Treatment advances in multiple myeloma

Autotransplant-supported high-dose therapy has improved the outcome of multiple myeloma for two-thirds of patients with normal metaphase cytogenetics, 40% of whom enjoy a survival of >10 years. This approach has been challenged by the introduction of non-myeloablative allotransplants and novel drugs also targeting the microenvironment. Non-myeloablative allotransplantation may increase the cure rate but needs evaluation in patients with high-risk myeloma (abnormal cytogenetics) because of its high risk and absence of mature data. Novel drugs may induce marked responses in 10–20% of patients. However, little is known about the quality and duration of these responses and their long-term toxicities. Given the marked cytoreductive potential of autotransplantation, it should be considered a platform for more effective future therapies. Incorporation of novel agents in autotransplant protocols appears appropriate for good risk disease (normal cytogenetics), whereas non-myeloablative allotransplants or other immunological manipulations should only be tested in high-risk myeloma (abnormal cytogenetics). Little progress had been made in the first 30 years after the introduction of melphalan-prednisone as palliative treatment for multiple myeloma in the early 1960s. Approximately 5% of patients attained a stringently defined complete response (CR; normal bone marrow and absence of monoclonal protein in blood and urine by immunofixation electrophoresis). The median survival had not been extended beyond 30–36 months, and no more than 5% of patients lived longer than 10 years (Barlogie et al, 1992). Considerable progress was made when, in the early 1980s, dose-escalated melphalan (100–140 mg/m2) resulted in a CR rate of 35% and a median overall survival (OS) of 48 months when applied to newly diagnosed patients (Sirohi et al, 2001). Treatment-related mortality (TRM) due to prolonged cytopenia after high-dose melphalan was reduced from 15% (Moreau et al, 1997; Sirohi et al, 2001) to 2–3% (Barlogie et al, 1995) with the introduction of mobilized peripheral blood stem cell (PBSC) support, which also allowed further dose-escalation of melphalan to 200 mg/m2. Thus, the TRM with autotransplants is not higher than with 6 months of conventional chemotherapy (Alexanian et al, 1990; Monconduit et al, 1992). The superiority of high dose over standard dose chemotherapy has been demonstrated in two randomized trials (Attal et al, 1996; Child et al, 2003): a comparison with matched historical controls (Barlogie et al, 1997) and a population-based study (Lenhoff et al, 2000). Consequently, in its evidence-based review, the American Society for Blood and Marrow Transplantation concluded that autotransplantation is the preferred treatment modality for myeloma and its application is recommended as de novo rather than as salvage therapy (Hahn et al, 2003). The Intergroupe Français du Myelome (IFM) randomized trial (IFM-94) reported a 7-year OS of 42% with tandem autotransplants versus 21% for a single transplant (P = 0·01), with a median follow-up of 5 years (Attal et al, 2002). An Italian randomized trial (Bologna-96) showed a superior event-free survival (EFS) for tandem compared with single autotransplantation (median 34 vs. 25 months; P = 0·05), with no difference yet in OS with a median follow-up of only 3 years (Cavo et al, 2002). Another French study by the Myèlome-Autogreffe group (MAG-95) failed to reveal a difference in outcome between single and tandem transplants, with a median follow-up of 40 months (Fermand et al, 2001). The Arkansas total therapy I phase II trial, which applied tandem autotransplants with melphalan 200 mg/m2, effected a median EFS and OS of 43 and 68 months, respectively (Barlogie et al, 1999). At 10 years, 15% of the patients remained event-free and 33% were still alive. In an analysis of 515 consecutive, newly diagnosed as well as previously-treated patients, one-fifth remained event-free beyond 7 years (Tricot et al, 2002). The absence of relapses past this point suggests that ‘cure’ is possible with autotransplantation. This supports the original hypothesis that a profound reduction in tumour mass results in prolonged disease control and, therefore, superior survival. The European Group for Blood and Marrow Transplantation reported on the outcome of 162 allotransplants performed between 1983 and 1993 (Gahrton et al, 1995). Forty-four per cent entered CR. The OS at 4 and 7 years was 32% and 28%, respectively, with the best outcome seen