Treatment of recurrent and refractory pediatric solid tumors with high-dose busulfan and cyclophosphamide followed by autologous bone marrow rescue

PURPOSE

The purpose of this study was to determine the toxicities of and responses to high-dose busulfan and cyclophosphamide with autologous bone marrow transplant (ABMT) in patients with recurrent or refractory pediatric solid tumors.

PATIENTS AND METHODS

We treated 18 patients (ages, 2 to 38 years; median, 14) who had tumors that were resistant to conventional chemotherapy and radiotherapy with busulfan 16 mg/kg and cyclophosphamide 200 mg/kg. Seventeen patients received bone marrow purged with 4-hydroperoxycyclophosphamide; one received unpurged marrow.

RESULTS

Despite extensive prior treatment, including radiotherapy in 16 patients, toxicity generally was acceptable. For seven patients with measurable disease, there were three partial responses of 2, 10, and 20 months' duration, three patients with stable disease (SD), and one early, toxic death. Of the 11 patients with no measurable disease at the time of transplantation, one patient with osteosarcoma continues in remission at 57+ months and one third of the patients survived for at least 16 months. Mucositis was the predominant nonhematopoietic toxicity.

CONCLUSION

Although the high-dose busulfan and cyclophosphamide combination showed modest activity, changes in the preparative regimen should be considered to improve the response rate in refractory tumors.

A simple method of estimating cerebrospinal fluid pressure during lumbar puncture

It is often difficult to measure cerebrospinal fluid (CSF) pressure in children. CSF flow through a spinal needle is described by the equation: Flow = pressure/(needle constant x relative viscosity). Thus, CSF flow rate during lumbar puncture can be used to estimate CSF pressure. Because the viscosity of CSF is approximately the same as that of normal saline, 0.9% NaCl was used to model CSF flow in vitro. Flow of saline through various spinal needles was measured as pressure and temperature were varied to determine needle constants and variation in viscosity with temperature. Counting periods for which the number of drops counted equals the pressure (in centimeters of H2O) then were determined for each needle size. At patient temperatures less than 40 degrees C, counting periods were calculated at 21, 39, and 12 seconds, for 22-gauge 1.5-inch, 22-gauge 3.5-inch, and 20-gauge 3.5-inch spinal needles, respectively. Viscosity decreased slightly above 40 degrees C, and counting periods became 20, 37, and 11 seconds. Finally, the method was tested prospectively in 12 patients by comparing drop count (over the calculated counting period) to manometric pressure measurement. Drop counts were within 15% of manometric pressure in all patients. This method allows simple and rapid estimation of CSF pressure during lumbar puncture.

Initiation in DNA synthesis by colony-stimulating factors in subsets of human CD34+ marrow cells

Initiation of DNA synthesis by recombinant colony-stimulating factors (CSFs) was assessed in normal human marrow blast cells isolated by expression of CD34 antigen (tritiated thymidine incorporation). Continuous exposure to CSF was required. A mild increase in DNA synthesis was initiated by granulocyte CSF (G-CSF; greater than or equal to 1 ng/ml), to approximately 1.5 times control levels. A greater increase was initiated by granulocyte-macrophage CSF (GM-CSF), with a threshold of approximately 0.1 ng/ml and a plateau increment 2.5 times control levels. CD34+ cells were stimulated by interleukin 3 (IL-3) over a wide concentration range: two times control at 0.1/ml, three times control at 1 ng/ml, and four times control at 10 ng/ml. Overlap between responding populations was analyzed. G-CSF plus GM-CSF induced DNA synthesis greater than GM-CSF alone and supported the growth of much larger granulocyte-monocyte colonies. At saturating IL-3 concentrations, neither G-CSF nor GM-CSF induced additional DNA synthesis; at lower concentrations of IL-3, however, GM-CSF recruited additional cells into DNA synthesis. Using CD10 and CD19 antibodies to separate B-lineage cells, the CD34+ cells responding to CSF were observed to be in the non-B-lineage subset. Therefore 1) the response of CD34+ cell subsets CSFs is IL-3 greater than GM-CSF greater than G-CSF, and the IL-3-responsive population is heterogeneous for dose requirement; 2) a CD34+ subpopulation responding to concurrent G-CSF and GM-CSF includes increased proliferative potential cells; 3) IL-3-responsive cells include GM-CSF- and G-CSF-responsive cells, but cells responding to lower IL-3 concentration do not respond to GM-CSF; and 4) B-cell precursors do not respond to GM-CSF or IL-3 in this assay.

