Role of accessory cells in B cell activation. III. Cellular analysis of primary immune response deficits in CBA/N mice: presence of an accessory cell-B cell interaction defect.

The effect of the X-linked CBA/N genetic defect on the ability of mice to generate primary responses to thymic-dependent and thymic-independent antigens was assessed by comparing the ability of abnormal (CBA/N x DBA/2)F1 male mice and normal (DBA/2 x CBA/N)F1 male mice to generate 2,4,6-trinitrophenyl (TNP)-specific plaque-forming cell responses to TNP-keyhole limpet hemocyanin (KLH), TNP-conjugated Ficoll (TNP-Ficoll), TNP-Brucella abortus (BA), and TNP-lipopolysaccharide (LPS). The reciprocal F1 combinations used in this study differ genetically only in the origin of their X chromosome, but differ immunologically in that (CBA/N x DBA/2)F1 male mice express all the CBA/N immune abnormalities, whereas (DBA/2 x CBA/N)F1 male mice are immunologically normal. Analysis of thymic-dependent responses to TNP-KLH revealed that abnormal F1 mice were capable of generating primary responses in vivo to high doses of TNP-KLH, but failed to generate responses to suboptimal doses of TNP-KLH that were still immunogenic for normal F1 mice. Furthermore, under limiting in vitro micro-culture conditions, the abnormal F1 mice failed to generate primary thymic-dependent responses to any dose of TNP-KLH, even though under the identical conditions normal F1 mice consistently responded to a wide antigen dose range. The cellular basis of the failure of abnormal F1 mice to respond in vitro to TNP-KLH was investigated by assaying the ability of purified populations of accessory cells, T cells, and B cells from these mice to function in responses to TNP-KLH. The results of these experiments demonstrated that helper T cells and antigen-presenting accessory cells from abnormal F1 mice were competent and functioned as well as the equivalent cell populations from normal F1 mice. Instead, the failure of CBA/N mice to generate primary in vitro responses to TNP-KLH was solely the result of a defect in their B cell population such that B cells from these mice failed to be triggered by competent helper T cells and/or competent accessory cells. Similarly, the failure of abnormal F1 mice to respond either in vivo or in vitro to TNP-Ficoll was not the result of defective accessory cell presentation of TNP-Ficoll, but was the result of the failure of B cells from these mice to be activated by competent TNP-Ficoll-presenting accessory cells. In contrast to the failure of B cells from abnormal F1 mice to be activated in vitro in response to either TNP-KLH or TNP-Ficoll, B cells from abnormal F1 mice were triggered to respond to TNP-BA and TNP-LPS, antigens that did not require accessory cell presentation. The specific failure of B cells fron abnormal F1 mice to be activated in responses that required antigen-presentation by accessory cells suggested the possibility that the X-linked CBA/N genetic defect resulted in B cell populations that might be deficient in their ability to interact with antigen-presenting accessory cells...

Constitutive activation of different Jak tyrosine kinases in human T cell leukemia virus type 1 (HTLV-1) tax protein or virus-transformed cells.

HTLV-1 infection causes an adult T cell leukemia in humans. The viral encoded protein tax, is thought to play an important role in oncogenesis. Our previous data obtained from a tax transgenic mouse model revealed that tax transforms mouse fibroblasts but not thymocytes, despite comparable levels of tax expression in both tissues. Constitutive tyrosine phosphorylation of a 130-kD protein(s) was observed in the tax transformed fibroblast B line and in HTLV-1 transformed human lymphoid lines, but not in thymocytes from Thy-tax transgenic mice. Phosphotyrosine immunoprecipitation followed by Western blot analysis with a set of Jak kinase specific antibodies, identified p130 as Jak2 in the tax transformed mouse fibroblastic cell line and Jak3 in HTLV-1 transformed human T cell lines. Phosphorylation of Jak2 in tax transformed cells resulted from high expression of IL-6. Tyrosine phosphorylation of this protein could also be induced in Balb/c3T3 cells using a supernatant from the B line, which was associated with induction of cell proliferation. Both phosphorylation and proliferation were inhibited by IL-6 neutralizing antibodies. Constitutive phosphorylation of Jak kinases may facilitate tumor growth in both HTLV-1 infected human T cells and the transgenic mouse model.

Transcriptional suppression of the human T-cell leukemia virus type I long terminal repeat occurs by an unconventional interaction of a CREB factor with the R region.

