14q32 Translocations and Monosomy 13 Observed in Monoclonal Gammopathy of Undetermined Significance Delineate a Multistep Process for the Oncogenesis of Multiple Myeloma¹

MGUS³ is a condition defined by the detection of a monoclonal protein (the M-component) in the serum without evidence of any causal disease, such as MM, Waldenstrom macroglobulinemia, primary amyloidosis, or any related disorder (1). This situation is not unusual, because it represents the first cause of monoclonal gammopathy, corresponding to 1% of the population over 50 years of age, and 3% of the population over 70 years of age (2). However, as an exclusion diagnosis, MGUS is probably not a unique entity. This heterogeneity is evident when analyzing the evolution of these individuals. Some may remain asymptomatic for decades, whereas others evolve to highly malignant conditions in less than 1 year. The overall incidence of MM transformation is estimated to be 1–2% per year (2, 3), but thus far, no clear-cut prognostic parameter has been identified.

The origin of MGUS is largely unknown. Whereas malignant PCs accumulate within the bone marrow in MM, bone marrow examination is frequently normal in MGUS. However, serial studies usually show a plasmacytosis increase, making it difficult to distinguish early-stage MM from advanced-stage MGUS. Recent immunophenotypic studies have shown that PCs displayed similar phenotypic abnormalities, and that the only difference between these two conditions was the persistence of normal PCs in MGUS (4). What about chromosomal changes? Although most (if not all) MM patients have large numbers of aneuploid PCs, chromosomal abnormalities are found in only 30–50% of the patients (5, 6, 7, 8). The situation is even more dramatic in MGUS, for which only one successful cytogenetic analysis has been reported thus far (9). This scarcity is explained mainly by the low proliferative index of PCs in MGUS. To circumvent this problem, some authors have used FISH on interphase cells (10, 11, 12). Using only four centromeric probes, they have demonstrated that the incidence of trisomy for at least one chromosome was between 43% and 68%. Moreover, serial analyses revealed that most individuals without trisomy at first evaluation acquired numerical chromosomal changes on subsequent examinations (12). These studies suggest that, as in MM, most patients with MGUS display numerical chromosomal changes in their clonal PCs.

More recently, we and others have demonstrated that structural chromosomal changes also occur in the PCs of individuals with MGUS (13, 14). Nishida et al. (14) first reported the presence of 14q32 abnormalities in three of five MGUS patients. More recently, we demonstrated in a preliminary analysis that 14q32 abnormalities were present in 11 of 19 patients with MGUS (15). These data suggest that structural chromosomal abnormalities could be as frequent in MGUS as in MM. To confirm this hypothesis, we conducted a study on a series of 100 individuals with either MGUS or SMM using interphase FISH to determine: (a) the incidence of 14q32 rearrangements; (b) the frequency of the main partner chromosomal regions; and (c) the incidence of deletion/monosomy 13 in MGUS. These results were then compared with those of 102 patients with overt MM at diagnosis (15).

We selected a cohort of 100 consecutive individuals with either MGUS or SMM seen at one of the three hematology departments participating in this study. Nineteen of these patients have been reported previously (15). Diagnosis of MGUS and SMM was performed according to the criteria described by Kyle and Lust (1) and Kyle and Greipp (16), respectively. MGUS was assessed when individuals had serum monoclonal protein lower than 30 g/liter and bone marrow plasmacytosis lower than 10% in the absence of clinical symptoms, lytic bone lesions, anemia, hypercalcemia, and renal function impairment. Patients with SMM fulfilled the same criteria, except that the M-component was higher than 30 g/liter and/or bone marrow plasmacytosis was higher than 10%. In this series, there were 79 patients with MGUS and 21 patients with SMM. The median age at diagnosis for individuals with MGUS (44 males and 35 females) was 64 years (range, 33–86 years), whereas the median age of SMM patients (11 males and 10 females) was 65 years (range, 37–78 years). Within the MGUS population, M-component was Gκ in 44 cases, Gλ in 26 cases, Aκ in 8 cases, and Aλ in only 1 case. In the SMM population, the corresponding numbers of cases were 12, 3, 4, and 2, respectively. No patient with only light chain excretion was analyzed in this series. The median follow-up of patients with MGUS and SMM was 25 months (range, 3–168 months) and 16 months (range, 6–180 months), respectively. The control population, i.e., patients with overt MM, was defined according to the criteria of Durie and Salmon (17). A total of 102 consecutive patients with stage I, II, or III MM were analyzed using the same strategy. These patients have been reported previously (15). There were 17 patients with stage I disease, 27 patients with stage II disease, and 58 patients with stage III disease.

