Correlation of P-glycoprotein expression and function in childhood acute leukemia: a children's cancer group study

Clinical drug resistance may be attributed to the simultaneous selection and expression of genes modulating the uptake and metabolism of chemotherapeutic agents. P-glycoprotein (P-gp) functions as a membrane-associated drug efflux pump whose increased expression results in resistance to anthracyclines, epipodophyllotoxins, vinca alkaloids, and some alkylating agents. This type of resistance occurs as both de novo and acquired resistance to therapy for leukemia. We have studied P- gp expression and function in childhood acute leukemias by developing a series of doxorubicin- and vincristine-selected CEM, T-cell lymphoblastoid cell lines that recapitulate the low levels of expression and resistance seen clinically. These cell lines have been used to develop flow cytometric assays for the semiquantitative measurements of P-gp expression with the MRK16 monoclonal antibody and P-gp function using the enhanced retention of rhodamine 123 in the presence of verapamil, a resistance modulator. Kolmogorov-Smirnov statistics, represented by the D measurement, are used to determine the difference in level of P-gp expression by comparing MRK16 staining to an IgG2a isotype control. When D is > 0.09, there is an excellent correlation (R = 0.82) between P-gp expression and function. The evaluation of 107 bone marrow specimens from 84 children with lymphoblastic or myelogenous leukemia showed a statistically significant (P = .004) increase in P-gp function at relapse. P-gp expression at relapse, however, approached but did not reach a significant level (P = .097). Using this methodology, we can identify patients with levels of P-gp expression and function that we can define clinically, as well as children with discordant multidrug resistance phenotypes. This study supports the role of P-gp-mediated drug resistance in childhood leukemia and confirms that P-gp expression and function are measurable in their leukemic blasts. These assays provide the means for the in vitro testing of resistance modulators and the monitoring of in vivo response to treatment with these agents.     

Characterization of the IgG-Fc receptor on human platelets

To determine quantitatively the number and avidity of receptors for the Fc portion of IgG on human platelets, we have measured the binding to platelets of human monomeric monoclonal IgG, and of small covalently crosslinked polymers of IgG1 labeled with 125I. The binding of labeled IgG1 monomers to platelets is too weak to permit quantitation. The binding of dimers or larger polymers of IgG1 is much more avid (greater at 4 degrees C than 37 degrees C), is readily reversible, and is saturable. The number of receptor sites ranges from 400 to 2000 per platelet and the mean equilibrium association constant (Ka) for the binding of dimers at 4 degrees C is 2.2 x 10(7) M-1 +/- 0.9 x 10(7) M- 1. The binding is specific for the Fc portion of IgG, and IgG1 and IgG3 bind to the receptors much more avidly than IgG2 or IgG4. Unlabeled IgG1 dimers are about 7--8-fold more potent in inhibiting binding than are IgG1 monomers, and larger polymers are even more potent than dimers. Thus, the Fc receptors on platelets bind human IgG1 with the same specificity and similar avidity as Fc receptors on polymorphonuclear leukocytes (PMNs), but PMNs have about 300-fold more receptors per unit of surface area than platelets.     

Molecular basis of the heterogeneity of expression of glycosyl phosphatidylinositol anchored proteins in paroxysmal nocturnal hemoglobinuria

The purpose of these studies was to determine the molecular basis of the phenotypic mosaicism that is a defining feature of paroxysmal nocturnal hemoglobinuria (PNH). Analysis of T cell clones from a female patient revealed four distinct phenotypes based on surface expression of glycosyl phosphatidylinositol-anchored proteins (GPI-AP). When PIG-A (the gene that is mutant in PNH) from these clones was analyzed, four discrete somatic mutations were identified. Analysis of X chromosomal inactivation among the abnormal T cell clones was consistent with polyclonality. Together, these studies demonstrate that the phenotypic mosaicism that is characteristic of PNH is a consequence of genotypic mosaicism and that, at least in this case, PNH is a polyclonal rather than a monoclonal disease. That four distinct somatic mutations were present in a single patient suggests that in conditions that predispose to PNH PIG-A may be hypermutable.     

