DIF-1 is a putative morphogen that induces stalk cell formation in the cellular slime mold Dictyostelium discoideum. We have previously discovered that DIF-1 suppresses cell growth and induces cell differentiation in vitro in some tumor cells, and also that relatively low concentrations of DIF-1 promote retinoic acid-induced cell differentiation in the human myeloid leukemia HL-60 cells. In this study, to verify cell biological and therapeutic potential of DIF-1, we have examined whether and how DIF-1 affects normal mammalian cells in vitro, using rat leptomeningeal (RLM) cells and rat gastric mucosal RGM-1 cells. In growing phase of both cells, DIF-1 at 5–40 μM suppressed cell growth in a dose-dependent manner. High concentrations (15–40 μM) of DIF-1 were toxic to the growing cells so that the cells showed unusual morphology, but many of them were still alive even at Day 3–4. Withdrawal of DIF-1 allowed the cells to grow. In confluent phase of the cells, DIF-1 at more than 15 μM promoted medium acidification that resulted in cell death, but DIF-1 itself did not markedly affect cell viability for 3 days. DIF-1 increased [Ca2 ]i in RLM cells but did not affect β-trace secretion (an index of cell function in RLM cells). A low concentration (5 μM) of DIF-1 in combination with retinoic acid (0.1 μM) showed no marked effects on cell growth, cell viability, cell morphology and β-trace secretion in RLM cells. The present results indicate that DIF-1 may be utilized as a tool for cell biology. Also, since these concentrations of DIF-1 kill some tumor cells within 2–3 days, the present results suggest that DIF-1 might be utilized in the treatment of some sorts of cancer with a bearable degree of side-effects.
The cellular slime mold Dictyostelium discoideum is a simple eukaryote that forms a fruiting body consisting of spores and stalk cells at the end of its development. DIF-1 (differentiation-inducing factor-1), 1-(3,5-dichloro-2,6-dihydroxy-4-methoxyphenyl)-1-hexanone (Fig. 1) (Morris et al., 1987), is a putative morphogen that induces stalk cell formation (Kay et al., 1989). This signal molecule has been suggested to function at least partly via an increase in [Ca2+]i in D. discoideum (Kubohara and Okamoto, 1994; Schaap et al., 1996; Azhar et al., 1997).
Recently, it was found that DIF-1 suppresses cell growth and induces erythroid differentiation in murine and human leukemia (B8 and K562) cells (Asahi et al., 1995). It was also found that DIF-1 raises [Ca2+]i and suppresses cell growth in rat pancreatic tumor, AR42J cells (Kubohara et al., 1995b), and that high concentrations of DIF-1 induce apoptosis in the tumor cells (Kubohara et al., 1995a). Moreover, it was shown with the human myeloid leukemia, HL-60 cells (Collins et al., 1977), that DIF-1 suppresses cell growth at least partly by arresting the cells in G1 phase of cell-cycle, and that relatively low concentrations of DIF-1 promote retinoic acid-induced cell differentiation (Kubohara, 1997). These findings indicate that DIF-1 may have therapeutic potential in the treatment of a wide range of cancer. In particular, since retinoic acid has already been utilized in differentiation-therapy for human acute promyelocytic leukemia (APL) (Huang et al., 1988), co-administration of retinoic acid and DIF-1 (or its artificial analogs) may improve chemoprevention/differentiationtherapy for APL and the other types of leukemia. However, it has not been examined if/how DIF-1 affects normal mammalian cells. Detailed analyses on the effects of DIF-1 on normal cell functions should be necessary for the therapeutic use of this agent. It should also be of great importance to verify the cell biological potential of DIF-1 in many types of cells, in terms of basic cell science.
In our laboratory, Ishikawa and co-workers have established an in vitro primary culture system with rat leptomeningeal cells and have been analyzing the functions of the cells (Ishikawa et al., 1995a,b). Leptomeninges are the cerebrospinal fluid-blood barrier covering the brain parenchyma (Smith and Shine, 1992; Tanno et al., 1993). In vitro studies have revealed that these cells secrete many interesting proteins such as β-trace protein (prostaglandin D synthase), IGF-II, IGF-binding protein-2, apolipoprotein E, β2-microglobulin and many others (Ishikawa et al., 1995b; Ohe et al., 1996) which are seen in cerebrospinal fluid in vivo, suggesting that meningeal cells play pivotal roles at least as an origin of the cerebrospinal fluid proteins. Since the meningeal cells can be cultured in vitro and also because the assay system for cell biological function (β-trace secretion) have been established (Kabeya et al., 1998), these cells should provide an ideal in vitro system suitable for the analysis of the effects of DIF-1 on normal cell functions.
