Deletion of IRF-1, mapping to chromosome 5q31.1, in human leukemia and preleukemic myelodysplasia.

One of the most frequent cytogenetic abnormalities in human leukemia and myelodysplasia is an interstitial deletion within chromosome 5q. A tumor suppressor gene has been hypothesized to lie in 5q31, the smallest commonly deleted region. IRF-1, a gene whose product manifests anti-oncogenic activity, was mapped to 5q31.1. IRF-1 lies between IL-5 and CDC25C and is centromeric to IL-3 and GM-CSF. Among these genes, only IRF-1 was consistently deleted at one or both alleles in 13 cases of leukemia or myelodysplasia with aberrations of 5q31. Inactivating rearrangements of one IRF-1 allele, accompanied by deletion of the second allele, were also identified in one case of acute leukemia. Thus, IRF-1 may be a critically deleted gene in human leukemia and myelodysplasia.

The human Pim-1 gene is selectively transcribed in different hemato-lymphoid cell lines in spite of a G + C-rich housekeeping promoter.

The expression of the Pim-1 proto-oncogene was studied by using the K562, Daudi, and Jurkat cell lines. In K562, Pim-1 mRNA levels were more than 20-fold higher than in Daudi and 50-fold higher than in Jurkat. Nuclear run-on assay data correlated directly with the steady-state mRNA levels, suggesting that the rate of transcription was responsible for the selective expression of this gene. Furthermore, the half-life of Pim-1 mRNA was shown to be 47 min in K562, 71 min in Daudi, and 35 min in Jurkat. This indicated that selective Pim-1 mRNA expression did not depend on posttranscriptional regulation. Therefore, 1.7 kilobases of the Pim-1 promoter was sequenced and studied in detail. The sequence showed that the region from nucleotide -1 to -873 was G + C rich (71%). Study of promoter deletions defined two major functional regions, a proximal element (nucleotide -104 to -1) and a distal element (nucleotide -427 to -336). DNase I protection assays identified binding sites for the Sp1 and AP2 proteins in these elements. A possible new transcription factor binds at position -348 in the distal element. In our study of the 1.7-kilobase Pim-1 promoter, we found no differences between K562 and Jurkat that could explain large differences in transcription. Therefore, the Pim-1 promoter appears to function constitutively, and we conclude that distant elements must regulate the tissue-selective expression of this gene. Although the Pim-1 gene has a G + C-rich housekeeping promoter, expression is carefully regulated at the level of transcription.

Characterization of a candidate bcl-1 gene.

The t(11;14)(q13;q32) translocation has been associated with human B-lymphocytic malignancy. Several examples of this translocation have been cloned, documenting that this abnormality joins the immunoglobulin heavy-chain gene to the bcl-1 locus on chromosome 11. However, the identification of the bcl-1 gene, a putative dominant oncogene, has been elusive. In this work, we have isolated genomic clones covering 120 kb of the bcl-1 locus. Probes from the region of an HpaII-tiny-fragment island identified a candidate bcl-1 gene. cDNAs representing the bcl-1 mRNA were cloned from three cell lines, two with the translocation. The deduced amino acid sequence from these clones showed bcl-1 to be a member of the cyclin gene family. In addition, our analysis of expression of bcl-1 in an extensive panel of human cell lines showed it to be widely expressed except in lymphoid or myeloid lineages. This observation may provide a molecular basis for distinct modes of cell cycle control in different mammalian tissues. Activation of the bcl-1 gene may be oncogenic by directly altering progression through the cell cycle.

Amplification and expression of the bcl-1 gene in human solid tumor cell lines.

The bcl-1 gene maps to chromosome 11q13 and has recently been shown to be a member of the cyclin gene family. Amplification of the chromosome region containing bcl-1 occurs frequently in breast cancer, squamous cell cancer, and other tumor types. We have hypothesized that amplification results in altered expression of the bcl-1 gene, contributing to carcinogenesis. In this work, we studied bcl-1 gene amplification and expression in a panel of human cell line. bcl-1 is expressed in all cell lines studied. The level of expression tends to be higher in amplified cell lines. We also screened these cell lines for int-2 and hst-1 expression, genes which are frequently coamplified with bcl-1. No int-2 expression was detected, and the two cell lines expressing hst-1 were unamplified. Our data provide support for the importance of bcl-1 in carcinogenesis.

