Phase I study of recombinant tumor necrosis factor in cancer patients.

Tumor necrosis factor is a cytokine derived from activated macrophages. This agent is cytostatic and cytolytic against transformed human cell lines in vitro and has in vivo activity against a variety of murine tumors. We report a clinical study of the pharmacokinetics, toxicity, and biological activity of i.v. and i.m. administered recombinant human tumor necrosis factor (rTNF). Twenty patients with metastatic cancer were given rTNF in doses ranging from 1 to 200 micrograms/m2 by alternating i.m. and i.v. bolus injections with a minimal intervening period of 72 h. Each patient received a maximum of eight treatments given twice weekly over a 4-week period. With i.v. bolus administration, serum concentrations of rTNF were detected by enzyme-linked immunosorbent assay at doses of 25 micrograms/m2 or greater. The clearance of rTNF in the serum was described by a monoexponential equation with a half-life calculated to be 14-18 min. After i.m. administration, serum concentrations of rTNF were consistently detected by enzyme-linked immunosorbent assay at doses of 150 micrograms/m2 or greater. Peak concentrations were observed within 2 h and rTNF was occasionally detected, at the lower limit of sensitivity of the assay, at 24 h postinjection. rTNF was well tolerated clinically in this dose range, and there was evidence of antitumor effect.

Analysis of the epidermal growth factor receptor gene in fresh human head and neck tumors.

Seventeen fresh uncultured tumors obtained from biopsies of patients with various forms of head and neck cancer and two squamous cell carcinoma cell lines were analyzed for RNA expression and structural alterations of the epidermal growth factor reception gene. We used cDNA probes for the external and the internal domains of the gene. Enhanced mRNA expression was found only in one squamous cell carcinoma cell line, which is known to have high levels of epidermal growth factor receptor. No amplification or structural rearrangement of the epidermal growth factor receptor gene was found.

Effect of phorbol ester treatment on drug-induced, topoisomerase II-mediated DNA cleavage in human leukemia cells.

Tumor-promoting phorbol esters such as phorbol 12-myristate 13-acetate (PMA) induce the monocytoid differentiation of HL-60 human leukemia cells. The cellular receptor for PMA is protein kinase C. However, cellular events distal to protein kinase C phosphorylation are also critical steps toward differentiation. These events may include specific programs of oncogene transcription that have been associated with phorbol ester-induced leukemic cell differentiation. Recently, it has been found that topoisomerase II could be activated by protein kinase C-mediated serine phosphorylation and that PMA treatment of HL-60 cells enhanced extractable topoisomerase II from these cells. Additionally, topoisomerase II-reactive antineoplastic drugs could block PMA-induced differentiation of HL-60. This enzyme has been implicated in gene regulation, and drug-induced, topoisomerase II-mediated DNA cleavage sites have been identified within cellular oncogenes. Thus, topoisomerase II could play a critical role in the signal transduction cascade leading from PMA-protein kinase interaction to monocytoid differentiation. We have examined this relationship between topoisomerase II and PMA-induced differentiation through measurements of drug-induced, topoisomerase II-mediated DNA cleavage (via alkaline elution) in PMA-treated HL-60 cells. Etoposide-induced DNA cleavage was reduced 10-fold in HL-60 cells treated with 10 nM PMA for 24 h. Neither dimethyl sulfoxide (which produces granulocytoid differentiation) nor non-differentiation-inducing phorbol esters could produce this effect. The decreased cleavage was not due to a PMA-induced inhibition of cell-associated etoposide and was demonstrable in nuclei isolated from PMA-treated cells. The decrease was not simply related to decreased cellular proliferation rate as reflected in the inhibition of DNA synthesis because conditions leading to marked inhibition of DNA synthesis did not necessarily inhibit etoposide-induced DNA cleavage. By contrast, lower concentrations of PMA inhibited etoposide-mediated DNA cleavage disproportionately compared with PMA effects on DNA synthesis. Interestingly, PMA reduced cleavage induced by the topoisomerase II-reactive DNA intercalator 4'-(9-acridinylamino)methanesulfon-m-anisidide by 2-fold, suggesting that specific drug-DNA interactions could partially overcome the PMA-induced effect that resulted in decreased etoposide-induced, topoisomerase II-mediated DNA cleavage. Nuclear proteins in 0.35 M NaCl extracts from untreated or PMA-treated HL-60 cells were virtually identical in topoisomerase II activity and in topoisomerase II-associated drug sensitivity.(ABSTRACT TRUNCATED AT 400 WORDS)

Amplified and overexpressed epidermal growth factor receptor gene in uncultured primary human breast carcinoma.

