Interferon-gamma-treated K562 target cells distinguish functional NK cells from lymphokine-activated killer (LAK) cells

In vitro incubation of the erythroleukemic cell line K562 with interferon-gamma (IFN-gamma) renders these cells relatively resistant to natural killer (NK) cell lysis. However, such treatment does not alter their sensitivity to LAK cell lysis. Thus, the lytic susceptibility of interferon-gamma-treated K562 (I-K562) cells to LAK cells as opposed to its relative resistance to NK cell lysis provides a functional assay to help distinguish these two types of effector cells. The relative resistance of I-K562 for NK cell-mediated lysis was not secondary to the release of soluble factors or the frequency of Leu-19+, CD3+ T cells, residual IFN-gamma, or expression of MHC Class I molecules. Coincubation of I-K562 cells with NK or LAK cells overnight did not appreciably change the pattern of lytic responses against K562 and I-K562 target cells. However, incubation of PBMC in vitro with I-K562 but not native K562 in the presence of r-IL-2 leads to a marked decrease in the generation of LAK cells. The inhibition of LAK cell generation was not secondary to differences in the consumption of bioactive levels of IL-2. Differences in the lytic capability of NK and LAK effector cells suggest heterogeneity among cells that mediate such non-MHC-restricted lysis. Use was made of cells from a patient with a large granular lymphocyte lymphoproliferative disease (greater than 85% Leu-19+) to determine if such cells could be used to distinguish clonal population of cells which would represent NK or LAK cell function. Of interest was the finding that such cells, even after incubation in vitro with IL-2, showed lytic function representative of NK cells but not LAK cells. Data concerning the inhibition of LAK cell generation by I-K562 cells have important implications for future therapeutic trials of IFN-gamma and IL-2 in the treatment of human malignancies.     

Immunotoxin-mediated elimination of clonogenic tumor cells in the presence of human bone marrow

An immunotoxin was synthesized with pokeweed antiviral protein and an IgG1 monoclonal antibody directed against human B and pre-B cells. The B43 murine monoclonal antibody does not react with normal human bone marrow precursor cells. The immunotoxin bound to all Burkitt's lymphoma cell lines that were tested but not to human peripheral blood T cells. The ability of antibody-toxin conjugate to inhibit human lymphoblast cell lines was checked in a clonogenic assay system. The immunotoxin in the presence of chloroquine elicited 5.8 logs of killing of Burkitt's lymphoma cells (B-ALL). The efficient inhibition of clonogenic growth of B-ALL cells was not affected by the presence of normal bone marrow cells. The immunotoxin was not very toxic to pluripotent stem cells; less than 50% of the stem cells were lost under conditions where 5.6 logs of clonogenic lymphoma cells were eliminated from a 100-fold excess of normal marrow cells. Further, when assayed by long-term human bone marrow cultures, immunotoxin treatment did not result in a significant loss of pluripotent precursor cells.     

Lack of apoptosis in Sendai virus-infected HEp-2 cells without participation of viral antiapoptosis gene.

Sendai virus (SeV) has been reported to induce apoptosis in many types of cells. In HEp-2 cells, however, it did not induce apoptosis in most of the infected cells under the conditions in which vesicular stomatitis virus induced massive apoptosis. The use of a novel technique, which allows the detection of viral antiapoptotic activity in the infected cells, showed that SeV does not have any antiapoptotic activity to interfere with the induction of apoptosis. Consistently, vesicular stomatitis virus-induced apoptosis was not interfered with by preinfection with SeV. These results indicate that the observed lack of apoptosis in these SeV-infected cells does not result from the suppression of apoptosis by viral antiapoptotic activity in the infected cells and suggest that, without activating a signaling pathway for the induction of apoptotic response in the infected cells, SeV can escape apoptosis of the cells, allowing long-term survival of the infected cells.     

