Aluminum fluoride stimulates inositol phosphate metabolism and inhibits expression of...

JOURNAL OF CELLULAR PHYSIOLOGY 148:10&115 (1991)

Aluminum Fluoride Stimulates lnositol Phosphate Metabolism and Inhibits Expression

of Differentiation Markers in Mouse Keratinocytes

EDMUND LEE AND STUART H. YUSPA* laboratory of Cellular Carcinogenesis and Tumor Promotion, National Cancer Institute,

National institutes of Health, Bethesda, Maryland 20892

Mouse keratinocytes are induced to differentiate in vitro by elevating the level of extracellular calcium from 0.05 mM, where keratinocytes express a basal cell phenotype, to >O. 10 mM, where they express the differentiated phenotype. This process has been associated with a rapid, sustained increase in inositol phosphate (InsP) turnover, which precedes the expression of differentiation-specific proteins. In 0.05 m M Ca2+ medium, aluminum and fluoride salts (AIFJ, which combine to activate nonspecifically heterotrimeric guanine nucleotide-binding (G) proteins, cause a concentration-dependent increase in lnsP metabolism in keratinocytes and generate elevated intracellular diacylglycerol levels. This is associated with an inhibition of cell growth. Treatment with both AIF, and Ca2+ >0.10 m M resulted in an additive increase in lnsP turnover, implying the presence of at least two responsive lnsP pools. AIF, inhibited the expression of differentiation markers induced by Ca2+ >0.10 m M and altered the morphology of keratinocytes from squamous to dendritic, which was reversible upon withdrawal of AIF,. Neoplastic keratinocytes, in which basal levels of lnsP metabolism are higher than in normal cells, do not differentiate in response to Ca2+. Neoplastic keratinocytes responded to AIF, treatment with an even greater rise in lnsP metabolism. AIF, also inhibited cell growth and reversibly altered morphology in neoplastic keratinocytes. These data suggest that InsP metabolism in keratinocytes is at least partially regulated by a G protein mechanism. Furthermore, an increase in lnsP metabolism is not sufficient to stimulate differentiation and may be inhibitory to differentiation if exceeding limited increases. However, these observations cannot exclude the possibility that other AIFG-stimulated pathways involving G or non-G proteins can also influence keratinocyte biology.

Mouse keratinocytes in vitro dis lay the differenti-

tion (Ca,) is raised from 0.05 mM to levels above 0.10 mM (Hennings et al., 1980, 1981; Dale et al., 1983; Hennings and Holbrook, 1982,1983; Brysk and Snider, 1982; 0 Keefe and Payne, 1983; Watt and Green, 1982; Watt et al., 1984). The immediate biochemical signals that control this process have not been defined with certainty, but the Ca, signal is associated with increased inositol hosphate (InsP) turnover (Tang et al., 1988; Jaken an B Yuspa, 1988; Moscat et al., 1989; Lee and Yuspa, submitted for publication) and elevated intracellular Ca2+ (MacLaughlin et al., 1990; Sharpe et al., 1989; Hennings et al., 19891, which occur within minutes and are sustained. An increase of intracellular K+ (Hennin s et al., 1983a,b) has been documented

specific proteins, i.e., keratins, haggrin, and other structural proteins, is transcriptional1 activated over a 48 hr period (Yuspa et al., 1989). Tzere is evidence that G proteins participate in transducing the in-

ated phenotype when the extracellu 'p ar Ca2+ concentra-

after severa B hours, and a grou of differentiation-

0 1991 WILEY-LISS, INC

creased InsP turnover. Fisher et al. (1989) showed that hydrolysis of phosphatidylinositol 4,5-bisphosphate by epidermal preparations in vitro is stimulated by the addition of GTP.

To ex lore the possible role of G proteins and InsP in

entiation, we used the combination of aluminum and fluoride salts, which pharmacologically activates guanine nucleotide binding proteins (G proteins) (Stern- weis and Gilman, 1982). Previously, G protein activation by AlF, was shown to increase the activity of phospholipase C (Sternweis and Gilman, 1982; Boyer et al., 1989). G proteins are activated when bound to GTP and inactive when bound to GDP (Casey and Gilman, 1988). AlF, is thought to activate G proteins by occupying the y-phosphate site of the inactive GDP-

intracel Y ular signaling on normal keratinocyte differ-

Received October 2, 1990; accepted April 1, 1991. *To whom reprint requestsicorrespondence should be addressed.

INOSITOL PHOSPHATE METABOLISM IN MOUSE KERATINOCYTES 107

bound protein thereby converting the G protein to the active GTP-bound conformation (Bigay et al., 1985). AlF, had been used to stimulate InsP metabolism in different cell types, in isolated membranes (Sasaki and Hasegawa-Sasaki, 1987; Guillon et al., 1986; Sternweis and Gilman, 1982; Fleming et al., 1989; Gabler et al., 1989; Hawkins et al., 1989) and also in human keratinocytes (Fisher et al., 1989; Talwar et al., 1989). In this re ort, we examine the effects of short- and long-term A h , t reatment on InsP and diacylglycerol metabolism and growth and ex ression of differentiation markers

MATERIALS AND METHODS Chemicals

Aluminum chloride, sodium fluoride, and bovine serum albumin were purchased from Sigma. Ionomycin was obtained from Calbiochem. For simplicity, in the text, the concentration of A1C1, is in p,M and the concentration of NaF is in mM but will be unitless and represented by the numerical ratio where 1:l is 1 FM AlCl, and 1 mM NaF. [3Hlmyoinositol (17 Ci/mmol) was obtained from Amersham. [3Hlinositol olyphos- phate standards, L3H]choline, and [,H]arachi dp onic acid were obtained from New En land Nuclear. Organic