Cisplatin, ara-C and etoposide (PAE) in the treatment of recurrent childhood brain tumors

Sixteen patients with recurrent childhood brain tumors were treated with intravenous cisplatin, cytosine arabinoside and etoposide (PAE), daily for three days every three to four weeks. Objective responses were observed in 6 of 15 evaluable patients and an additional six patients had stable disease for greater than 6 months. The tumor-specific response rate for astrocytoma/glioma was 3 of 7 and for medulloblastoma was 2 of 4. The mean progression-free interval was 11.0 months and the hazard rate for progression was 0.085 per patient-month of observation. The most common toxicities were neutropenia and thrombocytopenia. Clinically significant ototoxicity was identified in 7 patients. The activity of PAE chemotherapy for recurrent childhood brain tumors warrants further investigation.

Osteoblastic metastases from astrocytomas: a report of two cases

The spread of primary central nervous system (CNS) neoplasms beyond the confines of the neuraxis is a relatively uncommon observation. Extraneural metastases involving bone are quite rare, and have been observed almost exclusively in cases of medulloblastoma, high-grade (III-IV) astrocytoma, and glioblastoma multiforme. To date there has been only one case reported of bone metastasis from a "well-differentiated" astrocytoma. We now report two cases of astrocytoma in children with diffuse osteoblastic metastases and a fulminant clinical course. These cases demonstrate that due to the potential morphologic heterogeneity of these neoplasms, an initial biopsy diagnosis of low-grade astrocytoma does not always imply a benign clinical course.

Efficacy of postoperative chemotherapy using cisplatin plus etoposide in young children with brain tumors

Neuraxis radiation therapy (RT) for primary intracranial tumors is associated with major late effects if administered to very young children. To control residual tumor and to delay RT, we treated eight young children (median age 6.5 months) with primary central nervous system (CNS) tumors using combination chemotherapy: cisplatin, 20 mg/M2/day plus VP-16, 75 mg/M2/day i.v. for 5 days, given q. 3-6 weeks for 8 cycles. The tumors were medulloblastoma (one), malignant ependymoma (two), primitive neuroectodermal tumor PNET (two), malignant glioma (two), astrocytoma (one). Six had measurable disease; three had positive cerebrospinal fluid (CSF) cytopathology. All patients with measurable tumor had initial objective responses (three) complete response [CR], one partial response [PR], two minor response [MR], including cytopathology (three CR of three) and metastatic deposits (two CR of two). One patient relapsed during chemotherapy. Median time to disease progression was 17.5 months; median survival was 34 months. Three patients, none of whom received RT, have prolonged progression-free intervals of 47-67 months to date. Neurodevelopmental progress continued during and after chemotherapy. Chemotherapy toxicity was mild. Median neutrophil nadir was 312/mm3, platelets 72,000. Fever during neutropenia occurred in six of 61 courses. Moderate high-frequency auditory losses were detected in three patients, and mild renal injury (GFR less than 70 ml/min) was detected in two of seven evaluable children. This pilot study demonstrates the apparent efficacy and mild toxicity of 5 day courses of cisplatin plus VP-16, with delayed RT, in young children with CNS neoplasms. A POG treatment protocol that incorporates cisplatin plus VP-16 is evaluating primary chemotherapy with delayed radiotherapy in larger numbers of pediatric brain tumor patients.

Selection of normal human hematopoietic stem cells for bone marrow transplantation using immunomagnetic microspheres and CD34 antibody

Complete yet nontoxic removal of tumor cells from autologous marrow grafts has proved difficult. New methods for separating normal stem cells from tumor cells are needed. The CD34+ cells in bone marrow, 1-2% of the low-density leukocytes, include precursors of all lymphohematopoietic lineages and probably also the primitive cells responsible for engraftment. A nontoxic, inexpensive, reproducible, and clinically applicable method for positive selection of CD34+ cells was developed. Paramagnetic microspheres coated with goat anti-mouse IgG1 are used to partition the cells; brief incubation with chymopapain is used to release them from the beads. Chymopapain exposure does not injury colony-forming cells or delay engraftment in rodents. Clinical volumes of bone marrow can be processed rapidly. In pilot experiments, the resulting grafts have a purity of 85-99% CD34+ cells and 40% median recovery of the assayable colony-forming cells. These studies form the background for a Phase I trial of autologous BMT using CD34+ stem cells.