Measurement of natural killer activity and target cell binding by mouse metrial gland cells isolated by enzymic or mechanical methods

Cells of the metrial glands of mice were isolated by enzymic or mechanical dissociation procedures. Morphological observations indicated that up to half of the enzymically dissociated cells and nearly all of the mechanically dissociated cells were granulated metrial gland cells, but the presence of some fibroblast-like stromal cells among the latter population was not ruled out. Moreover, the granulated metrial gland cells had lost a substantial part of their granule content during isolation. Both cell preparations had little or no natural killer (NK) activity, indicating either that granulated metrial gland cells are not NK-like or that their NK activity was impaired by loss of granule-associated lytic substances or by other factors. Enzymically dissociated metrial gland cells did not bind significantly to the NK target cell YAC-1, nor did they develop granules, NK activity, or the ability to bind YAC-1 cells during culture in vitro, either in normal medium or with the addition of indomethacin or lymphokines. Mechanically dissociated metrial gland cells bound avidly to YAC-1 cells but not to P815 cells or adult thymus cells, which are not NK target cells. Since many if not most of the mechanically dissociated metrial gland cells appeared morphologically to be granulated metrial gland cells, their selective binding to an NK target cell suggests that granulated metrial gland cells may be related in some way to NK cells.     

Human T cells in cord blood: Abnormalities in Ia antigens induction by phytohemagglutinin and in autologous mixed lymphocyte reactions

About 25% of human T cells isolated from cord blood acquired Ia antigens following stimulation with phytohemagglutinin (PHA) for 72 hr. This percentage is markedly lower than that found in PHA-activated T-cell populations (PHA-T cells) isolated from peripheral blood of adults. The low expression of Ia antigens by human T cells from cord blood does not reflect abnormalities in the sensitivity to PHA stimulation and/or in the kinetics of induction of Ia antigens. PHA-T cells from cord blood display a low stimulatory activity in autologous mixed lymphocyte reactions (MLR). The defect does not reflect a nonspecific abnormality in the stimulatory activity of PHA-T cells from cord blood, since the latter do not differ from PHA-T cells from adults in their ability to stimulate allogeneic T cells from adults. Furthermore the defect does not reflect a nonspecific abnormality in the proliferative response of T cells from cord blood, since the latter display a normal proliferative response to PHA-T cells from adults. The defect in the proliferative response is not restricted to the autologous MLR with PHA-T cells, since it was found also in autologous MLR with non-T cells as stimulators. Correlation of the temporal evolution of the abnormalities of human T cells with the maturation of the immune system may contribute to our understanding of the role of Ia antigens in cell-cell interactions and of the biological significance of abnormalities of autologous MLR.     

Factors responsible for immune resistance to L1210 murine leukemia in hyperimmune mice

Summary The serum of mice hyperimmune to L1210 leukemia was cytotoxic to L1210 cells and, to a much lesser extent, to P388 cells in the presence of complement. However, it did not suppress in vitro growth of L1210 cells, nor did it endow a recipient mouse with immunity to inoculated L1210 cells. This indicates that the serum did not play a significant role, if any, in immune protection of hyperimmune mice.Spleen and peritoneal exudate cells of hyperimmune mice suppressed the in vitro growth of L1210 but not of P388 cells. This is consistent with the fact that hyperimmune mice did not survive the inoculation of P388 cells. The immunocytes failed to suppress the in vitro growth of L1210 cells when preincubated with anti-Thy-1.2 antisera and complement. This, together with the finding that cell populations not adherent to a plastic dish suppressed in vitro growth of L1210 cells, indicates that T cells of immune spleen and peritoneal exudate cell populations were the effectors that suppressed in vitro growth of L1210 cells. Hyperimmune mice lost their immune protection in vivo following the administration of anti-thymocyte antisera, but not with carrageenan or silica, which resulted in the lethal growth of the inoculated L1210 cells. This indicates that T cells were in vivo effectors in immune protection.Hyperimmune spleen T cells endowed a recipient with immunity to L1210 leukemia when transferred in vivo. This confirmed the above results and suggests the applicability of immune cells in an adoptive immunotherapy approach.     