We analyzed the epidermal growth factor receptor gene using a complementary DNA probe of the epidermal growth factor receptor gene in 21 uncultured primary breast carcinomas and found that the gene was amplified in three of these tumors. We further demonstrated by immunohistochemistry using a monoclonal antibody to the epidermal growth factor receptor that the receptor protein product of this gene was overexpressed and displayed elevated kinase activity. Our data indicate that one of the molecular mechanisms for overexpression of epidermal growth factor receptor in human breast cancer is epidermal growth factor receptor gene amplification without rearrangement in a subset of tumors.

Frequent loss of heterozygosity on chromosomes 6q, 11, and 17 in human ovarian carcinomas.

Recently, tumor-specific allele loss has been shown to be an important characteristic of some tumors. When such loss includes one or more growth-regulatory genes, it may allow the expression of tumorigenicity. Using Southern blots, we analyzed normal and tumor DNA samples from 19 ovarian cancer patients, using a series of polymorphic DNA probes that map to a variety of chromosomal loci. Of 14 informative cases, tumor-specific allelic loss was observed in nine (64%) at the estrogen receptor (ESR) gene locus on chromosome 6q. On chromosome 17p at the D17S28 and D17S30 loci, allelic losses were also detected in 6 of 8 (75%) and 9 of 14 (64%) cases, respectively. Allelic loss at the HRAS1 gene locus on chromosome 11p occurred in 5 of 11 (46%) informative cases. The relatively high incidence of these allelic losses observed on chromosome 6q represents the first implication by molecular genetic analysis of this chromosomal region in a human malignancy, and it thus appears to be a genetic change specific to ovarian carcinoma. DNA sequence losses on 11p and 17p, also reported for other cancers, may reflect the presence of tumor- or growth-suppressor genes on these chromosomes that are important in the genesis of many tumor types, including ovarian malignancies.

Allele loss at the c-Ha-ras1 locus in human ovarian cancer.

Recent reports have shown allele loss at the c-Ha-ras1 locus on the short arm of chromosome 11 in some types of tumors. To determine whether loss of heterozygosity occurs at the c-Ha-ras1 locus in uncultured human ovarian carcinomas we used Southern blot analysis to study DNA from 17 pairs of ovarian tumors and matched white blood cell samples from the same patients. In one of these 17 tumors, the c-Ha-ras1 locus was rearranged, and in five tumor DNAs from ten informative patients, a c-Ha-ras1 allele was lost. This loss, of relatively high incidence, appears to be an important characteristic of human ovarian cancer and may provide a useful tool for understanding its biological behavior.

Multiple restriction fragment length polymorphisms of the human epidermal growth factor receptor gene.

We have examined the epidermal growth factor (EGF) receptor gene for structural alterations in fresh human tumors. DNA samples from 92 patients with solid tumors (lung cancer, 37; breast cancer, 24; head and neck cancer, 17; other tumors, 14) were analyzed and compared with those from 22 leukemia patients and 14 individuals without malignant neoplasms. When DNA samples were digested with HindIII restriction endonuclease, Southern blot analysis demonstrated 3 distinct polymorphic bands (9.8, 11, and 12 kilobases) after hybridization to the HER-A64-1 probe and another 2 distinct polymorphic bands (4.9 and 5.2 kilobases) after hybridization to the HER-A64-3 probe. Pedigree analysis of 43 members of a single family and comparative analysis of tumor and normal DNA samples from the same patients demonstrated that the variations in fragment size observed were due to 2 independent restriction fragment length polymorphisms in the region of the EGF receptor gene. Amplification of the EGF receptor gene was detected in 3 cases of breast cancer, but not in other tumors studied. We conclude that the human EGF receptor gene has multiple restriction fragment length polymorphisms and that in fresh human tumor samples rearrangement and amplification of the gene occur infrequently, if ever, within the region encompassed by the 2 complementary DNA probes used.

c-erbB-2 amplification in node-negative human breast cancer.

