Hironobu Shibata

Osteoclasts enhance myeloma cell growth and survival via cell-cell contact: a vicious cycle between bone destruction and myeloma expansion

Multiple myeloma (MM) expands in the bone marrow and causes devastating bone destruction by enhancing osteoclastic bone resorption in its vicinity, suggesting a close interaction between MM cells and osteoclasts (OCs). Here, we show that peripheral blood mononuclear cell-derived OCs enhanced growth and survival of primary MM cells as well as MM cell lines more potently than stromal cells, and that OCs protected MM cells from apoptosis induced by serum depletion or doxorubicin. OCs produced osteopontin (OPN) and interleukin 6 (IL-6), and adhesion of MM cells to OCs increased IL-6 production from OCs. In addition, IL-6 and OPN in combination enhanced MM cell growth and survival. However, the effects of OCs on MM cell growth and survival were only partially suppressed by a simultaneous addition of anti-IL-6 and anti-OPN antibodies and were completely abrogated by inhibition of cellular contact between MM cells and OCs. These results demonstrate that OCs enhance MM cell growth and survival through a cell-cell contact-mediated mechanism that is partially dependent on IL-6 and OPN. It is suggested that interactions of MM cells with OCs augment MM growth and survival and, thereby, form a vicious cycle, leading to extensive bone destruction and MM cell expansion.

Ability of myeloma cells to secrete macrophage inflammatory protein (MIP)-1α and MIP-1β correlates with lytic bone lesions in patients with multiple myeloma

Macrophage inflammatory protein (MIP)‐1α and MIP‐1β have been identified as candidates for multiple myeloma (MM)‐derived bone‐resorbing factors. To validate the clinical relevance of these observations, we investigated correlations between the ability of MM cells to secrete these chemokines and the extent of MM bone lesions as well as levels of biochemical bone markers in patients with MM. Patients with multiple bone lesions exhibited higher MIP‐1α and MIP‐1β secretion from MM cells along with elevated urinary deoxypyridinoline (Dpd), without significant elevation of serum bone‐specific alkaline phosphatase (BALP) or osteocalcin compared with those with minimal bone lesions. MIP‐1α and MIP‐1β levels correlated positively with urinary Dpd and serum BALP but not with serum osteocalcin. These results provide further evidence for a causal role of MIP‐1α and MIP‐1β in the development of lytic bone lesions, and suggest that MM cells suppress osteoblastic bone formation to cause an imbalance of bone turnover and development of destructive bone lesions.

A Single-Chain Fv Diabody against Human Leukocyte Antigen-A Molecules Specifically Induces Myeloma Cell Death in the Bone Marrow Environment

Cross-linked human leukocyte antigen (HLA) class I molecules have been shown to mediate cell death in neoplastic lymphoid cells. However, clinical application of an anti-HLA class I antibody is limited by possible side effects due to widespread expression of HLA class I molecules in normal tissues. To reduce the unwanted Fc-mediated functions of the therapeutic antibody, we have developed a recombinant single-chain Fv diabody (2D7-DB) specific to the alpha2 domain of HLA-A. Here, we show that 2D7-DB specifically induces multiple myeloma cell death in the bone marrow environment. Both multiple myeloma cell lines and primary multiple myeloma cells expressed HLA-A at higher levels than normal myeloid cells, lymphocytes, or hematopoietic stem cells. 2D7-DB rapidly induced Rho activation and robust actin aggregation that led to caspase-independent death in multiple myeloma cells. This cell death was completely blocked by Rho GTPase inhibitors, suggesting that Rho-induced actin aggregation is crucial for mediating multiple myeloma cell death. Conversely, 2D7-DB neither triggered Rho-mediated actin aggregation nor induced cell death in normal bone marrow cells despite the expression of HLA-A. Treatment with IFNs, melphalan, or bortezomib enhanced multiple myeloma cell death induced by 2D7-DB. Furthermore, administration of 2D7-DB resulted in significant tumor regression in a xenograft model of human multiple myeloma. These results indicate that 2D7-DB acts on multiple myeloma cells differently from other bone marrow cells and thus provide the basis for a novel HLA class I-targeting therapy against multiple myeloma.

Malignant B-Lymphoid Cells with Bone Lesions Express Receptor Activator of Nuclear Factor-κB Ligand and Vascular Endothelial Growth Factor to Enhance Osteoclastogenesis