Modulation of natural cytotoxicity by alloantibodies. V. The mechanism of a selective augmenting effect of anti-H-2 antisera on the natural killer activity of mouse spleen cells

Treatment of mouse spleen cells with specific anti-H-2 antisera augments their natural killer (NK) activity against K562 cells but not against YAC target tumor cells. The same population of natural killer cells was found to lyse K562 as well as YAC target cells, since (a) depletion of YAC reactive NK cells by absorption on YAC monolayers resulted in a concomitant depletion of anti-K562 NK activity of mouse spleen cells, and (b) both K562 and YAC cells could inhibit their own as well as each others lysis in a cross-competition assay. Anti-H-2 antiserum could not induce anti-K562 NK activity in spleen cells previously depleted of NK cells by absorption on YAC monolayers, indicating that alloantiserum does not act by recruiting otherwise nonreactive cells to become cytotoxic toward K562 target cells. In a target-binding assay, K562 binding of NK cells (T-cell-, B-cell-, and macrophage-depleted spleen cells) increased five- to eightfold in the presence of anti-H-2 antiserum whereas YAC cells binding of NK cells was not increased. H-2 antigens per se did not appear to be involved in the alloantisera effect since anti-NK antiserum directed against a non-H-2 antigen selectively expressed on NK cells, showed a similar selective NK enhancing effect. Protein A, a reagent which binds to the Fc region of immunoglobulin molecules, completely blocked the alloantiserum induced augmentation of anti-K562 NK activity, but did not alter basal NK activity. Moreover, the F(ab)₂ fraction of alloantibodies failed to enhance anti-K562 cytotoxic activity of mouse spleen cells, indicating a crucial role for the Fc portion of the alloantibodies attached to the NK cells, in NK augmentation. Utilization of several target cell lines with or without membrane Fc receptors (FcR) revealed that alloantiserum enhanced the lysis of only FcR⁺ target cells. It is proposed that alloantibody-coated NK cells, as a result of a secondary interaction between attached alloantibody and Fc receptors on target cells, interact more readily with the target cells and thereby cause a higher level of lytic activity.     

Prethymic nylon wool-passed bone marrow cells, substituting for helper T cells, can augment the generation of cytotoxic T lymphocytes from their precursors.

Nylon wool-passed bone marrow (NW-BM) cells treated with anti-Thy.1 monoclonal antibody and complement were added to a mixed lymphocyte culture which contained a limiting number of lymph node cells, as responder cells, and a sufficient number of mitomycin-c-treated allogeneic spleen cells as stimulator cells. NW-BM cells of the same MHC haplotype as responder cells enhanced the generation of allo-specific cytotoxic T lymphocytes (CTL) not only at a relatively high dose (3 x 10(3) cells/well) of responder cells, but also at an extremely dilute dose (1 x 10(3) cells/well). NW-BM cells which had a third-party MHC haplotype, a haplotype different from both responder and stimulator cells, also enhanced the generation of CTL at relatively high doses, but not at low doses, of responder cells. NW-BM cells which had MHC haplotypes identical with those of responder cells induced CTL from helper T cell-depleted responder cells, but NW-BM cells which had the third-party haplotype did not. These results showed that the enhancing effects of NW-BM cells of the same MHC haplotype as responder cells might be due to a specific helper effect and the enhancing effect of NW-BM cells of the third-party haplotype might be due to a nonspecific filler effect, which only conditioned the cultured cells. It was also found that, to exhibit the helper effect, NW-BM cells had to possess MHC class II, but not MHC class I, molecules in common with CTL precursors. This study showed that in the induction of CTL, prethymic NW-BM cells had a capability comparable to that of mature helper T cells.     

