Myeloid dendritic cells are primed in allergic asthma for thymic stromal lymphopoietin-mediated induction of Th2 and Th9 responses.

BACKGROUND: Type 1 myeloid dendritic cells (mDCs) contribute to inception of allergic asthma (AA) and are regulated by epithelial-derived cytokines. OBJECTIVES: To evaluate whether mDCs from AA patients are primed for thymic stromal lymphopoietin (TSLP)-driven responses. METHODS: mDCs from 18 AA patients and 15 controls were purified using immunomagnetic sorting. Cells were pulsed with TSLP or Dermatophagoides pteronyssinus (Der p) allergen, before FACS phenotyping and co-culture with allogeneic CD4+ T cells. Bronchial biopsies from 15 AA patients and four controls were immunostained for CD1c and TSLP receptor (TSLPR). RESULTS: Allergic asthma patients had a higher proportion of TSLPR+ mDCs, in blood and bronchial mucosa. When compared to mDCs from controls, both TSLP- and Der p-pulsed blood mDCs from AA patients induced increased polarization of CD4+ T cells into Th2 cells (IL-5, IL-13, and GATA3+), while only TSLP-mDCs promoted Th9 cells (IL-9 and PU.1+ /IRF4+). In addition, OX40L was induced upon TSLP stimulation and was required for the induction of Th2, but not Th9, cells. In contrast, development of Th9 cells in this model depended on TGF-ß1. CONCLUSIONS: Our data indicate overlapping but partially distinct effects of TSLP and Der p allergen pathways, showing that DCs are primed in human asthma for TSLP-driven induction of both Th2 and Th9 cells. This novel TSLP/mDC/Th9 axis operates through a distinct, OX40L-independent pathway. These data further highlight the TSLP pathway as a relevant target in human asthma.

Improved [18F]FDG PET/CT Diagnostic Accuracy for Infective Endocarditis Using Conventional Cardiac Gating or Combined Cardiac and Respiratory Motion Correction (CardioFreezeTM).

Infective endocarditis (IE) is a serious and diagnostically challenging condition. [18F]FDG PET/CT is valuable for evaluating suspected IE, but it is susceptible to motion-related artefacts. This study investigated the potential benefits of cardiac motion correction for [18F]FDG PET/CT. In this prospective study, patients underwent [18F]FDG PET/CT for suspected IE, combined with a conventional cardiac gating sequence, a data-driven cardiac and respiratory gating sequence (CardioFreezeTM), or both. Scans were performed in adherence to EANM guidelines and assessors were blinded to patients' clinical contexts. Final diagnosis of IE was established based on multidisciplinary consensus after a minimum of 4 months follow-up and surgical findings, whenever performed. Seven patients participated in the study, undergoing both an ungated [18F] FDG-PET/CT and a scan with either conventional cardiac gating, CardioFreezeTM, or both. Cardiac motion correction improved the interpretability of [18F]FDG PET/CT in four out of five patients with valvular IE lesions, regardless of the method of motion correction used, which was statistically significant by Wilcoxon's signed rank test: p = 0.046. In one patient the motion-corrected sequence confirmed the diagnosis of endocarditis, which had been missed on non-gated PET. The performance of the two gating sequences was comparable. In conclusion, in this exploratory study, cardiac motion correction of [18F]FDG PET/CT improved the interpretability of [18F]FDG PET/CT. This may improve the sensitivity of PET/CT for suspected IE. Further larger comparative studies are necessary to confirm the additive value of these cardiac motion correction methods.

Long axial field of view PET/CT in critically ill patients: lessons from a case report.

The introduction of new long axial field of view (LAFOV) scanners is a major milestone in positron emission tomography/computed tomography (PET/CT) imaging. With these new systems a revolutionary reduction in scan time can be achieved, concurrently lowering tracer dose. Therefore, PET/CT has come within reach for groups of patients in whom PET/CT previously was undesirable. In this case report we discuss the procedure of a continuous bed motion (CBM) total-body [18F]FDG PET/CT scan in an intensive care patient. We emphasize the clinical and technical possibilities with this new camera system, a matched clinical protocol, and the added value of a dedicated team.

A gene linked to the Igh-C locus controls the production of rheumatoid factor in the mouse.

