Transforming growth factor β stimulation of parathyroid hormone-related protein (PTHrP): A paracrine regulator?

Molecular and Cellular Endocrinology, 92 (1993) 55-62 0 1993 Elsevier Scientific Publishers Ireland, Ltd. 0303-7207/93/$06.00

MOLCEL

55

02942

Transforming

growth factor p stimulation of parathyroid hormone-related (PTHrP): a paracrine regulator?

protein

Takeshi Kiriyama, Matthew T. Gillespie, Jane A. Glatz, Seiji Fukumoto, Jane M. Moseley and T. John Martin The University of Melbourne Department of Medicine, St. tincent S Hospital and St. Vincent S Institute of Medical Research, Fitzroy 3065, Victoria, Australia (Received

Key words: Transforming

growth

factor-p;

Parathyroid

24 June 1992; accepted

hormone-related

protein;

11 November

Gene

1992)

regulation

Summary The regulation of PTHrP expression and production by transforming growth factor-p (TGFP) has been investigated in an epidermal squamous cancer cell line COLO 16. TGF& treatment increased steady-state levels of PTHrP mRNA and concentrations of PTHrP immunoreactivity in conditioned medium in a time- and dose-dependent manner with a half-maximal effect at 40 pM. An effect of TGF& on PTHrP mRNA was observed first after 4 h treatment and continued to increase up to 48 h with a concomitant increase in PTHrP immunoreactivity in the culture medium. TGFP, was found to stabilize PTHrP mRNA as assessed by actinomycin C, experiments. In addition, a direct effect of TGFP to increase PTHrP transcription was indicated by nuclear run-on and transient transfection experiments using a CAT promoter/ expression construct encompassing the region - 1100 bp to - 20 bp from the initiating AUG of the human PTHrP gene. The conditioned medium from COLO 16 cells was also shown to contain both latent and active TGFP at concentrations of 160 pM and 16 pM, respectively, in 72 h conditioned medium. A neutralizing antibody to TGFP, (and TGF&) decreased the level of immunoassayable PTHrP in the medium.

Introduction Parathyroid hormone-related protein (PTHrP) is commonly produced by squamous cell carcinoma resulting in the syndrome of humoral hypercalcemia of malignancy (HHM) which exhibits clinical and biochemical features similar to those of hyperparathyroidism (Martin, 1990). The amino-terminal sequence shows limited homology with parathyroid hormone (PTH) (Moseley et al., 1987; Stewart et al., 1987; Strewler et al., 1987; Suva et al., 1987) and chromosoma1 localization studies and comparison of the PTH and PTHrP genes Wasicek et al., 1983; Mangin et al., 1989; Suva et al., 1989; Yasuda et al., 1989) indicate that these two genes may share a common ancestry. Unlike the PTH gene (Vasicek et al., 1983) which is

Correspondence to: Professor T.J. Martin, St. Vincent’s of Medical Research, 41 Victoria Parade, Fitzroy 3065, Australia. Tel. 61-3-288 2480; Fax 61-3-416 2676.

Institute Victoria,

composed of three exons and encodes a single protein species of 84 amino acids in length, the PTHrP gene comprises nine exons and the activities of three potential promoters together with alternate splicing events could potentially result in up to 12 mRNA species which are capable of coding for three forms of the mature protein comprised of 139, 141 and 173 residues, respectively (Suva et al., 1987; Mangin et al., 1988, 1989; Thiede et al., 1988; Yasuda et al., 1989). Regulatory consensus sequences for glucocorticoids, 1,25(OH),-vitamin D,, and CAMP can be identified within the 5’ flanking sequence of the PTHrP gene, and PTHrP production has been shown to be modulated by dexamethasone (Ikeda et al., 1989; Lu et al., 19891, 1,25(0H),-vitamin D, (Ikeda et al., 1989), progestins (Thiede, 19891, cyclic-AMP (Rizzoli et al., 19901, estrogen (Suva et al., 1991; Thiede et al., 1990, phorbol esters (Rodan et al., 1988; Deftos et al., 1989) and epidermal growth factor Werner et al., 1991). In addition to its PTH-like biological activity, PTHrP exhibits unique actions on placental calcium transport