Germ Cells Are Not Required to Establish the Female Pathway in Mouse Fetal Gonads

The fetal gonad is composed of a mixture of somatic cell lineages and germ cells. The fate of the gonad, male or female, is determined by a population of somatic cells that differentiate into Sertoli or granulosa cells and direct testis or ovary development. It is well established that germ cells are not required for the establishment or maintenance of Sertoli cells or testis cords in the male gonad. However, in the agametic ovary, follicles do not form suggesting that germ cells may influence granulosa cell development. Prior investigations of ovaries in which pre-meiotic germ cells were ablated during fetal life reported no histological changes during stages prior to birth. However, whether granulosa cells underwent normal molecular differentiation was not investigated. In cases where germ cell loss occurred secondary to other mutations, transdifferentiation of granulosa cells towards a Sertoli cell fate was observed, raising questions about whether germ cells play an active role in establishing or maintaining the fate of granulosa cells. We developed a group of molecular markers associated with ovarian development, and show here that the loss of pre-meiotic germ cells does not disrupt the somatic ovarian differentiation program during fetal life, or cause transdifferentiation as defined by expression of Sertoli markers. Since we do not find defects in the ovarian somatic program, the subsequent failure to form follicles at perinatal stages is likely attributable to the absence of germ cells rather than to defects in the somatic cells.     

Rapid B cell receptor-induced unfolded protein response in nonsecretory B cells correlates with pro- versus antiapoptotic cell fate

The adaptive unfolded protein response (UPR) is essential for the development of antibody-secreting plasma cells. B cells induced by lipopolysaccharide (LPS) to differentiate into plasma cells exhibit a nonclassical UPR reported to anticipate endoplasmic reticulum stress prior to immunoglobulin production. Here we demonstrate that activation of a physiologic UPR is not limited to cells undergoing secretory cell differentiation. We identify B cell receptor (BCR) signaling as an unexpected physiologic UPR trigger and demonstrate that in mature B cells, BCR stimulation induces a short lived UPR similar to the LPS-triggered nonclassical UPR. However, unlike LPS, BCR stimulation does not induce plasma cell differentiation. Furthermore, the BCR-induced UPR is not limited to cells in which BCR induces activation, since a UPR is also induced in transitional immature B cells that respond to BCR stimulation with a rapid apoptotic fate. This response involves sustained up-regulation of Chop mRNA indicative of a terminal UPR. Whereas sustained Chop expression correlates with the ultimate fate of the BCR-triggered B cell and not its developmental stage, Chop-/- B cells undergo apoptosis, indicating that CHOP is not required for this process. These studies establish a system whereby a terminal or adaptive UPR can be alternatively triggered by physiologic stimuli.     

Induction of neonatal tolerance to the Mls-1a self-super-antigen. Time kinetics and MHC restriction.

We examined the accessibility of the thymus to a self-super-Ag encoded by the Mls-1a region of chromosome 1 and the process by which this Ag establishes immunologic tolerance. Intravenously administered Mls-1a Ag accumulates quickly in peripheral organs of adult or newborn Mls-1a- recipients, where it mounts an immune response. The Ag does not enter the thymus in detectable amounts and does not induce an immune response of Mls-1a-responsive T cells present in this organ. Instead, the thymus of newborn Mls-1a- recipients of Mls-1a+ lymphoid cells continues for several days to export Mls-1a-reactive T cells, which respond to Mls-1a Ag when they encounter it in peripheral organs. This response peaks around day 3 or day 4 and declines very rapidly thereafter. The deletion of intrathymic Mls-1a-reactive T cells ensues simultaneously with this decline. It has previously been shown that Mls-1a Ag causes deletion or anergy of Mls-1a-reactive peripheral T cells, subsequent to their activation. We see the same time kinetics in producing deletion or anergy of Mls-1a-reactive T cells in the thymus of newborn animals, with the exception that the activation phase that precedes the deletion of Mls-1a-reactive T cells occurs in the periphery and not in the thymus. This observation indicates that thymic Mls-1a-specific T cells are not deleted through activation. Whether their deletion depends on a feed-back from the peripheral activation of Mls-1a-reactive cells, as the time relationship could suggest, is not clear. The finding establishes, however, that the deletion of functionally mature Mls-1a-reactive T cells and the activation of such cells are not necessarily related events, which may or may not utilize a common trigger mechanism, such as the engagement of the TCR. Concerning the trigger mechanism, we report that Mls-1a-specific deletion of T cells is an MHC-restricted process, whereas Mls-1a-specific activation of T cells is not MHC restricted.     