Lymphohematopoiesis: role of growth factors in leukemogenesis and therapy

The interrelationship between proliferation, differentiation, and activation responses of hematopoietic progenitor cells and mature blood cells is complex. Therefore, we are only now learning what role colony-stimulating factors play in the regulation of normal hematopoiesis in vivo and in the dysregulation of hematopoiesis in leukemia. Recent advantages in molecular hematology have opened the door to the therapeutic administration of recombinant growth factors. Through continued preclinical trials in animals and by clinical trials in humans, a better understanding of the precise target cells and mechanisms of action of hematopoietic growth factors will improve the therapeutic index of administering colony-stimulating factors. Better understanding of hematopoietic growth factors will, in turn, suggest novel approaches to therapy of acute lymphoblastic leukemia.

Two-phase [11C]L-methionine PET in childhood brain tumors

Thirteen children (1.8-15.8 years of age) with brain tumors were studied with [11C]L-methionine positron emission tomography (METPET). Patients were injected intravenously with tracer before a baseline PET scan was obtained. To assess the sensitivity of [11C]L-methionine uptake to competitive inhibition, 10 patients received oral L-phenylalanine (100 mg/kg); 1 hour later, a second METPET was obtained. Subjective assessment of [11C]L-methionine uptake closely paralleled results of quantitative examination (r = 0.81). [11C]L-methionine uptake in tumor-containing brain was increased in 11 patients (mean ratio of [11C] radioactivity in tumor to normal brain: 1.5 +/- 0.57; range: 1.13-2.98). Increased tracer uptake occurred in ependymomas (3), medulloblastoma (1), and astrocytomas (5), but was less intense in low-grade tumors. L-phenylalanine reduced L-methionine uptake (25-69%) in 70% of studies. L-methionine uptake was not sensitive to competitive inhibition in brain radiation injury. Two-phase METPET is of potential value in difficult clinical situations evident in children with brain tumors, including the differential diagnosis of tumor recurrence and cerebral radiation injury.

Invasive fungal disease in pediatric acute leukemia patients with fever and neutropenia during induction chemotherapy: a multivariate analysis of risk factors

We evaluated the courses of 115 consecutive cases of pediatric acute leukemia treated with induction chemotherapy. Seventy-two patients developed fever associated with neutropenia; 15 developed systemic fungal infections. We reviewed multiple demographic and treatment characteristics of these patients in an attempt to identify potential risk factors for the development of invasive fungal disease (IFD). Risk factors identified in a univariate analysis included duration of neutropenia after first fever (P less than .0001), diagnosis of acute nonlymphocytic leukemia (ANLL) (P = .003), onset of fever and neutropenia within 5 days of starting induction chemotherapy (P = .009), and multiple (greater than one) surveillance culture sites positive for fungal organisms (P = .02). In a multiple logistic regression analysis, duration of neutropenia (P less than .001) remained a significant risk factor. The study group of patients had a significantly higher risk of fungal infections than a matched group of leukemia patients developing fever with neutropenia due to postremission consolidation chemotherapy (P = .003). In the first 48 patients, 14 (29%) developed IFD. In the subsequent patients (n = 24), intravenous miconazole (5 mg/kg every 8 hours) was begun at the time of the first fever. One of the 24 patients (4%) given miconazole developed IFD. The use of miconazole was a negative risk factor for the development of IFD in univariate (P = .01) and multivariate (P = .05) analysis. We conclude that pediatric leukemia patients who develop fever associated with neutropenia during induction chemotherapy are at high risk for developing IFD. The role of intravenous miconazole at the time of the first fever in this group deserves further study.

