Understanding the Molecular Biology of Myeloma and Its Therapeutic Implications

Kevin D Boyd; Charlotte Pawlyn; Gareth J Morgan; Faith E Davies


Expert Rev Hematol. 2012;5(6):603-617. 

In This Article

Using Biological Information to Make Clinical Decisions

The goal of myeloma therapy is personalized medicine, whereby biomarkers are used to define a patient group that will derive benefit from a specific drug. This approach is being driven by two main processes: an increased understanding of the biology of the disease and a broadening of the antimyeloma pharmacopoeia. The challenge now is to incorporate this biological knowledge into a therapeutic strategy in order to improve patient outcome. Moving forward, it will also be important to incorporate disease biology into therapeutic trials. There are two main ways in which biological information can be utilized: a risk-adapted strategy and targeted therapy.

Risk-adapted Therapy

The aim of a risk-adapted therapeutic approach is to define patient groups associated with favorable or adverse clinical outcomes and tailor treatment accordingly. It is recognized that prognosis in myeloma can be linked to both tumor and patient variables. The most widely applied prognostic system in myeloma is the International Staging System (ISS), which stratifies patients into three groups based on serum albumin and β2-microglobulin.[124] These two factors reflect both patient and tumor factors, with β2-microglobulin being a measure of tumor bulk and renal function, while albumin is associated with patient performance status. A significant limitation of this system is that it does not take into account the role of intrinsic myeloma cell variability. As such, it has clinical utility for stratification of groups of patients within clinical trials, but has limited utility for assessing risk on an individual patient basis. This review has shown that clinical outcome in myeloma can be linked to genetic lesions of the myeloma tumor, and in order to carry out a risk-adapted approach, robust prognostic biomarkers are required. There are two main methods that have been used to define biological high-risk disease: iFISH and gene expression profiling (GEP).

Defining Prognostic Groups Using iFISH. The International Myeloma Working Group evaluated the role of iFISH for stratifying patients into genetic risk groups in 2009, and based on the published data available at that time, they concluded that there was currently evidence for testing for t(4;14), t(14;16) and del(17p).[125] Since that time, it has been recognized that testing for +1q and t(14;20) may also be clinically important.[37] These lesions are all associated with impaired survival and can be used in the definition of high-risk disease. No FISH lesions have definitively been linked to good clinical outcome, and favorable prognosis myeloma is defined genetically by the absence of unfavorable features. The major remaining groups after these poor prognostic associated lesions have been removed are t(11;14) and hyperdiploid myeloma, provided they lack additional adverse lesions such as +1q21 and del(17p).

Only one study has looked at the interaction of each of these variables.[37] In this study, based on outcome data from the MRC Myeloma IX trial, the FISH probes associated with short survival in multivariate analysis were +1q21, del(17)(p13) and an adverse IGH@ translocation group comprising t(4;14), t(14;16) and t(14;20). +1q21 and del(17)(p13) often occurred together with a recurrent translocation in the same tumor and this co-segregation of adverse FISH biomarkers carried clinical significance. There was a clear association between the number of adverse FISH lesions and a progressive impairment of progression free survival and OS. A favorable risk group can be defined by the absence of +1q, del(17p) or adverse IGH@ translocations, an intermediate risk group has just one of these lesions and a high-risk group has the co-segregation of multiple adverse lesions. The implication from this study is that no one genetic lesion by itself defines high-risk myeloma and that it is important to know the presence or absence of a panel of genetic lesions in order to properly identify an adverse prognostic group using FISH.

The ISS and FISH consider different factors and can be combined in a prognostic model, and this approach has been used by several groups. The MRC Myeloma IX trialists defined high-risk disease by the presence of multiple adverse FISH lesions combined with ISS II or III. In the Intergroupe Francophone du Myélome experience, the combination of t(4;14), del(17p) and a high β2-microglobulin level produced similar results.[14] Thus, at this point in time, there are good data to support the application of a combined ISS and FISH approach to define patient disease characteristics that could usefully be applied to patient management or trial design.

Defining Prognostic Groups Using Gene Expression Signatures. Global GEP represents an alternative technology for assessing tumor genetic risk factors, and there has been considerable work in the development of prognostic information based on the use of this technique. Gene expression microarrays quantify mRNA, which may capture information regarding alterations to DNA as well as taking into account some aspects of transcriptional regulation. The UAMS group was the first to define high-risk disease in terms of gene expression, with a 70 gene expression signature, which was subsequently refined to a 17 gene signature that performed as well as the original.[38] Since then, other gene expression signatures linked to short survival have been defined including a 15 gene signature trained on patients treated in the IFM trials, and a 6 gene signature trained on patients treated in the MRC Myeloma IX trial.[126,127] It is interesting to note that although all these signatures have included patients treated with high-dose therapy, the 17, 15 and 6 gene signatures do not share any common genes. This may be due to variation in methods, treatment strategies and patient selection between studies. Therefore, GEP is valid in the context of individual specialist centers, and the signatures can give valuable information regarding disease biology. However, more work is required to define common genes or common gene functionality before gene expression signatures can be more widely applied as a prognostic tool.

Using Prognostic Biomarkers to Guide Risk-adapted Therapy. While it is currently feasible to define high-risk disease using these data, there are currently no data to support a specific alternative treatment strategy in this group. The three therapeutic options for this group are treatment intensification, allogeneic transplantation and novel therapies; however, all of these strategies require evaluation in a clinical trial setting.