56

(Rodda et al., 1988), renal bicarbonate elimination (Ellis et al., 1990) and osteoclast activity (Fenton et al., 1991). Its widespread occurrence in epithelia of the developing fetus (Moseley et al., 19911, in normal skin keratinocytes (Merendino et al., 1986; Hayman et al., 19891, and in squamous cell cancers (Danks et al., 1989) has suggested a potential role for PTHrP in epithelial cell proliferation and/or differentiation. Since TGFP, is a potent inhibitor of cell growth in skin keratinocyte cultures (Matsumoto et al., 1990) and is also widely localized in developing fetal epithelia (Akhurst et al., 1990) we have examined the effect of TGFP, on PTHrP expression by a human epidermal squamous carcinoma cell line (COLO 16). Materials

and methods

Materials

TGFP, was a gift of Genentech (California, USA), 1,25_dihydroxyvitamin D, was obtained from Roussel UCLAF (Paris, France). Indomethacin was obtained from Sigma (Australia), culture media were from Gibco (USA), fetal calf sera were from Cytosystems (NSW, Australia) and plastic culture dishes were from Nunc (Denmark). PTHrP(l-34) was synthesized in this laboratory as described previously (Kemp et al., 1987), iodinated with Na’251 (1 mCi) in the presence of chloramine T and purified with QUSO 32 silica powder (North America Silica Co., Teterboro, NJ, USA) followed by high-performance liquid chromatography (HPLC) (Grill et al., 1991). Na’251 (100 Ci/liter) was purchased from Amersham (UK), aprotinin was from Sigma Chemical Co. (St. Louis, MO, USA) and Sac-Cel solid phase anti-rabbit IgG suspension was from IDS Bolton (Tyne and Wear, UK). Monoclonal antibodies against TGFP, were gifts of Genentech. 2G7 neutralizand 12H a non-neutralizing antibody as ing TGFP,,,,, control. All other chemicals were of reagent grade and obtained from standard suppliers. Cells and culture conditions

The human epidermal squamous carcinoma cell line, COLO 16, cells originally established from a tumor removed from a patient with hypercalcemia, was maintained in RPM1 1640 medium containing 10% fetal calf serum (FCS) as described previously (Moore et al., 1975; Martin et al., 1978). For experiments to examine the effects of TGFP,, cells were grown in 6-well plates in RPM1 1640/10% FCS until approximately 100% confluent. The medium was then changed to RPM1 1640 containing 0.1% bovine serum albumin (BSA) 24 h before the addition of treatments. Incubation was continued with or without TGFP,, for 48 h except as indicated in time-course experiments. The media were then removed and assayed for PTHrP in a specific radioimmunoassay (Law et al., 1991). In similar sepa-

rate experiments, cells were grown and treated in 175 cm’ tissue culture flasks, harvested and total RNA extracted for Northern blot analysis. TGFP, was diluted in 4 mM HCl/O.l% BSA, cycloheximide was diluted in ethanol, and actinomycin D in methanol. Final concentrations of methanol or ethanol were less than 0.1%. Controls contain vehicle(s) alone. To test the effect of indomethacin on the TGFP, stimulation of PTHrP immunoreactivity in the culture medium, cells were treated for 48 h in the presence or absence of 14 mM indomethacin. RNA extraction and Northern blot analysis

Total RNA was isolated from cells following experimental treatment according to the guanidine isothiocyanate/ phenol-chloroform extraction procedure (Chirgwin et al., 1979). For Northern blot analysis, 20 Kg of total RNA was electrophoresed through a 1.5% agarose gel containing 0.66 M formaldehyde and transferred to Hybond-N (Amersham). Blots were hybridized with 32P-labelled nick-translated cDNA probe for human PTHrP (pBRF61; Suva et al., 1987), human TGFP, (Genentech) or chicken glyceraldehyde-3-phosphate-dehydrogenase (GAPDH; Dugaiczyk et al., 1983). Hybridizations were carried out in 0.2 M sodium phosphate (pH 7.2); 1% BSA; 35% formamide; 1 mM EDTA; and 5% sodium dodecyl sulfate (SDS) for 24 h at 42°C. The high stringency post-hybridization wash was in 0.1 X standard saline citrate (SSC)-1% SDS at 42”C, and filters were exposed to Kodak XAR-5 film at - 80°C with intensifying screens. All hybridizing PTHrP mRNA species on autoradiograms were scanned by densitometer and the signal intensity was normalized to that obtained with GAPDH cDNA hybridization expressing data as the ratio of PTHrP mRNA versus GAPDH mRNA. Nuclear run-on assay