Plasticity of T cell function. Cloned EL-4 lymphoma cells may help or suppress a primary antibody response depending on their own concentration and the assay system

Cloned EL-4 lymphoma T cells were tested in limiting dilution experiments for their capacity to suppress or to help the primary humoral immune response of spleen cells (or T cell-depleted spleen cells) to the Ag SRBC and 4-hydroxy-3-iodo-5-nitro-phenyl-keyhole limpet hemocyanin. EL-4 clones are able to suppress up to 80% of the total IgM responses in both systems, as well as to help. Suppression and help fluctuate between high and low levels with the numbers of EL-4 cells placed into tissue cultures. In Poisson plots, this is reflected as a "typical curve": usually one or two frequencies are estimated (e.g., integral of 1/4 and approximately 1/1000 for suppression and approximately 1/6 and approximately 1/200 for help), which appear regulated with increasing numbers of cells seeded. Control experiments showed that EL-4 cells need to be alive to exert the effects. EL-4 cells do not serve as additional antigen, do not induce an isotype switch and are not cytotoxic. Help and suppression are not restricted by the MHC. Help requires the presence of a small number of normal T cells in the assay system, indicating that EL-4 cells do not replace specific helper T cells. When a number of control cell lines were analyzed under identical circumstances, similar effects were observed with most long term T cell lines or clones expressing a T cell receptor, whereas cells of non-T lineages and T cells not expressing a T cell receptor did not show the phenomena. The results suggest a functional plasticity of T cells, dependent on cell numbers and the assay system used, and expressed via T cell communication.     

Responses of spleen cells from mice with X-linked B-cell defect to polyclonal B-cell activators, purified protein derivative of tuberculin, and dextran sulfate

Polyclonal responses to LPS, PPD, and DxS of spleen cells from mice expressing a X-linked B-cell defect were examined. Spleen cells from young (CBA/N × BALB/c)F₁ male mice responded slightly lower to LPS, significantly lower to PPD than the cells from age-matched F₁ female mice, and showed no response to DxS stimulation. This hypo- or unresponsiveness of F₁ male cells to PPD or DxS could not be explained by a shift in the dose-response or time kinetics of the responding cells, and also could not be due to the defect in the function of T cells or macrophages. Suppressor T cells to polyclonal response to PPD or DxS could not be shown in F₁ male spleen cells. The response of F₁ male cells to PPD was dramatically improved with age but not to DxS. These results suggest that B cells responsive to DxS may belong to a distinct subpopulation from the cells responsive to LPS or PPD.     

Oncogene cooperation in lymphocyte transformation: malignant conversion of E mu-myc transgenic pre-B cells in vitro is enhanced by v-H-ras or v-raf but not v-abl.

Although transgenic mice bearing a c-myc gene controlled by the immunoglobulin heavy-chain enhancer (E mu) eventually develop B-lymphoid tumors, B-lineage cells from preneoplastic bone marrow express the transgene but do not grow autonomously or produce tumors in mice. To determine whether other oncogenes can cooperate with myc to transform B-lineage cells, we compared the in vitro growth and tumorigenicity of normal and E mu-myc bone marrow cells infected with retroviruses bearing the v-H-ras, v-raf, or v-abl oncogene. The v-H-ras and v-raf viruses both generated a rapid polyclonal expansion of E mu-myc pre-B bone marrow cells in liquid culture and 10- to 100-fold more pre-B lymphoid colonies than normal in soft agar. The infected transgenic cells were autonomous, cloned efficiently in agar, and grew as tumors in nude mice. While many pre-B cells from normal marrow could also be induced to proliferate by the v-raf virus, these cells required a stromal feeder layer, did not clone in agar, and were not malignant. Most normal cells stimulated to grow by v-H-ras also cloned poorly in agar, and only rare cells were tumorigenic. With the v-abl virus, no more cells were transformed from E mu-myc than normal marrow and the proportion of tumorigenic pre-B clones was not elevated. These results suggest that both v-H-ras and v-raf, but apparently not v-abl, collaborate with constitutive myc expression to promote autonomous proliferation and tumorigenicity of pre-B lymphoid cells.     

