Natural T-cell regulation of spontaneous immunoglobulin secretion.

Splenic lymphocytes from nude (nu/nu), heterozygous/nude (+/nu), or wild type (+/+) mice were examined for their capacity to secrete immunoglobulin (Ig) in the absence of exogenous antigenic stimulation. Using the reverse hemolytic plaque assay, which measures spontaneous Ig secretion in vitro, whole spleen populations from both heterozygous/nude (+/nu) and nude (nu/nu) mice were found to have significantly fewer numbers of plaque-forming cells when compared with spleen cells from +/+ mice. Analysis of highly purified populations of T and B lymphocytes showed that increased numbers of B cells from +/+ mice were stimulated to secrete Ig when as few as 10% syngeneic +/+ T cells were added in vitro. In contrast, the same number of thymocytes suppressed the identical B-cell function. A comparison of splenic T cells obtained from either +/+ or +/nu mice revealed that T cells from +/nu animals stimulated additional plaque-forming activity by B cells from wild type or nude mice. The cellular mechanism underlying enhanced help by T cells from +/nu mice is unclear but may reflect a functionally restricted population of T cells inherited by heterozygous/ nude mice.     

Induction of interferon production in mouse spleen cell cultures by corynebacterium parvum.

Spleen cells of CS7BL/6 mice produced considerable amounts of interferon (IF) in vitro when tested 5 to 20 days after injection of killed Corynebacterium parvum. Interferon was also produced when C. parvum was added in vitro to spleen cell cultures of previously untreated mice. High levels were detected after 1 day of culture with some increment during subsequent days. In a number of experiments IF was also produced in untreated control cultures but only after prolonged cultivation and not after 1 day. The highest levels of IF were usually obtained when spleen cells of C. parvum-treated mice were challenged with additional C. parvum in vitro. The IF induced by C. parvum shared certain physicochemical properties with a tested immune IF and was not neutralized by an antiserum raised against a type I IF. Spleen cells of nu/nu mice and spleen cells treated by anti-θ serum plus complement did not differ from their respective controls, indicating that production of IF did not require mature T lymphocytes. Removal of B lymphocytes by nylon wool columns abolished the capacity of spleen cells to produce IF. When spleen cells were freed of adherent cells by the use of plastic surfaces, they no longer produced IF. Peritoneal exudate macrophages (PEC), which by themselves did not produce IF, in small numbers reconstituted nonadherent spleen cells. Nylon column-treated spleen cells, however, could not be restored by PEC. It is concluded that IF upon challenge with C. parvum is produced by B lymphocytes and requires the help of macrophages.     

The in vitro effect of a thymic epithelial culture supernatant on mixed lymphocyte reactivity and intracellular cAMP levels of thymocytes and on antibody production to SRBC by Nu/Nu spleen cells.

Addition of supernatants from rat thymic epithelial cultures (TES) to rat thymocytes stimulated with T-cell-mitogens or allogeneic cells leads to an increase in ¹⁴C-TdR incorporation. Furthermore, in the presence of TES, spleen cells from athymic nu/nu mice exhibit an enhanced in vitro antibody production to SRBC, whereas TES has no such effect on spleen cells from T-cell-deprived mice. If TES is added together with thymocytes to T-cell-deprived spleen cell cultures, the number of plaque-forming cells to SRBC is enhanced, suggesting that TES induces a helper cell function in thymocytes which, if added alone, have no effect. TES also increases intracellular levels of cAMP in thymocytes in vitro and appears to act on a membrane site distinct from the β-adrenergic receptor. TES fails to affect mitogen responses, MLR and cAMP levels of lymphocytes from other lymphoid organs. The biological activity of TES as compared to that of thymic extracts is discussed.     

Interferon inhibition of the primary in vitro antibody response to a thymus-independent antigen

Partially purified and crude mouse L cell interferon preparations inhibited the in vitro plaque-forming cell (PFC) response of mouse C57B1/6 spleen cells to the T-cell independent lipopolysaccharide antigen of Escherichia coli 0127. PFC responses of 5-day cultures were inhibited approximately 70–90% by 100–200 NIH reference units of interferon/culture. A similar inhibitory effect was obtained with spleen cells from athymic (nude) mice homozygous for the nu/nu allele. Spleen cultures depleted of adherent cells were also inhibited in their anti-0127 PFC response by interferon. Interferon, then, appears capable of inhibiting the PFC response to E. coli 0127 via direct action on B cells. Heating experiments along with the use of interferon preparations of different specific activities suggest that the inhibition was due to the interferon in the preparations.     