Cell culture Primary cultures of l-3-day-old newborn BALBic

mouse epidermal keratinocytes were maintained in minimal essential medium (MEM) supplemented with 8% Chelex (Bio-Rad)-treated fetal bovine serum (FBS) (GIBCO) with 0.05 mM Ca2+ (low-calcium medium) and antibiotics as described (Hennings et al., 1980). Cell lines 308 and SP-1 were derived as described by Strickland et al. (1988) and were used at passages 12-18 in low-Ca2+ medium. These cells have an activated.c-rasHa gene with a mutation in codon 61 and produce papillomas in nude mouse grafts (Strickland et a1 1988). They do not differentiate in response to Ca2+">0.10 mM. Normal cells were used at days 6-7 when confluent, and papilloma cells were also used when confluent. For the growth assays, neoplastic cells were plated at 5 x lo5 cellsi35 mm dish, and normal cells at 1 x lo6 cellsi35 mm dish, and treatments were initiated at 24 hr after plating. Cell growth was assayed as total protein per dish using the Bradford assay (Bio-Rad). All experiments were performed at least twice, with consistent results. Data of individual experiments are presented in the figures.

InsP quantitation Cells were cultured on 35 mm dishes in medium

supplemented with [3H]myoinositol (5 pCiiml) for 3 days. Normal cells were labeled on day 4 and refed with fresh labeled medium on day 5, whereas neoplastic cells were labeled for 48 hr prior to treatment. Before use, the cells were rinsed three times with 0.05 mM Ca2+, HEPES-buffered MEM, pH 7, then incubated in 1 ml of 30 mM LiC1-supplemented rinse medium for 60 min. Aluminum chloride and/or sodium fluoride were added to the LiC1-supplemented MEM and then added to the cells for the appropriate period of time. For long-term

in both normal an cp neoplastic keratinocytes.

solvents were from Mallinkro fi t.

assays, cells were incubated in AlFi-containing medium for 24 hr then were switched to serum-free medium supplemented with HEPES and LiCl as above and containing 1:l AIF, for 60 min. Inositol phosphates were extracted with 1 m10.4 M HC104 containing 1 p,l ofphytate hydrolysate (Wreggett et al., 1987). The InsP extracts were neutralized with 500 p.1 of 0.72 M KOH, 0.6 M KHCO,, 5 mM Na-EDTA and separated by anion exchange high-performance liquid chromatography (HPLC) (DuPont Bioseries SAX) using an ammonium phosphate gradient generated by a dual pump HPLC (LKB 2150; Pharmacia-LKB Biotechnology). The ra- dioactive InsP metabolites were quantitated by radio- active flow detection (model F1B; Radiomatic Instru- ment Co.) in conjunction with scintillation fluid compatible with high salt concentrations (Pic0 Flow IV; Packard Instrument Co.). Identification of radiolabeled peaks in each sample was facilitated by the use of ATP and ADP (Sigma) as internal ultraviolet (UV) absor- bance standards and also by radiolabeled InsP standards. The data are presented as the total inositol phosphates relative to 0.05 mM Ca2+ control groups and include inositol mono-, bis-, and trisphosphates. Since changing medium resulted in a small increase in InsP metabolism independent of additives, all controls were time matched for each point. All experiments were conducted using duplicate or triplicate dishes and were performed at least twice. Ex erimental error bars

range of cpm as indicated in the figure legend.

Choline and arachidonic acid release Cells were cultured on 35 mm dishes in 2 ml medium

su plemented with [3Hlcholine (1 p.Ci/ml) or ["Jar-

stimulated in 1 ml serum-free medium with the indicated chemical for 1 hr (choline) or 3 hr (arachidonic acid), and an aliquot (100 pl) was removed and counted in a scintillation counter.

indicate the standard deviation o P the sample set or the

ac f: idonic acid (0.5 p.Ci/ml) for 24 hr. Cells were then

Diacylglycerol quantitation The mass of sn-l,2-diglycerides was measured ac-

cording to the method of Preiss et al. (1986). Briefly, the cells were extracted with methanol, which was further extracted with ch1oroform:methanol:NaCl and the chloroform phase dried under nitrogen (Preiss et al., 1987). Lipids were resuspended in 40 p.1 chloroform, half of which was dried and kept dessicated at 4°C for phosphate analysis. Total lipid phosphate was determined on the chloroform extract by molybdate assay (Van Veldehoven and Mannaerts, 1987). The remain- ing portion of the extracted sn-l,2-diglyceride was dessicated and stored or reacted with recombinant Escherichia coli diacylglycerol kinase (Lipidex), in the presence of [32P]ATP (NEN; specific activity 3,000 Ciimmol, 8 pciisample), to form [32Plphosphatidic acid (PA). These radiolabeled phospholipids were then prepared for thin-layer chromato raphy (TLC) separation. The chloroform phase was f ried under a stream of nitrogen, and the phospholipids were resuspended in 20 ~1 of chloroform for spotting on 20 x 20 cm silica gel 60 TLC plates (F-254; EM Science) prerun with acetone. Chromatograms were developed in ch1oroform:metha-

108 LEE AND YUSPA

no1:acetic acid (65:15:5, v:v). Iodine vapor was used to visualize the unlabeled PA (Sigma) standard on the developed TLC, and the radioactivity was detected and quantitated using a TLC scanner (System 200; Bio- scan). The data are triplicate samples from one of two representative experiments and are presented as relative to lipid phosphate. Error bars represent the standard deviation.