Histochemistry and morphology of porcine mast cells

Summary Mast cells have been described extensively in rodents and humans but not in pigs, and the objective of this study was to characterize porcine mast cells by histochemistry and electron microscopy. Carnoy's fluid proved to be a good fixative but fixation with neutral buffered formalin blocked staining of most mast cells. Alcian Blue stained more mast cells than did Toluidine Blue (pH 0.5), although Alcian Blue also stained goblet cells. In pigs, unlike rodents, the Alcian Blue method did not distinguish between mast cells in the intestinal mucosa and those in the connective tissue of the intestinal submucosa, tongue and skin. Mast cells were significantly larger in adult pigs than in piglets; in adult pigs and piglets, mast cells in the intestinal mucosa were significantly larger than those in submucosal connective tissue, and they were more varied in shape in piglets and adults. Granules in mast cells in the intestinal mucosa stained less intensely than those in mast cells in connective tissue of tongue, skin and intestinal submucosa. Mast cells in the connective tissue of the tongue, skin and intestinal submucosa fluoresced strongly when stained with berberine sulphate or with a mixture of berberine sulphate and Acridine Orange, but mast cells in the intestinal mucosa did not. All mast cells reacted positively in an enzyme-histochemical method previously used to detect human tryptase but not in a method previously used to detect human chymase. Mast cells in the medulla of thymus stained similarly to mast cells in the intestinal mucosa. Ultrastructural differences between mast cells were not detected.     

Lipopolysaccharide (LPS) and zymosan-resistant mutant isolated from a macrophage-like cell line, WEHI-3, with a defective response to LPS under serum-free conditions

A LPS-resistant mutant, W3SF-1, was isolated from a murine macrophage-like cell line, WEHI-3. The W3SF-1 mutant did not produce a significant amount of nitric oxide (NO) or TNF-alpha even with high concentrations of LPS in the presence or absence of FCS, whereas the parental WEHI-3 cells produced them in response to LPS. The parental cells expressed a significant level of TNF-alpha mRNA after LPS stimulation, whereas the mutant cells did not. This defective response of the mutant cells to LPS was neither dependent on the concentration or chemical structure of LPS, nor on the time of LPS treatment. The mutant cells also showed a defective response to zymosan, suggesting that the defect in the mutant cells is common to LPS and zymosan in the signal transduction pathways. The parental and mutant cells showed similar levels of Mac1, F4/80 and CD14, suggesting that these surface markers of macrophages are not linked directly to the defective responses of the mutant to LPS. The treatment of mutant cells with IFN-gamma did not restore the defect of NO or TNF-alpha production on LPS treatment. Binding experiments with 125I-labelled LPS showed a similar binding affinity for LPS in the parental and the mutant cells. These results suggest that the defect in the W3SF-1 mutant cells may not reside in the LPS binding but rather in the early step of signal transduction pathways in the cells after LPS binding.     

The control of avian cephalic neural crest cytodifferentiation. II. Neural tissues.

A series of neural crest transplantations has been performed to (1) analyze whether avian premigratory cranial neural crest cells are pluripotential or restricted to specific developmental pathways and (2) examine the ability of trunk neural crest cells to develop in an environment usually occupied by cranial crest cells. Quail embryos, the cells of which have a unique nuclear marker, were used as donors and chick embryos as hosts. Hindbrain crest cells grafted in the place of diencephalic crest cells failed to form neurons in all but one case, in which a small ectopic ganglion was found. In the reciprocal transplants, neural crest cells emigrating from a segment of forebrain crest tissue grafted in the place of metencephalic crest cells produced trigeminal and ciliary ganglia which were completely normal. Thus, crest cells which normally never form ganglionic neurons will do so if placed in a suitable neurogenic environment. These results prove that premigratory avian cranial crest cells are not restricted to specific developmental pathways, but are initially pluripotential. Trunk crest cells grafted in the place of metencephalic crest cells form neuronal ganglia along the proximal trigeminal motor roots but do not form normal trigeminal ganglia. These root ganglia do not display normal peripheral projections, and placode cells, a normal component of the trigeminal ganglion, form ganglia in ectopic locations. Thus, while trunk crest cells respond to the metencephalic environment and form neurons, their response is different from that of cranial crest cells in the same location. Whether this is due to differences in developmental potential or in initial population size is not known.     

