Bone marrow stromal-derived soluble factors and direct cell contact contribute to de novo drug resistance of myeloma cells by distinct mechanisms.

The tumor microenvironment plays a critical role in determining the fate of tumor cells. We have previously reported that adhesion of human myeloma and leukemia cell lines to the extracellular matrix protein, fibronectin, confers a multidrug-resistant phenotype. Mechanisms associated with this cell adhesion-mediated drug resistance are drug-type specific. In the present study, we examined the influence of bone marrow stromal cells (BMSCs) on myeloma cell response to the topoisomerase II inhibitor, mitoxantrone. Apoptosis was inhibited by more than 50% when cells were adhered to BMSCs as compared to myeloma cells maintained in suspension. To investigate the mechanisms contributing to the resistance of myeloma cells in contact with BMSCs, we examined the protective effects of BMSCs under four separate conditions: (1) direct cell contact; (2) BMSCs conditioned medium; (3) medium conditioned by coculturing myeloma cells in direct contact with BMSCs; and (4) medium conditioned by coculturing myeloma cells and BMSCs without direct physical contact. Conditioned medium from BMSCs alone was not sufficient to protect myeloma cells from drug-induced apoptosis; however, soluble factors produced during the myeloma-BMSCs interaction decreased the sensitivity of myeloma cells to mitoxantrone, suggesting a dynamic interaction between myeloma cells and BMSCs. We also found that myeloma cells in direct contact with BMSCs underwent growth arrest, whereas soluble factors produced by myeloma cells-BMSCs coincubation stimulated the proliferation of myeloma cells. These data show that both cell-cell adhesion of BMSCs with myeloma cells and soluble factors induced by this cell-cell interaction are involved in the protection of myeloma cells from mitoxantrone-induced apoptosis; however, the mechanisms contributing to the drug resistance are different.

The farnesyl transferase inhibitor, FTI-277, inhibits growth and induces apoptosis in drug-resistant myeloma tumor cells.

Mutations of the ras gene are among the most commonly identified transforming events in human cancers, including multiple myeloma. Farnesyltransferase inhibitors (FTI) were developed to prevent Ras processing and induce cancer cell death. Several FTIs are in phase II and one is in phase III clinical trials. Preclinically, most of the focus has been on solid tumors, and the effects of FTIs in multiple myeloma have not been investigated. In this study we examined the cytotoxic activity and inhibition of Ras processing in three myeloma cell lines with differing Ras mutation status. H929 cells with activated N-Ras were more sensitive to FTI-277 treatment than 8226 and U266 cells with activated K-Ras or wild-type Ras, respectively. A combination of FTI-277 and a geranylgeranyltransferase I inhibitor (GGTI)-2166 inhibited K-Ras processing and enhanced cell death in 8226 cells. U266 cells and Bcl-x(L) transfectants were equally sensitive to FTI-277 treatment. Similarly, 8226 cells selected for resistance to various chemotherapeutic agents, which resulted in either P-glycoprotein overexpression, altered topoisomerase II activity, or elevated glutathione levels, were equally sensitive to FTI-277. These preclinical studies suggest that prenylation inhibitors may represent new therapeutic agents for the treatment of refractory or drug-resistant multiple myeloma.

Inhibition of JAK kinase activity enhances Fas-mediated apoptosis but reduces cytotoxic activity of topoisomerase II inhibitors in U266 myeloma cells.

Our previous work demonstrated that the Janus kinase (JAK)-Stat3 pathway regulates expression of Bcl-x(L) in the U266 human multiple myeloma cell line and prevents Fas-mediated apoptosis. Inhibition of this pathway by the JAK selective kinase inhibitor AG490 or dominant-negative Stat3 protein results in down-regulation of Bcl-x(L) expression and enhanced sensitivity to Fas-mediated apoptosis. Because Bcl-x(L) has also been implicated in resistance to chemotherapeutic drugs, we investigated whether inhibition of the JAK-Stat3 pathway and subsequent reduction in Bcl-x(L) expression would also enhance cytotoxic drug activity. Contrary to this prediction, pretreatment of U266 myeloma cells with AG490, followed by exposure to topoisomerase II- inhibiting agents, antagonized drug-induced apoptosis. This effect correlated with reduced cyclin D1 expression and cell cycle arrest. The cell cycle arrest following AG490 pretreatment further correlated with reduced mitoxantrone-induced DNA double-strand breaks and reduced cell death, findings consistent with the critical requirement of DNA damage for drug cytotoxicity. These studies demonstrate that inhibition of the JAK-Stat3 pathway can result in paradoxical effects relative to cytotoxic drug response. These paradoxical responses may be explained by the findings that JAK-Stat3 signaling regulates the expression of multiple genes involved in controlling cell proliferation and apoptosis. Thus, understanding the cellular context of inhibiting signal transduction pathways is essential for the design of novel combination therapies for cancer.