Immunoblotting Cells were rown on 60 mm culture dishes, treated

mM Ca2+ (low-calcium) growth medium, then switched to medium containing 0.12 mM Ca2+ in addition to aluminum fluoride when appropriate. After 48 hr under high-Ca2+ conditions, the cells were solubilized with 200 ~1 of lysis buffer containing 5% sodium dodecyl sulfate (SDS), 20% P-mercaptoethanol in 0.25 M Tris HC1, pH 6.8. The lysates were centrifuged for 5 min (Hermle Z-230M) and the supernatants electro- phoresed on 8.5% polyacrylamide gels containing 0.1% SDS. The proteins were then electroblotted onto 0.2 pm nitrocellulose paper (Schleicher and Schuell) and probed for keratin K1 and filaggrin using rabbit monospecific antisera (Roop et al., 1985). Blots were visual- ized using goat antirabbit alkaline phosphatase as the second antibody (Bio-Rad). Newborn BALBic epidermal cytoskeletal extracts were included as the positive controls for K1 and filaggrin in the immunoblots. Note that filaggrin appears as a diffuse band due to extensive processing, with a molecular weight ranging from 21 to 200 kD (Yuspa et al., 1989).

for 3 days wit a aluminum fluoride supplemented 0.05

RESULTS Normal keratinocytes treated with the combination

of aluminum and fluoride salts (1:l AlF,, i.e., 1 ~J.M AlCl, and 1 mM NaF) under low-Ca2+ conditions show a time-dependent rise in InsP metabolites similar to that stimulated by elevated extracellular Ca2+ (Fig. 1A). The results reflect a change in the experi- mental grou s; time-matched control values were rel-

slightly above those at zero time due to the change of medium. Unlike the Ca2+ response, which is immediate (Jaken and Yuspa, 1988; Lee and Yuspa, submitted for publication), the A1F; response displays a lag of -16 min before InsP metabolism is significantly increased. Neoplastic keratinocytes also respond to AIF; (Fig. 1A). The neoplastic cell responses will be discussed in detail in subsequent sections. As in the Ca2+ response, normal cells incubated in the continued presence of 1:l AlF, have increased InsP metabolites at 24 hr (Fig. 1B). The graded Ca2+ response can also be seen in Figure 7B.

Since AlF; acts as a nonspecific activator of G proteins, the metabolism of other signaling lipids was examined in normal and neoplastic cells. AlF, exposure does not stimulate release of [,H]arachidonic acid from labeled normal keratinocytes a t 3 hr of treatment. A shift to higher concentrations of Cao (0.12 or 1.4 mM) also causes no significant effect on this lipid (Table 1). This is in contrast with TPA, which at a concentration of 1 pgiml was previously shown to stimulate release of

atively stab P e during the time course shown, rising

i

0 0 300i

308 I - % - - - I

I- I

0 + 0 20 40 60 80 100

Time, min

***

** *

0.05 0.12 1.4 AIFd-

Fig. 1. A: Time course of AlF; stimulated InsP metabolism in normal and 308 and SP-1 cells. Cells were labelled with T3H]myoinositol for 3 days (5 FCiiml). The radiolabel medium was washed off and replaced with serum-free medium supplemented with 30 mM LiCl and then incubation for 60 min. This medium was then replaced with medium containing 1 pM AICI, and 1 mM NaF. The data represent duplicate dishes. Error bars were omitted for clarity but the range was <lo% of the indicated values. A second experiment gave similar results. B: Cells were labeled with [3Hlinositol as described above and treated with serum-containing 0.05,0.12, or 1.4 mM Ca2 ' or 1:l A1F; medium for 24 hr, and the InsP was extracted and analyzed as described in Materials and Methods. For the 1 hr LiCl preincubation period, the cells were maintained in the indicated concentration of Caz+ or AlF,. ***P < 0.0025, **P < 0.01, *P < 0.05.

arachidonic acid (Belury et al., 1989) and ionomycin (6.5 FM), both of which stimulate L3H1arachidonic acid release at this time point. [3Hlcholine was not released upon stimulation of keratinocytes with either Ca2' or AlF,, but its release was stimulated in the presence of TPA. These data indicate that AlF; and Ca2+ specifi- cally stimulate inositol phospholipid metabolism in - - this cell ty e.

of phospholipase C activity, also increase after stimulating cells with AlF, (Fig. 2). Keratinocytes treated with AlF,, in serum-free medium, showed greater

The leve P s of diacylglycerol (DAG), the other product

NaF 454 (21) n 0 - A1F; 399 (67)

'[.!HI choline-labeled cells (48 hr before use) were washed with PBS (3 X 1 ml) then stimulated with 1 ml of 0.05 mM Ca2+ serum-free MEM containing the indicated agents for 3 hr; AIFT, 1 pM AlCl3,l mM NaF; TPA, 100 ng/ml; ionomycin, 6.5 pM. Medium was collected (100 p l ) and radioactivity determined by scintillation counting. [3H] arachidonic acid-labeled cells (48 hrbefore use) were washed with PBS (3 X 1 ml) then stimulated with 1 ml ofO.05mM Ca'+serum-freeMEMcontaining the indicated agents for 3 hr; A I S , 1 pM AlCA3,l mM N a F TPA, 1 pg/ml; ionomycin, 6.5 pM. Medium was collected from triplicate dishes in each assay and counted for radioactivity by scintillation counting. The standard deviation is indicated in parentheses. Each assay was repeated with similar results. *Significant difference from control, P < 0.05 as determined by Student's t-test.

: A

mass of DAG compared with control cells, which were simply refed with fresh serum-free medium. In control cells, there was a transient rise in DAG levels peaking at 30 min followed by a decrease to below basal levels by 2 hr. A1F;-treated cells showed a similar, but larger, transient rise in DAG. By 2 hr, however, DAG levels in A1F;-treated keratinocytes were still greater than that of basal values. These experiments were done in the absence of serum, since previous studies have shown that refeeding with serum-containing medium actually causes a decrease in total cellular DAG over an 8 hr period before returning to basal levels at 24 hr (Lee and Yuspa, submitted for ublication). Keratinocytes treated with A1F; for 48 Yl r in the presence of serum showed increased levels of DAG (135%). comDared with basal levels, indicating the sustained nacure of the stimulation (Fig. 2B).