Isolation of nuclei, the transcription reaction, hybridization and washing conditions were as according to Harrison et al. (1989). 32P-labelled transcripts were hybridized for 3 days at 42°C to target sequences (10 pg pBRF61,lO pg pUC119 and 5 pg GAPDH) immobilized on a nylon membrane. Plasmid construct

Plasmid pSMR38 was constructed by introducing a XhoI site -19 bp from the initiating AUG by directed mutagenesis, using the oligonucleotide obrf 15.14 (5’TCCGCTCGCTCGAGACCTGCAACA-3’), of a recombinant Ml3 bacteriophage clone containing the region -1100 bp (BarnHI) to +2.5 bp (Sal11 of the PTHrP gene (Suva et al., 1989). The BarnHI-XhoI fragment was inserted into the HindIII-Sal1 sites of the CAT reporter plasmic pCAT- (C.S.I.R.O. Division of Biotechnology); the insert BamHI and vector

Hind111 sites were end-filled resulting in the loss of both restriction endonuclease sites. Transfections

The transfections of COLO 16 cells were performed by the Ca,(PO,), coprecipitation method for 3-4 h (Gorman et al., 1982) using 5 pg of pSMR38, and CAT activity determined by liquid scintillation counting (Seed and Sheen, 1988) 48 h after transfection. Each transfection experiment was carried out in triplicate, and CAT activity was standardized to pSMR38 basal activity representing 100%.

B

Radioimmunoassay for PTHrP

Radioimmunoassays were carried out as previously described using a polyclonal antiserum raised in rabbits against synthetic PTHrP(l-34) (Law et al., 1991). Standard recombinant PTHrP(l-84) or conditioned medium from COLO 16 cell culture was incubated with antiserum (1: 10,000 final dilution) and 1251Tyr4’-PTHrP(l-34) (10,000 cpm/tube) in a total volume of 0.4 ml in 50 mM sodium barbitone, pH 8.6 containing 0.5% BSA, 0.02% Tween and 0.02% thiomersal. Standards were diluted in culture medium RPM1 1640 containing 1.0% BSA. Antibody bound radioactivity was separated from free tracer using anti-rabbit globulin (SAC cell). Bound radioactivity was counted in a Packard multigamma counter (70% efficiency). Assay for TGFP-like activity

TGFP-like activity in conditioned medium from COLO 16 cells was determined with the CCL64 mink lung cell growth inhibition bioassay (Danielpour et al., 1989). Briefly, CCL64 mink lung epithelial cells were seeded at lo5 per 0.5 ml of minimum essential medium (MEM)/lO% FCS in 24-well dishes. After 1 h, media were changed to MEM/O.2% FCS and conditioned media and TGF& standards (250 ~1) were added, 22 h later cells were pulsed with 0.25 PCi (40-60 Ci/mmol) of thymidine [methyl, 3Hl (Amersham) for 2 h at 37°C. Cells were then fixed with 1 ml of methanol-acetic acid 3 : 1 (v/v). The wells were washed 3 times with 1 ml of 80% methanol, after treatment for 60 min with 0.5 ml of 0.05% trypsin at room temperature, then 0.5 ml of 1% SDS was added and radioactivities counted. Results Effect of TGFP, on immunoreactive tioned medium from COLO 16 cells

PTHrP in condi-

Serial dilutions of conditioned medium from TGFP,-treated COLO 16 cells were compared with dilutions of standard PTHrP(l-84) in the radioimmunoassay. The standard curve was linear over the range of 80 pg/lOO ~1-2000 pg/lOO ~1, and 48 h conditioned medium from TGFP,-treated COLO 16

801 -2 % u

a

7c” 40 a

L

c

0

40 200 400 Cont 0.4 4 TGFB (PM)