Elimination of clonogenic malignant human T cells using monoclonal antibodies in combination with 2'-deoxycoformycin

2'Deoxycoformycin (dCF) specifically inhibits adenosine deaminase (ADA) and causes selective cytotoxicity of normal and malignant T cells. In clinical trials, dCF caused rapid lysis of malignant T lymphoblasts. Although dCF has been associated with dose-limiting nonhematopoietic toxicities, myelosuppression has not been observed. Since dCF is relatively nontoxic to hematopoietic stem cells, we tested dCF for utility in the ex vivo purging of malignant T lymphoblasts from remission leukemic bone marrow for autologous bone marrow transplantation. We found that T lymphoblast cell lines were sensitive to dCF (plus deoxyadenosine [dAdo]) under conditions that did not ablate human hematopoietic colony-forming cells. Moreover, combined pharmacologic (dCF plus dAdo) and immunologic (anti-T cell monoclonal antibodies [McAb] plus complement) purging resulted in additive reduction in clonogenic T lymphoblasts. These results provide the basis for a clinical trial of bone marrow transplantation using combined pharmacologic/immunologic purging of T lymphoblasts from patients' harvested autologous marrow.

Antigenic analysis of hematopoiesis. VI. Flow cytometric characterization of My-10-positive progenitor cells in normal human bone marrow

We have previously shown [1] that the anti-My-10 murine monoclonal antibody detected an epitope of a 115-kDa glycoprotein expressed specifically on KG-1a leukemia cells and a small subset of normal human bone marrow cells. This My-10+ marrow cell subset was shown to contain a highly enriched population of morphologic blast cells and hematopoietic colony-forming cells [1]. In this report, My-10+ cells were characterized, by flow cytometry, as an approximately 1% subpopulation of normal human bone marrow cells. My-10+ cells were slightly larger than lymphocytes and agranular, as determined by their fluorescence-activating cell sorting (-er) (FACS) light-scattering properties. In two-color immunofluorescence experiments, My-10+ cells coexpressed the HLA-DR antigen. However, there was no detectable cellular coexpression of My-10 with either the Leu 1-5, 7, 9, 11, 15, M3, or My-18 antigens. There was an average of approximately 50,000 My-10 molecules per My-10+ marrow cell. This provides further evidence that the My-10 molecule is expressed, at relatively low levels, selectively on early human marrow cells but not on mature lymphohematopoietic cells.

Antigenic analysis of hematopoiesis. IV. The My-11 hematopoietic cell surface antigen is expressed by myelomonocytic and lymphoid, but not erythroid, progenitor cells

The anti-My-11 murine monoclonal antibody reported on in this work identifies a KG-1a cell surface glycoprotein with apparent molecular mass of 210,000 daltons. Peripheral blood B-lymphocytes, and a novel subset of T-lymphocytes (not coinciding with helper or cytotoxic subsets) express My-11 antigen; granulocytes, red cells, and platelets are antigen negative. In normal bone marrow, lymphoid progenitors (TdT positive) and most granulocyte-monocyte progenitors express My-11, but erythroid and multilineage progenitors are My-11 negative. Approximately half of acute leukemia blast cell specimens are My-11 positive. The My-11 antigen distinguishes between lymphohematopoietic cells on the basis of lineage, and assists in the purification of hematopoietic progenitor cells and the subclassification of leukemias and normal lymphocytes.

Antigenic analysis of hematopoiesis. V. Characterization of My-10 antigen expression by normal lymphohematopoietic progenitor cells

The My-10 glycoprotein is an hematopoietic cell surface antigen expressed specifically by undifferentiated (blast) cells, constituting 1%-4% of normal adult bone marrow leukocytes. We used several immunological and in vitro culture methods to analyze the expression of this unique antigen on a variety of lymphohematopoietic progenitor cells. Colony-forming cells (CFC) for granulocyte-monocyte colonies (CFC-GM) and erythroid colonies (BFU-E) were predominantly My-10 positive. CFC with higher proliferative potential were more strongly My-10 positive than CFC with lower proliferative potential, and those for mixed-lineage and blast cell colonies were even more uniformly My-10 positive. Cells maintaining CFC-GM number in short-term marrow culture (pre-CFC) were found to be My-10 positive, as were lymphoid precursors defined by their content of intranuclear terminal deoxynucleotidyl transferase. More mature erythroid precursors (CFU-E) were heterogeneous for antigen expression and lost My-10 antigen progressively, in parallel with advancing maturational stage. The My-10 antigen permits rapid identification and purification of hematopoietic progenitor cells for further study or potential clinical application. The disappearance of the My-10 antigen, moreover, may be a probe for differentiation-linked cellular events.