To analyze regulation of the human T-cell leukemia virus type I (HTLV-I) long terminal repeat (LTR), cell lines were generated from LTR-tax x LTR-beta-galactosidase (beta-Gal) doubly transgenic mouse fibroblastic tumors. The HTLV-I LTR directs expression of both the tax and lacZ genes, and Tax up-modulates both promoters in primary cells. However, once cells were transformed by tax, beta-Gal but not tax expression was suppressed. Supertransformation of these cells with v-src suppressed both beta-Gal and tax expression. This suppression was reversed by treatment with the tyrosine kinase inhibitor herbimycin A or protein kinase A inhibitor H8. Electrophoretic mobility shift assays demonstrated augmented binding in the R but not U3 region. This binding was competitively inhibited by a high-affinity CREB oligodeoxynucleotide and super-shifted with a specific CREB antibody. Treatment of cells with the cyclic AMP analog dibutyryl cyclic AMP also transiently increased the R region binding dramatically. In vitro DNase I footprint analysis identified a protein-binding sequence in the R region which corresponded with suppression. However, this target sequence lacked a conventional CREB-binding site. A 70.5-kDa DNA-binding protein was partially purified by affinity chromatography, along with a 49-kDa protein which reacted with CREB-specific sera. These data demonstrate that HTLV-I LTR suppression is associated with CREB factor binding in the R region, probably by direct interaction with a 70.5-kDa protein, and provide a novel mechanism for maintenance of viral latency.

Adenovirus-mediated interferon-gamma transfer inhibits growth of transplanted HTLV-1 Tax tumors in mice.

Human T cell leukemia virus type 1 (HTLV-1) causes adult T cell leukemia (ATL), and the virus-encoded trans-activator, Tax, plays an important role in T cell transformation. In the HTLV-1 long terminal repeat (LTR)-Tax transgenic mouse model, Tax expression causes fibroblastic tumors. A tumor-derived cell line (B line) obtained from an explant of a Tax-transformed tumor, was established. This line expresses high levels of many cytokines as a consequence of Tax activation. However, the tumors are not immunogenic when transplanted into syngeneic mice. Because B line cells do not express the immunogenic cytokine interferon-gamma (IFN-gamma), a replication-defective adenoviral vector was used to deliver the IFN-gamma gene to tumor cells. The recombinant IFN-gamma adenovirus (IFN-gamma/Ad) can efficiently infect B line cells, resulting in high levels of IFN-gamma expression and secretion. Local secretion of IFN-gamma from B line cells caused both CD(4+)- and CD(8+)-positive T cell infiltration, and completely inhibited local tumor development in transplanted mice. Immunization with these cells significantly delayed tumor development after subsequent challenges of parental tumor cells. Expression of IFN-gamma in B cells also partially inhibited the highly expressed immune suppressive cytokine, transforming growth factor-beta 1 (TGF-beta 1). This system provides us with a valuable tumor immune therapy model to evaluate the effects of cytokines in induction or inhibition of specific antitumor immunity.

Absence of association of HTLV-I infection with type 1 neurofibromatosis in the United States or Japan.

An HTLV-I tax transgenic mouse model develops a syndrome with similarities to type 1 neurofibromatosis (NF-1). To investigate possible associations between this human retrovirus and NF-1, we have analyzed 67 neurofibromas from Japan (where HTLV-I infection is endemic) and compared them with 21 cases from the United States. We were not able to identify virus in tumor tissue in either group. This suggests that HTLV-I infection is not commonly associated with NF-1.

Degeneration of oxidative muscle fibers in HTLV-1 tax transgenic mice.

The HTLV-1 tax gene under control of the HTLV-1 long terminal repeat (LTR) was introduced into transgenic mice. Previously tax protein expression in the muscle and peripheral nerves of three independent mouse lines was reported. Here the localization of this transgenic protein at a cellular and subcellular level is described. Tax protein was expressed in oxidative muscle fibers that developed severe progressive atrophy. It localized to the cytoplasma where it was associated with structures resembling degenerating Z bands. This pattern of muscle fiber involvement is similar to that observed in human retroviral associated myopathy. This transgenic mouse model suggests that preferential expression of the HTLV-1 viral promoter in oxidative muscle fibers may explain the productive infection of these fibers in HTLV-1 myopathy.

Transgenic thymocytes are refractory to transformation by the human T-cell leukemia virus type I tax gene.

Previous transgenic work demonstrated transforming activity of the human T-cell leukemia virus type I Tax protein in fibroblasts. In the present study, a Thy-1-based vector was used to express Tax in thymocytes. These mice developed no functional or neoplastic abnormalities of T cells but developed fibroblastic tumors with a longer latency than in the previous model.

Rapid activation of astrocyte-specific expression of GFAP-lacZ transgene by focal injury.

Astrocytes form a key cellular component of the central nervous system. They respond vigorously to diverse neurologic insults by undergoing hypertrophy and increasing expression of the glial fibrillary acidic protein (GFAP) gene, but their functions are largely unknown. To analyze astrocytes in vivo we constructed a transgenic vector from GFAP gene sequences and monitored its efficiency by fusing it to lacZ. Injection of the GFAP-lacZ hybrid gene into the germline of mice yielded six different lines of transgenic mice. In all lines the expression of lacZ was astrocyte-specific. In unmanipulated transgenic animals beta-galactosidase activity was much more prominent in astrocytes of the hippocampal formation, selected white matter tracts, and glial limitans than in astrocytes of other areas. This pattern of expression illustrates the physiologic heterogeneity of astrocytes and probably reflects differences in functional demands placed on these cells in different brain regions. Upmodulation of transgene expression was used to determine the time frame within which astroglial activation and increased GFAP gene expression occur following a neurologic insult. Induction of GFAP-lacZ expression was detectable within 1 hour after focal mechanical trauma. This demonstrates that the response of astrocytes to neurologic injury is very rapid and implies that these cells could fulfill important early functions in wound healing within the central nervous system.