in patients transplanted after first-line therapy with chemosensitive disease. The 5-year relapse rate was 45%. A subsequent update showed a significantly reduced 2-year TRM (46% vs. 30%) in patients receiving an allotransplant after 1993 (Gahrton et al, 2001). However, disappointingly, no reduction was seen in relapse rate although the transplants after 1993 were performed earlier in the disease course. In a retrospective case-matched analysis of allo- versus autotransplantation, OS was superior for autotransplantation (34 vs. 18 months; P = 0·001) (Bjorkstrand et al, 1996). The main cause of poor survival with allotransplantation was the high TRM, while the ultimate relapse rate was similar in both groups, although somewhat delayed with allotransplantation. It is generally accepted that allotransplantation is the only curative approach for myeloma, mainly based on minimal residual disease (MRD) studies using polymerase chain reaction analysis. Although molecular remissions are rare after autotransplantation (7% vs. 50% after allotransplantation) (Corradini et al, 1999), there are no convincing clinical data that show a higher ‘cure’ rate with allotransplantation. To further reduce TRM, the concept of non-myeloablative allotransplantation – with its reduced toxicity – after a single autotransplant was introduced. In a study by Kroger et al (2002), the CR rate was 73% and the TRM in the first 100 d was 11%. With a median follow-up of only 13 months after allotransplantation, the 2-year EFS was 56%. In a Seattle study, CR after allotransplantation was attained in 57%, and the 100-d mortality was 2%; 46% of patients developed chronic graft-versus-host disease (cGvHD), requiring therapy (Maloney et al, 2003). The overall TRM was 16%, and the 2-year EFS was 55%, with a median follow-up of 18 months after allotransplantation. Thalidomide, introduced for the treatment of multiple myeloma in 1998, was the first new agent with reproducible activity in end-stage myeloma since the introduction of alkylating agents and corticosteroids. One-third of patients, the majority of whom had relapsed after autotransplantation, showed more than 50% reduction in serum or urine myeloma protein after treatment with thalidomide (Barlogie et al, 2001). Some attained their best response ever on thalidomide. Remarkably, although the median EFS was only 9 months, 28% of patients with normal metaphase cytogenetics remained event-free at 4 years. Another immunomodulatory drug, Revimid, lacking many of the toxicities of thalidomide but producing more severe cytopenia, was active in one-third of patients with advanced and refractory myeloma (Richardson et al, 2002). The proteasome inhibitor, bortezomib (velcade), effected a 35% PR (≥ 50% reduction in serum in M protein and/or ≥ 90% in urine M protein) rate, including a 9% CR or near CR (only immunofixation positive) rate among 193 evaluable patients with advanced myeloma, lasting a median of 12 months (Richardson et al, 2003). Toxicities included thrombocytopenia, fatigue and peripheral neuropathy. Adding thalidomide to velcade induced partial responses in over 50% of patients, of whom 20% achieved a CR or near CR. Unfortunately, abnormal cytogenetics, while not affecting response frequency with this combination therapy, were associated with inferior EFS (Zangari et al, 2003). These novel agents have multiple modes of action, including altering the microenvironmental response to myeloma invasion (Hideshima et al, 2001). In the majority of patients, multiple myeloma probably evolves from an indolent phase of monoclonal gammopathy of undetermined significance or smouldering myeloma to symptomatic disease (Hallek et al, 1998) (Fig 1). In the early stages of symptomatic disease, myeloma cell growth and survival are dependent on the microenvironment, explaining the virtual absence of informative metaphases and the difficulty to establish in vitro myeloma cell growth. Only in a severe combined immunodeficient (SCID)-human model, which recreates the human microenvironment by placing a human fetal bone under the skin of a SCID mouse, can primary myeloma cells be grown in a consistent manner (Yaccoby & Epstein, 1999). The absence of growth outside of the fetal bone underlines the dependence of myeloma on the human microenvironment. In the advanced stages of the disease, myeloma becomes stroma-independent, explaining the extramedullary manifestations usually associated with increased lactate dehydrogenase levels. It is currently unknown whether such ‘malignant myeloma’ always represents an evolution from a more benign phase, akin to large cell transformation in follicular lymphoma or chronic lymphocytic leukaemia, or a de novo condition. Multistep progression of multiple myeloma in relation to stroma dependence, drug resistance, cytogenetics and treatment outcome after autotransplantation. MGUS, monoclonal gammopathy of undetermined significance; FISH, fluorescence in situ hybridization. Drug resistance in the early stages of myeloma is thought to be mainly induced by the microenvironment (Fig 2). Ligands to receptors present on myeloma cell membranes [interleukin-6 (Klein et al, 1995), interleukin-15 (Tinhofer et al, 2000) and insulin-like growth factor (Qiang et al, 2002)] activate anti-apoptotic pathways (leading to drug resistance) and stimulate myeloma cell growth and angiogenesis. Through β-1 integrin-mediated adhesion of myeloma cells to the stroma, p27 is upregulated (Damiano et al, 1999; Hazlehurst et al, 2000). The resulting G1 arrest provides a cytokinetic sanctuary for most cytotoxic agents (cell adhesion-mediated drug resistance) except for some alkylating agents in high doses such as melphalan, busulphan and carmustine. Administration of these alkylators in high doses is also toxic to non-cycling normal haematopoietic stem cells, which are protected from apoptosis after exposure to conventional chemotherapy by the same mechanism of adhering to bone marrow stroma, and will therefore require PBSC support. Drug resistance can also be induced by DNA methylation or chromatin remodelling, resulting in gene expression changes without causing DNA nucleotide sequence changes (Verma & Srivastava, 2002). Hypermethylation of p15, p16 (Ng et al, 1997) and of the death-associated protein kinase (Ng et al, 2001) have been reported in myeloma. Such epigenetic changes are environmentally induced by radiation, tobacco smoke, stress, hormones and various other chemicals (Verma & Srivastava, 2002) and may already be present in the early stages of the disease (Guillerm et al, 2001). Multiple mechanisms of drug resistance in myeloma. MM, multiple myeloma; IAP, inhibitors of apoptosis; CAM-DR, cell adhesion-mediated drug resistance. In the malignant myeloma entity, drug resistance is mainly induced by a multitude of genetically acquired changes, such as point mutations in RAS oncogene (Liu et al, 1996) (rendering myeloma cells independent of interleukin-6) and p53 mutations (Imamura et al, 1994) (blocking apoptosis induced by high-dose alkylators). Despite its stroma-independence, malignant myeloma still derives growth-stimulatory and anti-apoptotic stimuli from the microenvironment, maybe even to a larger extent than stroma-dependent myeloma, explaining why the disease remains mainly confined to the bone marrow. Whether advanced myeloma induces irreversible changes in the microenvironment, thus explaining early relapse even after high dose therapy, is a subject of intense investigations. Metaphase cytogenetic analysis is an excellent surrogate marker to evaluate stroma dependency, because in this assay myeloma cells are removed from their protective microenvironment and are forced to grow in a non-supportive milieu. It is not surprising that stroma-dependent myeloma cells will not divide under these conditions. Therefore, in early myeloma, cytogenetics are normal with the normal metaphases deriving from remaining haematopoiesis, while advanced myeloma grows independent of its protective microenvironment and therefore will show abnormal cytogenetics. Our clinical studies have consistently shown that metaphase cytogenetics are the single most important prognostic factor for outcome (Tricot et al, 1995, 2002; Shaughnessy et al, 2003a). Such information cannot be derived from fluorescence in situ hybridization analysis (Shaughnessy et al, 2003b). With a median follow-up of 10 years, the Arkansas total therapy I trial produced a superior OS in two-thirds of patients without cytogenetic abnormalities (median of 80 vs. 51 months; 10-year survival of 43% vs. 10%; P = 0·0001; Fig 3). Total therapy II, providing more intensive induction and consolidation, and administering more dexamethasone post-transplantation, effected superior EFS and OS in comparison with total therapy I. However, this marked improvement was only observed in patients with normal cytogenetics and thus stroma-dependent myeloma (Fig 4). With current treatment modalities, the potential for long-term disease control and even cure appears good for stroma-dependent myeloma, while malignant myeloma