QUARTERLY

Synthetic analogs of vitamin D for potential use in differentiation therapy should selectively regulate genes necessary for differentiation without inducing any perturbations in calcium homeostasis. PRI-1906, an analog of vitamin D2, and PRI-2191, an analog of vitamin D3 bind nuclear vitamin D receptor (nVDR) with substantially lower affinity than 1,25-dihydroxyvitamin D3 (1,25-D3), but have higher differentiation-inducing activity as estimated in HL-60 leukemia cellmodel. To examine how their increased differentiation-inducing activity is regulated we tested the hypothesis that membrane-mediated events, unrelated to nVDR, take part in the differentiation in response to PRI-1906 and PRI-2191. The induction of leukemia cell differentiation in response to the analogs of vitamin D was inhibited by LY294002 (phosphatidylinositol 3-kinase inhibitor), PD98059 (inhibitor of MEK1,2, an upstream regulator of extracellular-signal regulated kinase) and rapamycin (p70S6K inhibitor) pointing out that activation of signal transduction pathways unrelated to nVDR is necessary for differentiation. On the other hand, inhibition of cytosolic phospholipase A2 accelerated the differentiation of HL-60 cells induced by either 1,25-D3 or by the vitamin D analogs suggesting possible existence of a feedback loop between extracellular-signal regulated kinases and phospholipase A2.

The main function of 1,25-dihydroxyvitamin D 3 (1,25-D 3 ) is regulation of calcium homeostasis of the organism.This compound stimulates intestinal calcium absorption and bone calcium mobilization.Since 1,25-D 3 cooperates with parathyroid hormone in a precisely regulated manner it is now considered to be a hormone rather than a vitamin (Jones et al., 1998).In the past two decades great progress has been made in understanding the role of 1,25-D 3 in differentiation of various cells, such as osteoclasts (Suda et al., 1995), keratinocytes (Fogh & Kragballe, 2000) and monocytes (Abe et al., 1981).It is presently well documented that it also induces differentiation of leukemia cells from patients and leukemia cell lines (Munker et al., 1996).
The mechanism of action of 1,25-D 3 is rapidly being uncovered, however, it is still not fully understood.It is believed that in order to mediate its biological responses 1,25-D 3 must bind to a specific nuclear receptor protein, named nVDR (nuclear vitamin D receptor).Activated nVDR acts as a ligand-activated transcription factor and either upregulates or downregulates transcription of target genes (Jones et al., 1998).However, during the past years experimental evidence for "nongenomic" signalling has been provided.It is well documented that 1,25-D 3 rapidly stimulates calcium transport in the perfused chick duodenum in vivo and formation of second messengers such as ceramides, inositols, cAMP and calcium and activation of a variety of protein kinases in the target cells in vitro (Norman et al., 1999).It is generally accepted now that 1,25-D 3 generates its biological effects by these two mechanisms.
The existence of a separate VDR located in the cell membrane (mVDR) responsible for the "nongenomic" signalling was postulated many years ago (Nemere & Szego, 1981), but such receptor has not been cloned as yet.A putative mVDR was isolated from chick enterocytes and was further biochemically described (Nemere et al., 1994;1998;Pedrozo et al., 1999;Nemere & Campbell, 2000), but the evi-dence that this particular receptor is responsible for all "nongenomic" signals is lacking.It should be noted that the physiological importance of the "nongenomic" signalling and its relation to nVDR-mediated signalling is not easy to elucidate.The open questions about membrane receptors are the same for all steroid hormones acting in animals (Pietras et al., 2001).