RGM-1 is an epithelial cell line established from gastric mucosa of adult Wistar rats and is well characterized (Hassan et al., 1996; Okada et al., 1996). This cell line also provides a useful model system for the analysis of the effects of DIF-1 on normal cells.
In this study, in order to verify both cell biological and therapeutic potential of DIF-1 and thus to elucidate if/how this agent affects normal cells, we examined the in vitro effects of DIF-1 on RGM-1 and rat meningeal cells. We show here that DIF-1 suppresses cell growth but is less toxic to the normal cells than to some tumor cells. It is also shown that cell growth, viability and β-trace secretion were not markedly affected in the presence of both retinoic acid and a low concentration of DIF-1, an expected combination for leukemia therapy. The potential of DIF-1 in mammalian cells will be discussed.
MATERIALS AND METHODS
DIF-1 was purchased from both Molecular Probes (USA) and Affiniti Research Products (UK) and stored as a 10 mM solution in EtOH at –20°C. All-trans retinoic acid was from Wako Pure Chemical Industries (Osaka, Japan) and stored as a 100 mM solution in EtOH at –20°C.
Examination of the effects of DIF-1 on growing cells
Rat leptomeningeal (RLM) cells and rat gastric mucosa RGM-1 cells were used in this study. Meningeal cells were prepared from the brains of 1–2-day-old rats as described previously (Ishikawa et al., 1995a,b; Ohe et al., 1996). The RLM cells and RGM-1 cells were plated on poly-D-lysine-coated multi(12)-well plates at a density of 2–4 × 103 cells/cm2, and incubated with 1 ml of DMEM-FBS (Dulbecco's modified Eagle's medium containing 10% fetal bovine serum) at 37°C (5% CO2) for 1–2 days. The incubation media were replaced each with 1ml of fresh DMEM-FBS containing 0–40 μM DIF-1 and/or 0.1 μM RA, and incubated at 37°C (5% CO2). Cell growth and morphology were observed by a phase-contrast microscope. At 24 hr of incubation, cell density was counted; living cell numbers in 3–4 different microscopic fields of each well were counted and the mean of the numbers was used as a cell density. Round cells detached from the bottom of the wells were regarded as dead cells.
Examination of the effects of DIF-1 on confluent state of cells and β-trace secretion
Cells were incubated at 37°C (5% CO2) in multi(12)-well plate as described above for 5–7 days until they reached a confluent state. The incubation media were replaced each with 1ml of fresh DMEMFBS containing 0–40 μM DIF-1 and/or 0.1 μM RA, and cells were incubated at 37°C (5% CO2) continuously, or during incubation, the media were replaced with fresh ones containing the reagents at every 24 hr. For the assay for β-trace secretion, confluent RLM cells were incubated with DIF-1/RA-containing media, and the incubation media were collected and replaced with 1 ml of fresh DMEM-FBS containing the reagents at every 12 or 24 hr. β-trace contents of the media thus collected were assayed using anti-β-trace protein antibody as described by Kabeya et al. (1998).
Assay for [Ca2+]i
RLM cells were incubated in culture flasks until cells reached a confluent state. The cells were collected by trypsin treatment (Ohe et al., 1996), and incubated with FuraPE3/AM (Wako Pure Chemical Industries, Osaka, Japan). FuraPE3 fluorometry was performed as described previously (Kubohara, 1997).
Effects of DIF-1 on RGM-1
We first examined the effects of DIF-1 on growing phase of the rat gastric mucosa cell line RGM-1 (Figs. 2 and 3). In the absence of DIF-1, RGM-1 cells grew as shown in Fig. 3 (Control) and reached a fully confluent state in 3–4 days. In the presence of 5 μM DIF-1 (Fig. 3), cell growth and cell morphology were not much affected so that cells reached a confluent state as control cells did. DIF-1 at 10 μM suppressed cell growth slightly but did not affect cell morphology (Fig. 3). Cell growth was greatly suppressed in the presence of 20–40 μM of DIF-1 where the cells showed characteristic morphology; there appeared some flattened vesicle-containing cells and bi-polar cells (Fig. 3). The cells started growing after withdrawal of DIF-1 (data not shown), which indicates that the effect of DIF-1 is reversible.