In certain specific pathogen-free colonies, mice, upon aging, produce autoantibodies (RF) specific for the Fc portion of their IgG. In our colony, 129/Sv mice (H-2bvl; Igh-1a) have 10-20 times higher RF levels than C5BL/6 (h-2b; Igh-1b). In addition, the 129 have mainly IgA anti-IgG2a, and the B6 have mainly IgM anti-IgGl. We analyzed the genetic factors that control these differences. The high RF-producer phenotype of strain 129 was inherited as a recessive trait as indicated by the low RF levels of (129 X B6) F1 mice. About 1 of 4 129 X F1 (129 X B6) backcrosses and 1 of 10 (129 X B6) F2 mice had high RF levels, suggesting the involvement of two recessive genes in the control of this RF production. All F2 mice and all but one backcross with high IgA anti-IgG2a levels were homozygous for the Ihg-1a allele of the 129 mouse. In contrast, the B6-type RF was eight times more frequent in Igh-1bb than in Igh-1ab or Igh-1aa mice. High RF titers of either type were suppressed in Igh-1ab mice.

T cell growth and differentiation induced by interleukin-HP1/IL-6, the murine hybridoma/plasmacytoma growth factor.

Interleukin-HP1 (HP1)/IL-6 is a 25-30-kD protein produced by macrophages, fibroblasts, and certain T cell lines. It was originally identified as a mouse growth factor for B cell hybridomas and plasmacytomas, and was recently shown to stimulate growth and differentiation of normal B cells. Here we demonstrate that, in the presence of lectins or anti-T cell receptor antibodies, HP1/IL-6 has a growth factor activity equivalent to that of IL-2 for mature thymic and peripheral T cells of both the L3T4+ and Lyt-2+ subsets. Contrary to IL-2 and IL-4, HP1/IL-6 was, however, not capable of supporting the growth of established T cell lines. In addition to its effects on T cell proliferation, HP1/IL-6 also enhanced the differentiation of mouse cytolytic T cell precursors in primary allogeneic mixed lymphocyte cultures. Fractionation of responding cell populations indicated that HP1/IL-6 was capable of restoring the response of accessory cell-depleted T cells to Con A. This observation suggests that the production of HP1/IL-6 by macrophages could, at least partly, explain their role in polyclonal T cell activation.

The effect of complement on the ingestion of soluble antigen-antibody complexes and IgM aggregates by mouse peritoneal macrophages.

Complement was found to stimulate markedly the ingestion of soluble antigen-antibody complexes by mouse peritoneal macrophages. This was shown indirectly by measuring the release of degradation products when the complexes were labeled with 125I, or directly when the antigen, that was human transferrin, was labeled with 59Fe. In this case, the metal which was released from human transferrin inside the cells was not excreted, and its accumulation in the macrophages was a direct index of the uptake of immune complexes. The decay of radioactivity in macrophages after ingestion of 125I-labeled complexes was similar when they were taken up with or without complement, indicating that complement acts primarily on ingestion and not on digestion or excretion. The ingestion of complexes was morphologically confirmed using fluorescein-labeled antigen in the immune complexes. The opsonic effect of complement was also observed with IgM aggregates indicating that soluble complexes can be ingested through complement receptors without involvement of Fc-receptors, as required for particulate antigen-antibody complexes.

Monoclonal anti-IgG autoantibodies derived from lipopolysaccharide-activated spleen cells of 129/Sv mice.