Purpose: Receptor activator of nuclear factor-κB ligand (RANKL) is a key mediator of osteoclastogenesis. Because certain types of tumor cells aberrantly express RANKL, and because bone destruction also develops in B-cell lymphomas of bone origin, we investigated RANKL expression and the mechanisms of osteoclastogenesis in B-lymphoid neoplasms. Experimental Design and Results: Immunohistochemistry of bone specimens resected from patients with primary B-cell lymphoma of bone with bone destruction revealed that lymphoma cells express RANKL as well as vascular endothelial cell growth factor (VEGF). The tumor cells isolated from the bone specimens enhanced osteoclastogenesis in vitro. In contrast, B-cell lymphoma infiltrating to the bone marrow without bone destruction did not express RANKL. Both RANKL and VEGF were expressed by a portion of B-lymphoid cell lines, including Daudi and IM-9. These RANKL-expressing tumor cells enhanced osteoclastogenesis from RAW264.7 cells and human monocyte-derived preosteoclasts in the absence of stromal cells/osteoblasts in a RANKL-dependent manner. Furthermore, conditioned media from Daudi cells enhanced transmigration of preosteoclasts that was inhibited by anti-VEGF antibody, suggesting that tumor cell–derived VEGF mediates recruitment of osteoclast precursors. Moreover, cocultures of B-lymphoid cell lines with osteoclasts enhanced the growth of B-lymphoid cells. Conclusions: Some malignant B cells aberrantly express functional RANKL as well as VEGF to enhance osteoclastogenesis. The coexpression of RANKL and VEGF may also contribute to the close cellular interactions with osteoclastic cells, thereby forming a vicious cycle between osteoclastic bone destruction and tumor expansion in bone.

Construction of a Fourier-transform phase-modulation fluorometer

We have constructed a Fourier-transform phase-modulation fluorometer (FT-PMF) by which a fluorescence decay waveform can be obtained. In the FT-PMF, the modulation frequency of the excitation light source is swept continuously from a direct current (dc) to a high frequency fmax with a time duration T. The resultant fluorescence signal waveform is Fourier transformed to obtain its amplitude and phase spectra. The ratio of the amplitude spectrum and the difference of the phase spectrum over those of the reference spectra from an excitation waveform are calculated, respectively, and the pair of both spectral data is inverse-Fourier-transformed again to obtain the fluorescence decay waveform. The light source used was an ultraviolet light-emitting diode (UV LED) whose operating condition was fmax = 50–120 MHz and T = 10 µs. To demonstrate the performance of the FT-PMF, we carried out (1) the measurement of a fluorescent decay waveform of YAG materials enclosed in a white LED and (2) determinations of fluorescence lifetimes of 10 ppm quinine sulfate in 0.1 N H2SO4 and 10 ppm rhodamine 6G in ethanol.

Factor VIII gene analysis in Japanese CRM-positive and CRM-reduced haemophilia A patients by single-strand conformation polymorphism

Haemophilia A is the most common X‐linked blood coagulation disorder; it is caused by deficiency of factor VIII activity (FVIII:C). Half of the affected patients do not have detectable levels of FVIII protein in their plasma, whereas about 5% have normal levels of the FVIII antigen (FVIII:Ag) (> 50 u/dl), and are called cross‐reacting material (CRM) positive (CRM+ or A+). About 45% of patients have reduced levels of the FVIII:Ag (1–50 u/dl), classified as CRM reduced (CRMR or AR). We screened the FVIII gene of 13 Japanese patients (five CRM+ and eight CRMR) by single‐strand conformation polymorphism, and identified 11 different mutations in 13 patients by analysing all 26 exons (Trp255Cys, Tyr473Cys, Gly479Arg, Arg531His, Thr667Arg, Arg1689Cys, Arg1941Gln, Arg2150His, Arg2159Cys, Thr2245Ala and Gly2285Val). Seven mutations were identified in the A domains (four in the A2 domain). All the mutations are point mutations resulting in missense codons. Four mutations (Trp255Cys, Thr667Arg, Thr2245Ala and Gly2285Val) have not been described previously.

Fas ligand-induced apoptosis of hepatocytes in natural killer cell leukaemia.

Neoplastic natural killer (NK) cells overexpress Fas ligand (FasL), which may cause damage of Fas‐bearing tissues. We report a patient with NK cell leukaemia who developed liver injury after pharyngitis. The NK leukaemic cells expressed functional FasL. In addition to soluble FasL, serum levels of interleukin‐6 and interferon‐γ were increased dramatically when liver injury was aggravated. Moreover, hepatocytes expressed Fas and apoptotic hepatocytes were detected in the portal areas. These findings are consistent with the notion that inflammatory cytokines enhance the sensitivity to FasL and trigger apoptosis of hepatocytes in NK cell malignancies.

A deconvolution procedure for determination of a fluorescence decay waveform applicable to a band-limited measurement system that has a time delay

We propose a deconvolution procedure used for deriving a true fluorescence decay waveform, which is applicable to a band-limited measurement system that has a time delay. In general, the fluorescence decay response has a time delay over the excitation light pulse mainly due to the wavelength dependence of the time delay of the response of a photomultiplier tube (PMT). Furthermore, the frequency band is limited instrumentally or often should be cut off purposely at a relatively low frequency to eliminate high frequency noise. In order to extrapolate such a band-limited response function and to perform a deconvolution procedure in a frequency domain, finally, we introduce a delay time parameter in the conventional Gerchberg–Papoulis (GP) algorithm. We also demonstrate that the Hilbert-transform (HT)-based extrapolation method that we have proposed quite recently is equivalent to the GP algorithm. Simulation and experimental results show that the new method is effective.

Bleeding tendency in chronic neutrophilic leukemia.