Because of the scarcity of PCs in individuals with MGUS (by definition less than 10%, and very often 0–2%), we chose to positively select them. This selection was based on CD138 expression, as described previously (13). Because CD138 is specific for PC within bone marrow, we used an immunomagnetic method with an anti-CD138 antibody (Miltenyi Biotec, Auburn, CA). Briefly, bone marrow mononuclear cells were separated using gradient density (Ficoll-Hypaque) and then incubated with anti-CD138-coated magnetic beads. Cells were passed through columns, allowing us to sort PCs. Recovery and purity of PCs were evaluated by morphology. PCs were then fixed in methanol:acetic acid (3:1 v/v and dropped on slides for FISH analysis.

FISH was performed as described previously (13). Probes used in this study have been reported previously (13). Briefly, the IGH gene was analyzed in a first attempt using the Y6 and Ig10 probes, which were kindly provided by Dr. Matsuda (Kyoto University, Japan) and Dr. Rabbits (Medical Research Council, Cambridge, United Kingdom), respectively. Each lot of probes was validated on 1000 bone marrow cells obtained from healthy donors, and cut-off values were determined for each lot. Illegitimate IGH rearrangements (defined by any situation other than two colocalized green and red signals) were considered when present in more than 10–16% of cells (mean + three SDs). Patients with illegitimate IGH rearrangements were then analyzed for the presence of the three main partner chromosomal regions, i.e., 11q13 (CCND1), 4p16 (FGFR3/MM-SET), and 8q24 (MYC). The CCND1 and FGFR3 probes have been reported previously (13). The MYC locus was analyzed using a YAC probe described recently (18). Patients were also analyzed using centromeric probes specific for chromosomes 3, 9, 11, and 15 (Vysis, Downers Grove, IL) to evaluate the percentage of clonal PCs. These four probes were selected based on the high frequency of trisomies 3, 9, 11, and 15 in MM and MGUS patients. The thresholds for trisomy were determined for each probe by analyzing 2000 nuclei from five normal bone marrow specimens (healthy donors). Cut-off values (mean + three SDs) were 1.7%, 2.4%, 2.1%, and 3.4% for centromere 3, 9, 11, and 15 probes, respectively. Finally, chromosome 13 was analyzed using a probe specific for the D13S319 locus at 13q14 (Vysis). The deletion cut-off was fixed at 5.3%, as described previously (15).

An abnormal configuration of 14q32 probes was observed in 46 (36 patients with MGUS and 10 patients with SMM) of 100 patients. This abnormal configuration was mainly a separation of the two probes, corresponding to probable translocations involving the IGH gene. These results were slightly lower than those of patients with overt MM, showing 61 of 102 (60%) patients with an illegitimate IGH rearrangement (P = 0.068; Ref. 15). Among the 46 patients with MGUS/SMM who had a 14q32 abnormality, the median percentage of cells displaying the abnormality was 72% (range, 19–99%). Twenty-nine of these patients were aneuploid for one of the tested centromeric probes. The median percentage of aneuploid cells was 70% (range, 18–96%), in good agreement with the percentage of cells with illegitimate IGH rearrangement. The results are described in Table 1. Specimens with a normal 14q32 configuration, i.e., two yellow signals corresponding to the fusion of the VH and CH probes, could be interpreted in two different ways. These patients could lack any illegitimate IGH rearrangement in their clonal PCs or the percentage of clonal PCs could be above the cut-off threshold for the detection of 14q32 abnormalities, i.e., 10–16%. To answer this important question, we analyzed these patients with chromosome 3, 9, 11, and 15 centromeric probes to estimate the percentage of clonal PCs. Among the 54 patients with a normal IGH configuration, at least one trisomy was detected in 32 patients (15 patients displayed two trisomies, 5 patients displayed three trisomies, and 2 patients were trisomic for the four probes). The median percentage of clonal PCs (defined as cells presenting at least one trisomy) was 58% (range, 6–97%). Combining the results of both analyses using centromeric probes, slightly but not significantly different results were found in patients with MGUS and patients with SMM: whereas the median percentage of clonal PCs was 71% in patients with MGUS, it was 88% in patients with SMM.

