Antiproliferative actions of insulin-like growth factor binding protein (IGFBP)-3 in human breast cancer cells

A number of lines of evidence suggest that IGFs are important mitogens in human breast cancer: (1) IGFs are the most potent growth factor in human breast cancer cells; (2) estrogen stimulates expression of IGF-II and the type 1 IGF receptor; and (3) stromal cells express IGFs, which may act in a paracrine manner. Numerous studies have demonstrated that IGFBPs modulate the mitogenic effects of IGFs in the local environment. In particular, we have recently demonstrated that IGFBP-3 inhibits the growth of Hs578T and MDA-MB-231 human breast cancer cells in an IGF-independent manner. Further studies revealed the existence of cell surface-associated IGFBP-3 receptors. Receptor binding and the subsequent antiproliferative action of IGFBP-3 was inhibited by IGFs, owing to the formation of an IGF-IGFBP-3 complex that prevents the binding of IGFBP-3 to its receptors. In addition, exogeneously added soluble heparin or heparan sulfate inhibited the binding of IGFBP-3 to the cell surface in a dose-dependent manner. However, when heparin and heparan sulfate linkages of glycosaminoglycans on the cell surface were enzymatically remove, IGFBP-3 binding was only minimally affected. These data suggest that soluble heparin or heparan sulfate forms a complex with IGFBP-3, thereby inhibiting receptor binding of IGFBP-3, rather than competing with cell-surface glycosaminoglycans for binding of IGFBP-3. Additionally, the role of IGFBP-3 in the antiproliferative effects of transforming growth factor (TGF)-beta and retinoic acid (RA) is supported by our observations that: (1) inhibition of IGFBP-3 gene expression using an IGFNBP-3 antisense oligodeoxynucleotide not only blocks TGF-beta and RA simulation of IGFBP-3 production by up to 90%m but also inhibits their antiproliferative effects by 40-60%; and (2) treatment with IGF-II and IGF-II analogs diminish TGF-beta effects by blocking TGF-beta induced binding of IGFBP-3 to the cell surface. Taken together, our results support the hypothesis that IGFBP-3 is an important antiproliferative factor in human breast cancer, acting in an IGF-independent manner in addition to its ability to modulate the binding of IGF peptides to IGF receptors.

Developments in mucosal influenza virus vaccines

Immunogenicity and efficacy of aerosol inactivated split influenza virus vaccines, which are threefold the strength of the vaccines for parenteral use, and cold-adapted reassortant live influenza virus vaccines were evaluated. Mucosal immune responses were evaluated by quantifying specific IgA antibody of the nasal swab solution, and systemic immune responses were evaluated by determining serum haemagglutination inhibition antibody levels. Efficacy of the aerosol inactivated vaccine was evaluated by a challenge test using live vaccine virus. It was concluded that mucosally administered inactivated influenza virus vaccine stimulated systemic and mucosal immune responses more strongly than live influenza virus vaccine and manifested a much stronger booster effect than live vaccine. Mucosal administration of inactivated influenza virus vaccine was effective in preventing infection by live vaccine virus.

Identification of insulin-like growth factor-binding protein-3 (IGFBP-3) fragments and IGFBP-5 proteolytic activity in human seminal plasma: a comparison of normal and vasectomized patients

