Autologous stem cell transplantation and multiple myeloma cancer stem cells

Abstract

It is well established that high dose therapy (HDT) combined with autologous stem cell transplantation (ASCT) produces superior response rates and progression-free survival compared to conventional chemotherapy in patients with multiple myeloma (MM). Accordingly, MM currently represents the most common indication for ASCT. Despite these clinical improvements, the impact of ASCT on overall survival is unclear since the vast majority of patients eventually experience disease relapse and progression. The continual risk of relapse suggests that malignant cells resistant to HDT possess the clonogenic growth potential to mediate tumor regrowth, and in several diseases cancer stem cells (CSCs) have been identified that are both highly tumorigenic and resistant to standard anti-cancer approaches. Putative CSCs have been identified in MM, and their characterization may lead to the development of novel maintenance strategies that inhibit the production of new tumor cells, prevent disease relapse, and improve overall survival.

AUTOLOGOUS STEM CELL TRANSPLANTATION IN MULTIPLE MYELOMA

ASCT has been considered the standard approach for frontline therapy in eligible MM patients based on several randomized trials comparing ASCT to conventional chemotherapy. Since 1996, seven randomized clinical trials have been reported and clearly demonstrated the superiority of ASCT in inducing disease responses, especially complete remissions, and prolonging event-free survival. In contrast, the impact of ASCT on overall survival has remained unclear. The IFM90, MRC VII, and Italian MMSG phase III trials demonstrated significant prolongation of median overall survival in newly diagnosed patients undergoing ASCT.(1–3) However, these results are balanced by the PETHEMA, MAG91, MAG95, and US intergroup S9321 trials that failed to detect a significant improvement in overall survival for patients undergoing ASCT compared to those receiving conventional chemotherapy.(4–7)_ENREF_6 Moreover, a recent meta-analysis of 2,411 patients failed to identify a significant overall survival advantage for ASCT.(8) The uncertain impact of ASCT on overall survival is largely due to disease relapse and progression that affects virtually all MM patients regardless of whether or not they have undergone ASCT. Accordingly, the development of maintenance strategies to prolong responses and improve overall survival has been the subject of great interest.

MAINTENANCE THERAPY FOR MULTIPLE MYELOMA IN THE POST-TRANSPLANT SETTING

Several strategies have been examined in hopes of extending disease responses following ASCT. Initial studies focused on the use of interferon-alpha based on its ability to improve overall survival when used as maintenance following conventional chemotherapy.(9) Patients with major reductions in tumor burden when initiating maintenance appeared to gain the greatest benefit. Therefore, the post-ASCT setting was thought to represent the optimal time to utilize maintenance interferon-alpha, and this concept was supported by a retrospective EBMTR analysis demonstrating improved progression-free and overall survival in patients receiving interferon-alpha.(10) Two randomized studies were subsequently undertaken but failed to demonstrate improved overall survival in administered interferon-alpha following ASCT.(7, 11) Although better tolerated post-ASCT than following conventional therapy, the relative toxicity of interferon-alpha combined with the unclear impact on survival has led to the abandonment of this approach.

The introduction of novel agents including thalidomide, its analogue lenalidomide, and the proteosome inhibitor bortezomib over the past decade have dramatically modified treatment approaches for multiple myeloma.(12, 13) All of these agents were initially tested and approved for use in patients with refractory and relapsed disease, but their relative efficacy has led to their examination and incorporation into both initial treatment paradigms as well as in the maintenance setting following ASCT. Maintenance thalidomide has been extensively studied and to date five randomized clinical trials have been reported.(7, 14–17) The IFM 99–02 study compared the use of thalidomide and pamidronate post-ASCT to pamidronate alone or no maintenance in patients with standard-risk myeloma.(14) Here, thalidomide demonstrated significantly greater response, progression-free survival, and overall survival rates. These results were essentially replicated in a randomized study from Australia (ALLG MM6) that compared thalidomide and prednisone with prednisone alone.(15) In contrast, three other large studies have failed to demonstrate improvements in overall survival despite higher response rates and prolonged progression-free and event-free survival. These studies have included Total Therapy 2 in which patients received thalidomide both during induction and as maintenance within a complex regimen that included induction chemotherapy, tandem ASCT, and interferon-alpha and dexamethasone post-ASCT,(18) the HOVON 50 study utilizing thalidomide during induction therapy followed by thalidomide or interferon as maintenance,(17) and the MRC Myeloma IX trial in which thalidomide was administered following either ASCT or conventional therapy.(16) Although an overview of these studies fails to provide definitive evidence for the use of thalidomide as maintenance following ASCT, distinct groups of MM patients appear to benefit. In the IFM 99–02 trial, thalidomide maintenance was found to be beneficial in patients failing to achieve less than a VGPR following ASCT, whereas a subsequent analysis of Total Therapy 2 at a median follow up of 72 months demonstrated a significant improvement in overall survival in patients with detectable metaphase cytogenetic abnormalities who received thalidomide.(19) Therefore, it is likely that specific subgroups of MM patients may benefit from thalidomide maintenance, but this potential benefit must be weighed against adverse effects of the drug, most notably cumulative peripheral neuropathy that is dependent on the dose and duration of administration.