As already mentioned, 1,25-D 3 induces differentiation of myeloid leukemia cells, such as HL-60, U937, THP-1 and M1.These cells, when incubated with 1,25-D 3 , acquire the functional properties and express cell surface differentiation markers of monocytes (Brackman et al., 1995;Oberg et al., 1993;Schwende et al., 1996;Abe et al., 1981).Differentiation therapy has attracted many researchers, but undesirable side effects of natural 1,25-D 3 caused that its new analogs have been synthesized for this purpose.Analogs that could potentially be useful for differentiation therapy are deprived of the calcemic activity of 1,25-D 3 , but have an even stronger differentiation-inducing potential.A series of side-chain modified analogs was studied in the past with respect to their differentiation-inducing and antiproliferative activity.Our previous studies revealed that some analogs with extended side-chain are either comparably effective (Marcinkowska et al., 1998a) or even more effective (Opolski et al., 1999) than 1,25-D 3 in the induction of leukemia cell differentiation.Surprisingly, the effectiveness of a particular analog in the induction of cell differentiation is not related to its affinity for nVDR.Further studies revealed that 1,25-D 3 within minutes activates extracellular-signal regulated protein kinases (erk) 1 and 2 and induces their translocation to the nucleus in serum-starved HL-60 leukemia cells (Marcinkowska et al., 1997).Pharmacological inhibition of the MEK/erk1,2 signal transduction pathway by PD98059, as well as inhibition of nuclear translocation by thapsigargin, reduce the cell-differentiating effects of 1,25-D 3 (Marcinkowska et al., 1998b;Marcinkowska, 2001).Another study revealed that also phosphatidylinositol 3-kinase (PI 3-K) is rapidly activated in human leukemia cells by 1,25-D 3 with the kinetics similar to that of erk1 and erk2 kinases (Hmama et al., 1999).Inhibition of PI 3-K by either of the two specific inhibitors results in a concentration-dependent repression of the differentiation process.Similar effect of inhibition of 1,25-D 3 -induced differentiation may be obtained with the use of an inhibitor of p70 S6K protein kinase, namely with rapamycin (Marcinkowska et al., 1998b).Possibly p70 S6K is downstream of PI 3-K in the same signal transduction pathway, but further studies are necessary to reveal if this is the right explanation and to verify if p70 S6K is actually rapidly activated by 1,25-D 3 .
The erk1 and erk2 kinases that are activated by 1,25-D 3 belong to the larger family of mitogen-activated protein kinases (MAPKs).Different members of this protein kinase family are regulated by feedback loops in order to maintain precise levels of transcription factors activity, that eventually allow the cell to proliferate, differentiate or die.As was reported in a separate paper, inhibition of p38 kinase, another member of the MAPK family, potentiates the 1,25-D 3 -induced differentiation through prolonged activation of erk1 and erk2 (Wang et al., 2000).The data summarized above point out that "nongenomic" sig-nalling is important for monocytic differentiation of leukemia cells.
For the studies presented in this paper we have selected side-chain modified analogs of vitamin D 2 (PRI-1906) and D 3 (PRI-2191) presented in Fig. 1, which are the most effective ones in inducing differentiation of the set studied by our group (Opolski et al., 1999).These analogs are deprived of the calcemic activity of natural 1,25-D 3 (Chodyñski et al., 1997;Hansen et al., 2000), possibly because of their lowered affinity for nVDR (Ikekawa & Ishizuka, 1992).It is therefore very probable that the increased differentiation-inducing activity of our analogs is mediated by mechanisms unrelated to nVDR.Thus, we wanted to find out if membrane-mediated activation of signal transduction pathways is necessary for the leukemia cell differentiation induced by PRI-1906 and by PRI-2191.Here we show that not only the 1,25-D 3 -induced leukemia cell differentiation, but also differentiation induced by synthetic analogs of vitamin D requires activation of both the PI 3-K and erk1,2 signal transduction pathways.We also report that cytosolic phospholipase A 2 (cPLA 2 ) is involved in the 1,25-D 3 -induced the differentiation and in the differentiation induced by vitamin D analogs, but in a negative manner.Differentiation was screened in our experiments by acquisition of the CD11b and CD14 cell surface markers (McCarthy et al., 1983).CD11b is the a subunit of CR3 integrin that associates noncovalently with a b subunit partner, CD18.CD11b/CD18 is present on the surface of mature myeloid cells (Remold-O'Donnell, 1988).CD14 is another differentiation marker of monocytes.This protein is involved in responses to bacterial cell wall products (Wright et al., 1990).

MATERIALS AND METHODS
Cell cultures.Human promyelocytic leukemia HL-60 cell line was obtained from the European Type Culture Collection.The cells were maintained in a suspension culture in RPMI1640 medium supplemented with 10% fetal calf serum (FCS, North American Origin; GIBCO BRL, European Division), 100 units/ml penicillin and 100 mg/ml streptomycin (both Polfa, Poland) and kept at standard cell culture conditions, i.e. humidified atmosphere of 95% air and 5% CO 2 at 37°C.
Chemicals.1,25-D 3 was obtained from the Pharmaceutical Research Institute (Warsaw, Poland).The analogs PRI-1906 and PRI-2191 were synthesized by the methods described earlier1 (Chodyñski et al., 2000).The compounds were aliquoted and stored in glass ampoules under argon at -20°C.The amount of the compound in an ampoule was determined by UV spectrometry at 264 nm.Prior to use, the compound was dissolved in absolute ethanol to 100 mM and subsequently diluted in culture medium to the required concentration.