The effect of AlF; on InsP metabolism was dose dependent. Higher concentrations of each of the component salts increased the amount of InsPs released (Table 2). Although the highest concentrations.(lO:lO) showed significant (20-25% dye-positive cells) cytotoxicity as determined by trypan blue dye exclusion by 8 hr, there were no dye-positive cells at 1 hr when InsP metabolism was assayed, nor at 2 or 4 hr of treatment. In addition to the dose dependence of AlF, in 0.05 mM Ca2+, there was an effect of dose when exposed in concert with higher Ca2+ (Fi . 3). In this situation,

the pool of inositol hospholipid had not been depleted

example, in the experiment shown, 1:l AlF, stimulated InsP metabolism to the same extent as 1.4 mM Ca2+ (125%), but together the mean value was 170%. Considerable variability was measured in individual

AlF, was additive with 1.4 m a Ca2+ suggesting that

by either agent a t t R ese respective concentrations. For

Time, min

2.0 48 hrs *

con AIF,

Fig. 2. A Cells were treated in tri licate in serum-free, 0.05 mM Ca2+ MEM supplemented with or witEout 1 pM AlCl, and 1 mM NaF for up to 2 hr. Lipids were extracted, and DAG was quantitated as indicated in Materials and Methods. Error bars indicate the standard deviation. B: Triplicate dishes were treated as in A but for 48 hr in the presence of 8% serum. DAG was quantitated as above. Error bars re resent the standard deviation. An asterisk indicates significant digerence from control, P < 0.05, as determined by Student's t-test. Both assays were repeated and similar results were obtained.

TABLE 2. AlFh concentration-dependent stimulation of InsP metabolism'

Cell tvue AlCl? (uM) NaF (mM) Total InsP % con (s.d.)

Keratinocytes 1 1 10 10

Neoplastic 308 1 1

SP-1 10 10 1 1

10 10

152 (19)* 300 (84)*

200 (57)* 574 (154)* 122 (8.9)* 441 (136)*

' [3H] inositol-labeled keratinocytes were prepared as indicated in Materials and Methods. Cells were then stimulated, in triplicate, with the indicated concentration of salts for 1 hr. InsPs were extracted and quantitated as described in Materials and Methods. The standard deviationisindicatedin parentheses(n = 4). Each assay was repeated with similar results. *Significant difference from control, P < 0.05, as determined by Student's t-test.

110 LEE AND YUSPA

300

200

100

Caz+ 0.05 1.4 0.05 1.4 0.05 1.4 AIFI- 0:o 1:l 3:3

Fig. 3. Extracellular Ca2+ and AIF; increase InsP metabolism additively. [3H]inositol-labeled cells were stimulated with 0.05 mM Ca2' and 1 p,M AlC1, plus 1 mM NaF or 3 pM AlC1, plus 3 mM NaF for 1 hr in duplicate dishes. A second group of cells were also shifted to 1.4 mM Ca2+ in addition to the AlF; treatment. InsPs were extracted and quantitated as described in Materials and Methods. The error bars represent the range of values. A repeat of this experiment gave similar results.

samples in these experiments, but re etitive experi-

sistently observed. The influence of 1:l AlF, on cell growth was exam-

ined by measuring total protein per dish in treated and control dishes. In normal keratinocytes, repetitive experiments showed a slight inhibitory effect of 1:l AlF, on cell growth. In Figure 3, there was no immediate effect for -3 days of treatment, but then the amount of proteinidish plateaued or diminished slightly (Fig. 4A; Fi . 4B and C are discussed below). This pattern in % icates that the treatment is likely to be inhibiting the increase in cell number rather than causing cell loss by cytotoxicity. Higher concentrations of AlF; (1O:lO) resulted in a loss of protein from the dish. Additional studies of L3H]1eucine incorporation by keratinocytes exposed to 1:l AlF, over a 5 day period showed no significant alteration of new protein synthe- sis (data not shown).

Continuous treatment with AlF, produced a stable change in the morphology of keratinocTtes (Fig. 5A). Normal keratinocytes in 0.05 mM Ca2 (Fig. 5A) are heterogeneous but generally have a cobblestone mor- hology. After exposure to AlF, for 72 hr, many cells

gecome more elongated and spindle shaped (Fig. 5B). This change in keratinocyte morphology was reversible on removal of AlF, from the medium. When keratinocytes were cultured for 4 days in medium containing 1:l AlF;, then switched to medium without AlF, for 6 days, the morphology of these cells returned to normal (Fig. 5C; Fig. 5D and E are discussed below).

The long-term effects of A1F; treatment on the morphology of keratinocytes suggested that the differentiation program was altered. To test their capacity to express differentiation markers, keratinocytes were

ments indicated that these additive e P fects were con-

0 1 1 I ' I ' I ' I ' I

0 1 2 3 4 5 Time, days

0.50 4

0.40

0.30

0.20

0.10

0 2 0 1 2 3 4 5

Time, days

1.00 4

0.75

0.50

0.25

0 0 1 2 3 4 5

Time, days

Fig. 4. Aluminum fluoride inhibits the growth of normal (A), 308 (B), and SP-1 (C) cells in culture. Normal cells were plated at 1 x lo6 per 35 mm dish and 308 and SP-1 cells at 5 x lo5 cells per 35 mm dish in 0.05 mM Ca2+ medium and grown for 24 hr. At that time, the cells were switched to medium containing 1 )*M AlC1, and 1 mM NaF. Some dishes were set aside as day 0 controls. Otherwise, two dishes from each group were taken each subsequent day for protein analysis. Error bars represent the range of protein determinations from the mean. This experiment was repeated, giving similar results. Solid squares, control; open circles, treated.

INOSITOL PHOSPHATE METABOLISM IN MOUSE KERATINOCYTES

Fig. 5 . Morphology of A1F;-treated cells. Normal cells were treated without (A) or with AlF, (1 pM AICI,; 1 mM NaF) (B) for 72 hr and then photographed. Note the greater number of spindle-shaped keratinocytes in B. C: Normal cells were grown in A1F;-containing medium as in B, then switched to unsupplemented medium for 6 days. 308 cells were treated without (D) or with AlF, (1 KM AICI,; 1 mM NaF) (E) for 72 hr then photographed under phase contrast ( x 100 magnification). These changes were observed in each of at least five experiments.