Data concerning the use of bortezomib and lenalidomide as maintenance following ASCT are limited given their more recent approval. The incorporation of bortezomib into both induction and post-ASCT therapy has been the subject of two large randomized phase III studies in Europe.(20, 21) Accrual has been completed on each of these trials and further follow up will better define the utility of bortezomib in the maintenance setting. Post-transplant lenalidomide is under study in several randomized studies, and preliminary findings have recently been reported. In the IFM 2005–02 trial 614 patients were administered lenalidomide as consolidation following ASCT then randomized to receive maintenance lenalidomide or placebo.(22) Progression–free survival was significantly greater in patients in the lenalidomide maintenance arm. However, 5-year survival rates for each group was identical at 83%. In the CALGB 100104 trial, 460 patients were randomized to receive maintenance lenalidomide or placebo following ASCT.(23) Patients receiving lenalidomide experienced significantly prolonged time to progression, although the impact on overall survival is less clear given the low number of deaths. Lenalidomide was well tolerated in both trials, but an increased incidence in the rate of secondary malignancies has been observed in each. Therefore, the role of lenalidomide maintenance following ASCT remains unclear, and the balance between beneficial effects on MM relapse and potential toxicity must be further evaluated.

CANCER STEM CELLS IN MULTIPLE MYELOMA

Disease relapse remains the major factor limiting overall survival, therefore, a better understanding of the processes responsible for tumor recurrence and regrowth may ultimately lead to improvements in long-term outcomes. Emerging data in a wide range of human malignancies, including MM, have demonstrated that cells within an individual tumor may be phenotypically and functionally heterogeneous despite their clonal origins.(24) The majority of cells appear to lack sufficient long-term replicative potential required for disease propagation. Instead, tumor growth arises from relatively rare populations of phenotypically distinct cancer stem cells (CSCs). In both hematologic malignancies and solid tumors, CSCs have been prospectively identified based on their ability to give rise to differentiated progeny that recapitulates the original tumor in the ectopic setting. Several studies have also found that CSCs are relatively resistant to standard anti-cancer agents compared to bulk tumor cells. Therefore, CSCs may play a central role in disease relapse since they have the ability to persist during treatment and give rise to new tumor cells.

Several approaches have been undertaken to identify cancer stem cells.(25) Determination of specific cell surface antigens expressed by tumorigenic cells has been most commonly employed, and candidate markers have consisted of markers characteristic of normal stem cells or progenitors within the tissue of origin (e.g., CD34 in myeloid leukemias), expressed by multiple normal stem cells (e.g., CD133 in brain tumors), or associated with poor prognosis (CD44 in solid tumors). In addition, flow cytometric assays indicative of drug resistance, such as the Aldefluor assay that measures aldehyde dehydrogenase (ALDH) activity or detection of side population cells with increased drug efflux capabilities have also been used to isolate tumor cells with relatively enhanced tumorigenic potential. In MM, each of these strategies has successfully identified candidate CSCs.