wnt3a but not wnt11 supports self-renewal of embryonic stem cells

wnt proteins (wnts) promote both differentiation of midbrain dopaminergic cells and self-renewal of haematopoietic stem cells. Mouse embryonic stem (ES) cells can be maintained and self-renew on mouse feeder cell layers or in media containing leukemia inhibitory factor (LIF). However, the effects of wnts on ES cells self-renewal and differentiation are not clearly understood. In the present study, we found that conditioned medium prepared from L cells expressing wnt3a can replace feeder cell layers and medium containing LIF in maintaining ES cells in the proliferation without differentiation (self-renewal) state. By contrast, conditioned medium from NIH3T3 cells expressing wnt11 did not. Alkaline phosphatase staining and compact colony formation were used as criteria of cells being in the undifferentiated state. ES cells maintained in medium conditioned by Wnt3a expressing cells underwent freezing and thawing while maintaining properties seen with LIF maintained ES cells. Purified wnt3a did not maintain self-renewal of ES cells for prolonged intervals. Thus, other factors in the medium conditioned by wnt3a expressing cells may have contributed to maintenance of ES cells in a self-renewal state. Pluripotency of ES cells was determined with the use of embryoid bodies in vitro. PD98059, a MEK specific inhibitor, promoted the growth of undifferentiated ES cells maintained in conditioned medium from wnt3a expressing cells. By contrast, the P38 MAPK inhibitor SB230580 did not, suggesting a role for the MEK pathway in self-renewal and differentiation of ES cells maintained in the wnt3a cell conditioned medium. Thus, our results show that conditioned medium from wnt3a but not wnt11 expressing cells can maintain ES cells in self-renewal and in a pluripotent state.     

Growth factors and hyperthermia. II. Viability of Chinese hamster ovary HA-1 cells during serum starvation and hyperthermia

We tested the possibility that hyperthermia kills HA-1 cells in a manner analogous to growth factor deprivation. HA-1 cells were inactivated by serum starvation when incubated in Eagle's MEM at a density of 40 cells/cm2 or less. Cells became resistant to the absence of serum when the cell density was greater than 400 cells/cm2 or when lethally irradiated HA-1 feeder cells were present. The feeder cells exerted their effect through a diffusible factor. In addition, a 1:1 mixture of Eagle's MEM and Ham's F-12 enabled HA-1 cells to remain viable without serum. Ten days growth in Eagle's MEM + Ham's F-12 without serum resulted in the formation of microcolonies of cells. This indicated that growth factor deprivation was not lethal to HA-1 cells, and it suggested that they may have been partially transformed. The presence of the growth factors insulin, transferrin, and fibroblast growth factor (FGF) reduced cell killing by a small amount during conditions of serum starvation. After hyperthermia, the presence of growth factors again diminished cell killing by a modest amount (approximately twofold). Feeder cells also improved cell survival after hyperthermia. The effect of feeder cells was greatest when cells were trypsinized immediately after hyperthermia. When cells were not trypsinized after heating, feeder cells increased survival less than twofold. In summary, the absence of growth factors was not lethal to HA-1 cells, and therefore the cytotoxic effects of hyperthermia could not be explained fully by the failure to bind growth factors. HA-1 feeder cells secreted undefined, growth-promoting substances, but feeder cells exerted only a small positive effect on cell survival after hyperthermia when cells were not trypsinized after heating.