requires new approaches. Event-free and overall survival of the 231 patients treated with total therapy I. No CA, normal metaphase cytogenetics; CA, abnormal metaphase cytogenetics; no cytogenetic results were obtained in nine patients; mo., months. Comparison of overall survival between patients treated with total therapy I (TT I) versus total therapy II (TT II), according to metaphase cytogenetics. No CA, normal metaphase cytogenetics; CA, abnormal metaphase cytogenetics. Preliminary results of a single autologous followed by a non-myeloablative allotransplant are encouraging in patients exhibiting cytogenetic abnormalities, presumed to have stroma-independent myeloma. However, the beneficial graft-versus-myeloma effect is typically directed against major or minor histocompatibility antigens and not against specific tumour-associated antigens. Strategies are needed targeting only myeloma cells, while leaving normal tissues intact. Cancer/testis antigens (CTA) are very immunogenic and are found in high frequency in myeloma cells with abnormal cytogenetics (Gupta et al, 2002) but are not expressed in normal tissues (except for testes, ovary and placental trophoblasts) (Old & Chen, 1998). The success of CTA vaccination trials in malignant melanoma (Marchand et al, 1999) have prompted investigations in myeloma, where a stage of MRD can be achieved after autotransplantation, which should represent the ideal setting for immunotherapy. Alternatively, natural killer (NK) cells can be infused, which recognize their target cells by a variety of activating and inhibiting receptors (Farag et al, 2002). In the autologous setting, NK cells are not effective for disease control, because inhibitory signals predominate. In contrast, NK cells lyse allogeneic cells if mismatched for one of the killer cell immunoglobulin-like receptors (KIR) ligands (Farag et al, 2002). Spectacular results were obtained in patients with myeloid leukaemia receiving transplants from a human leucocyte antigen-haploidentical, KIR-ligand mismatched donor, where NK cells completely prevented rejection, GvHD, and relapse (Velardi et al, 2002; Giebel et al, 2003). Myeloma cells are susceptible to NK-induced cell kill, whereas CD34+ haematopoietic cells are resistant (Frohn et al, 2002). In contrast to cytotoxic T cells, which require a continuous presence to be effective, NK cell killing is rapid so that short exposure may suffice. We are therefore evaluating autotransplants following infusion of KIR-ligand mismatched NK cells in high-risk myeloma. Myeloma is a heterogeneous malignancy. At present, metaphase cytogenetic analysis provides the best tool to discriminate good risk (stroma dependent) and poor risk (stroma independent) myeloma. The therapeutic challenge in good risk myeloma is to improve upon outcome without jeopardizing the accomplishments of high-dose melphalan therapy. As more than 40% of such patients will be alive after more than 10 years, such patients should not be subjected to yet unproven, high risk procedures, such as non-myeloablative allotransplants. Rather than testing novel immunomodulatory and proteasome-inhibiting agents upfront and thereby excluding the early application of high dose therapy, the post-transplant platform of marked tumour control appears a logical scenario for evaluating maintenance therapy with these novel agents. Well-designed translational research should help determine whether these myeloma and microenvironment targeting agents can sustain durable responses by interrupting the survival and growth signals provided to myeloma cells by the microenvironment. In malignant myeloma, characterized biologically by stroma-independence and reflected by the presence of metaphase cytogenetic abnormalities, rapid post-transplant relapse probably results from less profound subclinical tumour cytoreduction because of genetically acquired drug resistance and more rapid re-growth, possibly facilitated by a hyperactive stroma. DNA microarray studies of highly purified CD 138-positive myeloma cells and of whole bone marrow biopsy samples are in progress to investigate this question. Given the inferior outcome of these patients with tandem transplants, it is justified to test new and potentially riskier treatment strategies such as non-myeloablative allotransplants, CTA peptide vaccination or NK cell immunotherapy, although they would best be performed in combination with autotransplantation. This study was in part supported by grant CA55819 from the National Cancer Institute in Bethesda, Maryland.