The sesquiterpene lactone parthenolide induces selective apoptosis of B-chronic lymphocytic leukemia cells in vitro.

We have studied the in vitro actions of the sesquiterpene lactone parthenolide (PTL) on cells isolated from patients with chronic lymphocytic leukemia (CLL). Dye reduction viability assays showed that the median LD(50) for PTL was 6.2 muM (n=78). Fifteen of these isolates were relatively resistant to the conventional agent chlorambucil but retained sensitivity to PTL. Brief exposures to PTL (1-3 h) were sufficient to induce caspase activation and commitment to cell death. Chronic lymphocytic leukemia cells were more sensitive towards PTL than were normal T lymphocytes or CD34(+) haematopoietic progenitor cells. The mechanism of cell killing was via PTL-induced generation of reactive oxygen species, resulting in turn in a proapoptotic Bax conformational change, release of mitochondrial cytochrome c and caspase activation. Parthenolide also decreased nuclear levels of the antiapoptotic transcription factor nuclear factor-kappa B and diminished phosphorylation of its negative regulator IkappaB. Killing of CLL cells by PTL was apparently independent of p53 induction. This is the first report showing the relative selectivity of PTL towards CLL cells. The data here warrant further investigation of this class of natural product as potential therapeutic agents for CLL.

Role of double-stranded RNA-activated protein kinase in human hematological malignancies.

The double-stranded RNA (dsRNA)-activated protein kinase (PKR) is one of many genes induced by IFN. The PKR sequentially undergoes autophosphorylation and activation on binding to dsRNA. Previous studies have shown that PKR may be an important factor in the regulation of viral and cellular protein synthesis. Recent studies suggest that PKR may function as a tumor suppressor gene. The role of PKR in various human leukemic cells was therefore investigated. PKR mRNA levels by reverse transcription-PCR, protein expression by Western blot and FACScan analysis, and activity by phosphorylation status were studied. The expression of a known inhibitor of PKR, p58, was also investigated at mRNA and protein levels. A total of 24 samples from normal mononuclear cells (MNCs), 26 samples of acute lymphoblastic leukemia, 26 samples of acute myelogenous leukemia, 32 samples of chronic lymphocytic leukemia, and 5 samples of hairy cell leukemia was investigated. Mean mRNA levels were increased in acute lymphoblastic leukemia and acute myelogenous leukemia and decreased in chronic lymphocytic leukemia compared to normal MNCs. The mRNA levels in hairy cell leukemia were similar to those of normal MNCs. PKR protein was detectable in normal MNCs and leukemic cell extracts, and on FACScan analysis, more than 70% of cells stained positive for PKR. PKR activity was detectable in all samples investigated and was enhanced 4-23-fold in the presence of the synthetic dsRNA, poly(I) x poly(C). Protein expression of a known PKR inhibitor, p58, was barely detectable in normal MNCs and leukemic cells, with high expression in the HeLa cell line. These findings provide no evidence to support the hypothesis that PKR acts as a tumor suppressor in human leukemic cells.

CCK2 gastrin receptor as a potential target for therapy in leukaemia cell lines.

The gastrin CCK2 pathway has been implicated in the development of various cancers including leukaemia. An autocrine or intracrine pathway may exist in the leukaemia cell that is involved in stimulating proliferation. We tested four leukaemia cell lines, KU812, ML-1, MOLT-4 and U937 for the existence of the CCK2 receptor and gastrin precursor protein using immunoblotting. We also assessed the effect of CCK2 antagonist PD 135 and both gastrin 17 and glycine-extended gastrin on the proliferation of the cell lines. We found immunoreactive CCK2 and gastrin precursors present in all 4 cell lines. We also observed a stimulatory effect on proliferation by gastrin and glycine-extended gastrin on 2 and 3 of the cell lines respectively and an inhibitory effect of PD 135 on all 4 cell lines. These results demonstrate that the gastrin-gastrin receptor axis is a potential target for new therapeutic strategies.

Interferon-alpha inducible 2'-5' oligoadenylate synthetase transcripts in lymphoid and myeloid leukemias.