We then examined the effects of DIF-1 on cell viability and morphology in a confluent state of RGM-1 cells. In the absence of DIF-1, cells kept confluent and they were alive at Day 3 (Fig. 4); cell morphology did not change for more than a week (data not shown). In the presence of 5–15 μM DIF-1, they showed normal morphology at Day 3 (data not shown). In contrast, confluent cells were killed within 3 days when 30 μM of DIF-1 was present in the media (Fig. 4). However, repeated experiments showed that in the presence of 30 μM DIF-1, the culture media became yellow (indication of acidification) before cells died (data not shown). This seemed to indicate that DIF-1 would promote cell metabolism in some way resulting in the accumulation of metabolites and the acidification of the culture media, which might cause cell death; that is, the cytotoxic effect of DIF-1 was exerted at least in part via the acidification of the media and/or the accumulation of cytotoxic metabolites. To assess this possibility, the culture media were exchanged with fresh ones containing DIF-1 at every 24 hr and cell viability was observed. As expected, cells were viable at Day 3 even in the presence of 30 μM of DIF-1 (Fig. 4).
Effects of DIF-1 on rat leptomeningeal cells
The effects of DIF-1 on growing phase of the rat leptomeningeal (RLM) cells were examined (Figs. 2 and 5). In the absence of DIF-1, RLM cells grew as shown in Fig. 5 (control) and reached a full confluent state in 3–4 days. In terms of the inhibitory effect of DIF-1 on cell growth, we obtained almost the same results with RLM cells as seen with RGM-1 (Figs. 2, 3 and 5). Again, withdrawal of DIF-1 allowed the cells to grow (data not shown). It should be noted however that in the presence of 20–40 μM DIF-1, there appeared many bi-polar cells with long cytoplasmic projections distinct from normal (Fig. 5).
We then examined the effects of DIF-1 on cell viability and cell morphology in confluent RLM cells. We obtained almost the same results as with RGM-1, except that DIF-1 at 15 μM was much toxic to RLM cells (data not shown) than to RGM-1 (Fig. 4); confluent RLM cells were killed by continuous incubation with 15–30 μM DIF-1. But this toxicity was again reduced by medium exchange (data not shown).
Combinatory effects of DIF-1 and retinoic acid on rat leptomeningeal cells
Retinoic acid (RA) has been used in differentiation-therapy for APL (Huang et al., 1988). Recently, we have shown with myeloid leukemia HL-60 that relatively low concentrations of DIF-1 promote RA-induced cell differentiation, suggesting that co-administration of RA and DIF-1 may improve differentiation-therapy for leukemia (Kubohara, 1997). We therefore examined the effects of RA and DIF-1 on RLM cells (Figs. 2 and 5). RA at 0.1 μM showed no marked effects on both cell growth and morphology in RLM (Figs. 2 and 5). RA (0.1 μM) in combination with 5 μM DIF-1 slightly suppressed cell growth but did not affect cell morphology in RLM (Figs. 2 and 5).
Neither RA (0.1 μM) nor the combination of RA (0.1 μM) and DIF-1 (5 μM) affected cell viability and morphology in confluent RLM cells at least for 3 days (data not shown).
Effects of DIF-1 on β-trace secretion and [Ca2+]i
To assess if DIF-1 affects normal cell functions, we examined the effects of DIF-1 on the secretion of β-trace protein in confluent RLM cells, using the medium-exchange-system described above. When the culture media were exchanged at every 24 hr, low concentrations (5–10 μM) of DIF-1 and/or RA (0.1 μM) did not affect β-trace secretion but high concentrations (20–30 μM) of DIF-1 inhibited β-trace secretion (data not shown). In this case however, since the culture media had become yellow in 24 hr in the presence of high concentrations of DIF-1, it seemed possible that the medium acidification and/or the accumulation of some metabolites affected β-trace secretion although the cells were alive. We therefore assessed β-trace secretion by exchanging the culture media with fresh ones at every 12 hr (Fig. 6). Under the conditions, DIF-1 and/or RA did not affect β-trace secretion at least for 2 days (Fig. 6).
In order to elucidate the mechanism of DIF-1-action in RLM cells, we examined the effect of DIF-1 on [Ca2+]i. Expectedly, DIF-1 (20 μM) increased [Ca2+]i in RLM cells, suggesting that DIF-1 might exert its effects at least in part via an increase in [Ca2+]i (Fig. 7).