Chromosome t(11;14)(q13;q32) breakpoints in centrocytic lymphoma are highly localized at the bcl-1 major translocation cluster.

Chromosome 11q13 translocation breakpoints have been found dispersed over more than 100 kb of genomic DNA in centrocytic lymphoma and lymphocytic lymphoma of intermediate differentiation (mantle cell lymphoma). Approximately one-half of these translocations occur at the bcl-1 major translocation cluster (MTC) and appear tightly clustered by Southern blot restriction mapping. In order to specifically characterize these t(11;14)(q13;q32) translocations, six cases of centrocytic lymphoma with MTC rearrangements on Southern blot were studied. Genomic DNA was amplified by PCR (polymerase chain reaction) using a consensus immunoglobulin heavy-chain joining gene (JH) primer and two separate MTC primers. Sequencing of the PCR products revealed the MTC breakpoints in all six cases to be clustered within a 61 basepair span, very similar to those previously reported in t(11;14)-containing human B-cell lines. No two breakpoints were identical. JH breakpoints involved JH4 in five cases and JH6 in the other. Tight clustering of the MTC breakpoints thus permits PCR detection of the t(11;14) in many cases of centrocytic lymphoma, which may be of value in classifying this subtype of non-Hodgkin's lymphoma.

Cloning of the t(11;14)(q13;q32) translocation breakpoints from two human leukemia cell lines.

The t(11;14)(q13;q32) translocation has been associated with several subtypes of human leukemia and lymphoma. It has been proposed that this translocation activates a proto-oncogene designated BCL1. In an effort to better understand the mechanism by which this translocation leads to malignancy, we have studied this translocation in two human cell lines. MO1094 and MO2058 were derived from patients with prolymphocytic variants of chronic lymphocytic leukemia. Southern blotting of the MO2058 cell line documented that the translocation linked the Jh region in the immunoglobulin heavy chain gene to the previously described BCL1 major translocation cluster (MTC). Using the polymerase chain reaction, we cloned this translocation and showed that the chromosome 11 breakpoint was within 7 bp of two other samples reported previously. Southern blotting of the MO1094 cell line suggested that the translocation in this cell line might link Jh sequences to a new region in the BCL1 locus on chromosome 11. Therefore, the MO1094 breakpoint was cloned from a genomic library. Comparison with normal cloned DNA from the BCL1 locus showed that the chromosome 11 breakpoint occurred 24 kb telomeric of the MTC. This work reinforces the concept that translocation breakpoints in the BCL1 locus are scattered over at least 63 kb.

Adenoviral vectors efficiently target cell lines derived from selected lymphocytic malignancies, including anaplastic large cell lymphoma and Hodgkin's disease.

We have hypothesized that adenoviral vectors might mediate gene transfer into cell lines derived from human lymphocytic malignancies, such as lymphoma, lymphocytic leukemia, and myeloma. A panel of 33 cell lines was studied for their ability to be transduced by an adenoviral (AD) vector encoding the Escherichia coli beta-galactosidase gene (AD-betagal). A cytochemical assay and a flow cytometry assay both demonstrated that a subset of lymphocytic cell lines can be efficiently transduced by adenoviral vectors. In particular, three of three anaplastic large cell lymphoma lines, two of two Hodgkin's disease cell lines, two of seven Burkitt's lymphoma cell lines, and three of five myeloma cell lines exhibited efficient gene transfer. The ability of an AD vector expressing the thymidine kinase (tk) gene from herpes simplex virus-1 (AD-tk) followed by ganciclovir (GCV) to kill 11 of these lymphocytic cell lines was studied. In eight of the cell lines tested, more than 68% of the cells were killed by AD-tk/GCV. Similar results were obtained using an adenoviral vector expressing the wild-type p53 tumor suppressor gene (AD-p53). Thus, AD-tk/GCV and AD-p53 both demonstrated efficient killing of these cell lines. These data document that adenoviral vectors are valuable reagents for the introduction of genes into selected lymphocytic cell lines. These data also suggest that adenoviral vectors might be useful for gene therapy of subsets of lymphocytic malignancy.