Patients with an illegitimate IGH rearrangement were then analyzed using the CCND1, FGFR3, and MYC probes in combination with probes specific for the constant IGH domain, as reported previously (13). No specific translocation was found in patients lacking illegitimate IGH rearrangement, as determined by the interphase FISH assay. Translocation t(11;14) was observed in 15 patients (10 patients with MGUS and 5 patients with SMM). This incidence is similar to that observed in MM (13). In each case, only one CCND1-IGH fusion was observed in MGUS/SMM. This situation is strikingly different from that observed in MM, where multiple fusions are observed in a significant number of patients (13). In contrast to t(11;14), t(4;14)(p16;q32) occurs in a significantly lower number of patients: only two patients displayed this translocation [one individual with MGUS and one patient with SMM; P = 0.002 for difference with t(11;14)]. This incidence is significantly lower than that found in MM (12%; P = 0.013; Ref. 13). No patient displayed t(8;14)(q24;q32).

Using the 13q14 probe, significant percentages of PCs with only one signal were observed in 21 patients (5 patients with SMM and 16 patients with MGUS). The results are summarized in Table 2. Whereas 14q32 abnormalities were observed in the large majority of clonal PCs, deletion 13q14/monosomy 13 was usually present in a lower percentage of cells (median, 34%; range, 14–89%). Only seven patients displayed a high percentage of PCs with monosomy 13, i.e., 66–89% of PCs, as observed in patients with overt MM. Of note, four of these seven patients had SMM. Thus, whereas 4 of 21 patients with SMM presented 13q14 deletion in a majority of clonal PCs, only 3 of 79 patients with MGUS had this feature (P = 0.051). The median percentage of clonal PCs (defined by the percentage of PCs with another 14q32 abnormality or trisomy for one of the centromeric probes) could be evaluated in 13 of these patients and was 75% (range, 36–93%). In 7 of these 13 evaluable patients, the percentage of PCs with monosomy 13 was significantly lower (Table 2).

Chromosomal analysis of clonal PCs in MGUS is a difficult challenge, and very few cases have been reported. Several reasons may explain the poor literature on this topic. The two main explanations are the low number of clonal PCs within the bone marrow (usually less than 2%) and the low proliferation index of these PCs. Even in MM, in which the number of clonal PCs is much higher, chromosomal analyses remain a difficult art, and clonal chromosomal abnormalities are found in only 30–50% of patients (5, 6, 7, 8), although 90–100% of them display aneuploidy in tumor cells, as shown by DNA index or interphase FISH analyses (19, 20, 21, 22, 23). In MGUS, very few studies dedicated to chromosome analysis have been reported. Three main reports using interphase FISH with four centromeric probes identified numerical chromosomal changes in 50–84% of patients with MGUS (10, 11, 12). These percentages (although on a low number of patients) are similar to those found in MM. Very few data are available regarding structural chromosomal abnormalities. Using conventional cytogenetics, only one report described the occurrence of chromosome changes in 25% of patients with MGUS (9). However, these changes are strikingly different from those found in MM, especially by the lesser degree of complexity (usually a single chromosomal change). Apart from this study, no other cytogenetic study has been published. Using interphase FISH, Nishida et al. (14) reported the presence of 14q32 abnormalities in three of five patients with MGUS. We recently preliminarily reported 14q32 and 13q14 abnormalities in 11 of 19 MGUS patients (patients included in this study; Ref. 15). To our knowledge, no other study reporting structural chromosomal changes in MGUS has been published.

In this study, we looked for both types of chromosomal changes, i.e., numerical (chromosomes 3, 9, 11, and 15) and structural (13q14 and 14q32 chromosomal regions). Moreover, we also analyzed the most frequent specific chromosomal translocations involving the 14q32 region: (a) t(11;14)(q13;q32); (b) t(4;14)(p16;q32); and (c) t(8;14)(q24;q32). First of all, this study confirmed previous reports on the incidence of chromosome gains in MGUS/SMM (67% of patients; Refs. 10, 11, 12). Because the analysis was performed on purified PCs, a larger number of cells could be analyzed, and more precise percentages of clonal PCs could be defined. The median percentage of clonal PCs in patients with MGUS was 71% and 88% in patients with SMM. This percentage of clonal PCs in patients with MGUS is higher than those reported previously using FISH (10, 11, 12) but is similar to that described by Ocqueteau et al. (4) using an immunophenotypic analysis. This difference might be explained by the higher number of PCs analyzed and, moreover, by the fact that most of our patients were not studied at diagnosis, but at various times after diagnosis. This finding, associated with the even higher percentage of clonal PCs in patients with SMM, supports the hypothesis of a sequential increase of clonal PCs with MGUS evolution.