CD95 antigen mutations in hematopoietic malignancies.

The CD95 receptor, also known as Fas/Apo-1, is a member of the Tumor Necrosis Factor receptor (TNF-R) family of death receptors. Apoptosis mediated by CD95 plays a central role in maintaining homeostasis of the immune system. Dysregulation of the CD95 apoptotic pathway has been proposed as a mechanism of oncogenesis by providing a survival advantage to potentially malignant cells. This extended lifespan could allow the accumulation of further mutations leading to malignant transformation. Several mechanisms of resistance to CD95 mediated apoptosis have been identified, including reduced surface expression of the receptor, overexpression of anti-apoptotic molecules, and loss of function mutations. This review will focus on the potential role of the CD95-CD95 ligand system in the pathogenesis of hematological malignancies, with particular emphasis on recent work from our laboratory examining the expression of CD95 in B cell lymphomas. We demonstrate that CD95 mutations occur at low frequency in NHL tumors, however, surface expression of the CD95 protein varies with the subtype of lymphoma. Loss of surface CD95 is more likely to occur in lymphomas of aggressive histology, and is unrelated to the detection of CD95 mutations.

The tumor microenvironment as a determinant of cancer cell survival: a possible mechanism for de novo drug resistance.

The influence of the microenvironment in the pathogenesis and progression of human cancer has traditionally been considered in the context of solid tumors. More recently, evidence has been accumulating to support the role of the bone marrow microenvironment in hematologic malignancies as well, particularly in multiple myeloma. This review focuses on myeloma as a model to demonstrate that the bone marrow microenvironment provides a sanctuary against programmed cell death and promotes tumor cell survival and progression. Additionally, the protective effects of the bone marrow milieu may confer a protection from cytotoxic drugs, allowing the emergence of drug-resistant tumors. These advances may assist in the design of novel therapeutic approaches to enhance the efficacy of standard chemotherapeutic drugs.

Clonal variability in CD95 expression is the major determinant in Fas-medicated, but not chemotherapy-medicated apoptosis in the RPMI 8226 multiple myeloma cell line.

CD95 (Fas/APO-1) is a member of the TNFR superfamily that induces apoptosis following cross-linking with its cognate ligand, CD95L (FasL/APO-1L) or agonist antibody. The human myeloma cell line, RPMI 8226, has limited sensitivity to CD95-mediated apoptosis, with a maximum of 65% of the population responding. To determine the source of the limited sensitivity to CD95-mediated apoptosis, we isolated multiple clones from the RPMI-8226 cell line by limiting dilution. Analysis of these clones demonstrated that sensitivity to CD95-mediated cell death directly correlated with CD95 expression. Clones with high levels of CD95 expression had greater than 90% cell death, whereas cells with low levels of expression had less than 10% cell death. In contrast, no correlative differences were identified for other members of the DISC complex, or for members of the anti-apoptotic Bcl-2 family. We further examined the sensitivity of the 8226 clones to various cytotoxic agents. Although modest clonal variability was demonstrated in response to the chemotherapeutic drugs, doxorubicin, etoposide (VP-16), and vincristine, there was no correlation between CD95 function and sensitivity to chemotherapeutic drugs. These results indicate that in this cell line, receptor expression is rate limiting in CD95-mediated apoptosis, whereas CD95 expression was not a determinant in drug-induced programmed cell death.

Myeloma cells selected for resistance to CD95-mediated apoptosis are not cross-resistant to cytotoxic drugs: evidence for independent mechanisms of caspase activation.