Sequence requirements of ATF2 and CREB binding to the human T-cell leukemia virus type 1 LTR R region.

We have identified crucial transcription factor contact sites within the distal portion of the HTLV-1 LTR R region. Single base substitutions within this region greatly reduce formation of a complex which we previously described as a 70-kDa nuclear protein interacting with a CREB-like protein. Comparison of published sequences of HTLV-1 isolates obtained from scattered geographic locations revealed that clustered mutations in an 8-base segment near the U5 junction do, in fact, occur naturally. A single base substitution corresponding to a common naturally occurring mutation was introduced into the R region of an HTLV-1-LTR Cat construct. This resulted in derepression of the promoter in a cell line expressing high levels of the R region binding complex when compared to the wild-type LTR promoter. Affinity purification and electrophoretic mobility super-shift analysis identified a dominant 70-kDa DNA binding protein as ATF-2. Phosphorylated ATF-2 apparently interacts with CREB to form this downstream complex.

The usefulness of the Lukes-Collins classification in identifying subsets of diffuse histiocytic lymphoma responsive to chemotherapy.

Twenty-nine patients with Stage III and IV diffuse histiocytic lymphoma (DHL) were treated prospectively with cyclophosphamide, hydroxydaunorubicin, vincristine, prednisone and bleomycin (CHOP-B) or hydroxydaunorubicin, cyclophosphamide, vincristine, methotrexate with leucovorine rescue and cytosine arabinoside (ACOMLA). Twenty-six evaluable patients were reclassified blindly by the Lukes-Collins classification with five large noncleaved follicular center cell (FCC), six large cleaved FCC, three large cell unclassified, seven B-immunoblastic sarcoma and five T-immunoblastic sarcoma patients identified. There was no significant survival advantage between the two combination chemotherapy programs. Survival of the immunoblastic sarcoma patients was inferior to that of the FCC lymphoma patients (P = 0.02). There were no significant survival differences between the large cleaved FCC and noncleaved FCC subtypes. Immunoblastic sarcomas, B- and T-cell types, appear to be more resistant to standard combination chemotherapy programs and new approaches may warrant more aggressive therapy in future protocols. The large cell FCC lymphomas have an excellent prognosis.

Effect of phosphorothioate modification of oligodeoxynucleotides on specific protein binding.

Phosphorothioate modification of internucleoside linkages is widely used to prevent degradation of oligodeoxynucleotide (ODN) therapeutic agents in serum and cells. This modification generally increases ODN potency, but in many instances it is associated with an increase of poorly understood nonspecific effects. In this study, we have found that both cellular retention and nonspecific protein binding are dependent upon the extent of the oligonucleotide's modification. Flow cytometry of cells treated with fluorescein-labeled single-stranded (ss) or double-stranded (ds) ODNs demonstrated that fully phosphorothioate-modified ODNs exhibit much greater cellular association than 3'-terminally modified ODNs (with three 3'-terminal phosphorothioate linkages). Additionally, gel shift assays with either ss- or ds-probes showed that fully phosphorothioate-modified ODNs also exhibit much greater cytoplasmic and nuclear protein binding than either 3'-terminally modified or unmodified ODNs. However, gel shift competition assays showed that transcription factor binding by fully phosphorothioate-modified ds-ODNs was completely nonspecific relative to 3'-terminally modified and unmodified ds-ODNs. These results suggest that the benefits derived from full phosphorothioate modification of ODNs may be negated by increases of nonspecific protein binding and associated sequence-independent effects.

Randomized study comparing doxorubicin, cyclophosphamide, vincristine, methotrexate with leucovorin rescue, and cytarabine (ACOMLA) with cyclophosphamide, doxorubicin, vincristine, prednisone, and bleomycin (CHOP-B) in the treatment of diffuse histiocytic lymphoma.

Two combination chemotherapy programs, ACOMLA (doxorubicin, cyclophosphamide, vincristine, methotrexate and leucovorin rescue, and cytarabine) and CHOP-B (cyclophosphamide, doxorubicin, vincristine, prednisone, and bleomycin), were evaluated in 29 prospective randomized patients with advanced diffuse histiocytic lymphoma. A complete response was achieved in 13 of the 15 patients (87%) treated with CHOP-B and in nine of the 14 patients (64%) treated with ACOMLA. The overall complete response rate was 75%. Two patients treated with ACOMLA and none of the CHOP-B-treated patients have relapsed. Median followup is 32 months for ACOMLA patients and 26 months for CHOP-B patients. Actuarial freedom from relapse at 2 years is 49.9% for ACOMLA and 93.3% for CHOP-B (P = 0.04). Toxicity was substantial, with eight nonfatal episodes of sepsis and three drug-related deaths. There have been no central nervous system relapses. Although patients treated with CHOP-B have a better response rate, the small numbers of patients treated to date preclude definitive conclusions.