Deregulation of mouse antibody-forming cells by streptococcal pyrogenic exotoxin II. Modification of spleen T-cell-complemented nude mouse PFC responses

Streptococcal pyrogenic exotoxin (SPE), a toxic protein, secreted by Group A streptococci modifies antibody responses in two ways. It suppresses the early peak plaque-forming cell (PFC) and serum antibody responses to sheep erythrocytes (SE) and it engenders a late burst of PFC detected at 12–14 days. We have termed the late phase a deregulated response. This effect has been observed in rabbits and NIH (+/+ and +/nu) mice. NIH athymic nude (nu/nu) mice exhibit the early suppressed response but do not show the late phase. In reconstruction experiments to delineate the responsible target site of SPE we have conferred upon the nude or nude spleen cells in vitro, +/nu PFC responsiveness to SE by transfer of +/nu spleen cells in vivo or by supplementation with +/nu spleen cells in Marbrook cultures. When this is done, complementation of nude PFC responses and their ability to exhibit a deregulated response after SPE treatment is conferred coordinately. Pretreatment of donor cells with a B-cell inhibitory dose of X-ray or with a B-cell inhibitory dose of anti-Ig serum + C′ does not inhibit complementation of nude cells to +/nu responsiveness. Moreover, such donor suspensions when treated with SPE retain the ability to complement and to confer upon nude cells the ability to exhibit the late burst of PFC (a deregulated response). A similar pretreatment of the donor cell suspension with an anti-T-cell serum and C′, however, markedly inhibits both the adoptive complementation and the deregulation of the nude mouse PFC response. Thus, it is demonstrated that the target cell affected in this way by SPE is a T-cell. We presume from this evidence that SPE inhibits a T-cell which is involved in the regulation of antibody formation.     

Plasmodium berghei: The in Vitro immune response

Following primary in vitro Stimulation by Plasmodium berghei, IgM titers by the indirect fluorescent antibody test (IFAT) were negative in in vitro reconstituted syngeneic mouse spleen cultures containing T cells and macrophages, or B cells and macrophages, or macrophages alone, but IgM titers of 1:20 were obtained from cultures containing B cells, T cells, and macrophages. IFAT IgG titers were negative for cultures with T cells and macrophages together, or macrophages alone, but rose to 1:40 with cultures containing B cells and macrophages and 1,80 with cultures of B cells, T cells, and macrophages together. After a second in vitro challenge, IgM and IgG titers were similar to the stimulated cultures from immunized mice; the IgM titer reached only 1:20 but the IgG titer rose to 1:160. Total immunoglobulin was higher in the cultures from immunized mice than from in vitro primed cultures.     

Polyclonal Antibody Production in Murine Spleen Cells Induced by Staphylococcus

Polyclonal plaque-forming cell (PFC) responses in murine spleen cells induced by Staphylococcus aureus and S. epidermidis were studied. Injection of Balb/c mice with S. aureus strain 248βH resulted in the generation of anti-trinitrophenyl (TNP) and anti-sheep red blood cell PFC in their spleens. Cultures of Balb/c spleen cells in the presence of S. aureus 248βH, Cowan I, or a protein A-deficient mutant yielded many anti-TNP PFC. The larger the number of organisms that were added to the cultures, the better was the PFC response. Both living and killed organisms, were capable of inducing the response, but an excess of living 248βH organisms in the cultures abrogated the response. All of the organisms (12 strains of S. aureus and 11 strains of S. epidermidis) freshly isolated from patients had the ability to induce the polyclonal PFC response in cell cultures. These organisms stimulated cultured C3H/HeJ mouse spleen cells, which were unresponsive to bacterial lipopolysaccharide (LPS). Cultured cells from the spleens of athymic nu/nu mice also responded to these organisms, and the number of PFC in nu/nu cell cultures was always greater than that in nu/+ cells prepared from a haired litter mate. Moreover, the responses of nu/nu spleen cell cultures to which nylon wool column-filtered splenic nu/+ T cells were added were lower than expected. These findings suggest that the polyclonal PFC response to staphylococci is thymus independent, but that the magnitude of the response is regulated by mature T cells. Cultures of macrophage-depleted spleen cells responded to the organisms to an extent similar to that of the control. The 248βH organisms were less capable of stimulating spleen cells of 2-week-old mice (i.e., early maturing B cells) than LPS. However, spleen cells from adult (7-week-old) and aged (9-month-old) mice responded well to both the organisms and LPS. Previous sensitization with the organisms in vivo did not affect any polyclonal responses of spleen cells in vitro to either the organisms or LPS. The role of staphylococcal protein A in the polyclonal PFC response to staphylococci is discussed.     