Previous studies have demonstrated that insulin-like growth factor (IGF) peptides, IGF-binding proteins (IGFBPs), and IGFBP-3 proteolytic activity, are present in human seminal plasma (SP). In this study, we have further characterized the IGFBPs in SP using immunoprecipitation and Western ligand blotting, Western immunoblotting, affinity cross-linking and immunoprecipitation, and RIA of IGFBP-3 using two different assays and have identified additional proteolytic activities for IGFBP-4 and IGFBP-5 in SP. Immunoprecipitation with antibodies to IGFBP-2, IGFBP-3, and IGFBP-4, before and after affinity cross-linking, demonstrated that intact IGFBP-2 and IGFBP-4 are present in SP, but intact IGFBP-3 is absent. Low mol wt fragments of IGFBP-3, which did not bind to IGF-I or IGF-II on Western ligand blot and did not cross-link to IGF-II, were demonstrated on Western immunoblot and were measurable by two different RIAs. Proteolytic activities for IGFBP-4 and IGFBP-5 were demonstrated in SP by incubation with the respective iodinated IGFBPs. On comparing the proteolytic activity for IGFBP-4 by purified prostate-specific antigen (PSA; a known IGFBP-3 protease in SP) or by SP with measured equivalent concentrations of PSA, the dose response and fragment patterns were identical. With IGFBP-5, however, proteolysis by purified PSA was different from that by SP with measured equivalent concentrations of PSA: 1) proteolysis by pure PSA was less efficient than matched concentrations of SP; 2) the pattern of fragments after proteolysis by pure PSA was different from that after proteolysis by matched concentrations of SP; and 3) proteolysis by purified PSA was significantly inhibited by phenylmethylsulfonylfluoride and aprotinin, but proteolysis by SP was not. We conclude that human SP contains intact IGFBP-2 and IGFBP-4, but has only IGFBP-3 fragments with low affinity for IGF peptides; that PSA is able to proteolyze IGFBP-4 and IGFBP-5 (as well as IGFBP-3); and that an additional IGFBP-5 protease is probably present in SP. There was no significant difference in any of these findings in SP from normal volunteers, vasectomized patients, or patients with idiopathic azoospermia. The roles of IGFBPs and IGFBP proteases in the male reproductive system and male infertility remain to be further elucidated.

Alteration in insulin-like growth factor-binding proteins (IGFBPs) and IGFBP-3 protease activity in serum and urine from acute and chronic renal failure

The insulin-like growth factors, IGF-I and IGF-II, are proteins that promote cellular growth and differentiation of various organs, including the kidney. These peptides interact with high affinity cell surface receptors and bind to a family of IGF-binding proteins (IGFBPs). Altered serum and urinary IGFBP patterns in children with chronic renal failure have been previously described. In this study, we evaluated serum and urinary IGFBP profiles in acute renal failure patients (ARF; n = 10) and chronic renal failure patients (n = 10), using Western ligand blots. Most patients with acute or chronic renal failure showed decreased intact serum IGFBP-3 and increased serum IGFBP-2. Both groups displayed marked urinary IGFBP alterations, including increased urinary IGFBP-1 and totally absent urinary IGFBP-3, as detected by Western ligand blot. To evaluate altered IGFBP profiles, we performed IGFBP-3 protease assays with sera and urine from renal failure patients and normal controls. Although control urine had only minor protease activity (defined by the ability to degrade [125I]IGFBP-3), significant protease activity was found in urine from renal failure patients. The proteolytic pattern and susceptibility to protease inhibitors in most renal failure urine samples were the same as those seen in normal urine and with plasmin. Protease activity was completely inhibited by serine protease inhibitors. We speculate that urinary protease activity is mediated primarily by a serine protease(s), which may be involved in the modulation of renal IGF activity in health and disease.

The beta 1 subunit mRNA of the rat brain Na+ channel is expressed in glial cells

Although the molecular characteristics of glial Na+ channels are not well understood, recent studies have shown the presence of mRNA for rat brain Na+ channel alpha subunits in astrocytes and Schwann cells. In this study, we asked whether the mRNA for the rat brain Na+ channel beta 1 subunit is expressed in glial cells. We performed in situ hybridization using a complementary RNA probe for the coding regions of the rat brain Na+ channel beta 1 subunit mRNA and detected beta 1 subunit mRNA in cultured rat optic nerve astrocytes and sciatic nerve Schwann cells. The beta 1 subunit was amplified by reverse transcription-polymerase chain reaction in rat optic and sciatic nerves, which lack neuronal somata but contain astrocytes and Schwann cells, respectively. Doublet bands of the beta 1 subunit mRNA were amplified from both optic and sciatic nerves. Through the cloning and sequencing of these bands, we confirmed the amplification of a mRNA highly homologous to the previously cloned rat brain Na+ channel beta 1 subunit (beta 1.1) and a novel form of the beta 1 subunit mRNA (beta 1.2), which is closely homologous to beta 1.1 but contains an additional 86-nucleotide insert in 3' noncoding regions. Two beta 1 subunit mRNAs were also amplified from rat brain and skeletal muscle, but not from rat liver or kidney. These results indicate that rat brain Na+ channel beta 1 subunit mRNAs are expressed in glial cells as well as in neurons.