Treatment intensification with dose-dense chemotherapy that is rotated and maintained for a long period of time would induce consistent therapeutic pressure on the myeloma clone. However, although this approach has biological plausibility, there is evidence for the UAMS group that intensification of treatment does not benefit all patients. The 13% of patients defined as at high risk by the UAMS gene expression signature continue to have short survival, even with the very intensive chemotherapy used in the Total Therapy regimens.[38] It is possible, however, that the incorporation of novel agents into a dose dense schedule incorporating high-dose therapy may improve outcomes for some high-risk patients.

The biological rationale of offering allogeneic transplantation to high-risk myeloma patients is to eradicate residual clonal cells with a graft-versus-myeloma effect. An early IFM study assessed the efficacy of allogeneic transplant compared with autologous transplant specifically in high-risk patients defined by high β2-microglobulin and del(13q) detected by FISH.[128] Although the definition of high-risk disease has been improved since the inception of this trial, no difference in event-free survival was seen between the two transplant types.

A third option is to test promising new drugs in this patient group. A caveat of this approach is that potentially useful drugs may be discarded if they fail to show efficacy in a group of patients that are inherently refractory to treatment. A better approach may be to include high-risk patients in these Phase II trials alongside other patients, with a predetermined analysis of outcomes within high-risk subgroups. The implication from this is that genetic analysis is imperative for the effective analysis of patient outcomes in clinical trials.

Targeted Therapeutics

Targeted therapeutics involves using a biomarker to define a patient subgroup that may benefit from treatment with a specific drug. An example of how this could be applied to myeloma treatment is shown by the impact of bortezomib in myeloma with t(4;14). t(4;14) has consistently been shown to be associated with a poor therapeutic outcome in patients treated in a variety of clinical contexts including conventional chemotherapy, thalidomide-based induction therapy and high-dose melphalan with ASCT. A number of reports suggest that bortezomib may improve the outcome of this group.[17] Patients with t(4;14) treated with a nonintensive bortezomib-based approach in the VISTA study did not have inferior survival compared with patients without t(4;14).[16] t(4;14) was not associated with impaired survival in patients treated on the Total Therapy 3 programme, which incorporated bortezomib into an intensive chemotherapy regimen with tandem autologous transplantation and a prolonged maintenance phase.[129] Perhaps the most convincing evidence of improved outcome with bortezomib in patients with t(4;14) comes from a subgroup analysis of patients treated in the IFM trials.[17] In this analysis, t(4;14) remained a biomarker associated with poor prognosis in 106 patients with t(4;14) treated with four courses of bortezomib induction chemotherapy. However, these patients had superior survival when compared with patients with t(4;14) treated with non-bortezomib based therapy within the same time period. While this was not a randomized study, these data suggest that the short survival associated with t(4;14) may be partially overcome by bortezomib induction therapy. This suggests that t(4;14) may define a patient group whose outcome could be improved if targeted with a bortezomib-based therapeutic approach compared with conventional therapy.

In addition to identifying biomarkers that define sensitivity to existing therapies such as the examples of t(4;14) and bortezomib, some of the novel agents in development in myeloma have specific targets and, therefore, lend themselves to a more individualized approach. Several examples of the potential application of this approach have been previously outlined in this review, including MEK inhibitors in myeloma with increased MAF expression, and BRAF kinase inhibitors in patients with BRAF mutations.[130]

However, if this targeted approach is to be taken, it is important to ensure that the lesion being targeted is central to the disease process (e.g., a driver rather than a passenger mutation) and that the lesion is present in the majority/all of the myeloma cells. Recent reports have suggested that clonal heterogeneity is common in cancer.[73,131,132] Importantly, studies also suggest that clones respond differently to therapies, leading to the concept of clonal tides.[74,133]

Clinical Trial Design

The biological heterogeneity of myeloma has implications for the design and interpretation of clinical trials. In large, Phase III trials, it is now appreciated that it is important to have genetic information in order to properly interpret clinical outcome measures. An example of the importance of this is the effect of maintenance therapy in the MRC Myeloma IX trial.[134] In this trial, patients were randomized to maintenance thalidomide versus no maintenance until disease progression. When the results of this randomization were examined in the whole trial population, thalidomide was shown to be associated with improved PFS but no change in OS due to shorter survival after relapse. However, these clinical outcomes were found to be different in the context of different genetic risk groups. In high-risk patients, maintenance was not associated with improved PFS and there was a negative effect on OS, while in favorable risk patients, maintenance was associated with improved PFS and improved OS in the context of effective relapse therapy. Patients without high-risk genetic features therefore benefited from thalidomide maintenance, while patients with a high-risk genetic background who received thalidomide maintenance derived no benefit, and their outcome may even have been inferior to patients who received no maintenance. These conclusions would not have been possible if the tumors had not been characterized by an effective iFISH panel. In a similar way, in all large Phase III trials, it will be important to have predetermined analyses defined to examine the end points within the context of biologically defined groups.

Moving beyond large Phase III trials, the knowledge that patients can be stratified based on genetic risk factors may alter trial design. For high-risk subgroups with median OS times of less than 24 months, smaller, more focused studies are possible, as defined end points will quickly be reached. By contrast, low-risk subgroups may have median survival times of more than 8 years if treated intensively, and to gather sufficient outcome data, these groups require larger studies with longer follow-up. It is possible, therefore, to design smaller trials in order to answer specific therapeutic questions in defined biological groups. High-risk myeloma could be targeted in this way, by randomizing patients to a standard therapeutic approach versus a more intensive approach.

Disease biology may alter the design of trials of newer, more targeted therapeutic agents. The effect of drugs that are useful in a small subset of myeloma patients, determined by the biology of their disease, may not be seen if the drug is tested in unselected patients. If the target of the drug is well defined, and there is an appropriate biomarker to identify patients suitable for treatment, patients can be targeted in smaller trials that are more likely to show clinical benefit.