Alpha-interferon (IFN) is effective in the treatment of a proportion of patients with hairy cell leukemia (HCL). B-cell chronic lymphocytic leukemia (CLL), chronic myeloid leukemia (CML) and multiple myeloma (MM). One of the proteins induced by IFN is the enzyme 2'-5' oligoadenylate synthetase (2-5 AS). Peripheral blood or bone marrow samples treated with IFN in vitro, or from patients treated with IFN were studied for expression of the different 2-5 AS mRNA transcripts. A total of four normal individuals and 31 patients (nine HCL, five CLL, six MM, nine acute myeloid leukemia (AML) and two T-cell acute lymphoblastic leukemia (T-ALL) have been investigated. In normal peripheral blood lymphocytes, only the 1.8 kb transcript was induced with IFN in vitro. In HCL, CLL, and MM all four transcript sizes were induced by IFN in vivo and in vitro. The 1.6 and 1.8 kb forms were equally and predominantly expressed in HCL and B-CLL. On the other hand, the 1.8 kb transcript was predominantly expressed in MM and this increased expression was statistically significant. In acute leukemia, the majority of samples expressed all four transcripts equally but four of eleven samples expressed only the 1.8 kb transcript. These results suggest that the pattern of induction of specific 2-5 AS mRNA transcripts may be related to the underlying disease. Whether these different patterns of 2-5 AS induction have implications for response to IFN treatment, remains to be determined.

Expression and function of the FAS antigen in B chronic lymphocytic leukemia and hairy cell leukemia.

The expression and function of the FAS antigen was analyzed in 21 patients with B chronic lymphocytic leukemia (CLL) and four with hairy cell leukemia (HCL) using a specific IgM monoclonal antibody and FACS analysis. The FAS antigen was expressed in a minority (5-41%, mean 15.6%) of the CLL cells in 10 of 21 CLL patients and this expression was not modified during spontaneous or hydrocortisone-induced apoptosis of CLL cells. In contrast, culture with gamma-interferon (gamma-IFN) upregulated the expression of FAS in all CLL patients, with 65-100% (mean 84.8%) of the cells being positive after 2 days in vitro culture. Culture with alpha-IFN induced FAS expression in 15 of 19 CLL patients tested, with 15-74% (mean 34%) of the cells being FAS+ after 2 days culture. IL-4 and IL-10, lymphokines that inhibit and promote CLL apoptosis respectively, did not modify the expression of FAS. These results from FACS analysis were consistent with FAS mRNA analysis of fresh and cultured CLL cells, using a semi-quantitative reverse transcriptase (RT)-PCR technique. Although IL-4 and IFNs prevent apoptotic cell death of CLL cells in vitro, the present results show that IFNs induce the expression of the apoptosis-inducing protein FAS. However, FAS+ CLL cells were not killed in the presence of anti-FAS monoclonal antibody (while the FAS+ Jurkat and four lymphoblastoid cell lines were killed). This resistance is not due to a mutated FAS protein, since only wild-type FAS cDNA was demonstrated in the leukemic cells of three CLL patients. In four HCL patients 34-53% (mean 44.5%) of the leukemic cells were FAS+ and they were also resistant to the anti-FAS mediated cytotoxicity. The combination of high bcl-2 protein levels and resistance to anti-FAS mediated cytotoxicity may contribute to the extended in vivo survival of CLL and HCL cells.

Effects of interferon-alpha (IFN) on the expression of interleukin 1-beta (IL-1), interleukin 6 (IL-6), granulocyte-macrophage colony-stimulating factor (GM-CSF) and tumor necrosis factor-alpha (TNF) in acute myeloid leukemia (AML) blasts.

