Although high-dose chemotherapy and autologous bone marrow transplantation (ABMT) are increasingly being used for the treatment of relapsed and resistant Hodgkin's disease, there have been few large, single-center studies reported with adequate follow-up to allow full evaluation of this therapeutic modality. We present 155 poor-risk Hodgkin's disease patients who received high-dose BEAM (BCNU, etoposide, cytosine arabinoside, and melphalan) chemotherapy and ABMT who have been studied over a period of 8 years. All patients had either not attained a remission on mechlorethamine, vincristine, procarbazine, prednisone-type therapy and had poor prognostic features at presentation, not attained a complete remission or relapsed within 1 year of an initial alternating regimen, or not attained remission with two or more lines of treatment. At the time of ABMT the relapse status of the patients was as follows: 46 patients were primarily refractory to induction therapy, 7 were good partial responders, 52 were in first relapse, 37 in second relapse, and 13 in third relapse. Seventy-eight patients had chemoresistant disease, 33 had chemosensitive disease at the time of ABMT, and 44 were untested for chemosensitivity at latest relapse. The procedure related mortality in the first 90 days post-ABMT of 10% overall. At 3 months 43 patients (28%) were assessed as complete responders, 72 patients had a partial response (46%), and 24 patients (16%) had no response or progression of disease. However, by 6 months, 53 (24%) patients were assessed as complete responders and 51 (33%) patients had nonprogressive disease. Forty-five patients had received radiotherapy post-ABMT to residual masses (41 patients) or to previous sites of bulk disease (4 patients). The actuarialoverall and progression-free survival at 5 years was 55% and 50%, respectively. On multivariate analysis patients with bulk (masses >10 cm), heavily pretreated patients (those receiving three or more lines of treatment) and females had a significantly poorer prognosis. Relapse status was also significant for progression-free survival in that patients in second (60%) and third relapse (70%) had a better prognosis than those in first relapse (44%) or with primary refractory disease (33%). Response to prior chemotherapy did not predict for progression-free survival. These results enable comparisons to be made between high-dose chemotherapy with ABMT and conventional dose salvage therapy. Furthermore, although the results as a whole are highly encouraging, certain groups carry an unfavorable prognosis.

Exposure of neutrophils to a range of cytokines augments their response to subsequent agonist-induced activation of the respiratory burst. We have examined the effects of several of these factors, both singly and in combination, on the priming of f-met-leu-phe (FMLP) and complement C5a-stimulated neutrophil H2O2 production, using a whole blood flow cytometric assay designed to minimize artefactual activation. Both granulocyte-macrophage colony-stimulating factor (GM-CSF) and tumor necrosis factor alpha (TNF alpha) produced a similar degree of priming of the FMLP-stimulated burst in vitro (558% +/- 86%, n = 41, and 581% +/- 95%, n = 21, of the response seen with FMLP alone, respectively), but with markedly different kinetics (half-maximal response 20 minutes and 7 minutes, respectively). Preincubation with granulocyte colony- stimulating factor (G-CSF) alone caused only modest priming (202% +/- 39%, n = 14). Priming with cytokine combinations of the FMLP-stimulated burst showed that the combinations of G-CSF and TNF alpha and GM-CSF and TNF alpha are highly synergistic, with recruitment of neutrophils unresponsive to priming by single agents. Priming with the combination of GM-CSF and G-CSF was not significantly different to priming with GM- CSF alone. Similar results were obtained using C5a as the respiratory burst stimulus. Significant priming of the FMLP-stimulated respiratory burst was seen in vivo in patients receiving an infusion of GM-CSF (332% +/- 50% of preinfusion response to FMLP, P less than .005, n = 8). Priming was also seen in patients receiving G-CSF (152% +/- 58%, n = 5), although this did not reach conventional significance levels (.05 less than P less than .1). Although GM-CSF infusion caused priming in vivo, this was 48% less than predicted by preinfusion in vitro responses. This result was not due to inadequate GM-CSF levels as addition of further GM-CSF ex vivo did not correct the response. However, these neutrophils were still able to respond appropriately to ex vivo priming with TNF alpha, with a doubling in H2O2 production.