Fig. 1. Radioimmunoassay (RIA) for PTHrP. (A) ng equivalents of PTHrP(l-84) per lo6 cells in conditioned medium against time in response to 200 pm (5 ng/ml) TGFP,. (B) Dose response of PTHrP(l-84) equivalents with increasing concentrations of TGFP,. * p < 0.05.

cells examined at 4, 8, 16, 32, 64-fold dilutions showed a dilution curve parallel to that of the PTHrP(l-84) standard and contained about 23.8 ng/ml PTHrP(l-84) equivalents. Western blot analysis has indicated that the predominant secreted form of PTHrP by COLO 16 cells is approximately the size of PTHrP(l-108) (data not shown). The time-dependent effects of TGFP, treatment of COLO 16 cells on immunoreactive PTHrP levels are shown in Fig. 1A. Treatment with TGFP, resulted in a significant increase in PTHrP levels from 24 to 48 h. Results are expressed as ng PTHrP(l-84) equivalents per lo6 cells to compensate for any growth effects of TGFP,, although treatment had no significant effect on growth rate up to 24 h. At 48 h TGFP, showed a slight stimulation of cell numbers: mean 2.9 X 106/ 2 cm2 well in the treated compared with 2.7 X 106/2 cm2 well in the untreated. Effects of treatment with increasing doses of TGFP, for 48 h are shown in Fig. 1B. TGFP, increased immunoreactive PTHrP levels significantly and in a dose-dependent manner from 0.4 to 400 pM, with half-maximum stimulation of PTHrP secretion in response to 40 pM TGFP, (Fig. 1B).

58

Effect of TGFP,

Effect of TGFp,

on PTHrP mRNA levels

on the stability of PTHrP mRNA

TGFP, treatment increased steady-state levels of PTHrP mRNA in COLO 16 cells in a time- (Fig. 2A) and dose-dependent manner (Fig. 2B). Several PTHrP-specific mRNA bands were observed, consistent with multiple mRNA species resulting from the action of multiple promoters and alternate splicing events (Suva et al., 1987, 1989; Mangin et al., 1988, 1989; Thiede et al., 1988). Fig. 2A shows that TGFP, initially increased PTHrP mRNA levels between 4 and 10 h with a maximal effect achieved at 48 h. At 24 h the PTHrP mRNA level was more than 2-fold that at time 0 (Fig. 2 A). At 4 pM TGFP, had a significant stimulatory effect on PTHrP mRNA with maximal stimulation achieved at 200 pM TGFP, (Fig. 2B).

The effect of TGFP, on the stability of PTHrP mRNA using the RNA polymerase inhibitor, actinomycin C, (0.8 x 10V6 M) was also examined. Cells were pre-treated for 12 h with TGFP,, then washed and incubation continued in the absence or presence of actinomycin C, or TGFP, alone or together (Fig. 3A). There was a rapid decay of PTHrP mRNA in cells treated with actinomycin C,, which was prevented to some extent in the continued presence of TGFP, (Fig. 3A and 38). The estimated half-life of PTHrP mRNA was l-2 h (Fig. 3B) and TGFP, treatment, albeit weakly, did increase the half life of PTHrP mRNA indicating that TGFP, does, to some extent, stabilize PTHrP mRNA.

Effect of the protein synthesis inhibitor, cycloheximide, on TGFP, regulation of PTHrP expression

Nuclear run-on assays and transient transfection analysis

To assess whether de novo protein synthesis is required for the stimulatory effect of TGFP, on PTHrP mRNA, cultures were pre-treated with TGFP, (200 pM) for 1 h and then with cycloheximide (5 X lop5 M) in the continued presence or absence of 200 pM TGFP, for 6 h. The expression of PTHrP mRNA was enhanced by cycloheximide and TGFP, further increased PTHrP mRNA levels in the presence of cycloheximide in COLO 16 cells (Fig. 20. The magnitude of the response was increased over that observed in the absence of cycloheximide. This small effect was consistently observed in three separate experiments. Immunoreactive PTHrP was not detected in the medium of the cycloheximide-treated cells at any time during the time course of the experiment.