We have previously shown that selection for resistance to the anthracenes, doxorubicin or mitoxantrone, results in coselection for resistance to CD95-mediated apoptosis (Landowski et al: Blood 89:1854, 1997). In the present study, we were interested in determining if the converse is also true; that is, does selection for CD95 resistance coselect for resistance to chemotherapeutic drugs. To address this question, we used two isogenic models of CD95-resistant versus CD95-sensitive cell lines: 8226/S myeloma cells selected for resistance to CD95-mediated apoptosis; and K562 cells expressing ectopic CD95. Repeated exposure of the CD95-sensitive human myeloma cell line, 8226/S, to agonistic anti-CD95 antibody resulted in a cell line devoid of CD95 receptor surface expression and completely resistant to CD95-mediated apoptosis. Multiple clonal populations derived from the CD95-resistant cell line showed no difference in sensitivity to doxorubicin, mitoxantrone, Ara-C, or etoposide, demonstrating that cross-resistance between Fas-mediated apoptosis and drug-induced apoptosis occurs only when cytotoxic drugs are used as the selecting agent. Using the inverse approach, we transfected the CD95-negative cell line, K562, with a CD95 expression vector. Clones expressing variable levels of cell-surface CD95 were isolated by limiting dilution, and analyzed for sensitivity to CD95-mediated apoptosis and response to chemotherapeutic drugs. We show that CD95 surface expression confers sensitivity to CD95-mediated apoptosis; however, it does not alter response to chemotherapeutic drugs. Similarly, doxorubicin-induced activation of caspases 3 and 8 was identical in the CD95-sensitive and CD95-resistant cell lines in both isogenic cell systems. In addition, prior treatment with the CD95 receptor-blocking antibody, ZB4, inhibited CD95-activated apoptosis in 8226/S cells, but had no effect on doxorubicin cytotoxicity. These results show that CD95 and chemotherapeutic drugs use common apoptotic effectors, but the point of convergence in these two pathways is downstream of CD95 receptor/ligand interaction.

Constitutive activation of Stat3 signaling confers resistance to apoptosis in human U266 myeloma cells.

Interleukin 6 (IL-6) is the major survival factor for myeloma tumor cells and induces signaling through the STAT proteins. We report that one STAT family member, Stat3, is constitutively activated in bone marrow mononuclear cells from patients with multiple myeloma and in the IL-6-dependent human myeloma cell line U266. Moreover, U266 cells are inherently resistant to Fas-mediated apoptosis and express high levels of the antiapoptotic protein Bcl-xL. Blocking IL-6 receptor signaling from Janus kinases to the Stat3 protein inhibits Bcl-xL expression and induces apoptosis, demonstrating that Stat3 signaling is essential for the survival of myeloma tumor cells. These findings provide evidence that constitutively activated Stat3 signaling contributes to the pathogenesis of multiple myeloma by preventing apoptosis.

Dysregulation of CD95/CD95 ligand-apoptotic pathway in CD3(+) large granular lymphocyte leukemia.

CD95 (Fas)-induced apoptosis plays a critical role in the elimination of activated lymphocytes and induction of peripheral tolerance. Defects in CD95/CD95L (Fas-Ligand)-apoptotic pathway have been recognized in autoimmune lymphoproliferative diseases (ALPS) and lpr or gld mice and attributed to CD95 and CD95L gene mutations, respectively. Large granular lymphocyte (LGL) leukemia is a chronic disease characterized by a proliferation of antigen-activated cytotoxic T lymphocytes. Autoimmune features such as hypergammaglobulinemia, rheumatoid factor, and circulating immune complexes are common features in LGL leukemia and ALPS. Therefore, we hypothesize that expansion of leukemic LGL may be secondary to a defective CD95 apoptotic pathway. In this study, we investigated expression of CD95 and CD95L in 11 patients with CD3(+) LGL leukemia and explored the apoptotic response to agonistic CD95 monoclonal antibody (MoAb). We found that leukemic LGL from each patient expressed constitutively high levels of CD95/CD95L, similar to those seen in normal activated T cells. However, cells from 9 of these 11 patients were totally resistant to anti-CD95-induced apoptosis. Similarly, cells were resistant to anti-CD3-MoAb-triggered cell death. Lack of anti-CD95-induced apoptosis was not due to mutations in the CD95 antigen. Leukemic LGL were not intrinsically resistant to CD95-dependent death, because LGL from all but 1 patient underwent apoptosis after phytohemagglutinin/interleukin-2 activation. The patient whose leukemic LGL were intrinsically resistant to CD95 had an aggressive form of LGL leukemia that was resistant to combination chemotherapy. These findings that leukemic LGL are resistant to CD95-dependent apoptosis despite expressing high levels of CD95 are similar to observations made in CD95L transgenic mice. These data suggest that LGL leukemia may be a useful model of dysregulated apoptosis causing human malignancy and autoimmune disease.