Inhibitors.PD98059, LY294002 and AACOCF 3 were from Calbiochem (San Diego, CA, U.S.A.).Rapamycin was from Sigma (St. Louis, MO, U.S.A.).Inhibitors were dissolved in Me 2 SO, aliquoted and kept at -20°C.Before experiments the inhibitors were thawed and diluted in culture medium to the required concentrations.
Antibodies.CD11b and CD14 monoclonal antibodies, mouse IgG1 and IgG2a fractions, all FITC conjugated, were from Sigma (St. Louis, MO, U.S.A.).HL-60 differentiation assay.Cells were seeded in culture medium (supplemented with 10% FCS) on 24-well plates (Costar Cambridge, MA, U.S.A.) in a final volume of 1 ml.The cells were exposed for different times to 1,25-D 3 , PRI-1906 or PRI-2191 alone or together with inhibitors.After completion of the exposure time the cells were collected by centrifugation (1000 r.p.m., 5 min), washed in phosphate-buffered saline (NaCl/P i ) and counted in a hemacytometer.
To determine CD11b or CD14 expression by flow cytometry, 2.5 ´10 5 HL-60 cells in 40 ml of NaCl/P i (supplemented with 0.1% BSA and 0.01% sodium azide) were mixed with 3 ml of monoclonal antibody solution (prechilled to 4°C).The cells were incubated for 45 min on ice, subsequently washed twice with 0.5 ml of NaCl/P i (supplemented as above) and resuspended in 0.5 ml of the diluent.The cells were protected from light.As negative controls FITC labeled IgG1 for CD11b and IgG2a for CD14 were used.The cell surface fluorescence was measured using a FACS Calibur flow cytometer (Becton Dickinson, San Jose, CA, U.S.A.).The acquisition parameters were set for a negative control.Damaged cells were labeled by adding 5 ml of a propidium iodide solution (25 mg/ml) to each test tube just before the data aquisition.Data for damaged cells were not analysed.Data analysis was performed with the Cell Quest software (Becton Dickinson, San Jose, CA, U.S.A.). Figure 2 was prepared with the WinMDI 2.8 software (freeware by Joseph Trotter).MAP kinase assay.HL-60 cells (3106 cells/sample, starved for 3 h in the absence of FCS) were exposed for the required time to either 0.1 mM or 1 mM 1,25-D 3 or PRI-1906.The cells were washed and lysed for 15 min on ice (lysis buffer: 50 mM sodium chloride, 10 mM Tris, 5 mM EDTA, 50 mM sodium fluoride, 30 mM sodium pyrophosphate, 100 mM sodium orthovanadate, 1 mM PMSF, 1 mg/ml aprotinin and 1% Triton X-100, pH 7,4).Then the lysates were centrifuged at 14000 r.p.m. for 5 min.Supernatants were transferred to new tubes and 1.5 mg of anti-MAPK antibody was added.The lysates were incubated for 1 h under continuous shaking at 4°C.Then 40 ml of 50% protein G-Sepharose was added and incubated at 4°C for 2 h.The samples were centrifuged, washed twice with 1 ml of lysis buffer, once with 0.5 M sodium chloride and twice with 1 ml of kinase buffer (20 mM Tris, 10 mM magnesium chloride, 2 mM manganium chloride, 1 mM dithiothreitol, 1 mM EGTA, 100 mM sodium orthovanadate; pH 7.6).The pellets with immunoprecipitated erk1 and erk2 were soaked with 10 ml of kinase buffer.One ml of MBP solution (2 mg/ml) and 1 mCi of [g-32 P]ATP were added to each tube and samples were kept at 30°C for 15 min.The reaction was stopped by adding 10 ml of 2 ´SDS sample buffer and heating to 100°C for 5 min.The eluted proteins were separated by SDS/PAGE on a 14% resolving gel, fixed, dried and subjected to autoradiography, which was performed with an intensifying screen at -70°C with Kodak films.
Statistical evaluation.For statistical analysis experiments were repeated at least three times.Statistical analysis was performed with the Statistica software (StatSoft, Inc.).We used two-tailed Student's t-test for independ-ent samples to assess the significance of differences between groups.