Interferon-alpha (IFN) induces the enzyme 2-5 oligoadenylate synthetase (2-5 AS) in cells from patients with hairy cell leukemia and B-cell chronic lymphocytic leukemia and this is associated with a breakdown of certain species of cytokine messenger (m)RNA via the activation of a latent ribonuclease. We have studied the expression of the cytokines interleukin 1-beta (IL-1), interleukin 6 (IL-6), granulocyte-macrophage colony-stimulating factor (GM-CSF), and tumour necrosis factor alpha (TNF) as well as of the ribonuclease activator 2-5 AS in the presence and absence of IFN in acute myeloid leukaemia (AML) blast cells from 26 patients. Before monocyte and T-cell depletion there was no expression of IL-1, IL-6 or GM-CSF, and only three of 13 patients studied expressed TNF mRNA. After cell depletion one or more cytokine was expressed in 31-62% of the 26 patients. Expression of one or more mRNA for IL-1, IL-6, GM-CSF and TNF after 18 h incubation was detected in 16 of 26 patients (63%) and this was particularly so in French-American-British (FAB) subtypes M4 and M5. Eight of nine patients with IL-6 mRNA expression and seven of 10 with IL-1 mRNA expression were in the FAB subtypes M4 and M5. Twenty-two of 26 patients showed induction of 2-5 AS mRNA in response to IFN in vitro. Exposure to IFN resulted in reduction of IL-1 mRNA in nine of 12 cases, of IL-6 mRNA in eight of nine, and GM-CSF mRNA in five of seven cases. TNF mRNA was unaffected by IFN despite 2-5 AS induction in 12 of 13 patients expressing this cytokine. In the presence of exogenous IFN, cells from six of seven patients studied showed inhibition of 3H-thymidine incorporation into DNA. DNA synthesis could also be abrogated in six of seven patients with anti-IL-1 monoclonal antibodies (MoAb) and in two of seven with anti-IL-6 MoAb. This inhibitory effect could be reversed in all patients when anti-IL-1 or anti-IL-6 was given in combination with their corresponding cytokine. These data suggest that IFN may exert a therapeutic effect in a proportion of AML patients by blocking IL-1 and IL-6 mediated growth, consequent on activation of the ribonuclease activator 2-5 AS.

Expression of the multidrug resistance-associated protein (MRP) in acute leukaemia.

A semi-quantitative reverse transcription polymerase chain reaction (RT-PCR) was used to investigate and compare transcription levels of the human multidrug resistance gene (MDR1) and the recently described multidrug resistance-associated protein (MRP) in 105 samples from patients with acute leukaemia at presentation and relapse. MRP gene expression was significantly greater in samples from patients with acute lymphoblastic leukaemia (ALL) compared with samples from normal peripheral mononuclear cells (PBMC) and patients with de novo acute myeloid leukaemia (AML). MRP gene expression was found to be higher in patients with relapsed de novo AML compared to those at presentation but prior therapy did not affect MRP gene expression in ALL. MDR1 gene expression was significantly lower in ALL patients compared to normal PBMC and AML samples. Samples from patients with secondary AML had higher levels of MDR1 expression than those of de novo AML. No changes of MDR1 expression were observed in AML or ALL at relapse. No correlation was observed between MDR1 and MRP gene expression in this group of patients. Our results suggest that MRP expression may be of prognostic importance in AML but the significance of the increased levels we have detected remain unclear.

First intron and M-bcr breakpoints are restricted to the lymphoid lineage in Philadelphia positive acute lymphoblastic leukemia.

Knowledge of the level of commitment of the target cell in hematological malignancies may have important therapeutic and prognostic implications. Cell lineage involvement was investigated in two cases presenting with acute lymphoblastic leukemia diagnosed on clinical and immunological findings and having the Philadelphia translocation t(9;22)(q34;q11). DNA from cells separated into mononuclear (lymphoid) and granulocytic fractions was hybridized with Philadelphia breakpoint-specific probes. This revealed that the breakpoint giving rise to the Philadelphia chromosome in case 1 was within the major breakpoint cluster region and in case 2 was in the first intron of the BCR gene. Rearrangement was found in the lymphoid but not the granulocyte fraction in each case. It is therefore concluded that the target cell for chromosomal change in these cases was a lymphoid committed progenitor cell, irrespective of breakpoint location.

Expression of the human major vault protein LRP in acute myeloid leukemia.