111

112 LEE AND YUSPA

grown in 1:l AlF; for 3 days in 0.05 mM Ca2+ and switched to 0.12 mM Ca2+ for 48 hr (Fig. 6). Two markers of differentiation, keratin 1 (Kl) (Fig. 6A), a spinous cell marker, and filaggrin (Fig. 6B), a granular cell marker, were assessed by Western blotting. In control cultures, the switch to 0.12 mM Ca2' is associated with a major increase in immunodetectable K1 and fila grin. However, A1F; treatment completely

Ca2+. Filaggrin appears as a broad band (21-200 kD) due to the extensive processing to which it is subjected (Yuspa et al., 1989).

The block in keratinocyte differentiation observed after AlF, was similar to the absence of marker expression in neoplastic keratinocytes. Neoplastic keratinocytes differ from normal cells by their unrespon- siveness to the Ca2+ signal for differentiation (Strick- land et al., 1988). This is the earliest detectable phenotypic alteration in transformed keratinocytes and is expressed already in the initiated state when the cell potential is to produce only a benign tumor. Two

inhibite f the expression of both markers in 0.12 mM

characteristic cell lines of this type are 308 and SP-1. These cell are known to have a rate of InsP metabolism two- to five-fold greater than normal keratinocytes in 0.05 mM Ca2+ and respond to higher Ca2+ with an even greater InsP turnover. This increase could be relevant to the neoplastic state (Lee and Yuspa, submitted for publication). To determine if 308 or SP-1 cells remain sensitive to G protein activation, parallel studies with A1F; were performed as on normal cells.

As with normal cells, AlF, treatment did not stimulate significant L3H1choline release in SP-1 cells (Table 1). However, both neoplastic cell lines are sensitive to AlFl-stimulated InsP metabolism with SP-1 being somewhat more sensitive than normal or 308 cells (Fig. 1). Their InsP response to 1:l AlF; is similar to that of normal keratinocytes, displaying a lag of -15 min before significant InsP accumulation is observed. When neoplastic keratinocytes are cultured in medium containing 1:l AlF;, their growth is inhibited to a greater extent than that of normal cells when assayed by protein per plate (Fig. 4B,C). Again this profile

Fig. 6. Immunoblots of SDS-PAGE-separated cell homogenates. Nor- ma], 308, and SP-1 cells were cultured for 3 days in 0.05 mM Ca2' medium and treated with the indicated agent for an additional 3 days. The same medium was supplemented with or without 0.12 mM Ca2' for 48 hr more. The cells were solubilized and the proteins electro- phoresed on 8.5% polyacrylamide gels containing 0.2% SDS. The gels

were electroblotted onto nitrocellulose paper and probed with rabbit monospecific antiserum for mouse K1 (A) or mouse filaggrin (B). Goat antirabbit antibody conjugated to alkaline phosphatase was used to visualize primary antibody binding. A second experiment gave similar results. Filaggrin is processed extensively and appears as a broad band of 21-200 kD (Yuspa et al., 1989).


suggests growth inhibition rather than cytotoxic cell loss. Morphology is also altered in neoplastic cells by continuous ex osure to AlF; (Fig. 5D,E), although the

Four day treatment of 308 cells with AlF; led to a change from small cell size (Fi . 5D) to an enlarged,

trast to the spindle-like morphology of AlFi-treated normal keratinocytes. This type of change was also seen in SP-1 cells (not shown).

Neoplastic cells do not express certain markers of differentiation, e.g., K1, which may be related to the presence of the activated c-msHa in these cell lines (Strickland et al., 1988). Although neoplastic keratinocytes do not express K1, they do express filaggrin, a protein involved in keratin filament formation in the mature keratinocyte (Fig. 6B). A shift from 0.05 to 0.12 mM Cao resulted in a strong induction of filaggrin expression in 308 cells, with no change in SP-1 cells. However, treatment of 308 or SP-1 cells with AlF4- containing medium for 3 days inhibited constitutive and 0.12 mM Ca2+-induced expression of filaggrin in neoplastic cells as in normal cells.

DISCUSSION The regulation of differentiation by extracellular

Ca2+ in keratinocytes is closely associated with changes in metabolism of inositol phospholipids (Tang et al., 1988; Jaken and Yuspa, 1988; Lee and Yuspa, submitted for publication) and intracellular Ca2+ (Hen- nings et al., 1989). Biochemical data indicate that stimulation of keratinocytes with high-Ca2+ concentrations he. , 1.4 mM), 12-0-tetradecanoy1phorbo1-13-ace- tate (TPA), ionomycin, DAG, or bacterial phospholipase C (PLC) stimulates certain pathways associated with differentiation (Jeng et al., 1985). These agents also act on the InsP/protein kinase C signal transduction pathway either directly or indirectly (Thomas, 1988; Nishizuka, 1984; Beaven and Gonzaga, 1990). The InsP response to extracellular Ca2+ is intriguing in that it is dose dependent (Lee and Yuspa, submitted for publication). Lower concentrations of extracellular Ca2+, i.e., 0.12 mM, stimulate InsP metabolism and enhance the expression of differentiation markers, while excessive concentrations (1.4 mM) stimulate InsP metabolism substantially more than 0.12 mM Ca2 + and inhibit marker expression (Yuspa et al., 1989). Al- though the nature of the Ca2+ sensing mechanism in keratinocytes is unknown, many receptors are coupled to this particular signal transduction pathway (Ber- ridge and Irvine, 1989). A similar receptor for Ca2+ may exist in parathyroid cells, which control the level of Ca2+ in the blood via parathyroid hormone (Nemeth and Scarpa 1987). In these cells, high levels of extracellular Ca'+ result in a type of feedback inhibition of parathyroid hormone release, which is seen both in vivo and in vitro.