RESULTS
HL-60 cells may be exposed to 1,25-D 3 without inducing cytotoxicity for a very long time.1,25-D 3 -induced differentiation of HL-60 cells is concentration-and, up to some limit, time-dependent.Figure 2 shows how the expression of CD11b (A) and CD14 (B) in HL-60 cells changes with the time of exposure.HL-60 cells were exposed for two, four or six days either to 1,25-D 3 or to the most active analog in increasing concentrations.As presented earlier, the ethanol vehicle did not induce leukemia cell differentiation (Marcinkowska et al., 1998a;Opolski et al., 1999).Since a 48-h exposure of the cells to high concentrations of 1,25-D 3 or the side-chain modified analogs was sufficient to induce a significant increase in CD11b and CD14 and since some of the inhibitors used could be toxic after longer incubation time, all subsequent experiments were carried out under these conditions.On the other hand, the increased differentiation-inducing activity of PRI-1906 and PRI-2191 towards HL-60 cells could be seen after longer exposure and in the case of PRI-1906 is apparent even for very low concentrations (Opolski et al., 1999).
Past studies revealed that 1,25-D 3 activates PI 3-K in human leukemia cells THP-1 (Hmama et al., 1999).Another study demonstrated that pharmacological inhibition of PI 3-K in HL-60 cells reduces the 1,25-D 3 -induced cell differentiation (Marcinkowska et al., 1998b).Here we determined if the differentiation of leukemia cells induced by vitamin D side-chain modified analogs is also PI 3-K dependent.HL-60 cells were exposed for 48 h to either 1 mM 1,25-D 3 , as a positive control, or to 1 mM PRI-1906 and 1 mM PRI-2191.Additional cell samples were exposed simultaneously to either 1,25-D 3 or the analogs and to an inhibitor of PI 3-K, LY294002.The concen-tration of LY294002 was adopted from the published papers (Hmama et al., 1999;Marcinkowska et al., 1998b).The Me 2 SO vehicle alone did not inhibit HL-60 cell differentiation (Marcinkowska et al., 1998b).As presented in Table 1, the differentiation of HL-60 promyelocytic leukemia cells induced by vitamin D side-chain modified analogs is PI 3-kinase dependent.The presence of LY294002 in the culture medium markedly inhibited the induced expression of CD11b and CD14 cell dif-ferentiation antigens.In all flow cytometry experiments the data were collected only for viable cells (propidium iodide non-incorporating).Separate experiments, revealed that simultaneous exposure of the cells to a differentiating agent and to the inhibitor did not cause high cytotoxicity.Cell cycle analysis performed with the Modfit software (Becton Dickinson) showed that the percentage of cells in sub-G 0 phase was below 5% in all samples (not shown).P70 S6K protein kinase is a possible downstream element in the PI 3-K signal transduction pathway activated by various stimuli (Chung et al., 1994).An involvement of p70 S6K in the 1,25-D 3 -induced leukemia cell differentiation was suggested in the past (Marcinkowska et al., 1998b).We investigated if also the differentiation induced by sidechain modified analogs of vitamin D may be inhibited with rapamycin, which blocks the activity of p70 S6K .The results obtained in this experiment and presented in Table 1 indicate that activation of p70 S6K is required for differentiation of leukemia cells induced by the vitamin D analogs studied here.
The erk1 and erk2 kinases are activated by 1,25-D 3 (0.1 mM to 1 mM) within minutes in serum-starved HL-60 cells (Fig. 3).We show here that also a side-chain modified analog of vitamin D 2 is able to induce rapid activation of these kinases.Serum starved HL-60 cells were stimulated for 10 min with either 1,25-D 3 or PRI-1906.Then, the activity of erk1 and erk2 was assayed by measuring the ability of immunoprecipitated kinases to phosphorylate myelin basic protein (MBP).Moreover 1,25-D 3 -induced differentiation of HL-60 cells may be abolished by PD98059 (Marcinkowska, 2001), an inhibitor of MEK1 and MEK2, which are the only known and direct upstream activators of erk1 and erk2 (Robinson & Cobb, 1997).The question we examined next was whether the differentiation of leukemia cells induced by the vitamin D side-chain modified analogs is sensitive to PD98059.Concentration of PD98059 was adopted from previous experiments (Marcinkowska, 2001).The data presented in Table 1 show that expression of CD11b, and CD14 may be significantly inhibited by PD98059 in cells differentiated by the vitamin D side-chain modified analogs.The involvement of cPLA 2 in the 1,25-D 3 -induced differentiation was reported in the past, but was restricted to the differentiation of growth zone chondrocytes (Schwartz et al., 1988).In growth zone chondrocytes, regulation of PLA 2 is involved in the activation of protein kinase C (PKC), but it is still unclear if PLA 2 is an upstream positive regulator of PKC, or is downstream of PKC in the signal transduction pathway and regulates PKC by a feedback loop.To asses whether cPLA 2 is involved in the 1,25-D 3 -induced differentiation of promyelocytic leukemia cells, HL-60 cells were simultaneously exposed for 24 h to 1,25-D 3 and to an inhibitor of PLA 2 .The inhibitor AACOCF 3 targets only the cytosolic form of PLA 2 (Street et al., 1993).The concentration of AACOCF 3 was the same as previously used in experiments with chondrocytes (Helm et al., 1996).Surprisingly, the results of these experiments showed that cPLA 2 is involved in the 1,25-D 3 -induced differentiation of leukemia cells, but in a negative manner.As presented in Fig. 4 the inhibitor itself does not induce differentiation of HL-60 cells, but in concert with 1,25-D 3 it potentiates the acquisition of the CD11b differentiation marker.To investigate if this is also the case for the differentiation induced by side-chain modified analogs of  vitamin D, HL-60 cells were cultured with either PRI-1906 or PRI-2191 and AACOCF 3 at the same time.In these experiments 1,25-D 3 was included as a positive control.As shown in Table 2 the presence of AACOCF 3 in culture medium significantly increased the expression of CD11b, but not of CD14.