In some colonies, 129/Sv mice produce, upon aging, a rheumatoid factor (RF) that is specific for mouse IgG2a but fails to react with IgG2a of the b allotype. It is not known whether this narrow specificity is due to the absence of other RF specificities in the repertoire of these mice or to the selective activation of the production of anti-IgG2a autoantibodies by a specific stimulus. To analyze the RF repertoire of 129/Sv mice, we have derived hybridomas from their spleen cells 3 d after an intraperitoneal injection of lipopolysaccharide. We have obtained 68 hybridomas secreting a monoclonal IgM with RF activity. This represents approximately 3 percent of the total number of hybridomas generated in four hybridizations. In addition, one monoclonal IgA RF was derived from unstimulated 129/Sv spleen cells. The specificities of these monoclonal RF were examined by testing their ability to bind to a panel of homologous and heterologous IgG preparations. The majority of the IgM RF reacted exclusively with a single mouse IgG subclass: 58 with IgG1, and 1 with IgG2a. Eight bound preferentially to IgG1 but cross-reacted to some extent with IgG2a and one was specific for a determinant shared by IgG1, IgG2a, and IgG3. The IgA RF derived from unstimulated spleen cells was primarily directed against IgG2a but cross- reacted somewhat with IgG2b. Identical results were obtained with two different monoclonal IgG1 and IgG2a proteins of the a allotype. No allotypic specificity was found for the anti-IgG1 RF, which all reacted well with IgG1 of the b allotype. In contrast, the IgM anti-IgG2a antibody exhibited such allotypic specificity because it failed to react with IgG2a of the b allotype. When tested on heterologous IgG preparations, all anti-IgG1 RF reacted better with rat IgG1, rat IgG2c, bovine IgG2, goat IgG2, and rabbit IgG than with mouse IgG1, demonstrating a particular homology between these Ig. On the basis of additional cross-reactions with other IgG, including rat IgG2a, rat IgG2b, bovine IgG1, goat IgG1, human IgG, and chicken IgG, seven different anti-IgG1 clonotypes could be identified. However, despite their heterogeneity, nearly all antigenic determinants recognized by anti-IgG 1 RF appeared to be located in the hinge region of the molecule. Total lack of binding to IgG1 Fab fragments was indeed observed, and only one antibody reacted with IgG1 Fc fragments. Unlike the anti-IgG1 RF, the IgM and the IgA anti-IgG2a antibodies did not cross-react with any heterologous IgG of the same panel. Altogether, t 1 different RF clonotypes could be distinguished on the basis of their fine specificity. The anti-IgG2a specificity of the RF spontaneously produced by 129/ Sv mice is thus not due to the absence of other RF specificities in the repertoire of these mice.

Immunogenic variants obtained by mutagenesis of mouse mastocytoma P815. I. Rejection by syngeneic mice.

We have reported that it is possible to obtain variants that are incapable of forming progressive tumour in syngeneic mice (tum-) by mutagenesis and cloning of a teratocarcinoma and a Lewis lung carcinoma cell line. These observations were extended to the ascitic P815 mastocytoma of mouse strain DB/2. After a treatment with the mutagen N-methyl-N"-nitro-N-nitrosoguanidine, we obtained a high frequency (14%) of tum- clones. A colony assay indicated that after a period of rapid multiplication extending to approximately 10 d after injection, the P815 tum- cells were rejected by a process that was usually completed by day 15. No rejection was observed in sublethally irradiated animals. The immunological nature of the rejection of the P815 variants was further inferred because, upon rejection, the mice acquired a radioresistant specific protection that could be transferred adoptively with spleen cells. Cross-immunization patterns demonstrated the presence of singular antigenic specificities on three of the five variants that were examined. In addition, a common antigen was found on all the tum- variants and the original cells that were capable of forming progressive tumors in syngeneic mice (tum+). Mice injected with tum- cells were significantly protected against a tum+ challenge, even though no significant protection was generated by irradiated tum+ cells. A study of the T lymphocyte-mediated cytolysis against the P815 variants described here is presented in the accompanying report (9).

Rheumatoid factor (RF) production during anamnestic immune responses in the mouse. III. Activation of RF precursor cells is induced by their interaction with immune complexes and carrier-specific helper T cells.

IgG1 immune complexes were identified as the humoral stimuli responsible for the synthesis of IgG1-specific IgM rheumatoid factor (RF), which occurs in the mouse during the early stages of secondary immune responses to protein antigens. The specificity of this phenomenon was illustrated by the fact that complexes made with IgG1 F(ab')2 fragments or with antibodies of a different isotype failed to induce significant anti-IgG1 RF synthesis. The importance of immune complexes in the induction of RF was further underscored by the substantial increase in the titers of isotype-specific RF observed in the serum of mice immunized with IgG1- or IgG2a-complexed antigen rather than with antigen alone. The RF-inducing capacity of the complexes varied with the antigen/antibody ratio: it was maximal in antibody excess or at equivalence, but dramatically reduced in large antigen excess. The importance of T cell priming in RF precursor cell activation by immune complexes was demonstrated by the failure of T cell-deprived spleen cells to reconstitute the capability of irradiated mice to produce RF, and by the optimal RF responses observed after reconstitution of irradiated recipients with primed T cells and naive B cells. The involvement of T cells in this process could not be explained by the release of nonspecific B cell activators, because antigenic stimulation of primed T cells failed to enhance the activation of RF precursor cells by immune complexes of unrelated antigen.

