Although several studies have been devoted to the incidence and signification of activating Ras mutations in patients with plasma-cell disorders, and especially with multiple myeloma (MM), to date it remains difficult to conclude the exact significance of such mutations in these diseases. In this context, we have been interested in the work of Kalakonda et al1 showing for the first time that codon 61 mutations of the NRAS gene were universal in patients with MM at presentation in a subpopulation of 12% to 100% of malignant plasma cells. But we would like to comment on their findings with regard to the discrepancy existing between this study and some others (summarized in Table1), including our own recent study2 (Table 2) re-evaluating both NRAS and KRAS2 mutations in plasma-cell disorders.
Overall, in these studies, the percentages of NRAS mutations range from 0% to 26% with a majority involving codon 61. Higher percentages were recently found by our group,2 in agreement with those of the Eastern Cooperative Group Phase III Trial.3 The argument of Kalakonda et al for explaining the higher incidence of NRAS codon 61 mutations in their study is that less sensitive screening strategies have been used in earlier studies. This deserves some comment. Indeed, this argument cannot be applied to our study,2 for example. The one-stage allele-specific polymerase chain reaction (PCR) we developed was sufficient to detect the lower percentage of tumor cells with mutant allele (12%), which has been observed on purified plasma cells in Kalakonda et al's study. Thus there is no clear explanation for such a discrepancy.
Another point that deserves some comment is that, curiously, Kalakonda et al did not find any mutations (except one) in the KRAS2gene in 34 evaluated MM cases. This result contrasted with our data showing 35% of KRAS2 mutations at diagnosis, as compared with 25% of NRAS mutations. If we had included human multiple myeloma cell lines (HMCLs) in our discussion, we would have observed that 33% of cell lines tested were also positive for KRAS2 mutations. Thus clearly, KRAS2mutations were underrepresented in the patients series from Kalakonda et al1 without any evident explanation. A few number of mutations in KRAS2 were found in other studies, but in these cases a number of them used a low sensitive method. On the other hand, Liu et al3 found 16.5% of mutations using a more sensitive method. Therefore, Kalakonda et al's study is not really consistent with our own study and that of Liu et al.3 Of major interest, in our series KRAS2mutations were even more frequent than NRAS mutations, were antinomic, and were significantly associated with extramedullary disease, thereby associated with a more aggressive disease as in Liu et al's study (Table 2). NRAS mutations were less frequent and involved mainly cases restricted to medullary progression. This model is difficult to reconcile with a model presenting NRAS mutations of codon 61 without KRAS2mutations as a universal abnormality in MM regardless disease status.
Activating Ras mutations in multiple myeloma
By using a newly developed allele-specific amplification (ARMS) method with improved sensitivity, Bezieau et al report that in their patient series activating mutations affecting the KRAS2 gene are more prevalent than those involving the NRASgene1-1 (35% of cases and 25% of cases, respectively). This finding contrasts with all other published data, including our own,1-2 on RAS mutation frequency in multiple myeloma, which show that mutations in NRAS(specifically those at codon 61) are most prevalent. There is however, a wide variation in the reported frequency and spectrum ofRAS mutations in published studies. Whereas differences in sensitivities of various methods used for mutation detection would undoubtedly account for some of these disparities, the particular cohort(s) of patients studied by different groups may also be important, particularly as most studies (including our own) recruited relatively small numbers of patients. For example, Bezieau et al highlight an apparent association between the occurrence of KRAS2mutation and extramedullary disease, as alluded to in an earlier study.1-3 While our investigation did not address the issue of clinical associations, it may be significant that (on review of clinical data) none of the patients in our study had extramedullary disease, and this may well account for the alleged underrepresentation of cases harboring mutant KRAS2in our study. By similar reasoning, if the patient group studied by Bezieau et al included a disproportionate number of cases with more aggressive disease, compared with other published studies, then this might well account for the overrepresentation of KRAS2mutation–positive cases seen in their study.
It is puzzling why Bezieau et al did not detect a higher proportion of cases harboring mutations in the NRAS gene, specifically at codon 61. But aside from differences in patient cohorts, 3 considerations may be pertinent to this issue: (1) The quoted sensitivity of their ARMS method is 1 in 10,1-3 demonstrable by serial dilution of cells harboring a KRAS codon 12 mutation.1-1 This is an impressive figure for sensitivity, but it may not necessarily be extrapolated to detection of other mutations, for example in the NRAS gene, particularly at codon 61. (2) In our study, screening for RAS mutations was performed using BB4-purified myeloma cells from all patients.1-2 (3) As our study demonstrated, NRAScodon 61 mutation–positive cells represent only a subset of the malignant cell population in the vast majority of patients (range, 12%-100%).1-2 Although the lowest representation ofRAS mutation–positive cells in our study was 12%, there is no particular reason for believing that this figure represents an arbitrary lower limit; there may well be myeloma patients, albeit uncommon, in which RAS mutation–positive cells represent considerably less than 12% of the total tumor cell population. For these reasons, it is quite plausible that the study of Bezieau et al may have underestimated the frequency of NRAS mutations in their patient series. What is certainly clear from both our studies is that the use of more sensitive methods of detection leads to a higher rate of detection of RAS mutations than reported in earlier work. As an aside, we are skeptical about interpreting data on the prevalence of RAS mutations in studies of myeloma cell lines (in which of course, sensitivity of detection is not an issue).RAS mutation–positive hematopoietic cell populations are well known to be highly dynamic in vivo. For example, cases of acute myeloid leukemia with activating RAS mutation at presentation have been found to have lost the mutant RAScell population at relapse.1-4 It is not therefore difficult to envisage how, during in vitro culture, RASmutation–positive myeloma cells may be selectively lost from the population during establishment. We think the reported absence ofRAS mutations in cases of myeloma cell lines probably says more about the cell immortalization/establishment process than it does about the prevalence of RAS mutation in primary myeloma.
One final point concerning the subclonal nature of RASmutation–positive cells in myeloma (a central finding of our study1-2) deserves some comment in relation to the study of Bezieau et al.1-1 In relapsed disease, Bezieau et al reported an incidence of RAS mutation of 81%, that is, significantly higher than at presentation in their own study.1-1 This would intimate that there are a significant number of patients who would be RASmutation–negative at presentation (within the limits of sensitivity of their detection method) who subsequently becomeRAS mutation–positive at relapse. This being the case, the question then arises as to exactly when during disease course do theseRAS mutation–positive cells seen in relapse actually arise? Either they arise de novo at some stage subsequent to presentation, or, as we imply in our study, they may already be present in low levels at presentation. During subsequent disease course culminating in relapse, the RAS mutation–positive population gradually expands to become the predominant clone and, hence, more amenable to detection. Distinguishing between these possibilities through serial follow-up analysis of individual patients is of fundamental importance to understanding RAS mutation in the pathobiology of multiple myeloma. Clearly, the issue of the role of activating RASmutation in plasma cell disorders is worthy of further appraisal.