Evidence for molecular subtypes of HIV-associated lymphoma: division into peripheral monoclonal, polyclonal and central nervous system lymphoma.

The pathogenesis of the HIV-associated lymphomas is not well understood. In order to begin characterizing this class of lymphoma, we initiated a molecular genetic study of DNA extracted from 31 diagnostic biopsy specimens from patients diagnosed with AIDS-associated non-Hodgkin's lymphoma. Analysis of 25 peripheral lymphomas showed that 14 were monoclonal B-cell processes, while 11 appeared to be of polyclonal origin. Five of the 14 monoclonal lymphomas were found to have rearrangements of the c-myc gene. Epstein-Barr virus (EBV) genomes were found in seven out of 14 monoclonal samples, but only two out of nine polyclonal samples. The six primary central nervous system (CNS) lymphoma samples were more homogeneous than the peripheral samples and all were monoclonal, positive for EBV and lacked detectable c-myc gene rearrangements. This study allows us to subdivide the HIV-associated lymphomas into three major molecular subtypes: (1) monoclonal B-cell process frequently associated with c-myc rearrangement or detectable EBV genomes, (2) polyclonal B-cell process typically without evidence of EBV, and (3) monoclonal primary CNS process associated with EBV genomes and lacking detectable c-myc rearrangement.

AIDS-associated polyclonal lymphoma: identification of a new HIV-associated disease process.

High-grade non-Hodgkins B-cell lymphoma is one of the principle malignancies that occurs in individuals infected with the human immunodeficiency virus (HIV-1). Immunoblastic lymphomas that arise in immunosuppressed transplant patients have been described as both monoclonal and polyclonal, and occur in association with Epstein-Barr virus (EBV) infection. To test whether polyclonal lymphoma occurred in patients with AIDS we studied tumors from multiple sites in three patients who died with widespread AIDS-associated large cell or large cell immunoblastic lymphoma. All biopsy specimens contained invasive lymphoma. Tumor cells were mature IgM-positive immunoblasts by immunohistochemical analysis, with the same B-cell phenotype observed in all tumor sites. Only a minority of sites from all patients analyzed were monoclonal as measured by immunoglobulin gene rearrangements, with one case having several foci of monoclonal disease with other histologically identical metastases showing no evidence of monoclonal proliferation. Similar to the transplant-associated polyclonal B-cell proliferations. EBV gene sequences were present in multiple sites from one autopsy. In the other two autopsies, polyclonal B-cell proliferations occurred in the absence of EBV involvement except at one site, where a minor clone of EBV-infected cells was found. In contrast to HIV-associated Burkitt's lymphoma, no c-myc rearrangements were found at any site. These studies describe the occurrence of polyclonal lymphoma in AIDS and suggest that EBV-negative polyclonal lymphoma may be a distinct disease entity unique to HIV-infected individuals.

Selective elimination (purging) of contaminating malignant cells from hematopoietic stem cell autografts using recombinant adenovirus.