Several studies have examined MM CSCs within the context of normal plasma cell differentiation. In a simplistic schema of normal B cell development, naïve B cells harboring immunoglobulin V(D)J gene rearrangements undergo somatic hypermutation in germinal center reactions to produce unique antibody idiotypes and improve antigen binding specificity and affinity. Germinal center B cells subsequently give rise to post-germinal center memory B cells that maintain long-term humoral immunity and differentiate into antibody secreting plasma cells upon antigen re-exposure. The examination of immunoglobulin gene sequences in MM plasma cells have demonstrated that they are somatically hypermutated and remain constant throughout the entire clinical course suggesting that the disease arises from post-germinal center memory B cells or plasmablasts.(26) The unique immunoglobulin gene rearrangement also provides a highly tumor-specific means of establishing clonal relationships among different tumor cell populations, and several studies over the past two decades have detected phenotypic B cells sharing the identical immunoglobulin gene rearrangements as MM plasma cells within the bone marrow and circulation of MM patients.(27–30) The clinical relevance of these clonotypic B cells has been unclear, but functional studies have demonstrated that these cell are tumorigenic and can give rise to MM plasma cells both in in vitro colony forming assays as well as in vivo in immunodeficient non-obese diabetic severe combined deficiency (NOD/Scid) mice.(31–33) Moreover, clonotypic B cells isolated from the circulation of MM patients are capable disease propagation during serial transplantation studies demonstrating that they are capable of self-renewal.(33) Therefore, MM may be hierarchically organized and recapitulates normal plasma cell development. Studies have also demonstrated that clonotypic B cells are relatively resistant compared to plasma cells to several clinically utilized anti-myeloma agents including traditional cytotoxic chemotherapies and novel agents including bortezomib and lenalidomide in vitro and in vivo.(33, 34)

In addition to cell surface antigen expression, functional flow cytometric assays have been used to identify clonogenic MM cells. The expression of ALDH is characteristic of normal stem cells in a variety of adult tissues and can identify CSCs in both solid tumors and hematologic malignancies.(25) MM CSCs have been found to express relatively high levels of ALDH with tumorigenic potential both in vitro and in vivo.(33) The side population assay takes advantage of differential emission spectra following exposure to the DNA binding dye Hoechst 33342 and was initially developed to identify normal hematopoietic stem cells.(35) This assay has been used to study MM, and clonogenic tumor cells have been found to express the side population phenotype.(33, 36) Further analysis of MM side population cells has also demonstrated increased expression of membrane bound drug transporters (ABCG2 and MDR/Pgp) suggestive of relative drug resistance.(36)

Despite these findings, the existence and precise phenotype of MM CSCs remains unclear. Studies examining the B cell nature of tumorigenic MM cells have demonstrated that these cells lack expression of the characteristic cell surface antigen CD138 expressed by both normal and MM plasma cells.(33, 37) On the other hand, studies examining MM side population cells have demonstrated that these cells are both CD138^neg and CD138⁺.(33, 36) Moreover, the engraftment and growth of primary MM specimens directly injected into the ectopic human or rabbit bone fragments implanted into severe combined immunodeficiency mice (SCID-Hu, SCID-Rab) is restricted to CD138⁺ MM plasma cells.(38, 39) The reasons for these discrepancies are unclear, but they likely represent the intrinsic differences between the animal models or patient specimens studied, and similar variations in CSC phenotypes have been observed in several other human malignancies. An important consideration, consistent with data emerging now from several lines of independent research in different malignant and normal models is that some surface markers conventionally associated with the “stem cell” compartment of the tumor cell population may actually exhibit significant plasticity. Indeed, there is emerging evidence of substantial bi-directional inter-conversions between non-stem and stem-like compartments in a population of malignant or normal cells.(36, 40–42) These results provide another possible explanation for some variations in the observations of different groups studying the phenotypic features of putative cancer stem cells. Importantly, the studies on bi-directional inter-conversions also indicate the critical significance of re-focusing the investigation on putative cancer stem cell compartments towards functional assays, as this may provide a better approach to define the biological and clinical implications of cancer cell populations with stem cell features.