Important components of the InsPiprotein kinase C signaling pathway are G proteins (Casey and Gilman, 1988). As others have found, G proteins can be stimulated with GTP or aluminum fluoride salts (Sasaki and Hasegawa-Sasaki, 1987; Bigay et al., 1985). In particular, Fisher et al. (1989) have found that hydrolysis of inositol phospholipids is enhanced by the addition of

changes are c f istinct from the effect on normal cells.

flattened, and rounded morpho 7 ogy (Fig. 5E) in con-

GTP in human epidermal extracts. In hepatocytes, or their membranes, treatment with AlF, or GTP stimulated InsP metabolism (Blackmore et al., 1988). Direct evidence for the involvement of G proteins in InsP metabolism comes from a study of the phospholipase C-mediated urinergic receptor system in turkey eryth-

(1989) reconstituted these membranes with py G protein subunits from human placenta and found that InsP metabolism was potentiated in the presence of purines. Boiled py-subunits had no activity.

As seen in Figure lA, AlF4 treatment led to a rise in InsP levels in normal and neoplastic keratinocytes after an initial lag of -15 min. The lag may be attributed to movement of the aluminum and fluoride ions across the plasma membrane. The A1F;-induced increase in InsP metabolism is maintained over 24 hr as was observed with Ca2+ (Fig. 1B). This A1F;- stimulated rise in InsP metabolism was concentration dependent as well (Table 2). Trypan blue assays indicate that, at 1 hr of 1O:lO AlF,, cells are viable and membrane function is intact. Thus the dose-dependent InsP increase is likely due to the stimulatory effects of AlF4 on G protein-dependent inositol phospholipid turnover. Another feature of the rise in InsP caused by AlF4 treatment was the additive nature with extracellular Ca2+ (Fig. 3). Whereas A1C1, (1 pM) alone had no significant stimulatory effect, NaF (1 mM) was somewhat stimulatory (data not shown). However, it is the combination of the salts that yields significantly higher InsP turnover.

Keratinocytes undergo a change in morphology when cultured in medium containing AlF,. Within 24 hr, normal keratinocytes displayed an increase in the number of elongated, spindle-shaped cells (Fig. 5A). A similar type of morphologic change is also seen with TPA treatment (Jeng et al., 1985). While TPA treatment leads to cornification and squame formation, AlF, treatment does not. The A1F;-induced morphologic changes were stable but reversible on removal of AlF4 from the medium (Fig. 5B).

The effects of various chemical modifiers of keratinocyte differentiation on lipid metabolism are complex. Although elevated extracellular Ca2+ stimulates InsP metabolism, there is little effect on [3Hlcholine (Jaken and Yuspa, 1988) or L3H1arachidonic acid release into the culture medium (Table 1). TPA, on the other hand, inhibits InsP metabolism (Lee and Yuspa, submitted for publication) yet stimulates both choline and arachidonic acid release. Ionomycin stimulates metabolism of InsP (Jaken and Yuspa, 1988) and arachidonic acid. AlF, stimulates InsP metabolism and has little effect on choline and arachidonic acid release. These dispar- ate data reflect the complex interactions taking place at the plasma membrane in response to these modula- tors of differentiation.

A common element among these agents may be the direct or indirect activationidown-regulation of protein kinase C (PKC). Several studies suggest that PKC is important in regulating the state of differentiation of keratinocytes (Dlugosz et al., 1990). Elevated extracellular Ca2 and AlF4 increase the mass levels of DAG, a potential activator of PKC; phorbol esters and exog- enous DAG produce similar biological effects on kera-

rocyte mem i: ranes (Boyer et al., 1989). Boyer et al.

114 LEE AND YUSPA

tinocytes (Jeng et al., 1985); ionomycin indirectly raises DAG by elevating intracellular Ca2+ levels and stimulating Ca2+-dependent PLC activity. Thus, each is potentially able to activateidown-regulate PKC. TPA, ionomycin, and high concentrations of Ca2+ in the medium (1.4 mM) all inhibit expression of K1 (Roop et al., 1987; Lee and Yuspa, submitted for publication). As is shown in Figure 6, A1F; treatment also inhibited the 0.12 mM Ca2'-stimulated K1 expression. Thus, it may be that these agents cause an excessive rise in DAG levels, relative to that caused by 0.12 mM Ca'', resulting in the activation and subsequent down-regulation of PKC. Related studies on keratin 8, which is expressed in simple epithelia in vivo but can be induced in cultured keratinocytes, show a similar, although opposite, response pattern. Where K1 expression is inhibited by the above agents, KS expression is enhanced by the same pharmacologic agents as well as by introduction of a v-rasHa oncogene, suggesting that a common pathway is modulating the expression of K1 and K8 (Lee and Yuspa, in preparation).

Neoplastic keratinocyte cell lines 308 and SP-1 are Ca2+ resistant, possess an activated c-rusHa gene (Strickland et al., 198S), and have high levels of DAG relative to normal keratinocytes (Lee and Yuspa, 1990). Treatment of the two neoplastic cell lines with AlF; stimulated InsP metabolism and caused them to uniformly flatten and become larger (Fig. 5E). As with normal keratinocytes, these mor hologic chan es were

InsP metabolism in neoplastic cell lines may be linked to the absence of expression of differentiation markers of the s inous cell layer (K1 and K10). This is supported

when AlF, stimulates this pathway even further. The results presented in this report indicate that the

stimulation of InsP metabolism by pharmacological methods, such as AlF, a t noncytotoxic concentrations, is not sufficient to induce differentiation and, if excessive, may inhibit differentiation. Of course, we cannot rule out an additional G protein action that alters the differentiation program. Nevertheless, AlF, is rela- tively specific for the InsP lipid as opposed to other li id

have shown that cyclic nucleotides do not alter keratinocyte differentiation, nor are cyclic nucleotide levels changed during the process of differentiation (Hen- nings et al., 1983a). Thus AIF4-mediated changes in that pathway seem unlikely to be involved in the biological effects observed here. In other studies, we have found that the introduction of a v-rasHa oncogene into normal keratinocytes increases InsP turnover and blocks differentiation. The G protein nature of the rus gene and the similar responses of keratinocytes to an activated ras gene and AlF, suggest that the effects on the InsP signaling pathway are fundamental to their influence on keratinocyte biology.