The effects of cortisol on the primary response of mouse spleen cell cultures to heterologous erythrocytes

Cell viability and the production of direct PFC were studied in mouse spleen cell cultures after cortisol treatment in vivo or in vitro at various times relative to primary stimulation with SRBC in vitro.Cortisol treatment in vivo reduced spleen cell numbers by 88% after 48 hr, but cultures of the remaining cells produced as many PFC in vitro as did cultures of equal numbers of normal spleen cells.In normal spleen cell cultures incubated with cortisol for 4 hr prior to the addition of antigen, peak responses of PFC/culture and PFC/10⁶ cells occurred 24 hr later than in controls and averaged, respectively, 27% and 141% of control values. Minimum viable cell numbers were observed in cortisol-treated cultures after 3 days; thereafter cell numbers gradually increased. These results were not significantly altered when cultures were treated simultaneously with cortisol and antigen.The response was not suppressed if the addition of antigen preceded that of cortisol by more than 4 hr. Suppression was also considerably reduced if fetal calf serum was used when preparing cells for culture.     

T-dependent suppression of the primary antibody response to sheep erythrocytes in mice infected with Trichinella spiralis.

Mice infected for 20 days with the parasitic mematode Trichinella spiralis had significantly reduced numbers of splenic antibody-forming cells (AFC) and decreased serum hemagglutinin titers following intraperitoneal immunization with sheep erythrocytes (SE). Similarly, when immunized in vitro to SE, cultures of splenocytes from infected mice developed fewer AFC than cultures of normal cells. Splenocytes from infected mice actively suppressed the in vitro response of normal cells to SE, and this in vitro suppression was abolished by lysis with anti-thy 1 antiserum and enhanced by lysis with anti-immunoglobulin antiserum. The addition of supernatant fluids from cultures of splenocytes from infected mice to cultures of normal cells on Day 0 of culture reduced by 70% the number of AFC produced by these cultures. These results indicate the presence of T-suppressor cells and suggest that antigen-induced suppression (antigenic competition) is one mechanism of Trichinella-induced suppression.     

Transplantation antigen expression on murine trophoblast detection by induction of specific alloimmunity

Cell-mediated immune responses to murine embryonic trophoblast cells were investigated using lymphocyte trophoblast cultures (LTC) and cell-mediated lympholysis (CML). Spleen cells from CBA (H-2^k) or C57BL/6 (H-2^b) mice hyperimmunized with 3.5-day-old Balb/c (H-2^d) blastocysts did not undergo DNA synthesis after in vitro exposure to Balb/c blastocyst outgrowths nor were cytotoxic lymphocytes (CTL) generated against H-2^d alloantigens. Splenocytes from Balb/c mice presensitized with semiallogeneic (Balb/c female × C57BL/6 male) trophoblast cells derived from 17- to 20-day placental tissue expressed a weak proliferative response in the presence of semiallogeneic placental trophoblast and produced a moderate number of CTL against H-2^b (paternal strain) alloantigens when compared to mixed lymphocyte cultures (MLC) between Balb/c responder and semiallogeneic (stimulator) spleen cells. CTL were also generated in vitro after splenocytes from Balb/c mice hyperimmunized with semiallogeneic spleen cells were restimulated in vitro with placental trophoblast cells. These studies showing that early-stage trophoblast cells fail to evoke transplantation immunity and placental trophoblast is capable of generating alloimmunity only after combined in vivo hyperimmunization with in vitro restimulation suggest that these trophoblast cells are poorly immunogenic due in part to the relatively weak functional expression of major transplantation antigens.     