The binding of human lactoferrin to mouse peritoneal cells.

Human iron-saturated Lf (FeLf), which was labeled with 125I or 50Fe, was found to combine with the membrane of mouse peritoneal cells (MPC) which consisted of 70% macrophages. The following experimental data suggested the involvement of a specific receptor. (a) The binding of FeLf to MPC reached a saturation point. (b) The binding of radioactive FeLf was inhibited by preincubating the cells with cold FeLf but not with human Tf, human aggregated and nonaggregated IgG, or beef heart cytochrome c (c) Succinylation and carbamylation of FeLf resulted in a loss of its inhibiting activity on the binding of radioactive FeLf. Removal of neuraminic acid from FeLf increased its inhibitory activity. (d) The ability of apoLf to inhibit the binding of FeLf to MPC was significantly lower than that of FeLf. The existence of a Lf receptor capable of concentrating Lf released from neutrophils on the membrane of macrophages could explain the apparent blockade of the release of iron from the reticuloendothelial system, which accounts for the hyposideremia of inflammation. A receptor for FeLf was also found on mouse peritoneal lymphocytes. The affinity constant of FeLf for both lymphocytes and macrophages was 0.9 X 12(6) liter/mol. Howerver, macrophages bound three times more FeLf molecules (20 X 10(6)) per cell than did lymphocytes (7 X 10(6)).

Age-dependent production of IgA and IgM autoantibodies against IgG2a in a colony of 129/Sv mice.

Although much of the basic immunological work has been done with mice, little is known about anti-IgG autoantibodies in this species. Dresser (1, 2) has reported the occurrence, in CBA mice, of anti-IgG antibody (Ab)(1) detected by a hemolytic-plaque assay after stimulation with endotoxin or immunization against sheep erythrocytes. IgM rheumatoid factor has also been described in various strains of mice with a systemic lupus erythematosus-like disease (3). Recently, we have tried to induce anti-IgG in mice of the 129/Sv strain by inoculating autologous IgG. To our surprise, we found that the sera of all the animals had, before any inoculation, anti-IgG detectable by agglutination of particles coated with autologous IgG. The possibilities to investigate the mechanism of production and the biological role of this kind of Ab prompted us to undertake a study of the nature and specificity of the mouse anti-IgG.

Incidence and specificities of IgA and IgM anti-AgG autoantibodies in various mouse strains and colonies.

Mice, greater than 20 wk old, were tested for the presence of anti-IgG autoantibodies by agglutination and radioimmunoassay. IgA and IgM anti-IgG were found in the 129/Sv, C57BL/6, and DBA/2 strains from the local colony at the International Institute of Cellular and Molecular Pathology (ICP), at the Institut Pasteur de Paris (IP), and in the endotoxin-resistant C3H/He strain of ICP. These strains were negative at Iffa Credo (IC), and at The Jackson Laboratory (JL). Among 33 strains from the latter colony, 129/J, AKR/J, CBA/J, C57L/J, and NZB/BinJ were positive. All were specific pathogen-free and, excepting the NZB/BinJ, are not known to develop systemic autoimmune disorders. These differences between colonies suggest an influence of the environment on the production of anti-IgG. Evidence for the role of an infectious agent was provided by the fact that germ-free DBA/2 were negative in contrast to their SPF relatives. Strains which were positive at ICP and IP for anti-IgG had four-times higher serum levels of total IgA and two-times higher levels of total IgG than the corresponding negative strains from IC and JL. The anti-IgG titers differed markedly from one strain to the other in the same environment; e.g., in mice from ICP, BALB/c mice produced 40-times less anti-IgG than 129/Sv. IgA anti-IgG occurred only in high producers of anti-IgG. In these animals, the proportion of IgA vs. IgM anti-IgG was very different from one group to the other; C57BL/6 had mainly IgM anti-IgG, DBA/2 mainly IgA anti-IgG, and 129/Sv both IgM and IgA anti-IgG. The IgA anti-IgG from 129/Sv, 129/J, NZB/BinJ, C57L/J, DBA/2, and C3H/He had restricted hetero-, iso-, and allotypic specificities. It reacted only with mouse IgGa2, but not with the Ig-1b allotype. C57BL/6 also had IgA anti-IgG with a narrow specificity, but directed against IgG1 without allotypic restriction. In contrast to the specificity of IgA anti-IgG, the antibody activity of IgM anti-IgG was much broader, except in the 129/Sv and 129/J strains where IgM anti-IgG shared the same narrow specificity with IgA.