Overexpression of a 110-kD protein (lung resistance-related protein [LRP]) may predict a poor response to chemotherapy in patients with acute myeloid leukemia (AML) and ovarian carcinoma. The LRP gene has recently been mapped to chromosome 16, close to the multidrug resistance-associated protein (MRP) gene. Seventy-seven samples from 67 patients with AML were examined for expression of LRP, MRP, and multidrug resistance (MDR1) mRNA using a semiquantitative reverse transcription polymerase chain reaction (RT-PCR) assay. Results were compared with 29 normal samples (11 normal peripheral blood and 18 normal bone marrow). Thirty-three patients with untreated AML were evaluable for response to chemotherapy. Levels of LRP, but not of MRP or MDR1 mRNA, were significantly higher in eight patients who failed to achieve complete remission (CR) compared with 25 patients who achieved CR (p = 0.033). A positive correlation was demonstrated between LRP and MRP (R = 0.368, p = 0.001) and between MRP and MDR1 mRNA levels (R = 0.301, p = 0.01) in the 77 clinical samples analyzed. In AML samples, a significant difference in MDR1 mRNA levels was found between presentation (47 samples) and relapse (30 samples) (p = 0.031). No significant difference was seen in LRP mRNA levels between these two groups or in eight patients studied sequentially at both presentation and relapse. Thirteen samples (10 at presentation, 3 at relapse) were analyzed for LRP protein expression by flow cytometry. Eight (5 at presentation, 3 at relapse) displayed greater than 10% positive cells (range 15-86%). These data suggest that LRP gene overexpression may constitute a novel mechanism of multidrug resistance.

Absence of Kirsten-ras oncogene activation in B-cell chronic lymphocytic leukemia.

By using a combination oligonucleotide probe hybridization and restriction enzyme polymorphism analysis, a series of 48 cases of B-cell chronic lymphocytic leukemia were investigated for activating point mutations at codons 12, 13 and 61 of the K-ras proto-oncogene. A small series of acute leukemias (seven with acute lymphoblastic leukemia (ALL), 11 with acute myeloid leukemia (AML)) were examined in parallel. None of the cases of B-CLL contained detectable activating mutations of the K-ras gene at codon 12 (GGT-gly----GCT-ala) was detected at presentation. In both cases of acute leukemia, the mutation was restricted to one allele and could not be detected in remission samples. Those data suggest that activation of members of the ras oncogene family, typified by K-ras, may be less important in disease pathogenesis in leukemias such as B-CLL that arise from a more committed progenitor.

Therapeutic and prognostic implications of glucocorticoid receptors and terminal deoxynucleotidyl transferase in acute leukemia.

Glucocorticoid receptors (GR) have been reported to predict clinical responsiveness to glucocorticoid therapy and to possess prognostic significance in leukemia. A raised level in terminal deoxynucleotidyl transferase (TdT) also seems to suggest clinical responsiveness to prednisone and vincristine therapy. We have investigated these two biochemical markers in the leukemic blasts in 23 patients with acute myeloid leukemia (AML) and in 19 patients with acute lymphoblastic leukemia (ALL) and have compared the binding data with clinical glucocorticoid sensitivity, with response to chemotherapy, with survival and with TdT. In 25 of these patients we have also investigated the in vitro glucocorticoid sensitivity by measuring dexamethasone inhibition of radiolabelled uridine and thymidine incorporation into the leukemic blasts. In patients with AML, GR levels were mostly high but did not correlate with the in vitro sensitivity, response to chemotherapy or survival. On the other hand, in the 19 patients with ALL there seemed to be a correlation between GR and in vitro sensitivity and between GR and clinical responsiveness to glucocorticoid therapy. Furthermore, patients with low levels of either TdT or GR seemed to be associated with poor prognosis. TdT activity and GR, however, did not associate with one another. TdT is a useful marker for the identification of lymphoblasts but, by itself, is not connected with glucocorticoid sensitivity. Gr, when applied to ALL, could supplement TdT determination by predicting response to therapy.

Enzyme activities of leukemic cells and biochemical changes induced by deoxycoformycin in vitro--lack of correlation with clinical response.