ACKNOWLEDGMENTS

reversed on removal of AlF;. T R e very high K evels of

by the f oss of filaggrin expression in 308 and SP-1 cells

turnover pathways in keratinocytes. Previous stu ct! ies

This work was funded in art by a grant from the

Carson Loomis of Duke University was instrumental in introducing to us the diacylglycerol assay. We thank Dr. Joel Moss for his critical review of the manuscript.

Sterling Drug Division of 8 astman-Kodak, Inc. Dr.

LITERATURE CITED Beaven, M.A., and Gonzaga, H.M.S. (1990) Stimulatory signals for

secretion in a cultured mast cell line: Evidence for activation of phospholipase C through more than one G-protein. Adv. Second Messenger Phosphoprotein Res., 24:387-392.

Belury, M.A., Patrick, K.E., Locniskar! M., and Fischer, S.M. (1989) Eicosapentaenoic and arachidonic acid: Comparison of metabolism and activity in murine epidermal cells. Lipids, 24:423429.

Berridpe. M.J.. and Irvine. R.F. (1989) Inositol Dhowhates and cell - I

signilling. Nature, 341:197-205. Bigay, J., Deterre, P., Pfister, C., and Chabre, M. (1985) Fluoroalu-

minates activate transducin-GDP bv mimicking the .r-phosphate of ~- GTP in its binding site. FEBS Let t , 191:181-f85.

Blackmore, P.F., Lynch, C.J., Uhing, R.J., Fitzgerald, T., Bocckino, S.B., and Exton, J.H. (1988) Role of guanine nucleotide regulatory proteins and inositol phosphates in the hormone induced mobilization of hepatocyte calcium. Adv. Exp. Med. Biol., 232:169-182.

Boyer, J.L., Waldo, G.L., Evans, T., Northup, J.K., Downes, C.P., and Harden, T.K. (1989) Modification of A1F;;- and receptor-stimulated phospholipase C activity by G-protein PT subunits. J . Biol. Chem., 264:13917-13922.

Brysk, M.M., and Snider, J.M. (1982) The effect of state of differentiation on labeling of epidermal cell surface glycoproteins. J . Invest. Dermatol., 78:36&370.

Casey, P.J., and Gilman, A.G. (1988) G protein involvement in receptor-effector coupling. J. Biol. Chem., 263:2577-2580.

Dale, B.A., Scofield, J.A., Hennings, H., Stanley, J.R., andYuspa, S.H. (1983) Identification of filaggrin in cultured mouse keratinocytes and its regulation by calcium. J. Invest. Dermatol., 81:90~-95s.

Dlugosz, A.A., Pettit, G.R., and Yuspa, S.H. (1990) Involvement of Protein kinase C in Ca2' -mediated differentiation on cultured primary mouse keratinocytes. J. Invest. Dermatol., 94519419.

Fisher, G.J., Baldassare, J.J., and Voorhees, J.J. (1989) GTP-dependent hydrolysis of phosphatidylinositol-4,5-bisphosphate by soluble phospholipase C from adult human epidermis. J. Invest. Dermatol., 925331436.

Fleming, N., Sliwinski-Lis, E., and Burke, D.N. (1989) G regulatory proteins and muscarinic receptor si nal transduction in mucous acini of rat submandibular gland. Lik Sci., 44:1027-1035.

Gabler, W.L., Creamer, H.R., and Bullock, W.W. (1989) Fluoride activation of neutrouhils: Similarities to formvlmethionvl-leucvl- phenylalanine. Inflimmation, 13:47-58.

Guillon, G., Mouillac, B., and Balestre, M.-N. (1986) Activation of aolvohosDhoinositide Dhosuholipase C bv fluoride in WRKl cell mgibranes. FEBS Lett., 2?)4:183-188. '

Hawkins, D., Enyedi, P., and Brown, E. (1989) The effects of high extracellular calcium and magnesium concentrations on the levels of inositol 1,3,4,5-tetrakisphosphate in bovine parathyroid cells. Endocrinology, 124:838-844.

Hennings, H., and Holbrook, K.A. (1982) Extracellular calcium reg- ulates growth and terminal differentiation of cultured mouse epidermal cells. In: Ions, Cell Proliferation and Cancer. A. Boynton, W.L. McKeehan. and J.F. Whitfield. eds. Academic Press. New York, pp. 499-516.

Hennings, H., and Holbrook, K.A. (1983) Calcium regulation of cell-cell contact and differentiation of eDiderma1 cells in culture. An ultrastructural study. Exp. Cell Res., i43:127-142.

Hennings, H., Holbrook, K.A., and Yuspa, S.H. (1983a) Factors influencing calcium-induced terminal differentiation in cultured mouse epidermal cells. J. Cell. Physiol., 116:265-281.

Hennings, H., Holbrook, K.A., and Yuspa, S.H. (1983b) Potassium mediation of calcium-induced terminal differentiation of epidermal cells in culture. J . Invest. Dermatol., 81 :51~-55s.

Hennings, H., Kruszewski, F.H., Yuspa, S.H., andTucker, R.W. (1989) Intracellular calcium alterations in response to increased external calcium in normal and neoplastic keratinocytes. Carcinogenesis,

Hennings, H., Michael, D., Cheng, C., Steinert, P., Holbrook, K., and Yuspa, S.H. (1980) Calcium regulation of growth and differentiation of mouse epidermal cells in culture. Cell, 19:245-254.

Hennings, H., Steinert, P., and Buxman, M.M. (1981) Calcium induction of transglutaminase and the formation of (T-glutamyl) lysine crosslinks in cultured mouse epidermal cells. Biochem. Biophys. Res. Commun., 102:739-754.