Mutations in the Fas antigen in patients with multiple myeloma.

Programmed cell death, or apoptosis, is well documented as a physiological means of eliminating activated lymphocytes and maintaining immune homeostasis. Apoptosis has also been implicated in the targeting of tumor cells by cytotoxic T lymphocytes and natural killer cells. One of the two primary mechanisms used in cell-mediated cytotoxicity is the Fas/FasLigand system. Activated or transformed cells expressing the Fas antigen on their surface are susceptible to killing by immune effector cells that express the Fas ligand. Many neoplastic cells, including those derived from patients with multiple myeloma, express Fas antigen on their surface, but do not undergo apoptosis in response to antigen crosslinking. One possibility for the lack of Fas-mediated apoptosis includes mutations in the Fas antigen. Loss of function mutations in the Fas antigen have been associated with congenital autoimmune disease in humans, and have been defined as the genetic defect the in lpr mice. Mutations in the Fas antigen have not been previously described in cancer patients. In this study, we show that mutations occur in the Fas antigen which may cause loss of function and contribute to the pathogenesis of the neoplastic disease, multiple myeloma. Using reverse transcriptase-polymerase chain reaction (RT-PCR), single-stranded conformation polymorphism (SSCP) analysis, and DNA sequencing, we examined the cDNA structure of the Fas antigen in 54 bone marrow (BM) specimens obtained from myeloma patients. Six patient specimens (11%) did not express detectable levels of Fas antigen mRNA. Of the 48 BM specimens which did express Fas antigen, 5 (10%) displayed point mutations. All of the mutations identified were located in the cytoplasmic region of the Fas antigen known to be involved in transduction of an apoptotic signal. Two separate individuals demonstrated an identical mutation at a site previously shown to be mutated in the congenital autoimmune syndrome, ALPS. One patient exhibited a point mutation at a site only two amino acids removed from the documented lesion of the lprcg mouse. Although the functional status of these point mutations remains to be determined, we propose that Fas antigen mutations may contribute to the pathogenesis and progression of myeloma in some patients.

Selection for drug resistance results in resistance to Fas-mediated apoptosis.

Recent evidence has supported the hypothesis that chemotherapeutic drugs and radiation induce an apoptotic pathway that requires the active participation of the cell. One pathway of apoptosis in malignant lymphoid cells is mediated by the Fas antigen. We studied the human myeloma (8226) and T-cell leukemia (CEM) cell lines selected for resistance to the anthracenes, doxorubicin or mitoxantrone, by continuous culture in the presence of either agent. We found that these drug-resistant cell lines were also resistant to Fas-mediated apoptosis in a dose-dependent manner. The degree of resistance to Fas-mediated apoptosis correlated directly with the level of resistance to chemotherapeutic drugs. These observations indicate that, as cancer cell lines develop mechanisms of drug resistance, they may also develop mechanisms of resistance to physiologic signals of apoptosis. Two mechanisms of resistance to Fas-mediated apoptosis were observed in these cell lines. One mechanism was associated with a dose-dependent reduction in the surface expression of Fas antigen. Analysis of RNA by reverse transcriptase-polymerase chain reaction assays showed that the reduction of Fas antigen expression occurred at the level of transcription. A second mechanism of drug resistance showed no decrease of Fas antigen expression; however, the apoptotic response was diminished. In this situation, removal of the chemotherapeutic agent resulted in a partial reversion to chemosensitivity and re-expression of the Fas antigen, but these cell lines did not regain the ability to undergo apoptosis in response to cross-linking by anti-Fas antibody. These findings support the hypothesis that apoptosis mediated by both chemotherapeutic agents and physiologic stimuli may share a common downstream effector. The demonstration that selection for drug resistance in hematopoietic cell lines results in a simultaneous resistance to Fas-mediated apoptosis may have clinical implications in the development of strategies for the treatment of resistant disease. Further analysis of the molecular mechanisms of Fas expression and function will facilitate the design of biological response modifying agents for the treatment of malignancy.

