Abs allowed the detection of a spot in the expected region for cyclin D1 form [b] (MW=31 kDa, pI=8.2). We generated the BD1b-A1 cell line derived from the B-cell BaF3 cell line engineered to express cyclin D1 form [b] upon ponasterone A stimulation. The expression system and the generation of conditionally expressing cells have been described elsewhere in detail.⁵ By RT-PCR analysis, an amplified fragment 615 bp long, corresponding to the cyclin D1 form [b], was detected in BD1b-A1 cells treated for 24 h with 10 M ponasterone A. The identity of PCR amplification product from [b] transcript was confirmed by direct sequencing. This amplified product was absent in the BaF3 parental cell line (data not shown). Proteins from uninduced and induced BD1b-A1 cells were analyzed by 2-D electrophoresis and immunoblotted with the DCS-6 and sc-8396 Abs (Figure 2d–f). Mainly one signal was detected at the expected size (black arrows). It was more acidic than the pI deduced from the amino-acid sequence of cyclin D1 form [b] (ie 6.3 vs 8.2), most probably reflecting post-translational modifications. This signal was present in the noninduced extracts (Figure 2d) indicating some leakage in the expression system, dramatically enhanced after a 24 h ponasterone A treatment (Figure 2e and f) and detected by the two DCS-6 and A-21 Abs (Figure 2e and f). These combined results unambiguously demonstrate the absence of the cyclin D1 form [b] protein in Karpas 620 cells, although they did contain significant amounts of the corresponding mRNA. To our knowledge, the presence of cyclin D1 form [b] has been reported in the Karpas 620 cell line⁶ and in three t(11;14)(q13;q32)-positive cell lines.³ Analysis of cyclin D1 protein was performed with the DCS-6 Ab, which displays no cross-reactivity with other cyclin proteins, but in experimental conditions generating high background.^3,6 This probably leads to a misinterpretation of the results. By contrast, cyclin D1 protein form [b] was unambiguously detected in glioma cells expressing ectopic cyclin D1 form [b],⁷ and we detected the cyclin D1 form [b] protein in BaF3-derived cell lines engineered to express this form. We can conclude that cyclin D1 protein form [b] is not present in B-cell malignancies or expressed at an undetectable level.

Figure 2.

2-D electrophoresis and immunoblotting. Proteins (3 mg) extracted from Karpas 620 cells (a–c), BD1b-A1 cells (d) and BD1b-A1 cells treated for 24 h with 10 M ponasterone A (e and f) were separated by 2-D electrophoresis (pH 3–10 gradient in the first dimension, and 14% acrylamide SDS-PAGE in the second dimension), and then transferred onto PVDF membranes. Immunoblots realized as described⁸ were revealed with the ECL chemiluminescence kit (Amersham Pharmacia Biotech.). The following antibodies were used to discriminate between the two alternative forms [a] and [b] of cyclin D1: sc-718 (a), DCS-6 (b, d and e) and sc-8396 (c and f). White arrows: cyclin D1 form [a]; black arrows: cyclin D1 form [b]. sc-718 and sc-8396 came from Santa Cruz Biotech. and DCS-6 from Pharmingen.

Full figure and legend (259K)

In normal B-lymphoid cells, whatever the stage of differentiation, the CCND1 gene is silent. By contrast, in some hematological B-malignancies such as MCL, HCL and MM, the CCND1 gene is activated and leads to the expression of cyclin D1 mRNA and protein. In this study, we have examined the presence of the alternative cyclin D1 transcript referred to as form [b] in cells isolated from MCL, HCL patients and MM cell lines and investigated the presence of the corresponding protein. By direct or seminested RT-PCR, we have demonstrated that cyclin D1 transcript form [b] is present in all tumoral cells expressing cyclin D1 form [a] irrespectively of the presence of t(11;14)(q13;q32). Then, the generation of an alternative splice does not depend on the mechanism of CCND1 gene activation. The alternative splice form [b] is generated by an A/G polymorphism within the splice donor sequence at position 870,² but both alleles can splice to form both [a] and [b] transcripts. Whatever the mechanism of CCND1 gene activation, the cyclin D1 open reading frame as well as the upstream regulatory sequences remain intact; thus, the presence of transcripts [a] and [b] depends only on the A/G polymorphism. We cannot exclude that form [b] transcript could play a role by itself, but our data support the notion that the cyclin D1 protein form [b] does not participate in the transformation process in mature B-cell malignancies. In good agreement with us, a recent report has evaluated the impact of the A/G polymorphism on the survival of a group of MCL patients and found no significant influence.⁸