Rat brain Na+ channel mRNAs in non-excitable Schwann cells

The expression of rat brain voltage-sensitive Na+ channel mRNAs in Schwann cells was examined using in situ hybridization cytochemistry and RT-PCR. The mRNAs of rat brain Na+ channel subtype II and III, but not subtype I, were detected in cultured Schwann cells from sciatic nerve and in intact sciatic nerve, which contains Schwann cells but not neuronal cell bodies. These results indicate that rat brain Na+ channel mRNAs, which have been considered as mainly neuronal-type messages, are also expressed in glial cells in vitro and in vivo.

Insulin-like growth factor-binding protein-2 concentrations in cerebrospinal fluid and serum of children with malignant solid tumors or acute leukemia

Many tumor cell lines express insulin-like growth factors (IGFs) as autocrine growth factors and IGF-binding protein-2 (IGFBP-2) as a major IGFBP, which, in turn, regulates the bioavailability and bioactivity of IGFs. The aim of our study was to investigate 1) whether children with malignancies have elevated IGFBP-2 levels in cerebrospinal fluid (CSF) and serum, and 2) whether IGFBP-2 levels in these biological fluids could be useful markers for the diagnosis and follow-up of certain tumor types. We, therefore, measured IGFBP-2 levels in the CSF and serum of children with malignancies by Western ligand blot analysis; RIA with alpha IGFBP-2, a polyclonal antibody for human IGFBP-2; and immunoprecipitation with alpha IGFBP-2 and alpha Hec-1a, a polyclonal antibody that recognizes IGFBP-2 and -3. Furthermore, the expression of IGFBP-2 messenger ribonucleic acid in tumor tissue from three central nervous system (CNS) tumor patients was analyzed by Northern blot analysis. We examined CSF from 21 children with malignant CNS tumors, 25 patients with acute leukemia, and 4 patients with peripheral solid tumors and compared the IGFBP-2 levels with those in CSF from 21 patients who received a lumbar puncture to exclude meningitis. Serum was obtained from 7 patients with solid tumor, 12 patients with malignant CNS tumor, and 16 patients with acute leukemia. The serum IGFBP-2 levels were compared to serum levels in 5 patients with sarcoma who had reached complete remission and 13 normal control children. CSF and serum were collected at the same time, before initiation of therapy. Patients with malignant CNS tumors showed elevated IGFBP-2 levels in CSF (P < 0.001), whereas patients with solid peripheral tumor or acute leukemia had normal IGFBP-2 levels in CSF. CNS tumor patients with microscopically detectable malignant cells in the CSF had the highest CSF IGFBP-2 levels. Serum IGFBP-2 levels were increased in patients with solid peripheral tumors (P < 0.05), whereas patients in complete remission had normal serum IGFBP-2 levels. In summary, IGFBP-2 was elevated specifically in CSF from patients with CNS tumor, whereas IGFBP-2 serum levels were elevated in children with various peripheral tumors. We conclude that IGFBP-2 in CSF could be a specific marker for malignant CNS tumors. We detected high IGFBP-2 messenger ribonucleic acid expression in 1 of 3 CNS tumor tissues analyzed.

In situ hybridization localization of the Na+ channel beta 1 subunit mRNA in rat CNS neurons

Localization of Na+ channel beta 1 subunit (Na beta 1) mRNA was examined in adult rat hippocampus, cerebellum and spinal cord by in situ hybridization histochemistry. In hippocampus, Na beta 1 mRNA was strongly expressed by CA3 followed by CA1 pyramidal cells and dentate granule cells. In cerebellum, strong Na beta 1 mRNA expression was observed in Purkinje cells and moderate expression in granule cells and scattered cells of the molecular layer. In spinal cord, neurons in gray matter exhibited moderate to strong expression of Na beta 1 mRNA. These results provide the first localization study of Na beta 1 mRNA in the CNS, demonstrating a differential expression in different neurons.

Insulin-like growth factor binding protein-3 concentrations and insulin-like growth factor binding protein-3 protease activity in sera of patients with malignant solid tumors or leukemia