Variable region sequences of murine IgM anti-IgG monoclonal autoantibodies (rheumatoid factors). II. Comparison of hybridomas derived by lipopolysaccharide stimulation and secondary protein immunization.

We have obtained the complete variable region mRNA sequences of 11 LPS-derived and 14 secondary immunization-derived monoclonal IgM anti-IgG antibodies (rheumatoid factors, RFs). A comparative analysis of these sequences showed that monoclonal RFs derived after polyclonal activation are structurally very similar to RFs derived after secondary protein immunization. This study was undertaken to evaluate the potential relationship between two previously described phenomena: (a) during a secondary response to a protein antigen, RF is produced in quantities that equal or exceed the immunogen-specific antibody; and (b) the frequency of B cells that make RF after polyclonal activation is quite high; 3-10%. It has been unclear whether LPS-stimulated cells that produce IgM anti-IgG that is detected by an in vitro assay are related to the cells that produce RF after in vivo stimulation. The similarity of the antigen receptors found in the two types of RF, however, suggests that most or all of the RF-producing B cells detected after LPS stimulation would also be stimulated during the secondary immune response. Thus, the presence of relatively large number of B cells that can make RF after nonspecific stimulation provides an explanation for the magnitude of RF production accompanying the secondary immune response.

Interleukin-HP1, a T cell-derived hybridoma growth factor that supports the in vitro growth of murine plasmacytomas.

We have recently described the purification and NH2-terminal amino acid sequence of a T cell-derived hybridoma growth factor that was provisionally designated interleukin-HP1 (IL-HP1). Here we report that a T cell supernatant containing high titers of this hybridoma growth factor considerably facilitated the establishment of primary cultures of murine plasmacytomas. Most plasmacytoma cell lines derived from such cultures remained permanently dependent on IL-HP1-containing T cell supernatant for both survival and growth in vitro. These cell lines, however, retained their ability to form tumors in irradiated pristane-treated mice. Analytical fractionation of a T cell supernatant rich in IL-HP1 by either gel filtration, isoelectric focusing, or reversed-phase HPLC revealed the existence of only one plasmacytoma growth factor activity that strictly copurified with IL-HP1, strongly suggesting the identity of both factors. This conclusion was further supported by the finding that IL-HP1 purified to homogeneity supported the growth of both B cell hybridomas and plasmacytomas. For half-maximal growth, plasmacytomas, however, required a concentration of IL-HP1 of approximately 30 pM, which is approximately 200 times higher than that required by B cell hybridomas. A clear difference in the specificity of IL-HP1 and B cell stimulatory factor 1 (BSF-1) was demonstrated by the finding that IL-HP1-dependent plasmacytomas did not survive in the presence of BSF-1, whereas helper T cell lines that proliferated in the presence of BSF-1 failed to respond to IL-HP1.

Isotypic and allotypic specificity of mouse rheumatoid factors.

The specificity of polyclonal mouse rheumatoid factors (RF) was analyzed by competition experiments with heat-aggregated mouse IgG subclasses. The RF spontaneously produced by three normal mouse strains (129/Sv, CBA/Ht, and C57Bl/6) and by two strains with autoimmune diseases (MRL/l and NZB) were found to consist of distinct non-cross-reactive antibody subpopulations each specific for one IgG subclass. The sera of the normal strains contained IgG1- and IgG2a-specific RF. The autoimmune strains produced an additional variety of RF that was specific for The autoimmune strains produced an additional variety of RF that was specific for IgG2b. Also, the RF secreted by spleen cells of various normal strains after in vitro polyclonal activation with lipopolysaccharide could be resolved into distinct subpopulations specific for IgG1 or IgG2a. These results were confirmed by the analysis of monoclonal RF derived from BALB/c, C57Bl/6, CBA/Ht, and 129/Sv mice: of 73 hybridomas with RF activity, 71 displayed a strict subclass specificity. The subclass predominantly recognized depended on the origin of the spleen cells used to generate the hybridomas. After polyclonal activation in vitro, a broad spectrum of different specificities was obtained with 16 RF specific for IgG1, 13 for IgG2a, and 4 for IgG2b. In contrast, 27 of 28 monoclonal RF derived from 129/Sv and BALB/c mice without prior polyclonal activation were specific for IgG2a, and of these 75% were allotype specific since they failed to react with IgG2a of the b allotype. These results demonstrate the importance of subclass specificity in the production of RF in vivo. With the exception of the IgG2b-specific clones, all these monoclonal RF reacted preferentially with heat-aggregated or antigen-bound IgG. Among the hybridomas generated by the fusion of in vitro polyclonally activated spleen cells of 4-wk-old mice, the frequency of clones with RF activity was at least 40 times higher than that of clones specific for mouse IgM, human IgG, ovalbumin, and hen lysozyme.