To define the clinical and pathologic features and natural history of chronic neutrophilic leukemia (CNL), as defined by the World Health Organization (WHO), Elliott critically reviewed the literature and identified a total of 34 patients meeting strictly defined WHO criteria for CNL [1]. Of the 34 patients, 9 died of intracranial hemorrhage and 3 demonstrated easy bruising. Elliot indicated that a disproportionate number of reports of intracranial hemorrhage must be considered in the context of an association with chemotherapy-induced thrombocytopenia during treatment of progressive disease. Furthermore, he described that excess hemorrhagic events were not a significant finding in chronic stable-phase disease in the absence of chemotherapy-induced thrombocytopenia. Here, we describe platelet dysfunction in a patient with CNL complicated by bleeding tendency. At the time this patient showed bleeding symptoms, she had not received chemotherapy and her platelet count was normal. On 20 August 2003, an 83-year-old Japanese woman was referred to our department for examination of bleeding tendency. She had a 3-year history of recurrent purpura and occasional nasal bleeding. Neither the patient nor any members of her family had a history of bleeding tendencies. On physical examination, purpura was detected on the face, back and extremities. A CT scan of the abdomen demonstrated splenomegaly. Peripheral blood showed: hemoglobin, 9.7 g/dl; platelets, 284 9 10/l; leukocytes, 29.0 9 10/l with 0.2% myelocytes, 0.6% metamyelocytes, 2.0% stabs, 90.8% segments, 1.0% eosinophils, 2.2% basophils, 1.2% monocytes and 2.0% lymphocytes. Neutrophil alkaline phosphatase (NAP) score was 352 (normal, 190–370). The serum concentration of granulocyte colonystimulating factor (G-CSF) was under 10.0 pg/ml. Immunoelectrophoresis did not show any M protein in serum or urine. The bone marrow was grossly hypercellular with marked myeloid hyperplasia and 1.0% blasts. There were no dysplastic features, minimal fibrosis, normal megakaryocytes and erythropoiesis. Chromosomal analysis demonstrated a normal karyotype. Reverse transcriptasepolymerase chain reaction (RT-PCR) studies failed to demonstrate the presence of BCR-ABL fusion. The patient was diagnosed as having CNL. Routine hemostatic tests demonstrated: bleeding time, more than 8 min (normal, 1–5 min); prothrombin time, 14.9 s (normal, 10.0–13.5 s); activated partial thromboplastin time, 39.8 s (normal, 25.0–40.0 s); plasma fibrinogen level, 295 mg/dl; plasma FDP concentration, 4.1 lg/ml (normal, below 10 lg/ml). Factor VIII activity was 180% (normal, 62–145%), von Willebrand factor antigen was 136% (normal, 50–150%), and ristocetin cofactor activity was 133% (normal, 50–150%). Activities of factor XIII and a2-plasmin inhibitor were 76% (normal, over 70%) and 86% (normal, 85–118%), respectively. Platelet aggregation induced by ADP, collagen, epinephrine or ristocetin was decreased when compared to that in the control. In contrast, arachidonic acid induced platelet aggregation was normal (Fig. 1). Similar patterns of abnormal platelet aggregation were repeatedly demonstrated on three different occasions. Flow cytometric analyses of her platelets for CD41 and CD42b expression showed normal patterns. These findings suggested that this patient had storage pool disease. In this disease, platelet aggregation induced by ADP, collagen, or epinephrine is impaired, but that induced by arachidonic T. Shigekiyo (&) J. Miyagi M. Chohraku K. Kawauchi E. Sekimoto A. Shirakami H. Shibata Department of Internal Medicine, Tokushima Prefectural Central Hospital, 1-10-3 Kuramoto-cho, Tokushima 770-0042, Japan e-mail: shigekiyo@tph.gr.jp

Extrapolation of band-limited frequency data using an iterative Hilbert-transform method and its application for Fourier-transform phase-modulation fluorometry

We propose a new band-extrapolation method using an iterative Hilbert-transform procedure, which is applicable for band-limited complex-frequency data obtained from a Fourier-transform phase-modulation fluorometer (FT-PMF). Without the band limitation, the real and the imaginary parts of the complex-frequency data obtained from the FT-PMF form a Hilbert-transform pair because of causality. However, the Hilbert-transform relation cannot actually hold because of the inevitable band limitation in the instrument. In order to alleviate this problem, we propose the band-extrapolation method. First, the real part of the complex-frequency data is Hilbert-transformed and the partial frequency data inside the band are replaced by the imaginary ones actually measured while those outside are kept as they are. Next, the corrected imaginary-frequency data are inverse-Hilbert-transformed again and the real-frequency data inside the band are replaced by those measured while those outside are kept. Such a procedure is iterated until changes in shapes of respective frequency waveforms are sufficiently small. Finally, we can obtain a desired time-domain waveform by the inverse Fourier transform of the pair of the extrapolated complex-frequency data. Numerical simulations showed that the proposed method was valid. A good agreement between two fluorescence decay waveforms of fluo-5N in calcium ion buffer, one obtained from the proposed method and the other obtained from the conventional time-correlated single-photon counting method, showed that the proposed method was useful.