We have previously shown that DIF-1, a putative morphogen in D. discoideum, has anti-tumor activities in vitro in several tumor cells tested so far (Asahi et al., 1995; Kubohara et al., 1995a,b; Kubohara, 1997). However, it has not been verified whether and how DIF-1 affects normal cells. In the present study, to assess this question, we have examined the effects of DIF-1 on normal cell functions, using RGM-1 and RLM cells; in this study, we have observed the DIF-1-treated cells for 3–4 days, since DIF-1 shows cytotoxic effects on tumor cells within 2–3 days (Kubohara et al., 1995a,b; Kubohara, 1997). Table 1 summarizes the present and previous results on the effects of DIF-1.
Effects of DIF-1 and retinoic acid on RLM and RGM-1 cells
Effects of DIF-1 on normal cells
DIF-1 at 5–40 μM suppressed cell growth in a dose-dependent manner (Figs. 2, 3 and 5) in both RLM and RGM-1 cells. High concentrations (20–40 μM) of DIF-1 affected cell morphology but the cells survived at Day 3–4; there appeared many bi-polar cells with long cytoplasmic projections, indicating that DIF-1 induced re-arrangement of cytoskeletal proteins possibly via a signalling system (Figs. 3 and 5). Characterization of the bi-polar cells should be of interest for the elucidation of DIF-1-signalling system (if any) in the mammalian cells.
In confluent RLM and RGM-1 cells, high concentrations of DIF-1 promoted the acidification of the culture media, which might result in cell death. Although it is unknown how DIF-1 promotes medium acidification and what triggers cell death under the in vitro conditions, it should be of great importance to note that both RLM and RGM-1 cells survived at least for 3 days (Fig. 4 and not shown data) and RLM cells secreted β-trace proteins normally when the culture media were exchanged with fresh ones (Fig. 6). These results would rather reflect in vivo situation where the blood-lymph systems circulate continuously, and the present results thus lead to an expectation that DIF-1 may not seriously damage normal confluent cells in vivo.
Considering that the high concentrations (20–40 μM) of DIF-1 kill human myeloid leukemia HL-60 (Kubohara, 1997) and rat pancreatic tumor AR42J cells (Kubohara et al., 1995a,b) within 2–3 days, DIF-1 seems to be much more toxic to the tumor cells than to normal cells (Table 1).
It seems likely that DIF-1 functions via similar machinery involving increases in [Ca2+]i in both normal cells (Fig. 7) and tumor cells (Kubohara et al., 1995a,b; Kubohara, 1997) as well as in Dictyostelium cells (Kubohara and Okamoto, 1994; Schaap et al., 1996; Azhar et al., 1997). However, the fine mechanisms of DIF-1-action in both mammalian and Dictyostelium cells are still unclear.
Combinatory effects of retinoic acid and DIF-1 on normal cells
RA has been used in differentiation-therapy for APL which results in a high score of complete remission (Huang et al., 1988). We have recently shown with human myeloid leukemia HL-60 cells that DIF-1 promotes RA-induced cell differentiation (Kubohara, 1997), suggesting that co-administraion of RA and DIF-1 may improve chemoprevention/differentiation therapy for APL and other leukemia. In this study, we have therefore verified the effects of DIF-1/RA on the normal cells.
RA (0.1 μM) or DIF-1 (5 μM) alone showed no marked effect on all the indexes examined here (Table 1). DIF-1 (5 μM) in combination with RA slightly suppressed cell growth, but did not affected cell viability, morphology (Figs. 2 and 5), and β-trace secretion (Fig. 6) in RLM cells (Table 1). These results suggest that RA and DIF-1 may not damage normal cells.
Most importantly, it was suggested here that DIF-1 alone or DIF-1 in combination with RA might be utilized in the treatment of some sorts of cancer cells with a bearable degree of side-effects, although the in vivo effects of DIF-1 should be examined in the future. Apart from the therapeutic potential of DIF-1, since DIF-1 affected cell growth and morphology in RGM-1 and RLM cells, it is also expected that DIF-1 may be a useful tool for the analysis of normal mammalian cell functions.
We should like to thank Ms. Kazue Takagi for skillful technical assistance. This work was supported in part by Grants (Y.K. and K.I.) from the Ministry of Education, Science, Sports and Culture of Japan.