c-erbB-2 gene analysis by Southern and DNA dot blot methods was done in 66 tumor samples from patients with histologically node-negative breast cancer. The c-erbB-2 gene was amplified 2- to greater than 8-fold in 13 tumors (20%). None of 59 tumors that were examined by the Southern method showed c-erbB-2 gene rearrangement. c-erbB-2 amplification was analyzed in relation to other prognostic factors. The c-erbB-2 gene was amplified in five of 36 (14%) diploid and eight of 30 (27%) aneuploid tumors. Thirteen of 54 (24%) tumors with nuclear Grade 1 or 2 displayed c-erbB-2 amplification, whereas none of 12 tumors with nuclear Grade 3 did. No correlation was observed with estrogen receptor content, tumor size, histological type, or age of patients. The median follow-up date for these patients was 85+ mo. Of 13 patients whose tumors showed c-erbB-2 amplification, six patients (46%) developed recurrence, and five patients (38%) died of metastatic disease. In contrast, of 53 patients whose tumors did not show c-erbB-2 amplification, 15 patients (28%) developed recurrence, and seven patients (13%) died of disease. In conclusion, our results show that c-erbB-2 gene amplification was more frequent in aneuploid tumors and tumors with poor nuclear grade. c-erbB-2 amplification may be considered a possible prognostic factor in node-negative breast cancer.

Transforming growth factor-alpha production and autoinduction in a colorectal carcinoma cell line (DiFi) with an amplified epidermal growth factor receptor gene.

The DiFi colorectal carcinoma cell line, derived from a patient with familial adenomatous polyposis, was examined for gene expression and production of the autocrine growth factor transforming growth factor alpha (TGF-alpha) and for epidermal growth factor receptor (EGFR) gene expression and gene copy number. DiFi cells expressed TGF-alpha transcripts as identified on Northern (RNA) blots. Addition of TGF-alpha (10 ng/ml) or EGF (10 ng/ml) to DiFi cell cultures (lacking EGF or serum) up-regulated DiFi cell basal TGF-alpha mRNA levels, suggesting that autoinduction of TGF-alpha occurs in these cells. DiFi cell cultures in log phase growth secreted measurable amounts of TGF-alpha (347 pg/10(6) cells/24 h) into their culture medium, as determined by radioimmunoassay. DiFi cells showed strong overexpression of the EGFR gene on Northern blots relative to three other colon cancer cell lines examined. Immunoperoxidase staining showed enhanced EGFR expression in a cell subpopulation among the original (uncultured) ascitic fluid cells from which the DiFi cell line was established. This cell subpopulation was observed to expand dramatically between passages 1 and 25. Immune complex kinase assay of DiFi cells showed that EGFR were functional as determined by their ability to autophosphorylate. The EGFR gene in these cells was not found to be rearranged or genetically altered using Southern blot analysis. Dot blot analysis of DiFi cell DNA revealed EGFR gene amplification in the range of 60-80 copies/cell, which is approximately twice the copy number seen in A-431 epidermoid carcinoma cells. To our knowledge DiFi cells represent the first example of EGFR gene amplification in a colorectal adenocarcinoma. Because DiFi colorectal cancer cells uniquely show production and auto-induction of TGF-alpha in addition to amplification and overexpression of the EGFR gene, these cells represent a valuable tool for studying the role(s) of the EGFR in the regulation of tumor cell growth.

Alternative 5' end of the bcr-abl transcript in chronic myelogenous leukemia.

Philadelphia chromosome positive acute lymphocytic leukemia and chronic myelogenous leukemia are strongly associated with two distinct forms of bcr-abl chimeric protein, known as P190 and P210, respectively. By studying cDNA clones obtained from the cell line KBM-5, we identified two new bcr-abl transcripts. These are formed by alternative splicing of at least two exons to the known bcr exon 2. One novel transcript can encode a protein kinase of approximately 190 kd, while the other can direct the synthesis of a larger protein whose amino terminus remains to be defined. The alternative exons can be spliced also to the two normal bcr transcripts, reflecting the activation of a cryptic promoter. These messages were present at low abundance in two cases of blastic crisis but were not detected in the chronic phase. It is conceivable that the proteins encoded by the new bcr-abl mRNAs are involved in the transformation to the acute phase in some cases of chronic myelogenous leukemia.

SIS/PDGF-B expression in benign and malignant human breast lesions.