Effects of Dual Targeting of Tumor Cells and Stroma in Human Glioblastoma Xenografts with a Tyrosine Kinase Inhibitor against c-MET and VEGFR2

Anti-angiogenic treatment of glioblastoma with Vascular Endothelial Growth Factor (VEGF)- or VEGF Receptor 2 (VEGFR2) inhibitors normalizes tumor vessels, resulting in a profound radiologic response and improved quality of life. This approach however does not halt tumor progression by diffuse infiltration, as this phenotype is less angiogenesis dependent. Combined inhibition of angiogenesis and diffuse infiltrative growth would therefore be a more effective treatment approach in these tumors. The HGF/c-MET axis is important in both angiogenesis and cell migration in several tumor types including glioma. We therefore analyzed the effects of the c-MET- and VEGFR2 tyrosine kinase inhibitor cabozantinib (XL184, Exelixis) on c-MET positive orthotopic E98 glioblastoma xenografts, which routinely present with angiogenesis-dependent areas of tumor growth, as well as diffuse infiltrative growth. In in vitro cultures of E98 cells, cabozantinib effectively inhibited c-MET phosphorylation, concomitant with inhibitory effects on AKT and ERK1/2 phosphorylation, and cell proliferation and migration. VEGFR2 activation in endothelial cells was also effectively inhibited in vitro. Treatment of BALB/c nu/nu mice carrying orthotopic E98 xenografts resulted in a significant increase in overall survival. Cabozantinib effectively inhibited angiogenesis, resulting in increased hypoxia in angiogenesis-dependent tumor areas, and induced vessel normalization. Yet, tumors ultimately escaped cabozantinib therapy by diffuse infiltrative outgrowth via vessel co-option. Of importance, in contrast to the results from in vitro experiments, in vivo blockade of c-MET activation was incomplete, possibly due to multiple factors including restoration of the blood-brain barrier resulting from cabozantinib-induced VEGFR2 inhibition. In conclusion, cabozantinib is a promising therapy for c-MET positive glioma, but improving delivery of the drug to the tumor and/or the surrounding tissue may be needed for full activity.     

Induction of an antibody response in vitro against synthetic antigens in guinea pigs. I. Culture and assay conditions

Guinea pig spleen and lymph node cells were found to produce anti-2,4-dinitrophenyl (Dnp) oligolysine PFC in vivo against 2,4-dinitrophenyl-β-alanyl glycyl glycyl (Dagg-SRBC) but not against trinitrophenyl-SRBC target indicator cells. Furthermore, when sensitized spleen cells or their purified B-cell fractions were cocultured with primed peritoneal exudate lymphocytes (PEL) but not splenic T cells they were able to generate a secondary PFC response in vitro to the synthetic antigens, Dnp oligolysines. PFC were not induced in vitro if these same cultures were pulsed with short-chain peptides (five lysines) or the complex antigen, dinitrophenyl-bovine γ-globulin (DnpBGG). Con A was able to substitute for PEL in triggering spleen cells to mount a secondary in vitro PFC response to homologous Dnp oligolysines. More importantly, the Con A-aided spleen cell cultures were not induced above background values when challenged in vitro with heterologous Dnp oligolysines. This study suggests that spleen cells may lack a nonspecific signal for the development of a secondary in vitro PFC response.     

Role of the H-2I region in the generation of an antiviral cytotoxic T-cell response in vitro

The participation of H-2I gene products in generating virus-specific proliferative and/or cytotoxic T-lymphocyte (CTL) responses was investigated. Spleen cells from mice infected with vaccinia virus were restimulated secondarily in vitro with syngeneic virus-infected peritoneal exudate cells (PEC) and then restimulated in tertiary cultures with virus-infected PEC from syngeneic and partially histoincompatible strains of mice. Based on the finding that comparable proliferative responses resulted when stimulating the responding cells were histocompatible at the H-2K, I, or D region of the major histocompatibility complex (MHC), the additively enhanced, but not potentiated, proliferation caused by identity at two or three H-2 regions was analyzed. Enhancement of proliferation followed when the H-2K/D components plus virus and the H-2I components plus virus were either on the same, or alternatively on two, stimulating cells. This suggests that H-2K, D, and I plus virus trigger distinct T-cell subsets. A virus-specific CTL response was generated in vitro when spleen cells from virus-primed mice and even unprimed mice were stimulated with cells sharing only H-2K and/or H-2D of the MHC. Identity at the H-2I region did not stimulate a CTL response, nor did it influence the magnitude of the $\text{K}\text{D}$ restricted response. Nevertheless, the presence of anti-Ia antiserum in cultures of syngeneic stimulators and responders inhibited the cytotoxic response to a great extent. Therefore, H-2I region products seem to participate in the generation of virus-specific CTL in vitro.     