IGF binding proteins (IGFBP) regulate the bioavailability and bioactivity of IGF. The major IGFBP in serum is IGFBP-3. We investigated whether sera from children with malignancies show alterations in levels of IGFBP-3 as measured by Western ligand blot analysis (WLB) and RIA with alpha IGFBP-3gl, a specific rabbit polyclonal antibody. Furthermore, IGFBP-3 proteolysis was quantified by densitometric analysis of [125I]IGFBP-3 protease assays, and IGFBP-3 fragments were visualized by Western immunoblot with alpha IGFBP-3gl. We examined sera from 21 children with solid tumors, five patients with sarcoma who had reached complete remission, and 13 children with acute leukemia. Serum samples were collected at diagnosis, before initiation of therapy. Sera of 10 healthy children served as normal controls. Children with solid tumor or leukemia had significantly higher (p < 0.001) IGFBP-3 protease activity in serum than did normal controls or patients with sarcoma in complete remission. Corresponding to this finding, densitometry of WLB showed lower IGFBP-3 levels in sera of children with malignancies in comparison with normal controls. The negative correlation (p < 0.001, r = -0.80) between IGFBP-3 proteolysis, as measured by [125I]IGFBP-3 protease assay, and IGFBP-3 band density on WLB indicates that proteolysis is the probable reason for reduction of IGFBP-3 on WLB. IGFBP-3 concentrations measured by RIA were in the normal range for most patients, further indicating that differences in serum IGFBP-3 levels measured by WLB reflect protease activity.(ABSTRACT TRUNCATED AT 250 WORDS)

Sodium channel mRNAs in cultured spinal cord astrocytes: in situ hybridization in identified cell types

The expression of rat brain sodium channel alpha-subunit mRNAs I, II and III and a putative glial cell-specific sodium channel (NaG) mRNA was examined in cultured astrocytes from P-0 rat spinal cord by RNA blot hybridization and by non-isotope in situ hybridization cytochemistry utilizing two independent sets of isoform-specific RNA probes. Sodium channel mRNA I was not detectable in the cultured astrocytes by RNA blot or in situ hybridization. Sodium channel mRNA II showed negligible-to-low levels of expression in flat, fibroblast-like and 'pancake' astrocytes at 4 days in vitro (div), while stellate, process-bearing astrocytes exhibited low-to-moderate levels of mRNA II expression. At 7 div, mRNA II expression ranged from low-to-moderate in flat astrocytes and was moderately high in most process-bearing astrocytes. In RNA blots, a weak band was observed at 9.5 kb. Sodium channel mRNA III expression was negligible in flat astrocytes and was detectable in low-to moderate levels in stellate astrocytes beginning at 4 div; by 7 div, mRNA III was detectable in low levels in flat astrocytes and low-to-moderate levels in stellate astrocytes. RNA blots showed two bands of nearly equal intensity, one at 9.0 kb and one at 7.2 kb. NaG mRNA showed increased expression with time in culture, being detectable in flat and stellate astrocytes at 4 div and becoming very prominent in flat astrocytes at extended times in culture. In RNA blots of cultured astrocytes at 7 div, a strong hybridizing signal with the NaG probe was observed. These observations demonstrate that flat and stellate astrocytes cultured from rat spinal cord express rat brain sodium channel mRNA II and III, and NaG, and suggest that astrocytes in vitro may co-express multiple forms of sodium channel mRNA.

Clinicopathological study of poststreptococcal glomerulonephritis in the elderly

In order to characterize the clinical features in elderly patients with poststreptococcal glomerulonephritis (PSGN), 31 patients, who were both histologically and immunologically proven to be PSGN, were divided up into 3 groups according to age; elderly patients being 55 years or older (n = 7), middle-aged patients being 40 to 54 years old (n = 7) and younger patients being 20 to 39 years old (n = 17). Renal functional impairment as indicated by serum creatinine levels of over 2.0 mg/dl, developed in 4 of the elderly patients and later completely improved at the end of the follow-up period (178.9 +/- 150.7 days). On the other hand, none of the middle-aged and younger patients revealed any renal function impairment. Hypertension was observed more frequently in elderly patients than in younger patients, and was 86% and 6% at the time of admission, respectively. In addition, 43% of elderly patients remained hypertensive at the time of discharge. There was no difference in total protein, ASO, CH50, the degree of proteinuria or proliferative and exudative features in renal histology among the three groups. None of the elderly patients with PSGN died or developed persistent renal failure. In conclusion, elderly patients with PSGN had a high incidence of renal functional impairment and hypertension compared to the younger patients on admission, however, their short-term prognosis seems to be favorable.

Risk factors for the development of encapsulated filtering blebs. The role of surgical glove powder and 5-fluorouracil

PURPOSE

To study risk factors for the development of encapsulated filtering bleb, especially the effect of surgical glove powder and 5-fluorouracil on the development of encapsulated blebs.