Northern blots with poly(A)-selected RNAs were prepared from four primary human breast carcinomas and hybridized with a SIS/PDGF-B gene probe. High expression of a normal-sized 3.7 kb SIS/PDGF-B transcript was observed in three samples. These three carcinomas were all T3 (greater than 5 cm) lesions and were histologically infiltrating ductal carcinomas accompanied by moderate to marked desmoplasia. Furthermore in situ hybridization with photobiotinylated probes was performed and demonstrated that the SIS/PDGF-B expression was localized within the epithelial cells of malignant as well as benign lesions. It is possible that SIS/PDGFB expression within the epithelial cell components of nonmalignant as well as malignant breast lesions may be in part responsible for the stromal reaction.

A phase I trial of intravenously-administered recombinant tumor necrosis factor-alpha in cancer patients.

We report a phase I clinical investigation of 30-minute and four-hour intravenous (IV) infusions of recombinant tumor necrosis factor (rTNF)-alpha. Thirty-nine patients with disseminated cancer received escalating doses of rTNF-alpha for five consecutive days every 2 weeks for a total duration of 8 weeks. Dose escalations followed a modified Fibonacci scale with a minimum of four patients entered at each dose level: 5, 10, 25, 50, 75, 100, 150, 200, and 250 micrograms/m2/d. Toxicities consisted mainly of constitutional symptoms including fever, chills, headache, and fatigue, increasing in severity with dose escalation. No significant differences in dose-limiting toxicities were seen between the two rates of IV infusion. The maximum tolerated dose (MTD) was 200 micrograms/m2 with dose limiting toxicity being constitutional symptoms and hypotension. Hematologic changes included median decrease in both granulocyte and platelet counts of 38% and 41%, respectively (range, 16% to 85%), although clinically significant granulocytopenia and thrombocytopenia were not observed. Hematological parameters returned to baseline within 72 hours after rTNF-alpha was stopped. rTNF-alpha induced changes in lipid metabolism were manifested by median fasting triglyceride elevations above baseline (median, 103 micrograms/dL) of 157% (range, 16% to 389%) after five days of therapy with doses greater than 75 micrograms/m2, associated with a median increase in very-low-density lipoprotein (VLDL) of 80%. Serum rTNF peak levels exceeding 10 ng/mL were observed 30 minutes following rTNF-alpha infusions at MTD dose. Twelve of 34 patients had no change in their evaluable disease for a median duration of 18 weeks (range, 8 to 30 weeks), and 22 patients showed progressive disease. This study forms the framework for phase II trials of IV administered rTNF-alpha.

Philadelphia-negative chronic myelogenous leukemia with breakpoint cluster region rearrangement: molecular analysis, clinical characteristics, and response to therapy.

We have detected rearrangement in the breakpoint cluster region (bcr) on chromosome 22 in cells derived from seven chronic myelogenous leukemia (CML) patients who had no cytogenetic evidence of a chromosome abnormality. These Philadelphia (Ph)-negative, bcr rearrangement-positive CML patients had clinical features and laboratory parameters that bore a strong resemblance to those of Ph-positive CML; all patients have shown a favorable response to hydroxyurea, busulphan, or alpha interferon (IFN-alpha) therapy. In one patient, because of the deletion of distal 3' sequences, detection of bcr rearrangement required a large probe that recognized proximal 5' sequences. Cells obtained from five patients were studied by Northern blotting and showed an aberrant 8 kilobase (kb) mRNA indistinguishable from the bcr-abl transcript that is felt to be a pathogenetic factor in Ph-positive CML. In three patients with a normal karyotype, bcr rearrangement was detected at the time of hematologic remission, and represented the only evidence for persistent malignancy. Our results suggest that: (1) the presence of bcr rearrangement in CML is associated with clinical features of Ph-positive disease, even in the absence of the Ph chromosome; (2) deletions occur within bcr and necessitate the use of probes covering both 5' and 3' DNA segments for accurate diagnosis; (3) molecular analysis may provide a useful approach to the follow-up of leukemia therapy in some patients; and (4) these patients respond to hydroxyurea, busulphan, and IFN-alpha therapy.

Suppression of c-myc and c-myb expression in myeloid cell lines treated with recombinant tumor necrosis factor-alpha.