In this work we have explored the use of adenoviral vectors for the purging of cancer cells from hematopoietic stem cell (HSC) autografts. We showed that a recombinant adenovirus expressing the herpes simplex-1 thymidine kinase gene (AD-tk) plus ganciclovir (GCV) killed HELA cells more effectively than did GCV alone. HELA cells were then mixed with human HSCs and exposed to AD-tk/GCV. AD-tk/GCV reduced the number of HELA colonies to 4% of control values, with no detectable reduction in the hematopoietic progenitor, colony forming unit-granulocyte/monocyte (CFU-GM). Similar studies of the JB6 non-Hodgkins lymphoma cell line showed a reduction to 5% of controls; studies of MCF-7, a breast carcinoma cell line, showed a reduction to 30% of controls, with no CFU-GM toxicity. Thus, AD-tk mediated selective killing of contaminating tumor cells. We also evaluated a recombinant adenovirus encoding the tumor suppressor gene p53 (AD-p53). AD-p53 was able to selectively kill all three cell lines (reducing tumor colonies approximately 100-fold) without any toxicity to CFU-GM. Although both AD-tk/GCV and AD-p53 were effective in these experiments, AD-p53 seemed to be more potent. Adenoviral vectors show promise for selectively targeting cancer cells that contaminate HSC autografts.

Primary central nervous system lymphomas: natural history and response to radiation therapy in 55 patients with acquired immunodeficiency syndrome.

The incidence of primary central nervous system (CNS) lymphoma has increased rapidly in patients with acquired immunodeficiency syndrome (AIDS) and is predicted to exceed 1800 cases annually by 1991. To characterize the natural history and response to radiation therapy (RT) of these lesions, the authors have reviewed the clinical histories of 55 AIDS patients with biopsy-proven primary CNS lymphomas. The tumors responded both clinically and radiologically to whole-brain RT consisting of 4000 rad in 267-rad fractions over 3 weeks or an equivalent neuroret dose. The mean duration of survival from the appearance of symptoms consistent with the mass lesion was significantly greater in patients who received RT than in those who did not (42 vs. 134 days, p less than 0.5; median 27 vs. 119 days). Autopsy findings showed that patients who did not receive RT died from tumor progression, whereas those who completed RT died of opportunistic infections. Patients with AIDS who are suspected of having primary CNS lymphoma should therefore immediately undergo biopsy and, if the diagnosis is confirmed, whole-brain RT. With early diagnosis and treatment, these tumors respond to, and patients benefit from, RT. Survival of such patients may in future be prolonged by more effective treatments for systemic opportunistic infections.

A unique human B lymphocyte antigen defined by a monoclonal antibody.

We produced a hybridoma designated 4G7 from a mouse immunized with chronic lymphocytic leukemia cells. The 4G7 hybridoma secretes an IgG1 antibody that is specific for normal and malignant B lymphocytes. Using dual color immunofluorescence staining, this antibody reacted with all immunoglobulin-positive cells but no T cells in normal peripheral blood. There was no detectable 4G7 antigen on monocytes, platelets, red cells, granulocytes, or phytohemagglutinin-activated T cells. When PBL were depleted of 4G7 positive cells and stimulated with pokeweed mitogen, secreted immunoglobulin levels fell to less than 10% of control values on Day 5 and less than 1% of control on Day 7. This antibody was reactive with 155 of 176 B lineage neoplasms on which it was screened. Thirty-five cases of myeloid or T-lymphoid malignancy were negative. Our studies show that the 4G7 antigen modulates in the presence of excess antibody. Free 4G7 antigen was not found circulating in human serum. The cell surface antigen identified by 4G7 was sensitive to pronase proteolysis but resistant to trypsin and chymotrypsin digestion. A comparison of 4G7 with other known B-cell antibodies indicates that the 4G7 antigen has not been previously identified. This antibody is of use for the identification of normal B lymphocytes, the study of B-cell differentiation, and the characterization of lymphoid malignancies.

The t(5;14) chromosomal translocation in a case of acute lymphocytic leukemia joins the interleukin-3 gene to the immunoglobulin heavy chain gene.

Chromosomal translocations have proven to be important markers of the genetic abnormalities central to the pathogenesis of cancer. By cloning chromosomal breakpoints one can identify activated proto-oncogenes. We have studied a case of B-lineage acute lymphocytic leukemia (ALL) that was associated with peripheral blood eosinophilia. The chromosomal translocation t(5;14) (q31;q32) from this sample was cloned and studied at the molecular level. This translocation joined the immunoglobulin heavy chain joining (Jh) region to the promotor region of the interleukin-3 (IL-3) gene in opposite transcriptional orientations. The data suggest that activation of the IL-3 gene by the enhancer of the immunoglobulin heavy chain gene may play a central role in the pathogenesis of this leukemia and the associated eosinophilia.