The response of red cell hexose monophosphate shunt after sulfhydryl inhibition

In this investigation, we studied the importance of cellular glutathione (GSH) in the hexose monophosphate shunt (HMPS) activity of unstimulated human erythrocytes and the mechanism by which pyruvate stimulates the HMPS. The rate of HMPS activity was measured by the production of radioactive CO2 from 14C-1-glucose or 14C-1-ribose using a vibrating reed electrometer and ionization chamber. HMPS activity was not significantly impaired by N-ethylmaleimide (NEM) in concentrations which bound all red cell GSH. Red cells incubated under carbon monoxide (CO), an experimental condition which eliminates peroxide production, still had HMPS activity which was 44% of the value under air. Pyruvate stimulation of the HMPS was unaffected by doses of NEM which bound all cellular GSH or by incubation under CO. These data indicated that pyruvate stimulation of the HMPS occurs by pathways which do not involve peroxide formation, GSH, or oxygen. This study indicates that sulfhydryl blockade of GSH does not necessarily inhibit HMPS activity and that HMPS activity in red cells may respond to reactions not linked directly to glutathione reduction.     

The biochemical and clinical consequences of 2'-deoxycoformycin in refractory lymphoproliferative malignancy

A deficiency of adenosine deaminase, an enzyme important in purine nucleoside catabolism, is associated with a severe combined immunodeficiency disease in children. Inhibition of this enzyme in vitro and in vivo results in an impairment in lymphoblast proliferation. We have investigated the pharmacologic inhibition of this enzyme by 2'-deoxycoformycin in 15 patients with hematologic malignancies. Biochemical consequences of the administration of this agent were closely monitored in erythrocytes, nucleated peripheral blood and bone marrow cells, serum, and urine. A marked rise in erythrocyte dATP was accompanied by a depletion of ATP in those patients exhibiting toxicity. Most patients excreted large amounts of deoxyadenosine but not adenosine in the urine. Serum deoxyadenosine rose in patients demonstrating a marked decrease in cell mass. The biochemical disturbances and clinical toxicity, including hepatic, renal, and conjunctival abnormalities, were usually reversible. Central nervous system toxicity, which potentially was the most serious consequence, was associated with high erythrocyte dATP/ATP ratios and high levels of cerebrospinal fluid deoxyadenosine. In patients with lymphoma and leukemia, objective responses were observed but were short- lived. Patients with chronic lymphocytic leukemia receiving weekly low doses of the drug demonstrated minimal toxicity and some efficacy. The chemotherapeutic potential o 2'-deoxycoformycin, as either a single agent or in combination with Ara-A, merits further exploration.     

Changes in protein turnover, IGF-I and IGF binding proteins in children with cancer

Changes in insulin-like growth factor-I (IGF-I) and insulin-like growth factor binding proteins (IGFBPs) were correlated with protein synthesis and breakdown using [1- ¹³C]leucine before chemotherapy and during subsequent febrile neutropenia (FN) in eight children with cancer, aged 6.3–17.5 y. IGF-I levels were similar to age-matched controls before chemotherapy (mean ±SEM: 250 ±28 and 228 ±22 μg l^-1, respectively). During FN, IGF-I fell to 156 ±22 /ng l ^-1(p= 0:02), and rose to 276 ±27 μ g l ^-1 with recovery at 6 months (p = 0:004). Similarly, IGFBP-3 decreased from 4.0 ±0.2mgl^-1 before chemotherapy to 3.0 ±0.3 mgl^-1 during FN (p= 0:01), and returned to 4.1 ±0.2mgl ^-1 at 6 months (p= 0:01). IGF-I correlated with IGFBP-3 (r=+0:7, p <0:001). Scanning densitometry showed a decrease in IGFBP-3 from 94 to 54% during FN, when the presence of IGFBP-3 protease activity was observed. Compared with normal human serum, IGFBP-2 was elevated throughout the study. IGFBP-1 increased from 14.6 ±3.5 to 30.6 ±2.8/ngl^-1 (p = 0:004), whereas serum insulin decreased from 26.5 ±6.8 to 7.8 ±0.8 mUl^-1 (p= 0:03) before and during FN, respectively. Whilst IGF-I and IGFBP-3 fell, daytime growth hormone increased from 3.3 ±0.6 to 6.7±0.8mUl ^-1 (p= 0:01), and cortisol from 197 ±48 to 594±98nmoll ^-1 (p = 0:005). Albumin decreased from 47 ±2 to 38 ±2gl^-1 (p= 0:004) and improved to 47 ±2gl^-1 with recovery (p= 0:003). Protein synthesis increased from 4.5 ±0.4 to 5.0 ±0.6gkg^-1 d^-1 before chemotherapy and during FN, while protein breakdown rose from 5.4 ±0.4 to 6.3 ±0.4kg^-1d^-1. Increasing protein breakdown was related to falling IGF-I and IGFBP-3 levels. Modification of IGFBP-3 by circulating proteolytic activity may alter IGF bioavailability, allowing protein synthesis to increase during periods of severe catabolic stress.