Granulocyte-macrophage colony-stimulating factor (GM-CSF) causes upregulation of neutrophil surface CD11b/CD18 expression, and enhances the adhesion of neutrophils to cultured human endothelial cells in vitro. Systemic administration of GM-CSF results in a rapid, transient decrease in circulating phagocyte numbers. Using a nonhuman primate model (Cynomolgus), we provide histologic evidence that this transient leukopenia is associated with the margination of neutrophils in the pulmonary microcirculation. In four animals receiving 2 to 15 micrograms/kg recombinant human GM-CSF (rhGM-CSF), light microscopic sections of lung contained 36 +/- 8, 17 +/- 7, 21 +/- 6, and 15 +/- 8 (mean +/- SD, n = 20) neutrophils within a graticule grid, as compared with two control animals receiving saline injections whose lung sections contained 2.1 +/- 1.6 and 3.1 +/- 2.1 (mean +/- SD, n = 20) neutrophils within the same grid. Scanning electron microscopy shows activated leukocytes adherent to pulmonary vascular endothelium, but no morphologic evidence of endothelial damage, and no migration of cells into the extravascular space. Margination is associated with an increase in surface expression of CD11b/CD18 on circulating phagocytes, which could contribute to the adhesion to capillary endothelial cells, but CD11b/CD18 levels remain elevated even when demargination is complete. In vitro, monoclonal antibodies (MoAbs) to CD18 and CD11b were able to inhibit neutrophil aggregation and adhesion to endothelium. FMLP-induced neutrophil aggregation was inhibited by 39.8% +/- 11.5% and 44.8% +/- 12.3%, respectively, by MoAbs to CD18 and CD11b (P less than .0005, n = 4 for both); a similar effect was demonstrated on TPA-induced aggregation. MoAb CD18 reduced the adhesion of unstimulated neutrophils to endothelium by 44% (P less than .01, n = 7), and inhibited the amount of GM-CSF-stimulated adhesion by 74% (P less than .001, n = 7), while MoAb to CD11b produced a reduction of unstimulated neutrophil adhesion by 30%, and of GM-CSF-stimulated adhesion by 40% (P less than .01, n = 5, for both). However, when administered in vivo, MoAb CD18 produced only a small, albeit significant, amelioration of GM-CSF-induced margination in vivo, while MoAb CD11b was without effect. These results show that GM-CSF-induced transient leukopenia is associated with enhanced neutrophil adherence to pulmonary vascular endothelium, but suggest that the beta 2 leukocyte integrins CD11/CD18 play only a minor role in this process.

Plasma macrophage colony-stimulating factor (M-CSF) levels were measured by enzyme immunosorbent assay (ELISA) using horse and rabbit polyvalent antibodies raised against human M-CSF purified from urine (hM-CSF). Plasma M-CSF levels in nonpregnant female controls were 364 +/- 69 U/mL (mean +/- SD, n = 20). Pregnancy results in significant elevation of circulating M-CSF levels (541 +/- 164 U/mL, n = 46, P < .0005). M-CSF levels were increased by 28 weeks' gestation and did not increase further in later pregnancy. M-CSF levels were also measured in 20 female controls before and after commencing on the oral contraceptive pill. There was no effect of the contraceptive pill on plasma M-CSF levels (364 +/- 69 U/mL before v 373 +/- 66 U/mL after commencing on the pill). In 28 nonpregnant patients with untreated immune thrombocytopenic purpura, (ITP), plasma M-CSF levels were significantly increased (797 +/- 402 U/mL, n = 28, v 364 +/- 69 U/mL in controls, N = 20, P < .0005). Pregnant ITP patients had higher levels of plasma M-CSF (929 +/- 327 U/mL, n = 25) than nonpregnant patients, but this difference was not significant. Elevated levels of M-CSF in ITP may reflect activation of the reticuloendothelial system (RES), which could result in positive feedback to increase the destruction of platelets. The increase in M-CSF associated with pregnancy could contribute to the exacerbation of latent ITP in pregnancy.

Human granulocyte-macrophage colony-stimulating factor (GM-CSF) increases neutrophil surface expression of the cellular adhesion molecule CD11b and primes the respiratory burst stimulated by the bacterial peptide f-met-leuphe (FMLP). We have examined the effects of the isoquinolinesulfonamide protein kinase inhibitors H7 and H8 on these functions of GM-CSF using whole blood assays. Concentrations of H7 and H8 that inhibited the 12-O-tetradecanoyl-phorbol-13-acetate (TPA) stimulated upregulation of CD11b expression and activation of the respiratory burst, both augmented the effects of GM-CSF. H7 and H8 enhanced the GM-CSF-stimulated increase in CD11b expression to 215% +/- 10% (P less than .05) and 233% +/- 45% (P less than .05), respectively, of the value obtained with GM-CSF alone. The GM-CSF priming of the FMLP- stimulated oxidative burst was increased to 190% +/- 44% (P less than .01) by preincubation with H7 and to 172% +/- 25% (P less than .01) with H8. Preincubation with H8 did not affect overall binding of 125I- GM-CSF to neutrophils, but inhibited GM-CSF receptor internalization after ligand binding (P less than .05). These data indicate that the effects of GM-CSF are not mediated by protein kinase C and that a phosphorylation event down-modulates the neutrophil response to GM-CSF. It suggests that internalization of the receptor-ligand complex is not a rate-limiting step in signal transduction, and that regulation of the rate of internalization may be an important level of control of the activity of GM-CSF.