For the first time, we showed on a large series of patients that structural chromosomal abnormalities are present in a high proportion of patients with MGUS/SMM. We focused our study on two of the most recurrent changes found in MM, i.e., 14q32 rearrangements and 13q14 deletions. Illegitimate IGH rearrangements were observed in 36 of 79 patients with MGUS (46%) and 10 of 21 patients with SMM (48%). These incidences are lower than that found in overt MM (61%). This difference could reflect a continuous mutational process in PCs, as shown by VH mutation analysis. However, it could also be interpreted as not significant, because of the possible inclusion of patients with very few clonal PCs, leading to a misclassification of these patients as nonprearranged (false-negative patients). The analysis of clonality using centromeric probes in patients lacking 14q32 abnormalities showed that some informative specimens contained only a small percentage of clonal PCs (6%, and possibly less). Because the cut-off for determination of illegitimate IGH rearrangements varied between 10% and 16%, we may have missed some patients with a 14q32 abnormality. Thus, the conclusion of this 14q32 study is that at least 46% of patients with MGUS/SMM display an illegitimate IGH rearrangement. We cannot rule out the hypothesis of an identical incidence in MGUS, SMM, and overt MM.

The second part of this 14q32 analysis was the search for specific translocations. We have previously shown that two translocations were highly recurrent in MM, i.e., t(11;14) and t(4;14), respectively, found in 16% and 12% of patients. To a lesser degree, t(8;14) was also recurrent (3% of MM patients). Surprisingly, t(11;14) was observed in 15 patients, an incidence similar to that of overt MM. Based on published data (24, 25), this translocation was described as a poor prognostic cytogenetic abnormality and thus was not expected in this series of patients with MGUS/SMM. Other poor prognostic abnormalities (such as monosomy 13) might be responsible for the outcome of the patients reported previously. This similar incidence favors the hypothesis of a similar incidence of 14q32 abnormalities in MGUS/SMM and overt MM, rather than an ongoing acquisition during PC evolution.

In contrast, the second highly recurrent 14q32 abnormality, i.e., t(4;14), was observed in only two patients. This lower incidence might be interpreted in two ways: (a) either it occurs later in PC oncogenesis; or (b) this specific translocation directly precipitates the PCs in a true myeloma cell. Hypothesizing a similar incidence of 14q32 abnormalities in MGUS/SMM and MM, we favor the second theory. Interestingly, the two patients bearing the t(4;14) have monosomy 13 in a large number of PCs. This association between the two chromosomal abnormalities is in agreement with the results found in MM. Apart from these 2 patients, deletion 13q14 or monosomy 13 was observed in 19 other patients. In contrast to the 14q32 abnormalities observed in the majority of clonal PCs, deletion 13q14 was often present in a subset of clonal PCs. Among the 21 patients with deletion 13q14, 13 were evaluable for clonality. Six patients displayed deletion 13q14 in the majority of clonal PCs, and seven patients presented deletion 13q14 only in subclones of PCs (Table 2). Of note, this latter finding was observed in six of nine evaluable patients with MGUS and in only one of four evaluable patients with SMM. This situation differs from MM, in which deletion 13q14 is usually observed in the large majority of clonal PCs. The presence of deletion 13q14 in patients with MGUS/SMM was also unexpected. This abnormality has been clearly associated with a poor outcome in MM (23, 26) and represents one of the most powerful prognostic factors in MM. However, the incidence of deletion 13q14 is twice as low in MGUS/SMM as in MM. Moreover, we have recently shown that deletion 13q14 was tightly associated with post-MGUS MM. These data, together with the higher percentage of clonal PCs bearing deletion 13q14 in SMM than in MGUS, support the hypothesis of a possible acquisition of deletion 13q14 during MGUS evolution. This chromosomal change could be involved in the transition between MGUS and MM. In this regard, the evolution of these patients with deletion 13q14 will be very interesting.

In conclusion, the analysis of this large series of patients demonstrated for the first time the presence of structural chromosomal changes in MGUS/SMM. Moreover, the comparison of the results obtained in MGUS, SMM, and overt MM (Tables 3and 4) enables us to construct a theoretical model of PC oncogenesis. Whereas 14q32 abnormalities are early events, deletion 13q14 is involved in the transition of MGUS to MM. Although other oncogenetic changes could be implicated in this transition, a close survey of our MGUS/SMM patient cohort will be of great value in validating (or invalidating) this model.