Deoxycoformycin (DCF) is a specific inhibitor of adenosine deaminase (ADA) and has been shown to be active in lymphoid neoplasms. Cytotoxicity is thought to be mediated by the accumulation of deoxyadenosine (AdR) and deoxyadenosine triphosphate (dATP) which inhibits ribonucleotide reductase and DNA synthesis in rapidly proliferating cells. Others suggested mechanisms leading to cell death particularly in non-dividing cells include depletion of ATP and NAD pools, inhibition of S-adenosylhomocysteine (SAH) hydrolase and induction of DNA strand breaks. In patients with high leukemic counts who were subsequently treated with DCF, we have studied (a) the levels of ADA, ecto-5'-nucleotidase (5NT), deoxyadenosine kinase (AdR-kinase) and SAH-hydrolase in the leukemic cells; [b) the in-vitro effects of DCF on dATP, ATP, NAD, SAH-hydrolase levels and on DNA strand breaks; and (c) the correlation between these parameters with clinical response to DCF. No significant difference in ADA, 5NT, AdR-kinase and SAH-hydrolase activities could be found between responders and non-responders. Incubation of the leukemic cells in vitro with DCF caused an inhibition of ADA, an accumulation of dATP, a moderate reduction in ATP and NAD levels, a suppression of SAH-hydrolase activity and an increase in DNA strand breaks in practically all the leukemic samples, irrespective of clinical response. Our results show that neither measurement of these enzymes nor studies of these biochemical sequelae of ADA inhibition in vitro predicts clinical responsiveness to DCF therapy.

Biochemical mechanisms of deoxycoformycin toxicity in chronic leukemias.

The in-vitro effects of deoxycoformycin (dCF) on dATP, NAD, ATP and DNA strand breaks have been evaluated in the cells from 42 patients with various types of chronic lymphoid leukemia. These included 18 with B-cell chronic lymphoid leukemias of different types (BCL); 10 with hairy cell leukemia (HCL) and 14 with T-cell chronic lymphoid leukemias of different types (TCL). The dATP concentrations in HCL, BCL and TCL increased from means of 2.9, 1.8 and 3.0 to 100.3, 68.2 and 51.3 pmol/10(6) cells respectively after 2 h with 10(-5) M dCF and 10(-4)M deoxyadenosine. After 18-24 h, the NAD levels and total double-stranded DNA decreased to 37 and 12.5% (HCL) to 36 and 21.6% (BCL) and 40 and 20.5% (TCL) of control values respectively. Similar decreases were observed in ATP levels. The results do not suggest that these measurements in vitro would predict which patients with these disorders will respond to dCF therapy. Although HCL responds particularly well to dCF in vivo, no difference in the in-vitro effects of dCF studied here could be detected between cases of HCL and the other types of chronic leukemia.

Multidrug Resistance in Acute Leukaemia: A Comparative Study of Immunocytochemical and RNA Slot Blot Detection of P-glycoprotein Expression.

We have analysed and compared the applicability of mRNA quantitation and immunocytochemistry to the detection of the multidrug resistant gene mdr1 and its protein product P-glycoprotein (PGP) in patients with acute leukaemia. Elevated mdr1 mRNA was detected in 12 of 55 patients and PGP expression by immunocytochemistry in 10 of 27. Raised mdr1 transcript was associated with immunocytochemical positivity in 6 of 8 cases studied. 4 of 10 cases of immunocytochemical positivity were not associated with elevated mdr1 mRNA. Immunocytochemistry showed PGP expression to be heterogeneous within blast populations. In 4 cases PGP positivity was most marked in differentiated leukaemic subpopulations. PGP expression was increased in secondary acute myeloid leukaemia (AML) compared to other disease groups. No association with disease stage was detected except within the secondary leukaemias. In 40 patients studied mdr3 mRNA was raised in one patient with AML and in two remission marrows. These results indicate that further prospective studies in acute leukaemias are required using combined immunocytochemical and RNA quantitation of PGP expression.

Flow cytometric assessment of multidrug resistance (MDR) phenotype in acute myeloid leukaemia.