Rearrangement of the chromosome 11 bcl-1 locus in centrocytic lymphoma: analysis with multiple breakpoint probes.

Centrocytic lymphoma is a B-cell non-Hodgkin's lymphoma (NHL) composed of lymphocytes resembling cleaved follicular center cells (centrocytes). Previous studies have suggested an association between t(11;14) chromosomal translocations and bcl-1 rearrangement in centrocytic and related intermediate lymphocytic lymphomas. To further characterize the association between bcl-1 and centrocytic lymphoma, Southern blot analysis was performed on samples from 23 patients using four separate bcl-1 breakpoint probes spanning 63 kb of the chromosome 11 bcl-1 locus. Rearrangements were identified in six patients with the major translocation cluster (MTC) probe and in another six with probe p94PS, located about 24 kb 5' of MTC. Eleven of these 12 cases showed comigration of rearranged bcl-1 and Ig heavy chain-joining genes, consistent with the t(11;14) chromosomal translocation. No rearrangements were observed with the bcl-1 locus probes p210 or p11EH located 5' of p94PS, nor with bcl-2 or c-myc oncogene probes. No bcl-1 rearrangements were identified in B-cell follicular NHL (15), small noncleaved cell (Burkitt's and non-Burkitt's) NHL (8), T-cell NHL (4), multiple myeloma (14), and pre-B-cell acute lymphoblastic leukemia (9). One of 23 B-cell NHL of large cell type and one of 19 chronic lymphocytic leukemias or small lymphocytic NHL had MTC rearrangement. Thus, bcl-1 rearrangement occurred at MTC or p94PS in 12 of 23 centrocytic lymphomas (52%), confirming a nonrandom association and suggesting a pathogenetic role for the bcl-1 locus in this immunohistologic subtype of NHL.

Monoclonal antibody therapy of lymphoid malignancy.

Monoclonal antibodies which bind to tumour cell surface antigens have produced regressions of malignancies in an increasing number of clinical trials. The largest experience to date is in the treatment of refractory B and T lymphoid tumours using a variety of intravenously administered mouse monoclonal antibodies. Treatment with antibodies against common differentiation antigens or very specific anti-idiotype antibodies has been effective in both cases. Toxicity has been acceptably low. A number of problems which limit the application and efficacy of monoclonal antibody therapy of lymphoid malignancy have been identified. Most prominent among these are tumour heterogeneity, which allows non-antibody binding subpopulations of the tumour to escape therapy, and the patient's immunological response to the monoclonal antibody-tumour cell complex. As more experience is accumulated, solutions to these problems will be found.

Anti-idiotypes in B-cell tumor therapy.

Thirteen patients with B-cell lymphomas were treated with mouse monoclonal anti-idiotype antibodies. All but 1 of the patients in this study had received extensive prior treatment with conventional therapy for lymphoma. The treatment protocol initially included an escalating dose schedule which was intended to help us evaluate toxicity and pharmacokinetics and, eventually, to achieve appreciable levels of free mouse antibody in the circulation. The last 4 patients received substantial initial doses. Tumor sampling was performed before and during therapy for evaluation of tissue penetration by antibody. Patients received antibodies of gamma 1, 2a, or 2b isotype. None of the patients had serum paraproteins by routine clinical testing, but 6 had an idiotype protein detectable by a sensitive immunoassay at levels greater than 1 microgram/ml, two of which were greater than 200 micrograms/ml. These levels were temporarily reduced by plasma-pheresis. However, the presence of serum idiotype increased the requirement for mouse antibody to achieve tumor penetration. Another obstacle to treatment was immune response to mouse Ig that occurred in 5 of the 13 patients. Once an immune response had begun, further infusions of antibody failed to reach the tumor or induce tumor regression and were associated with toxicity. Our initial patient remains in an unmaintained complete remission 50 months after receiving antibody. Six of 12 additional patients have had objective remissions which also were clinically significant. However, these remissions were not complete. This therapy shows promise as an alternative modality for the treatment of B-cell lymphoma. We will need further studies to determine the mechanisms of the antitumor effect and to improve the clinical results.