METHODS

One hundred forty-eight patients who underwent trabeculectomy were prospectively studied. After patient of each surgeon were divided into two subgroups of 5-fluorouracil versus no 5-fluorouracil, each subgroup was then randomized for powder versus powder-free surgical gloves. All patients were followed for at least 2 months (mean, 3.5 months; range, 2-10 months).

RESULTS

The overall incidence of encapsulated bleb was 11.2% when an encapsulated bleb was defined by clinical characteristics with a peak intraocular pressure of 26 mmHg or higher. The group that received 5-fluorouracil had a lower incidence of encapsulated blebs despite a higher rate of previous surgeries, whereas the group with powdered gloves had increased incidence. The subgroup in which powder-free gloves were used in conjunction with postoperative injections of 5-fluorouracil had the lowest incidence of encapsulated blebs (2.6%) compared with the other subgroups (12.0%-17.9%) (P = 0.01).

CONCLUSION

Combined use of powder-free surgical gloves and 5-fluorouracil may reduce the incidence of encapsulated bleb.

The expression of rat brain voltage-sensitive Na+ channel mRNAs in astrocytes

Astrocytes from various regions of CNS have been shown to express voltage-activated Na+ currents. To date, three distinct subtypes (I, II and III) of Na+ channels have been cloned from rat brain. We have applied a combined technique of reverse transcription and polymerase chain reaction (RT-PCR) to examine the expression of rat brain Na+ channels in rat astrocytes in vivo and in vitro. Five PCR primer sets were used to amplify coding or 3' non-coding regions of subtype I, II, and III Na+ channels. We were able to amplify all three of these rat brain Na+ channel subtypes from rat optic nerve, which does not have neuronal cell bodies but does contain astrocytes known to express voltage-sensitive Na+ channels. In studies on cultured spinal cord astrocytes, we were also able to amplify all three subtypes of rat brain Na+ channel mRNAs. In control experiments, RT-PCR was performed on RNAs prepared from several rat tissues, including brain, skeletal muscle, and liver. Rat brain was shown to express the three Na+ channel subtypes as expected. In rat skeletal muscle, subtype I and III Na+ channel mRNAs, but not subtype II, were amplified. In rat liver, Na+ channel messages were not detectable. The present study provides the first direct evidence that astrocytes in vivo and in vitro express rat brain voltage-sensitive Na+ channel mRNAs, which have been considered as mainly neuronal-type Na+ channel messages.

An important role of glomerular segmental lesions on progression of IgA nephropathy: a multivariate analysis

The independent predictors of progression of IgA nephropathy (IgAN) were investigated by multivariate life table analysis, using Cox's proportional hazard model, in 225 patients with IgAN diagnosed by renal biopsy (Bx). There were 105 men and 120 women. Mean age at Bx was 32.5 years. The follow-up period following Bx was 4.0 +/- 2.6 yrs, ranging from 5 months to 11 yrs. The clinical parameters analyzed were age at the time of discovery of the disease, age at Bx, intervals from discovery to Bx, presence or absence of macrohematuria, and clinical data at Bx such as presence or absence of hypertension, the degree of hematuria, the amount of urinary protein excretion, serum creatinine and serum IgA concentration. The following immunopathological parameters were also examined; glomerular hypercellularity index, percentage of glomeruli associated with segmental lesions such as tuft adhesions, crescents and segmental sclerosis, percentage of obliterated glomeruli by global sclerosis, severity of interstitial infiltration, fibrosis, arterial wall thickening, arterial hyaline changes, and intensity of the depositions of IgG, IgA, IgM, C3, C1q and fibrinogen by immunofluorescent study. Among all clinical and pathologic parameters examined, the following parameters were proved to be significant independent predictors of progression of IgAN: serum creatinine exceeding 1.5 mg/dl in men and 1.3 mg/dl in women, proteinuria over 2 g/day, segmental lesions involving more than 25% of glomeruli and interstitial fibrosis occupying more than 25% of cortical area.(ABSTRACT TRUNCATED AT 250 WORDS)

Demonstration of receptors for insulin-like growth factor binding protein-3 on Hs578T human breast cancer cells