Analysis of histiocytosis X infiltrates with monoclonal antibodies directed against cells of histiocytic, lymphoid, and myeloid lineage

Surface markers expressed on histiocytosis X (HX) cells were studied using 18 monoclonal antibodies (McAbs) and an in situ indirect immunoperoxidase technique. Specimens of skin (five biopsies from three cases) and lymph node (one case) were studied. Our study confirmed previous findings of an OKT6+ HLA DR+ Leu 3a+ phenotype of HX cells in skin and indicated that lymph node HX cells can also have this phenotype. A mixture of cells expressing T markers/T subset, monocyte/macrophage, and killer/natural killer (K/NK) markers was also present in two cases. However, one case with an aggressive course showed no detectable HNK 1+ (K/NK) or Leu 2A (suppressor) cells in the cutaneous infiltrates and no detectable HNK 1+ lymphocytes in an affected lymph node. The significance of these findings in relation to prognosis and therapy is discussed.

Development of a simplified counterflow centrifugation elutriation procedure for depletion of lymphocytes from human bone marrow

Graft-versus-host disease (GVHD) remains a major complication of allogeneic bone marrow (BM) transplantation. Techniques that effectively purge BM of mature T lymphocytes should reduce the incidence of GVHD and improve survival. We have developed a simplified, two-flow rate, fixed rotor speed counterflow centrifugation-elutriation (CCE) procedure that reproducibly depletes 99% of lymphocytes from Ficoll-Hypaque(F/H)-separated BM or BM buffy-coat. Two predetermined flow rates (24 and 28 ml/min) were used to purge small and intermediate-to-large lymphocytes, respectively, whereas faster sedimenting cells were recovered at the termination of the run. Lymphocyte depletion was substantiated by pan-T monoclonal antibody analysis as well as by complete loss of responsiveness to alloantigens and mitogens. Despite the lack of mature T cells, the depleted marrow fraction retained lymphoid colony-forming ability. Lymphocyte-purged marrow was obtained in high yield (72%), and retained high viability (greater than 97%) and hematopoietic colony-forming ability (greater than 99%). The ratio of total myeloid/erythroid colony-forming cells to T lymphocytes was 73-fold higher in the lymphocyte-depleted fraction than in unseparated BM. We concluded that a two-step CCE procedure can be used to rapidly deplete lymphocytes from both F/H-separated and buffy-coat BM inocula without altering hematopoietic capacity as measured by the in vitro clonogenic assays. It may be possible to adapt this procedure to the separation of the large number of marrow cells required for human BM transplantation.

Two primary brain tumors in one child

We studied a 4-year-old boy with symptoms and signs of a posterior fossa tumor. CT showed two separate intracranial tumors: a fourth ventricle choroid plexus papilloma and a frontal subependymal giant-cell astrocytoma. This case emphasizes that, even in the absence of special genetic predisposition to CNS tumors, two separate intracranial masses may not represent CSF metastasis of a single primary tumor.

Use of lectins in the study of histogenesis of renal cysts

Two peroxidase-labeled lectins, from Tetragonolobus lotus and from peanut (Arachis hypogaea), were used in the study of five cases of polycystic kidney of different types. The two lectins bind with high selectivity of proximal tubules and collecting ducts, respectively. It was found that in infantile polycystic kidneys the cyst walls consistently stained with peanut lectin but not with Tetragonolobus lectin, confirming their origin from collecting ducts only. In adult polycystic kidneys there was mixture of two types of cysts: one positive for peanut, the other positive for Tetragonolobus. The two lectins were, however, mutually exclusive. Moreover, some cysts did not stain for either lectin. The findings suggest that in this type of disease a cyst can originate from any portion of the nephron. The data are also consistent with reports in the literature which used microdissection techniques or chemical analysis of the cyst content. The use of lectins will probably be useful in investigating other types of cystic kidneys, such as dysplastic kidneys and experimental models. The technique is relatively easy to use, but must be applied under strictly controlled conditions, especially when using surgical or postmortem material.