DISCUSSION
The design of new vitamin D analogs for potential use in differentiation therapy should lead to the synthesis of analogs that could selectively regulate genes necessary for differentiation without inducing any perturbations in calcium homeostasis.Since the mechanism of action of 1,25-D 3 is not fully understood, the current approach in developing vitamin D analogs involves full biological testing of all newly designed compounds.According to Brown (2000) the potential mechanisms through which this selectivity could be achieved are: (1) altered systemic transport, (2) altered metabolism, (3) altered conformational change in nVDR, and finally (4) altered binding to mVDR.The three-dimensional structure of the ligand binding domain (LBD) of nVDR has only very recently become known from computational models (Yamada et al., 2001) and from the crystal structure (Rochel et al., 2000;Tocchini-Valentini et al., 2001).This means that the design of the presently studied analogs was not based on their binding to nVDR LBD.Moreover, the crystal structures of nVDR complexed to 20-epi analogs of 1,25-D 3 revealed that LBD conformation is the same as for 1,25-D 3 (Tocchini-Valentini et al., 2001).Although the increased differentiation-inducing activity of these 20-epi analogs was ascribed exclusively to the higher stability of the nVDR-agonist complex, augmented activation of the mVDR signalling should also be considered (Yan et al., 2002).
The potential involvement of mVDR in the activity of analogs complicates the situation.mVDR has been isolated only from chick intestinal cells (Nemere et al., 1994;Nemere & Campbell 2000) and rat osteoblastic cells (Nemere et al., 1998;Pedrozo et al., 1999).However, many studies have demonstrated the physiological importance of membranemediated effects for the 1,25-D 3 -induced leukemia cell differentiation (Marcinkowska et al., 1998b;Hmama et al., 1999;Wang et al., 2000;Marcinkowska, 2001;Yan et al., 2002).The relation of the nVDR-mediated "genomic" signalling to the mVDR-mediated "nongenomic" signalling is still unclear.In our understanding a cross-talk between these two mechanisms is necessary for cell differentia- tion.One possibility to consider is that the transcriptional activity of liganded nVDR could be modulated by transcription factors either activated or repressed downstream of mVDR.A second possibility is that phosphorylation of nVDR, necessary for its activation, may be differentially regulated by membrane-mediated events.
Here we studied two active analogs of vitamin D 2 and vitamin D 3 .Of a series of the previously obtained vitamin D analogs we selected for this study PRI-2191 as the most active analog of vitamin D 3 , and PRI-1906 as the most active analog of vitamin D2.Since our vitamin D analogs are more effective than 1,25-D 3 in in vitro induction of leukemia cell differentiation, their increased activity could not be mediated by an altered systemic transport.Therefore we hypothesized that enhanced activation of "nongenomic" signalling could contribute to the biological effect of our analogs.