Proto-oncogenes are thought to be involved in cellular differentiation and proliferation. Tumor necrosis factors (TNFs) are specific cytokines that have cytostatic and cytotoxic effects in vitro against a wide range of human tumor cells. We have previously demonstrated that recombinant TNFs (rTNFs) have an antiproliferative effect on certain human leukemic cell lines (HL-60, KBM3, KBM5) and no effect on others (K562). To study the possible role of the c-myc and c-myb oncogenes in this antiproliferative effect of TNF, we examined their expression in cell lines HL-60, KBM3, KBM5, and K562 before and after incubation with rTNF-alpha. Expression of c-myc and c-myb was elevated in all cell lines prior to incubation with rTNF-alpha. In the sensitive cell lines HL-60, KBM5, and KBM3 expression of c-myc and c-myb decreased rapidly 8-, 16-, and 4-fold, respectively, by 24 hr. K562 cells, insensitive to rTNF-alpha, exhibited no change in c-myc or c-myb expression over 24 hr. These studies demonstrated that down-regulation of c-myc and c-myb expression were associated with antiproliferative effects of rTNF-alpha on these cell lines.

Proto-oncogene expression in human normal bone marrow.

We and other investigators have previously reported our findings on oncogene expression in human leukemia in an attempt to study the possible involvement of these genes in the leukemic state. An important shortcoming of these studies has been the lack of information on the expression of these genes in normal hematopoietic cells. To address this question we analyzed both the transcript size and level of expression of six oncogenes in fresh hematopoietic cells obtained from hematologically normal individuals and compared the results to those found in fresh samples obtained from patients with various forms of leukemia (acute myelogenous leukemia, acute lymphocytic leukemia, and chronic myelogenous leukemia). We found low level expression of c-myc, c-myb, c-fes, and c-raf in normal bone marrow in sharp contrast to the high levels of expression found in some forms of leukemia. C-fos was highly expressed in both normal bone marrow and certain leukemias. We were unable to detect c-sis expression in our normal samples. With the exception of c-fes, there was no variation in transcript size when comparing normal and leukemic samples. Having defined the transcript sizes and levels of expression for these proto-oncogenes in normal hematopoietic cells, we know that aberrant transcript size for the genes we have studied is not a common event in leukemias. The levels of expression, however, vary widely between normal hematopoietic cells and leukemia as well as between different types of leukemia.

Bcl-1 gene rearrangements in B cell lymphoma.

We analyzed 50 B cell lymphoma samples by Southern blot analysis, using the bcl-1 and heavy chain immunoglobulin (JH) probes with two or more restriction endonucleases. All samples showed JH rearrangement, and three samples (two diffuse small lymphocytic lymphomas and one diffuse large cell lymphoma probably transformed from a diffuse small lymphocytic lymphoma) demonstrated rearranged bcl-1 sequences. The three samples showed the t(11;14)(q13;q32) chromosome translocation, and all three contained rearranged JH fragments that comigrated with the rearranged bcl-1 fragment. The breakpoint of the translocation occurred within a 1.6-kb region on chromosome 11 in the three cases. Two of the three patients had primary refractory disease. Two of the three patients had gastrointestinal involvement. Bcl-1 rearrangement may identify an unusual subset of patients with primary refractory disease with gastrointestinal involvement. It may also describe a unique subset of large cell lymphoma patients transformed from diffuse small cell histology.

Clinical evaluation of a DNA probe assay for the Philadelphia (Ph1) translocation in chronic myelogenous leukemia.

We report the clinical evaluation of an improved DNA probe assay for the characteristic genetic marker of human CML, observed by cytogenetics and designated the Philadelphia chromosome (Ph1). The Ph1 chromosome results from the fusion of c-abl proto-oncogene sequences from chromosome 9 to phl gene sequence on chromosome 22. (The phl gene is often referred to as bcr. However, for clarity we prefer to reserve the designation "bcr" for the region within the phl gene in which translocation breakpoints have been found to occur. We also find it useful to distinguish between two such regions in phl, bcr-210 and bcr-190, named after the 210- and 190-kDa phl/abl fusion proteins resulting from translocations with breakpoints in the respective regions. We refer to the corresponding chromosomal translocations as Ph1(bcr-210) and Ph1(bcr-190).) DNA, extracted from peripheral blood (PB) or bone marrow (BM) and digested with restriction endonuclease BglII, is hybridized with a probe (phl/bcr-3) spanning a breakpoint cluster region within phl. Rearrangements are revealed by the presence of one or two novel junction fragments. Clinical specimens from leukemic patients with active disease were compared by cytogenetic and DNA probe analysis at seven centers in the United States and Europe. The probe assay identified the phl rearrangement in 190 of 191 cases of Ph1-positive CML, as well as in 12 of 27 clinically diagnosed CML specimens lacking a typical Ph1 chromosome. DNA rearrangements also were seen in two of six cases of Ph1-positive ALL. No false positive results were obtained among 93 non-leukemic controls. Mixing experiments showed that the DNA probe assay can detect as few as 1% leukemic cells in a specimen. A preliminary study of CML patients in remission after allogeneic BM transplantation revealed a small fraction of residual Ph1-positive leukemic cells in a significant number of such patients.