B 0

A TGFfJ treatment

1.5 4

--II--+-+-+-+-+

To determine whether TGFP, altered the transcriptional activity of the human PTHrP gene, or whether up-regulation of PTHrP steady-state the TGFP, mRNA levels was solely the result of stabilization of the mRNA, transcription run-on assays and transient transfection assays were performed. Cells were exposed to medium containing vehicle or 200 pM TGFP, for 6 h and 32P-labelled nascent transcripts were prepared from isolated nuclei and allowed to hybridize with probes for PTHrP (pBRF61), an internal control (GAPDH), and with vector (pUC119) DNA which had been immobilized onto a nylon filter (Fig. 4A). Negligible hybridization to the vector DNA was observed with RNA prepared from either the control or the TGFB,-treated cells, nor was any signifi. cant difference noted between the control and

C

10 24 48 h -285 PTHrP

PTHrP

-“’

PTHrP TGFB Cycloheximide 18s GAPDH Ratio

Fig. 2. Effect of TGF/3, on PTHrP mRNA production in COLO 16 cells. (A) Total RNA from COLO 16 cells cultured in the presence or absence of TGFP, (200 PM) for the times indicated and hybridized with cDNAs for PTHrP and GAPDH. (B) Total RNA from COLO 16 cells cultured in increasing concentrations of TGFP,. With increasing concentrations of TGF@t, pre-spliced mRNA (i.e., below the 28s ribosomal marker) are more apparent. (C) Northern blot analysis for PTHrP mRNA of COLO 16 cells treated with 200 pM TGFp and/or cycloheximide. Values at the bottom of A, B and C represent the ratio between densitometric analysis of the combined PTHrP mRNA species relative to GAPDH mRNA normalized to the untreated control state which has a value of 1 .O.

59

PTHrP

TGFB Actinomycin GAPDH Ratio

\ -“:: IO-

--+,

52

\ \

\

\\

a0

4 0.5&L

’

__/--

'1

\ \ \

z z

TGFB 0

-4

\

0

\

b_N._

0.15I

01

TGFp+Act

\

62

I

_.‘0

I

9

I

2

3

4

Actinomycin

Time(h) Fig. 3. Analysis of PTHrP mRNA stability. (A) Total RNA from COLO 16 cells treated with actinomycin C, and/or TGFP, for the times indicated and hybridized with probes for PTHrP and GAPDH. Ratios at the bottom indicate the relative abundance of the combined PTHrP mRNA species to GAPDH mRNA, as determined by densitometric analysis of hybridization signal, normalized to the TGFP, stimulated track which has been assigned the value of 1.0. (B) Graphical plot of ratios of PTHrP mRNA to GAPDH mRNA against time of treatment as determined in A.

TGF/I,-treated in the transcription of GAPDH mRNA (Fig. 4A). In contrast, TGF& increased PTHrP gene transcription approximately 3-fold (Fig. 4A). The PTHrP-CAT reporter construct pSMR38 which contains the second and third promoter regions of the PTHrP gene along with sequences of exons III, IV and V, had high promoter activity (Fig. 4B). In the presence of 200 pM TGFP,, promoter activity was up-regulated approximately 2.5-fold (Fig. 48). This increase in transcription activity is concordant with the observed

increase in PTHrP transcription, steady-state mRNA and protein levels following TGFP, treatment of COLO 16 cells. TGF@,-like activity in conditioned medium from COLO 16 cells

We tested whether the COLO 16 cells secreted TGFp,-like activity using a mink lung epithelial cell assay. In this assay TGFP, is detected by its ability to cause a dose-dependent decrease in 13H]thymidine in-

60

corporation into mink lung epithelial cells. Conditioned medium which was assayed with and without prior acidification (which activates latent TGFP) showed dilution curves parallel to the standard (Fig. 5A). 72 h conditioned medium from COLO 16 cells contained about 3.2 ng/ml(160 PM) of TGFP, equivalent activity, of which 10% was already in the active form. TGFP, mRNA was also detected in COLO 16 cells and was itself stimulated by TGFP, in a dose-dependent manner (data not shown). Since these cells produce TGFP, some of which is in the active form, a neutralizing TGFP antibody (2G7) which recognizes was used to determine whether the TGFP, TGWL2&3 produced by COLO 16 cells modulated PTHrP production in these cells. Treatment with the neutralizing