In this communication we show that activation of membrane-mediated effects are necessary for the completion of leukemia cell differentiation induced not only by natural 1,25-D 3 ,  HL-60 cells were exposed for 48 h to a differentiation inducing factor alone (1 mM 1,25-dihydroxyvitamin D 3 (1,25-D 3 ), 1 mM PRI-1906 or 1 mM PRI-2191) or together with an inhibitor of phospholipase A 2 (2 mM AACOCF 3 ).Then expression of the CD11b and CD14 on surface of viable cells was determined by flow cytometry.MC (mean channel) of fluorescence was calculated using CellQuest software.Positive cells [%] fall into channels of fluorescence higher than the fluorescence of irrelevant isotype-matched IgG.Results marked by asterisks differ significantly from the results obtained for samples treated with respective the differentiation inducing factor alone (*P < 0.05; **P < 0.01).

Mean channel of fluorescence
but also by synthetic analogs with lowered affinity for nVDR, as long as they retain the cell differentiating activity equivalent to or higher than that of 1,25-D 3 .We report that activation of the PI 3-K and the MEK/erk1,2 signal transduction pathway is necessary for the process of differentiation induced by vitamin D analogs.From our earlier studies we could deduce that these two pathways are separate (Marcinkowska et al., 1998b).The activation of p70 S6K is also necessary, but it is unclear if this is a direct membrane-mediated effect of vitamin D analogs, since evidence that 1,25-D 3 rapidly activates this kinase is still lacking.Here we show that also cPLA 2 is involved in the regulation of leukemia cell differentiation induced by either 1,25-D 3 or by the side-chain modified analogs.The nature of this regulation is purely a matter of speculation now.However, it is well documented that cPLA 2 may be upstream regulated by erk1 and erk2 (Geijsen et al., 2000).On the other hand, 1,25-D 3 activates erk1 and erk2, which  Activation of the putative membrane vitamin D receptor leads to the rapid activation of PI-3 kinase and its downstream effector p70 S6K , which is necessary for differentiation (Hmama et al., 1999;Marcinkowska et al., 1998b).It also activates MAPK cascades.PKC mediated activation of erk1 and erk2 leads to their translocation to the cell nucleus (Marcinkowska et al., 1997) and is necessary for differentiation (Marcinkowska, 2001;Wang & Studzinski, 2001).Activation of p38, by a negative loop, downregulates 1,25-D 3 -induced activation of erk1 and erk2 (Wang et al., 2000).Another negative loop could possibly exist between products of PLA 2 and an upstream regulator of erk1 and erk2.
Although the inhibition of cell proliferation is a long term effect accompanying cell differentiation, Wang & Studzinski (2001) in their detailed study presented data supporting the hypothesis that differentiation could be divided into two phases in respect to cell proliferation.In the first phase, which is defined by activation of erk1 and erk2, the cells continue normal cell cycle.In the second phase, the cell cycle is blocked in G 1 phase and the activity of erk1 and erk2 is suppressed.It should be noted that increased expression of the CD11b and CD14 monocytic differentiation markers can be detected in both phases of cell differentiation process.It is therefore possible that activation of "nongenomic" signals is necessary for completion of the first, proliferative phase of differentiation.

Figure 2A .
Figure 2A.Expression of differentiation cell-surface markers on HL-60 cells.HL-60 cells were incubated in RPMI 1640 containing 10% FCS for the indicated times with either 1,25-D 3 or PRI-1906 in various concentrations.At the end of the incubation, expression of CD11b (A) and CD14 (B) was determined by flow cytometry.Cells were exposed to the compounds at the concentration of 1 mM (shadowed areas) and 0.1 nM (areas under thick black line).Graphs representing untreated cells (dashed line) are in the left column.Graphs were prepared with the WinMDI 2.8 freeware.

Figure 2B .
Figure 2B.Legend on the previous page.

Figure 3 .
Figure 3. Activation of erk1 and erk2 in response to 1,25-D 3 or PRI-1906.Serum starved HL-60 cells were treated for 10 min with 1,25-D 3 or PRI-1906 at two concentrations.Control cells were left untreated.Activation of erk1 and erk2 was assayed by their ability to phosphorylate meylin basic protein (MBP).
in turn could lead to activation of cPLA 2 .Inhibition of cPLA 2 may cause disruption of a feedback loop and overactivation of erk1 and erk2, leading to increased differentiation of HL-60 cells.The same mechanism as for 1,25-D 3 is involved in the differentiation induced by side-chain modified analogs of vitamin D. A hypothetical mechanism of cellular signalling activated in HL-60 cells during differentiation induced either by 1,25-D 3 or by its analogs is presented in Fig.5.There are still many open questions concerning the complex process of cell differentiation.

Figure 5 .
Figure 5. Hypothesis for the mechanism of differentiation induced by 1,25-D 3 or its analogs in HL-60 cells.