Jaken, S., and Yuspa, S.H. (1988) Early signals for keratinocyte differentiation: Role of Ca2 ' -mediated inositol lipid metabolism in normal and neoplastic epidermal cells. Carcinogenesis, 9:1033- 1038.

10:777-780.


Jeng, A.Y., Lichti, U., Strickland, J.E., and Blumberg, P.M. (1985) Similar effects of phospholipase C and phorbol ester tumor promot- ers on primary mouse epidermal cells. Cancer Res., 45:5714-5721.

Lee, E., and Yuspa S.H. (1990) Keratinocytes infected with a virus containing v-ras'" have elevated inositol phosphate metabolism. Pr. Am. Assoc. Cancer Res., 31:76-76 (abstract).

MacLaughlin, J.A., Cantley, L.C., and Hollick, M.F. (1990) 1,25(OH),D, increases calcium and phosphatidylinositol metabolism in differentiating cultured human keratinocytes. J. Nutr. Biochem., 1:81-87.

Moscat, J., Fleming, T.P., Molloy, C.J., Lopez-Barahona, M., and Aaronson, S.A. (1989) The calcium signal for BalbiMK keratinocyte terminal differentiation induces sustained alterations in phosphoi- nositide metabolism without detectable protein kinase C act.ivation. J . Biol. Chem., 264:1122&11235.

Nemeth, E.F., and Scarpa, A. (1987) Rapid mobilization of cellular Ca2+ in bovine parathyroid cells evoked by extracellular divalent cations. Evidence for a cell surface calcium receptor. J. Biol. Chem., 262.5188-5196.

Nishizuka, Y. (1984) The role of protein kinase C in cell surface signal transduction and tumour promotion. Nature, 308:693-698.

O'Keefe, E.J., and Payne, R.E. (1983) Modulation of the epidermal growth factor receptor of human keratinocytes. J. Invest. Dermatol., 81 :231-235.

Preiss, J.E., Loomis, C.R., Bell, R.M., and Neidel, J.E. (1987) Quan- titative measurement of sn-1,2-diacylglycerols. Methods Enzymol., 141:29&300.

Preiss, J., Loomis, C.R., Bishop, W.R., Stein, R., Neidel, J.E., and Bell, R.M. (1986) Quantitative measurement of sn-1,2-diacylglycerols present in platelets, hepatocytes, and ras- and sis-transformed normal rat kidney cells. J. Biol. Chem., 261:8597-8600.

Roop, D., Stanley, J.R., Cheng, C.K., Toftgard, R., Steinert, P. M., and Yuspa, S. H. (1985) The use of monospecific antibodies as probes to monitor keratin gene expression. Ann. N.Y. Acad. Sci., 4,55:409- 414.

Roop, D.R., Huitfeldt, H., Kilkenny, A,, and Yuspa, S.H. (1987) Regulated expression of differentiation-associated keratins in cultured epidermal cells detected by monospecific antibodies to unique peptides of mouse epidermal keratins. Differentiation, 35:1,43-150.

Sasaki, T., and Hasegawa-Sasaki, H. (1987) Activation of polyphos-

phoinositide phospholipase C by guanosine 5'-0-(3-thio)triphosphate and fluoroaluminate in membranes prepared from a human T cell leukemia line, Jurkat. FEBS Lett., 218237-92.

Sharpe, G.R., Gillespie, J.I., and Greenwell, J.R. (1989) An increase in intracellular free calcium is an early event during differentiation of cultured human keratinocytes. FEBS Lett., 254:25-28.

Sternweis, P.C., and Gilman, A.G. (1982) Aluminum: A requirement for activation of the regulatory component of adenylate cyclase by fluoride. Proc. Natl. Acad. Sci. USA, 79:48884891.

Strickland, J.E., Greenhalgh, D.A., Koceva-Chyla, A,, Hennings, H., Restrepo, C., Balaschak, M., and Yuspa, S.H. (1988) Development of murine epidermal cell lines which contain an activated rasHa oncogene and form papillomas in skin grafts on athymic nude mouse hosts. Cancer Res., 48:165-169.

Talwar, H.S., Fisher, G.J., Harris, V.A., and Voorhees, J .J . (1989) Agonist-induced hydrolysis of phosphoinositides and formation of 1,2-diacylglycerol in adult human keratinocytes. J. Invest. Derma-

Tang, W., Ziboh, V.A., Isseroff, R., and Martinez, D. (1988) Turnover of inositol phospholipids in cultured murine keratinocytes: possible involvement of inositol triphosphate in cellular differentiation. J. Invest. Dermatol., 90:3743.

Thomas, A.P. (1988) Enhancement of the inositol 1,4,5-trisphosphate releasable Ca2+ pool by GTP in permeabilized hepatocytes. J. Biol. Chem., 263:2704-2711.

Van Veldehoven, P.P., and Mannaerts, G.P. (1987) Inorganic and organic phosphate measurements in the nanomolar range. Anal. Biochem., 161 :4548.

Watt, F.M., and Green, H. (1982) Stratification and terminal differentiation of cultured epidermal cells. Nature, 295:434-436.

Watt, F.M., Mattey, D.L., and Garrod, D.R. (1984) Calcium-induced reorganization of desmosomal components in cultured human keratinocytes. J . Cell Biol., 992211-2215,

Wreggett, K.A., Howe, L.R.. Moore, J.P., Irvine, R.F. (1987) Extrac- tion and recovery of inositol phosphates from tissues. Biochem. J., 243:933-934.

Yuspa, S.H., Kilkenny, A.E., Steinert, P.M., and Roop, D.R. 11989) Expression of murine epidermal differentiation markers is tightly regulated by restricted extracellular calcium concentrations in vitro. J . Cell Biol., 109:1207-1217.

tol., 93:241-245.

Aluminum fluoride stimulates inositol phosphate metabolism and inhibits expression of...

Documents

Transcript of Aluminum fluoride stimulates inositol phosphate metabolism and inhibits expression of...