antibody was carried out for 3 days. The levels of immunoreactive PTHrP in the conditioned medium were significantly reduced by 50% compared to levels in cultures incubated with non-neutralizing antibody (Fig. 5B). The effect of indomethacin munoreactivity

treatment on PTHrP

im-

Since TGFa and TGFP can stimulate prostaglandin production, for example in mouse calvaria (Tashjian et al., 1985>, the effect of TGFP, treatment was studied in the presence and absence of indomethacin (14 mM). No significant difference was seen in PTHrP production by the two groups over the dose range investigated (0.4-200 PM) (data not shown).

TGFB

Control

pBRF61

puc119

GAPDH

Relative

CAT Activity

Normalized

to pSMR38

No DNA Promoterless Vet tor pSMR 38 pSMR 38+TGFp Fig. 4. (A) TGFP, effects on PTHrP gene transcription. “‘P-labelled run-on TGF/3,-treated cells were hybridized to membranes with target sequences for obtained in two separate experiments. (B) Transfection of COLO 16 cells with mock transfected: no DNA) is expressed relative to that of pSMR38. The data transfections performed with different plasmid

transcripts prepared from nuclei isolated from control and 6 h PTHrP (pBRF61), pUC119 and GAPDH. Similar results were pSMR38. The promoter activity of each reporter construct (and are shown as the mean k SD from one of at least ten separate preparations on different days.

61

Discussion These studies demonstrate both the co-production of PTHrP and TGFp and the modulation of PTHrP production by TGFP, at the level of transcription in an epithelial tumor cell line. Indomethacin treatment of COLO 16 cells showed no indication of prostaglandinmediated effects. The time course of TGFPl stimulation of PTHrP mRNA levels, with an effect detectable within 2 h, is similar to that of a recent study using cultured uterine myometrial and endothelial cells (Casey et al., 19921, and implicates a direct effect of TGF& at the level of gene transcription. In experiments where protein synthesis was inhibited by culture of cells in medium containing cycloheximide, both in the absence and in the presence of TGFP,, PTHrP steady-state mRNA levels were enhanced, and this effect was seen within 2 h. Enhancement of PTHrP mRNA by cycloheximide alone has been described previously by Ikeda et al. (19901, and was interpreted to indicate the role of an additional protein factor in PTHrP expression. Our results suggest that the action of TGF& on PTHrP production may include a requirement for de novo protein synthesis, for example by the synthesis of an additional labile transcriptional regulator or degradative enzyme. TGFP, appears to exert an effect, albeit weakly, by stabilizing PTHrP mRNA and this stabiliza-

A 807 m ‘,o X 3

40-

\3 E, -O

o-

0.5

5

I

50 TGFBVpMI I

11256 1164 1116

”

Cont

114 CM diln

TGFf.!Ab

Fig. 5. TGFP, activity of COLO 16 cells. (A) Dilution cwve of conditioned medium (0) and acidified conditioned medium (M) against standard TGFD, (0). (B) Effect of a neutralizing antibody to TGFp (2G7) against a non-neutralizing antibody to TGFB (12H5); * p < 0.05.