Hs578T human breast cancer cells are from an estrogen receptor-negative breast cell line derived from a highly aggressive mammary tumor. Our previous insulin-like growth factor binding protein-3 (IGFBP-3) binding studies (Oh, Y., Müller, H. L., Lamson, G., and Rosenfeld, R. G. (1993) J. Biol. Chem. 268, 14964-14971) have demonstrated specific binding of IGFBP-3 on the Hs578T cell surface and a significant inhibitory effect of IGFBP-3, itself, on monolayer growth. In this study, we have demonstrated cell surface association proteins that are specific for IGFBP-3 by showing: 1) detection of 20-, 26-, and 50-kDa proteins by affinity cross-linking with 125I-IGFBP-3E. coli and immunoprecipitation of cell monolayers and cell lysates with anti-IGFBP-3 antibodies; 2) dose-dependent competition of 125I-IGFBP-3E. coli by unlabeled IGFBP-3E. coli; 3) inhibition of IGFBP-3 binding to these cell surface proteins by EDTA and by coincubation with native insulin-like growth factor II (IGF-II), but not by coincubation with [Gln6,Ala7,Tyr18,Leu19,Leu27]IGF-II, an IGF-II analog with decreased affinity for IGFBP-3; and 4) partial purification of 20- and 26-kDa species by IGFBP-3.anti-IGFBP-3 antibody immunoaffinity membranes. Characteristics of these specific IGFBP-3 cell surface association proteins are identical to those observed in our previous monolayer binding assay and monolayer growth assay experiments. The specificity of binding and the inhibitory effect of IGFBP-3 binding on Hs578T cell growth suggest that these cell surface proteins are IGFBP-3-specific receptors or receptor subunits mediating the direct inhibitory effect of IGFBP-3 on monolayer growth of Hs578T cells.

Concentrations of insulin-like growth factor (IGF)-binding protein-3 (IGFBP-3), IGF, and IGFBP-3 protease activity in cerebrospinal fluid of children with leukemia, central nervous system tumor, or meningitis

Insulin-like growth factor-II (IGF-II) is the major IGF in human cerebrospinal fluid (CSF), whereas IGF-I is only detectable in trace amounts. The major IGFBPs in CSF are IGFBP-2 and IGFBP-4. Normally, IGFBP-3 is a minor component in CSF of healthy subjects, but may be increased in pathological states. We investigated IGF-I, IGF-II, and IGFBP-3 levels by specific RIAs in CSF from patients with central nervous system (CNS) tumor or leukemia and compared them with values in patients with meningitis. Further, as proteolysis of IGFBP-3 is part of the modulation of IGF activity, IGFBP-3 fragmentation was quantified by densitometric analysis of [125I]IGFBP-3 protease assays. We examined CSFs of 23 children with malignant CNS tumors, 18 children with leukemia, and 13 children with meningitis. The CSF from 38 children who received lumbar punctures to exclude meningitis was used to define the normal range for IGF-I, IGF-II, IGFBP-3, and IGFBP-3 protease activity in CSF. CNS tumor and leukemia patients had normal levels of IGF-I and IGF-II in CSF, whereas the IGF-II concentration in CSF of meningitis patients was elevated (P < 0.0001). Only 2 of 13 (15%) meningitis patients had elevation of CSF IGFBP-3 concentrations, despite high numbers of inflammatory cells. By comparison, elevated IGFBP-3 concentrations were found in the CSF of 16 of 23 (70%) CNS tumor patients and 6 of 7 (86%) CNS tumor patients with microscopically detectable malignant cells in CSF. Twelve of 13 (92%) patients with medulloblastoma or ependymoma and all 7 medulloblastoma/ependymoma patients with malignant cells in CSF had elevated IGFBP-3 concentrations. The IGFBP-3 protease activity in CSF was elevated in 15 of 16 (94%) patients with CNS tumors of high grade histological malignancy. Five of 6 patients (83%) with acute leukemia and microscopically detectable malignant cells in CSF at the time of diagnosis showed elevated IGFBP-3 concentrations, with normalization after chemotherapy. Leukemia patients without malignant cells in CSF had normal IGFBP-3 concentrations. We conclude that in CSF of children with highly malignant CNS tumor or CNS leukemia, IGFBP-3 is elevated. This phenomenon could be caused by disruption of the blood-CSF barrier and entry of IGFBP-3 from serum, although this appears unlikely, especially for CNS leukemia. More likely possibilities are 1) local production of IGFBP-3 by CNS tumor tissue and secretion into the CSF, or 2) local production of IGFBP-3 by malignant cells within the CSF.