Role of T Cells in the Mitogen-Induced Proliferation and Polyclonal Antibody Response of Murine B Cells

The effect of thymus-derived lymphocytes (T cells) on the responsiveness of bone marrow-derived lymphocytes (B cells) to bacterial lipopolysaccharide (LPS) was determined in in vitro experiments. Radiation resistant splenic T cells obtained from euthymic nu/+ mice increased the number of proliferating cells in the cultures of splenic B cells from athymic nu/nu mice even in a nonstimulated state. The radiation resistant T cells augmented significantly the responsiveness of B cells to LPS, as determined by an increase in proliferating cells and polyclonally induced anti-sheep erythrocyte (SRBC) IgM hemolysin plaque-forming cells (PFC). Addition of the T cells to B cell cultures not only augmented the responsiveness of B cells to suboptimal doses of LPS but also enabled B cells to respond to supraoptimal doses of LPS. As is well documented, the radiation resistant T cells were unable to induce the generation of anti-SRBC PFC in B cell cultures, unless the cultures were simultaneously stimulated with SRBC. Colcemid, a specific inhibitor of cell mitosis, blocked almost completely the exponential generation of anti-SRBC PFC in B cell cultures responding to SRBC with the aid of radiation resistant T cells. In contrast, colcemid did not affect the exponential generation of anti-SRBC PFC of a polyclonal nature in B cell cultures responding to LPS, either in the presence or absence of radiation resistant T cells.     

Nude mice--a model system for studying the cellular basis of the humoral immune response

Mice homozygous for the nu gene fail to develop a thymus. In comparison to spleen cells from +/nu mice spleen cells from nu/nu mice have a deficient 19S PFC response to SRBC when tested in culture or in vivo. This deficiency is due to a lack of “helper” T cells in nu/nu spleen; A cells and B cells appear to be normal. The capacity of nu/nu spleen cells to produce a PFC response in culture can be corrected by the addition of T cells obtained from either the thymuses or the spleens of +/nu mice. In contrast to “helper” T cells obtained from the spleen, “helper” T cells obtained from the thymus appear to require the capacity for proliferation during the response to SRBC.     

Microvesicles Derived from Human Wharton's Jelly Mesenchymal Stem Cells Promote Human Renal Cancer Cell Growth and Aggressiveness through Induction of Hepatocyte Growth Factor

In our previous study, microvesicles (MVs) released from human Wharton''s jelly mesenchymal stem cells (hWJ-MSCs) retard the growth of bladder cancer cells. We would like to know if MVs have a similar effect on human renal cell carcinoma (RCC). By use of cell culture and the BALB/c nu/nu mice xeno-graft model, the influence of MVs upon the growth and aggressiveness of RCC (786-0) was assessed. Cell counting kit-8 (CCK-8) assay, incidence of tumor, tumor size, Ki-67 or TUNEL staining was used to evaluate tumor cell growth in vitro or in vivo. Flow cytometry assay (in vitro) or examination of cyclin D1 expression (in vivo) was carried out to determine the alteration of cell cycle. The aggressiveness was analyzed by Wound Healing Assay (in vitro) or MMP-2 and MMP-9 expression (in vivo). AKT/p-AKT, ERK1/2/p-ERK1/2 or HGF/c-MET expression was detected by real-time PCR or western blot. Our data demonstrated that MVs promote the growth and aggressiveness of RCC both in vitro and in vivo. In addition, MVs facilitated the progression of cell cycle from G_0/1 to S. HGF expression in RCC was greatly induced by MVs, associated with activation of AKT and ERK1/2 signaling pathways. RNase pre-treatment abrogated all effects of MVs. In summary, induction of HGF synthesis via RNA transferred by MVs activating AKT and ERK1/2 signaling is one of crucial contributors to the pro-tumor effect.     

An in vitro model for induction of immunological unresponsiveness to turkey gamma-globulin in primed mouse spleen cells. I. The secondary in vitro response and induction of unresponsiveness.