tion may account for the delayed time-dependent induction of PTHrP mRNA. However, nuclear run-on and transient transfection analyses revealed that TGFPl does act directly at the level of gene transcription, up-regulating PTHrP transcriptional activity 2fold. There are several reports (Rizzino, 1988) indicating that TGFP, affects production of a number of proteins by transcriptional or post-transcriptional mechanisms or both (Rossi et al., 1988; Kim et al., 1990). Rossi et al. (1988) demonstrated that TGFP, directly activates transcription of the mouse type I collagen gene and that the effect is mediated by a nuclear factor 1 binding site located in the a2(1) collagen promoter. Another report demonstrates that auto-induction of TGFP, is mediated by the AP-1 complex (Kim et al., 1990). Examination of the 1100 bp genomic sequence 5’ to the initiating AUG of PTHrP revealed consensus sequences for nuclear factor 1 and a AP-1 binding site (Suva et al., 1989). This fragment of the gene is present in the CAT construct pSMR38, the CAT activity of which in transfection experiments could be up-regulated by TGFj3,. It is possible that either or both NF-1 and AP-1 are responsible for the stimulator-y effect of TGFP, on PTHrP production. Our results demonstrate directly the presence of TGFP mRNA in COLO 16 cells which responds to exogenous TGFP,. Furthermore TGFP is produced by COLO 16 cells in substantial amounts in both active and inactive forms. Although many cultured cell lines have been shown to produce TGFP which can be further regulated by TGFP (Obberghen-Schilling et al., 19881, in most cases it is latent inactive TGFP which is produced and requires activation for biological responses to occur. Since incubation with a neutralizing antibody to TGFP resulted in significant inhibition of PTHrP production this suggests that the active TGFp which we have demonstrated in COLO 16 conditioned medium can itself regulate the production of PTHrP. Regulation of the PTHrP gene is complex, involving alternate splicing events and the production of multiple messenger RNA species with alternate untranslated regions (UTR) in many tissues. The presence of the mRNA instability motif AUUUA in each of the variable UTRs is reminiscent of genes coding for a number of cellular cytokines (Miyata et al., 1980; Shaw and Kamen, 1986) and has led to speculation that PTHrP might function as a growth factor. To date increased PTHrP production has been shown to be associated with differentiation (Kremer et al., 1991) and regulation of its production is influenced by a number of factors which have potent growth effects, in particular epidermal growth factor (Rodan et al., 1988) and dexamethasone (Ikeda et al., 1989; Lu et al., 1989), which act at the level of transcription, in addition to CAMP (Rizzoli et al., 1990) and phorbol esters (Rodan

62

et al., 1988; Deftos et al., 1989), the mechanisms for which have not yet been evaluated. These studies thus demonstrate regulation of PTHrP expression by TGFP, and provide support for a possible feedback regulatory mechanism of PTHrP production by epithelial cells. PTHrP is a potential paracrine regulator of epithelial cell growth and differentiation, a hypothesis which is supported not only by its production by epithelial cells in many tissues, particularly in the developing fetus (Moseley et al., 19911, but also by recent experiments which demonstrate growth effects of PTHrP on epithelial cells in culture (Henderson et al., 1991). TGFP, like PTHrP is also a common product of squamous cancers (Derynck et al., 19871, and epithelial cells (Shipley et al., 1986; Akhurst et al., 1988; Rizzino, 1988) and it has been shown to promote differentiation of skin keratinocytes (Shipley et al., 1986; Akhurst et al., 1988; Rizzino, 1988; Matsumoto et al., 1990). The direct stimulation of PTHrP expression by TGFP and the implication of possible control of PTHrP production by constitutive TGFP might be relevant to the growth and maintenance of both normal and malignant epithelial cells. Further investigation of the interactions of PTHrP and TGF/3 at the gene level and in studies of epithelial growth and differentiation will assist in defining the role of PTHrP in these processes. References Akhurst, R.J., Fee, F. and Balmain, A. (1988) Nature 331, 363-365. Akhurst, R.J., Lehnert, S.A., Gatherer, P. and Duffie, E. (1990) Ann. NY Acad. Sci. 593, 259-271. Casey, M.L., Mibe, M., Erk, A. and MacDonald, P.C. (1992) J. Clin. Endocrinol. Metab. 74, 950-952. Chirgwin, J.M., Przybyla, A.E., MacDonald, R.J. et al. (1979) Biochemistry 18, 5294-5299. Danielpour, D., Dart, L.L., Flanders, K.C. et al. (1989) J. Cell. Physiol. 138, 79-86. Danks, J.A., Ebeling, P.R., Hayman, J.A. et al. (1989) J. Bone Miner. Res. 4, 273-277. Deftos, L.J., Gadzar, A.F., Ikeda, K. and Broadus, A.E. (1989) Mol. Endocrinol. 3, 503-508. Derynck, R., Goeddel, D.V., Ulrich, A. et al. (1987) Cancer Res. 47, 707-712. Dugaiczyk, A., Haron, J.A., Stone, E.M. et al. (1983) Biochemistry 22, 1605-1613. Ellis, A.G., Adam, W.R. and Martin, T.J. (1990) J. Endocrinol. 126, 403-408. Fenton, A.J., Kemp, B.E., Kent, G.N. et al. (1991) Endocrinology 129, 1762-1768. Gorman, C.M., Moffat, L.F. and Howard, B.W. (1982) Mol. Cell. Biol. 2, 1044-1051. Grill, V., Ho, P., Body, J.J. et al. (1991) J. Clin. Endocrinol. Metab. 73, 1309-1315. Harrison, J.R., Petersen, D.N., Lichtler, A.C. et al. (1989) Endocrinology 125, 327-333. Hayman, J.A., Danks, J.A., Ebeling, P.R. et al. (1989) J. Pathol. 58, 293-296. Henderson, J., Rabbini, S.A., Rhim, J. et al. (1991) J. Bone Miner. Res. 6 (SuppI. l), A588.