The ingestion and digestion of human lactoferrin by mouse peritoneal macrophages and the transfer of its iron into ferritin.

Human lactoferrin (Lf) labeled with 125I and/or 59Fe was found to be ingested in vitro by mouse peritoneal macrophages (MPM). The uptake measured after 15 h incubation reached a saturation point at a concentration of 200 microgram/ml in the culture medium, whatever was the iron content of Lf. In such conditions, the uptake of transferrin (Tf) used as a control was 10 times lower. At a concentration of 80 microgram/ml in the medium, one cell picked up about 0.7 X 10(6) molecules of Lf per hour, and 0.13 X 10(6) molecules of Tf per hour. Iron-saturated Lf disappeared from MPM with a half life of 14.5 h, whereas the halflife of iron-free Lf was 4.2 h. Concomitant with the intracellular digestion of Lf, the iron was transmitted to ferritin. These data provide additional support for the hypothesis that Lf plays a key role in iron turnover, especially at the level of the reticuloendothelial system where iron is recovered from the catabolism of erythrocytes.

The involvement of lactoferrin in the hyposideremia of acute inflammation.

The hyposideremia of inflammation was found to be based on a three-step mechanism involving lactoferrin, the iron-binding protein from the specific granules of neutrophilic leukocytes. (a) Lactoferrin is Released from Neutrophils in an Iron-Free Form. When phagocytosis was induced in neutrophils by zymosan or bacteria, lactoferrin was recovered in the incubation medium together with other constituents of the specific granules, such as alkaline phosphatase and lysozyme. Lactoferrin extracted from leukocytes was able to bind the amount of iron corresponding to its theoretical iron-binding capacity. After injection of endotoxin into rats, lactoferrin was detected in various tissues where it was normally absent, or in the plasma when the reticuloendothelial system (RES) had previously been blocked by injections of India ink or aggregated albumin. (b) Lactoferrin is Able to Remove the Iron from Transferrin. Significant exchange of iron from transferrin to lactoferrin was observed in vitro only at a pH below 7.0 or in the presence of a high concentration of citrate. However, the fast elimination of lactoferrin in vivo, when saturated with iron, might account for the observed transfer of iron to endogenous or administered apolactoferrin. Intravenous injection of human apolactoferrin into rats caused a marked decrease of the plasma iron level. The kinetics of this process, as well as controls with other proteins, ruled out the possibility of a secondary inflammatory effect due to phlogogenic contaminants. (c) Fe-Lactoferrin is Taken-up by the RES. By immunofluorescence, lactoferrin was shown to be bound and ingested by monocytes. The rate of elimination of human Fe-lactoferrin injected into rats was particularly fast when compared to that of human apolactoferrin, succinylated Fe-lactoferrin, or other human proteins. Blockade of the RES slowed down the rate of clearance of Fe-lactoferrin and was also found to retard the elimination of endogenous rat lactoferrin released by endotoxin. These experiments suggest the existence of specific receptors for Fe-lactoferrin on the membrane of macrophages.

Autonomous growth and tumorigenicity induced by P40/interleukin 9 cDNA transfection of a mouse P40-dependent T cell line.

To test the transforming potential of deregulated P40/Interleukin 9 expression, we transfected a mouse P40-dependent T cell line with P40 cDNA, and examined the tumorigenicity of the resulting transfectants. When the cells, which grew autonomously in vitro, were injected intraperitoneally or subcutaneously into syngeneic mice, a very high tumor incidence was observed with as few as 10(4) cells per inoculum. Animals died as a result of widespread dissemination of lymphomatous tissue to abdominal and thoracic organs. The same P40-dependent cell line transfected with a control construct did not form tumors even after injection of 10(7) cells. These results indicate that uncontrolled expression of P40 can support T cell proliferation in vivo, and may be a transforming event involved in the development of certain T cell tumors.