Cloning and characterization of the human PIM-1 gene: a putative oncogene related to the protein kinases.

The mouse PIM-1 gene has been implicated in the evolution of retrovirus-associated mouse lymphomas. We have initiated a study of the human PIM-1 gene because of its potential importance as a human oncogene. We have isolated genomic and cDNA clones for this gene and characterized this locus in detail. The predicted PIM-1 protein is 313 amino acids in length. It has homology to a number of the protein kinases but does not have a transmembrane region. The amino acid corresponding to tyrosine-416 of pp60v-src is a tyrosine (position 198), which is consistent with the hypothesis that PIM-1 is a tyrosine kinase rather than a serine-threonine kinase. The PIM-1 gene was found to have six exons and five introns derived from 5 kb of genomic DNA. The site of transcription initiation was localized by S1 nuclease protection studies which indicated that the mature PIM-1 mRNA was approximately 2.7 kb in length. The promotor of this gene had no TATA or CAAT box but did have multiple GC boxes (CCGCCC) that might bind the Sp1 protein. The PIM-1 gene was expressed in myeloid and B lymphoid cell lines, but not in T lymphoid and nonhemopoietic lines. This initial characterization of PIM-1 will allow us to define its role in normal and malignant hematolymphoid differentiation.

Activation of the interleukin-3 gene by chromosome translocation in acute lymphocytic leukemia with eosinophilia.

The t(5;14)(q31;q32) translocation from B-lineage acute lymphocytic leukemia with eosinophilia has been cloned from two leukemia samples. In both cases, this translocation joined the IgH gene and the interleukin-3 (IL-3) gene. In one patient, excess IL-3 mRNA was produced by the leukemic cells. In the second patient, serum IL-3 levels were measured and shown to correlate with disease activity. There was no evidence of excess granulocyte/macrophage colony stimulating factor (GM-CSF) or IL-5 expression. Our data support the formulation that this subtype of leukemia may arise in part because of a chromosome translocation that activates the IL-3 gene, resulting in autocrine and paracrine growth effects.

A unique antigen on mature B cells defined by a monoclonal antibody.

A novel 42,000 dalton antigen (MB-1) expressed by mature human B cells in blood and tonsil was identified and characterized by utilizing a hybridoma monoclonal antibody. A comparison of MB-1 with other known B cell antigens suggests that the MB-1 antigen has not been previously identified. From one-and two-color immunofluorescence studies, it appears that the MB-1 antigen is found on all normal immunoglobulin (Ig)-expressing cells, but not on T cells, thymocytes, granulocytes, or platelets. Studies of malignant B cell tumors reveal that the antigen is expressed by virtually all Ig-expressing B cell tumors but only 10% of SIg- B-lineage leukemias. Data from these studies suggest that the MB-1 antigen is expressed late in B cell ontogeny before the expression of SIg.

Lack of detectable somatic hypermutation in the V region of the Ig H chain gene of a human chronic B lymphocytic leukemia.

Monoclonal human B cell tumors are a model system for the study of somatic hypermutation of the Ig genes of humans. It was previously shown that a number of B cell lymphomas exhibited striking V region point mutation, hypothesized to result from the somatic hypermutation mechanism. In this study we have extended the analysis to chronic lymphocytic leukemia. We have cloned and sequenced the productive Vh representing five different cells from a monoclonal chronic lymphocytic leukemia. All five Vh sequences were identical. Therefore, the Vh region in this leukemia was not the subject of detectable somatic mutation. These data suggest that chronic lymphocytic leukemia might lack the mechanism for somatic hypermutation and represent a stage of normal B lymphocyte differentiation in which the somatic hypermutation mechanism is not active.