TARGETING MYELOMA STEM CELLS

Given the unique functional attributes of CSCs, efforts to identify potential targeting strategies are currently underway. The impact of the tumor microenvironment on the proliferation and survival of MM plasma cells has been well documented, and these interactions have provided the scientific rationale for specific therapeutic strategies. The influence of the microenvironment on MM CSCs has recently been reported, and both the proliferation and relative proportion of side population MM cells can be enhanced by co-incubation with bone marrow stromal cells.(36) Moreover, lenalidomide, but not thalidomide, has been found to significantly inhibit the number and colony forming potential of MM side population cells associated with alterations in the phosphorylation of several important signaling molecules including Akt, GSK-3α/β, MEK1, c-JUN, p53, and p70^S6K. These intriguing results suggest that recent clinical data demonstrating improved progression and event free survival in patients receiving post-ASCT lenalidomide is mediated by its impact on MM CSCs expressing the side population phenotype.

Self-renewal appears to be a unique property required for the maintenance of both normal and cancer stem cells. The factors regulating MM CSCs are not fully understood, but the functional similarities between CSCs and normal stem cells suggest that shared cellular pathways may be involved. During development of the normal embryo, several highly conserved signaling pathways are required for cell fate specification, such as Notch, Wnt, and Hedgehog, and these have been increasingly implicated in the pathogenesis of a wide variety of human cancers. Hedgehog signaling has been studied in MM and may have multiple effects that depend on the precise cell type examined. In plasma cells, aberrant Hedgehog signaling primarily promotes cell survival, whereas it regulates cell fate decisions of MM CSCs.(43, 44) In CD138^neg cells, pathway activation by Hedgehog ligand induces self-renewal and cell expansion, whereas the inhibition by antagonists of SMOOTHENED, a positive regulator of Hedgehog signaling, induces plasma cell differentiation and the loss of clonogenic potential. Telomerase may represent another potential means of targeting MM CSC self-renewal a recent study has reported that the inhibition of telomerase activity within CD138^neg cells using specific inhibitors leads to decreased tumorigenic potential both in vitro and in vivo.(45) Therefore, shared self-renewal pathways may serve as novel therapeutic targets.

The potential of immune based therapies to target MM CSCs is evident through the allogeneic graft versus tumor effect that may result in long-term remissions, and the tumor-specific nature of the M protein suggests that it may serve as an ideal antigenic target. A previous study found that the number of circulating clonotypic B cells was significantly diminished following vaccination of MM patients with purified M protein.(46) Interestingly, M protein levels were not significantly altered in these patients, suggesting that this strategy may preferentially target malignant precursors rather than differentiated plasma cells. Aberrant expression of the embryonic stem cell associated transcription factor SOX2 was initially identified by screening reactive antibodies present in patients with monoclonal gammopathy of unknown significance (MGUS) and may represent another potential immune target preferentially expressed by MM CSCs.(47) In MGUS, the expression of SOX2 was restricted to CD138^neg cells, and anti-SOX2 immunity limited the clonogenic growth of primary specimens in vitro. Tumor infiltrating lymphocytes may exhibit anti-tumor activity, and the use of T cells derived from the bone marrow of MM patients is emerging. These marrow-infiltrating lymphocytes (MILs) can be harvested from the bone marrow, efficiently activated and expanded ex vivo, and safely reinfused in patients with MM (Personal Observation, I.B.).(48) Moreover, functional analyses of MILs have demonstrated that they display greater specificity for tumor cells than normal peripheral blood lymphocytes and are capable of inhibiting both mature plasma cells and CD138^neg CSCs. Thus, immune-based approaches hold promise as clinically effective MM CSC targeting strategies.

CONCLUSIONS

Evidence for both functional and phenotypic cellular heterogeneity in MM is emerging. However, the precise phenotype of tumorigenic MM cells has not been definitively established and controversy remains. Whatever their phenotype, it is likely that definitive proof of the existence of MM CSCs will be provided by actual improvements in long-term outcomes using clinical strategies based on CSC biology rather than further studies of various model systems. Increasing knowledge of the basic biology of MM CSCs should produce candidate approaches that can serve as the basis for novel therapies. The implementation of these strategies into clinical scenarios that allow MM CSC functions to be assessed, such as maintenance following tumor debulking by ASCT, may ultimately provide evidence for their clinical relevance.