Forty one patients with acute myeloid leukemia (AML), including 27 at presentation and 14 relapsed or resistant cases, were assessed for laboratory evidence of the MDR phenotype. Leukaemic cells from all 41 cases were studied by immunocytochemistry using the JSB-1 monoclonal antibody and simultaneously by reverse transcription polymerase chain reaction (RT-PCR) to evaluate expression of the mdr 1 gene. Cells from 32/41 cases were also assessed for daunorubicin (DNR) accumulation and retention by flow cytometry (FC). Nineteen of the 41 (46%) patients were positive for MDR by JSB-1 immunocytochemistry (11/27 [41%] at presentation and 8/14 [57%] relapsed or resistant cases). Nine of the 19 (47%) P-gp positive, de novo patients achieved complete remission. 22 patients were negative by JSB-1 immunocytochemistry (16/27 [59%] at presentation and 6/14 [43%] of the relapsed or resistant cases) and 11/22 (50%) P-gp negative patients achieved a complete remission. Of the 32 patients assessed by FC, 7 (22%) were positive for the MDR phenotype with increased DNR accumulation and retention in the presence of the MDR reversing agent verapamil (VPM). 6/7 of the FC positive cases were also JSB-1 positive, and 6 had additional poor risk features. Of the twenty five FC negative patients, 6 had received previous chemotherapy and 15 (60%) achieved complete remission. Mdr 1 mRNA levels were increased in all seven FC positive cases whereas only 7/19 JSB-1 positive cases had raised mdr 1 mRNA levels. These results suggest that the assessment of MDR status by immunocytochemistry using JSB-1 is not predictive of response to chemotherapy. Flow cytometric analysis of blast cells appears to correlate well with mdr 1 mRNA levels and may be a better predictor of treatment outcome.

Inability of poly-ADP-ribosylation inhibitors to protect peripheral blood lymphocytes from the toxic effects of ADA inhibition.

We have evaluated the effectiveness of two inhibitors of poly-ADP-ribosylation, nicotinamide and 3-aminobenzamide as rescue agents in resting and PHA-stimulated lymphocytes damaged by the combination of deoxycoformycin (dCF) plus deoxyadenosine (dAdo). Incubation with dCF (10(-5)M) and dAdo (10(-4)M) for 18 hours, inhibited protein and RNA synthesis in unstimulated lymphocytes and impaired the ability of the cells to respond to PHA stimulation or to give rise to T-cell colonies in methyl-cellulose. Predominantly dead cells using trypan blue exclusion were observed at day 4, in both unstimulated and PHA-stimulated lymphocytes, whether or not the drugs were removed at 18 hours. The number of viable cells at day 4 increased from 13.7% to 41.1% with the addition of 5 mM nicotinamide, and to 28.8% with 5 mM 3-aminobenzamide added with dCF and dAdo. Although nicotinamide was able to prevent a fall in NAD concentration for 24h (but not for 48h) and to reduce the fall of cell ATP concentration, the inhibition by dCF and dAdo of protein synthesis, RNA synthesis, ability of cells to form colonies or to respond to PHA was not reversed. We conclude that inhibition of NAD utilisation by inhibiting ADP-ribosylation with nicotinamide or 3-aminobenzamide does not protect cells in vitro from deoxyadenosine toxicity with ADA inhibition and is not likely to give significant clinical benefit in ADA deficiency.

Selective toxicity of purine nucleosides to human leukaemic cells.

The in vitro cytotoxicity of various purine nucleosides and purine enzyme inhibitors, alone or in combination, and of the alkylating agent mafosfamide (Asta Z7557), incubated for 4 and 24 h have been studied in 17 leukaemic cell lines and normal bone marrow (BM). The purine nucleosides and their analogues included: 2'chlorodeoxyadenosine (CdA), 2'deoxyadenosine (AdR), 3'deoxyadenosine (3'AdR) (cordycepin), adenosine (AR), adenine arabinoside (Ara-A), deoxyguanosine (GdR) and guanine arabinoside (Ara-G). Purine enzyme inhibitors included 2-deoxycoformycin (dCF) and 8-aminoguanosine (8-AG). Cytotoxicity was based on inhibition of (i) incorporation of 3H-leucine into cell proteins and (ii) colony forming units--granulocytic/monocytic (CFU-GM) and for mixed cell colonies (CFU-GEMM). Marked and selective inhibition of T-cell growth was shown by the combinations dCF with either AdR or Ara-A, 8-AG and GdR and by CdA or Ara-G alone; these compounds even at high concentrations produced only partial inhibition of the growth of normal bone marrow CFU-GM and CFU-GEMM except for CdA which completely inhibited the formation of CFU-GEMM colonies. The combination dCF + cordycepin and alkylating agent mafosfamide were, however, toxic to all the cell lines at the concentrations employed, as well as to CFU-GM and CFU-GEMM. The high therapeutic index of some of the purine nucleosides with a relatively short exposure time makes them candidates for selective in vitro removal of residual neoplastic cells in autologous bone marrow transplantation (ABMT) for T-ALL.