Effect of retinoids on the release and gene expression of tumor necrosis factor-alpha in human peripheral blood monocytes.

The effect of retinoic acid (RA) and retinol (ROH) on the release of tumor necrosis factor (TNF) by human peripheral blood monocytes (HPBM) was determined. HPBM were cultured for various periods of time in either 5% complete (cAB) or delipidized (DLS) AB serum. TNF release (L929 cytolytic assay) in the presence of cAB occurred during the first 3 days of in vitro culture. Delipidization of AB serum completely inhibited the lipopolysaccharide (LPS)-induced release of TNF by HPBM. Addition of RA (0.5 microM) to DLS restored LPS-induced TNF release by HPBM, and supplementation with ROH (1.0 microM) resulted in release of TNF-like activity, but only after 3 days of in vitro culture. The maintenance of TNF release by the addition of exogenous RA after 3 days of in vitro culture suggested that depletion of endogenous RA was partially responsible for loss of TNF-like activity. The levels of endogenous TNF protein and mRNA were not influenced by delipidization of serum and were found to be similar to those of HPBM cultured in the presence of AB serum. TNF protein and mRNA were undetectable in HPBM ROH-treated cell lysates, although cytolytic activity was observed in culture supernatants. These results suggest that retinoids are required for the release of cytolytic factors from HPBM and that non-TNF cytolytic factors may be released by these cells at different stages of maturation.

Lithium chloride stimulates human monocytes to secrete tumor necrosis factor/cachectin.

The purpose of this study was to examine the effect of lithium chloride (LiCl) on human monocytes. Patients undergoing lithium therapy have elevated white blood cell counts. Since both tumor necrosis factor alpha (TNF alpha) and interleukin 1 (IL-1), which are secreted by monocytes, can stimulate endothelial cells to produce granulocyte-macrophage colony-stimulating factor (GM-CSF), we determined whether lithium-stimulated monocytes produced TNF alpha and/or IL-1. Normal human monocytes were incubated for 24 h with medium (negative control), lipopolysaccharide (positive control), or LiCl (0.05-50 mM). The supernatants were removed and assayed for IL-1 and TNF alpha secretion using the D10.G4.01 and L929 assays, respectively. Lithium did not stimulate IL-1 secretion but did stimulate TNF alpha secretion (5-10 U/ml of TNF alpha per 2 x 10(5) monocytes). The increased secretion of TNF alpha was associated with a fourfold increase in TNF alpha mRNA. TNF alpha activity in the supernatants was neutralized by a monoclonal antibody against human TNF alpha but not by antibody against human albumin. Other alkali metals such as rubidium and cesium did not stimulate monocytes to secrete TNF alpha. These data indicate that one mechanism by which Li may cause granulocytosis is through a transcriptional enhancement of TNF production and subsequent secretion by monocytes.

Clonal nature of Philadelphia chromosome-positive and -negative chronic myelogenous leukemia by DNA hybridization analyses.

Evidence for the clonal nature of chronic myelogenous leukemia (CML) has been obtained primarily from studies of black females expressing polymorphic glucose-6-phosphate dehydrogenase (G6PD) isoenzymes where, instead of the heterozygous pattern normally found as a result of random X chromosome inactivation, exclusive expression of only one G6PD allele has been demonstrated in leukemic cell populations. We report here the use of two other molecular approaches to examine clonality of peripheral blood cells in patients with CML. The first of these is based on the analysis of consistent differential methylation patterns associated with active and inactive X chromosomes within the region spanned by a BamHI restriction fragment length polymorphism (RFLP) at the hypoxanthine phosphoribosyltransferase (HPRT) locus. By this method, three heterozygous females gave results consistent with monoclonal origin of the disease, including one patient lacking the Philadelphia chromosome (Ph1) normally associated with CML. In the other two patients, both of whom had Ph1-positive CML, clonality was confirmed by the demonstration of simple gene rearrangements by Southern hybridization with a breakpoint cluster region (bcr) probe from chromosome 22.