Hallucinogenic amphetamine selectively destroys brain serotonin nerve terminals

(+/-)-3,4-Methylenedioxyamphetamine (MDA), an amphetamine analog with hallucinogenic activity, produced selective long-lasting reductions in the level of serotonin, the number of serotonin uptake sites, and the concentration of 5-hydroxyindoleacetic acid in rat brain. Morphological studies suggested that these neurochemical deficits were due to serotonin nerve terminal degeneration. These results show that MDA has toxic activity for serotonin neurons in rats and raise the question of whether exposure to MDA and related hallucinogenic amphetamines can produce serotonin neurotoxicity in the human brain.

Disparate differentiation in hemopoietic colonies derived from human paired progenitors

We analyzed the differentiation of hemopoietic colonies derived from human paired daughter cells. Candidate progenitor cells were isolated by use of a micromanipulation technique from cultures of My-10 antigen-positive cord blood cells. Then nine to 36 hours later, the paired daughter cells were separated with a micromanipulator and allowed to form colonies in methylcellulose medium containing erythropoietin, phytohemagglutinin leukocyte-conditioned medium, and platelet-poor plasma. The cellular composition of the colonies was determined by differentiating all of the cells of the May-Grünwald-Giemsa-stained preparation. Of a total of 75 evaluable pairs of colonies, 35 consisted of 28 types of disparate pairs revealing nonhomologous lineage combinations. Forty pairs were homologous in lineage expression. However, the proportions of the individual cell lineages were significantly different in the members of some of the homologous pairs. Some pairs revealed significant differences in colony size. These observations are similar to those reported for murine paired progenitors and are consistent with the stochastic model of human stem cell differentiation.

Attenuation changes of the normal and ischemic canine kidney. Dynamic CT scanning after intravenous contrast medium bolus

The potential of CT scanning to explore total and regional renal blood flow was evaluated in a dog model with unilateral renal artery stenosis (n = 7, reduction of renal blood flow: 32-75% of base line flow). Attenuation versus time curves were generated for the renal cortex and medulla, as well as for the aorta and renal vein. A fast CT scanner was used which allowed for up to 24 scans/minute at the same level (slice thickness: 10 mm). A total of 10 ml contrast medium was injected into a peripheral vein for each scan series taken. During baseline conditions, the curve of the renal cortex and medulla demonstrated 2 peaks. The first peak was mainly related to early vascular enhancement, whereas the second peak corresponded mainly to the appearance of contrast medium in the distal convolutes and collecting ducts. Ischemia of the kidney resulted in a reduction of the first peak and a flattening of the leading edge slope. Transport of contrast medium through the extravascular compartments of the kidney was delayed during ischemia. Relative renal blood flow was obtained from the CT data by dividing peak enhancement by rise-time as assessed from the cortical curve. All measurements were related to baseline flow and validated by flow measurements using radioactive labeled microspheres (n = 5). Correlation was found to be r = 0.97.

Localization of hematopoietic progenitor cells in tissue with the anti-My-10 monoclonal antibody

Putative hematopoietic progenitor cells were localized in human lymphoid and nonlymphoid tissues with the use of the anti-My-10 antibody and avidin-biotin immunoperoxidase. Groups of My-10+ round mononuclear cells were evident in the splenic marginal zone near the follicular mantle. My-10+ cells were also evident in the hepatic portal triads and rare cells were seen near Peyer's patches. Except for My-10+ endothelial cells, there was no staining of the thymus, lymph node, skin, or kidney specimens.

[1 year's experience with a 57-cm image intensifier in a universal roentgen work station]

3342 patients were examined by means of a 57 cm image intensifier during one year. The installation permits examination with the patient standing or lying. The resolution was measured with a lead grid under conditions resembling actual practice and were found to be similar to those obtained with angiographic equipment. Using a medium format technique for examining the gut and vascular system resulted in radiographic quality which was not significantly different from the use of a large format technique. For examination of the skeleton, radiographic quality, using the intensifier, was markedly poorer than that resulting from Bucky examinations. There was no diagnostic difference when carrying out lymphograms. The large intensifier window of 53 cm was particularly advantageous for DSA of the pelvic and lower limb vessels when compared with smaller intensifiers. Possible savings of film and contrast medium costs are discussed.