Spleen cells from mice previously immunized with turkey γ-globulin (TGG) were shown to give a vigorous secondary response in vitro when challenged in Mishell-Dutton cultures with TGG covalently coupled to pig erthrocytes (TGG-PRBC). However, 90–100% of the response could be abrogated by the incorporation of soluble TGG (sTGG) into the culture medium at concentrations greater than 1 mg/ml. Unresponsiveness, as measured by the absence of plaque-forming cells (PFC) in cultures receiving sTGG, was found to be antigen specific in that these cultures were still able to give normal PFC responses to sheep or burro erythrocytes. Spleen cells incubated with sTGG for short periods of time were shown to remain unresponsive after removal of sTGG from the culture and addition of TGG-PRBC. A 1-hr exposure period resulted in greater than 70% Unresponsiveness and a complete unresponsive state required only 8 hr of exposure. In contrast to the continued Unresponsiveness of sTGG-treated cells in vitro, spleen cells incubated with sTGG for 24 hr were fully responsive to an immunogenic challenge with alum-precipitated TGG when they were transferred into irradiated syngeneic mice. These data suggest that the readily induced unresponsive state in cultures of TGG primed cells may involve either a reversible antigen blockade of antigen-sensitive lymphocytes or a peripheral inhibition of reactive cells by suppressor lymphocytes.     

Muscle development in vitro following X irradiation

Myogenic cells obtained from 12-day-old embryonic chicken hind limb and breast muscle were exposed to 5000 rads of X irradiation. Although 10% of the initial cell dissociates were killed by irradiation, the remaining cells were comparable to controls in plating efficiency and light microscopic morphology. Moreover, there was no increase or loss of cells for at least 72 hr in vitro when plated at a density of 2 × 10⁶ cells/60-mm plate. It was found that muscle cell fusion after irradiation proceeded at the same rate and to the same relative extent as in control cultures. Myotubes developed normally; cross-striations were prominent by 5 to 7 days of culture and the cells maintained a well-differentiated state for periods of at least 3 weeks in vitro. In control cultures continuously labeled with 1 μCi/ml of [³H]TdR, 75% of the nuclei within myotubes were heavily labeled by 118 hr; less than 15% of the nuclei within syncytia of irradiated cultures were labeled. Quantitative microphotometry of Feulgen-stained cultures demonstrated that all nuclei within control and irradiated myotubes contained the 2C complement of DNA. Similar experiments conducted with cells released from limbs and breasts of 10-day-old embryos revealed lower absolute levels of cytoplasmic fusion in both control and irradiated samples, however, there was slightly more cell death after exposure to X rays in 10-day-old than 12-day-old material. Nevertheless, considerable cell fusion occurred in irradiated limb and breast cell cultures, consistent with the conclusion that the commitment to myogenesis of prefusion myoblasts is extremely stable even in the face of massive ionizing radiation and that neither cell division nor replication of DNA is an obligatory prerequisite for the in vitro fusion and subsequent differentiation of skeletal muscle obtained from 10- and 12-day-old chick embryos.     

In vitro studies of the rabbit immune system: I. Hemolytic plaque-forming response of dissociated spleen cells from normal and primed rabbits

The conditions for the in vitro generation of primary and secondary immune responses by rabbit spleen cells to sheep red blood cell (SRBC) antigen have been examined. Spleen cells from many normal and all previously immunized rabbits are capable of producing in vitro plaque-forming cell (PFC) responses when cultured as dissociated cell suspensions in the presence of antigen. Primed spleen cells generate approximately 100 times the number of PFCs obtained in normal cultures with a shorter lag period. Both types of cultures demonstrate a period of exponential increase in PFCs during which the doubling time is 12–14 hr. This increase occurs after 1 day of culture of spleen cells from primed rabbits and after 4 days of culture of spleen cells from unprimed rabbits. The PFCs which arise in cultures of primed cells appear not to be the progeny of those generated in vivo but to be derived from an increased number of PFC precursors. Repeated immunization of the spleen cell donor is required to produce significant numbers of indirect (IgG) PFC or indirect precursors; most of the PFC found after a single immunization in vivo or in vitro are direct (IgM). There is no evidence for conversion of IgM to IgG PFC in vitro. This system should provide a means for further identification of the cellular interactions involved in the immune response of the rabbit.