Studies of the structure of the metastasis-associated 67 kDa laminin binding protein: fatty acid acylation and evidence supporting dimerization of the 32 kDa gene product to form the mature protein.

The level of expression of the 67 kDa high-affinity laminin binding protein (LBP) correlates with the progression of many solid tumors. The cDNA clone for the 67 kDa LBP is sufficient to encode a polypeptide of only 32 kDa, and there is no readily identifiable mechanism for membrane association. We have overexpressed the transfected 67 kDa hamster LBP in quantities that have enabled us to analyze the membrane-bound form of the protein. Treatment of the purified LBP with methyl transesterification reagents, followed by GC-MS, identified the covalently bound fatty acids palmitate, stearate, and oleate. The fatty acid modification may provide a mechanism for membrane association. Molecular mass determination by MALDI-TOF MS demonstrated the true molecular mass of the protein to be 66.7 kDa, compatible with the SDS-PAGE observation of 67 kDa. Treatment of the LBP with neuraminidase, O-glycanase, or Endo-F glycosidase has no detectable effect on the apparent molecular mass of the protein, and the MALDI-TOF MS did not show evidence of mass heterogeneities typically observed with glycosylated proteins. Reduction with dithiothreitol or beta-mercaptoethanol had no effect on the apparent molecular mass on SDS-PAGE or on the relative quantities of molecular mass species on MALDI-TOF MS. The experimentally determined amino acid composition, however, was found to be consistent with the 67 kDa form being a homodimer of the 32 kDa precursor. Preliminary experiments also suggest that the high-affinity laminin binding characteristic of the protein may be modulated by an, as yet, unidentified membrane accessory molecule.

Control pathways of the 67 kDa laminin binding protein: surface expression and activity of a new ligand binding domain.

A number of papers have been published on the clinical correlation of the expression of the 67 kDa laminin binding protein (LBP) with the metastatic potential of solid tumors. Both mRNA and protein expression levels have been reported, but both the relationship between them and the molecular nature of the 67 kDa surface product remain unclear. We have utilized a homotypic overexpression system to investigate the cell surface presentation of the 67 kDa LBP and the contribution of this protein to the invasive phenotype of cultured cell lines. We report here that the cellular mRNA levels do not directly reflect the levels of the 67 kDa LBP observed on the cell surface in this overexpression system. Methotrexate amplification of transfected plasmids expressing the 67 kDa LBP leads to an initial elevation of both the LBP mRNA and surface protein levels. This is accompanied by an altered, more flattened, cell morphology. Later, apparent adaptation of the cells to methotrexate is accompanied by a down-regulation of the surface expression of the protein. mRNA levels, however, remain elevated. A nine amino acid sequence, CDPGYIGSR (peptide 11), within the beta chain of laminin 1 has been identified as a probable binding domain for the 67 kDa LBP. Previous studies have identified a region of the 67 kDa LBP which may be involved in laminin interaction, although not necessarily via the peptide 11 domain. We have identified a second site within the amino acid coding sequence of the 67 kDa LBP which also shows biological activity both in vitro and in vivo. A peptide with this sequence, LBP residues 205-229, binds laminin-1 in a peptide 11 inhibitable manner. The receptor-derived peptide modulates invasion of basement membrane matrix in vitro and inhibits experimental lung colony formation when injected along with B16BL6 mouse melanoma cells. However, pretreatment of the melanoma cells with the peptide enhances lung colony formation. Thus, the interaction of the 67 kDa LBP with basement membrane matrix appears to involve a complex series of events including multiple adhesive sites and tight regulation of cell surface expression.