Ikeda, K., Lu, C., Weir, E.C. et al. (1989) J. Biol. Chem. 264, 15 743-15 746. Ikeda, K., Lu, C., Weir, E.C. et al. (1990) 3. Biol. Chem. 265, 539885402. Kemp, B.E., Moseley, J.M., Rodda, C.P. et al. (1987) Science 238, 1568-1570. Kim, S.J., Angel, P., Lafyatis, R. et al. (1990) Mol. Cell. Biol. 10, 1492-1497. Kremer, R., Karaplis, A.C., Henderson, J. et al. (1991) J. Clin. Invest. 87, 884-893. Law, F.M.K., Moate, P.J., Leaver, D.D. et al. (1991) J. Endocrinol. 128, 21-26. Lu, C., Ikeda, K., Deftos, L.J. et al. (1989) Mol. Endocrinol. 3, 2034-2040. Mangin, M., Webb, A.C., Dreyer, B.E. et al. (1988) Proc. Natl. Acad. Sci. USA 85, 597-601. Mangin, M., Ikeda, K., Dreyer, B.E. and Broadus, A.E. (1989) Proc. Natl. Acad. Sci. USA 86, 2408-2412. Martin, T.J. (1990) Q. J. Med. 280, 771-786. Martin, T.J., Nahorski, R., Hunt, N.H. et al. (1978) Clin. Sci. Mol. Med. 55, 23-29. Matsumoto, K., Hashimoto, K., Hashiro, M. et al. (1990) J. Cell. Physiol. 145, 95-101. Merendino, T.J., Insogna, K.L., Milstone, L.M. et al. (1986) Science 231, 7328-7332. Miyata, T., Yasunoga, T. and Nishida, T. (1980) Proc. Natl. Acad. Sci. USA 77, 7328-7332. Moore, G.E., Merrick, S.B., Woods, L.K. and Arabasz, N.M. (1975) Cancer Res. 35, 2684-2688. Moseley, J.M., Kubota, M., Diefenbach-Jagger, H. et al. (1987) Proc. Natl. Acad. Sci. USA 84, 5048-5052. Moseley, J.M., Hayman, J.A., Danks, J.A. et al. (1991) J. Clin. Endocrinol. Metab. 73, 478-484. Obberghen-Schilling, E.V., Roche, N.S., Flanders, K.C. et al. (1988) J. Biol. Chem. 263, 7741-7746. Rizzino, A. (1988) Dev. Biol. 130, 411-422. Rizzoli, R., Sappino, A.P., Aubert, M.L. and Bonjour, J.-P. (1990) Calcif. Tissue Int. 46 (Suppl. 21, A223. Rodan, S.B., Wesolowski, G., Ianacone, J. et al. (1988) J. Endocrinol. 122, 219-227. Rodda, C.P., Kubota, M., Heath, J.A. et al. (1988) J. Endocrinol. 117, 261-271. Rossi, P., Karsenty, G., Roberts, A.B. et al. (1988) Cell 52, 405-414. Seed, B. and Sheen, J.-Y. (1988) Gene 67,271-277. Shaw, G. and Kamen, R. (19861 Cell 46, 659-667. Shipley, G.D., Pittelkow, M.R., Willie, J.J. et al. (1986) Cancer Res. 46, 2068-2071. Stewart, A.F., Wu, T., Goumas, D. et al. (1987) Biochem. Biophys. Res. Commun. 146, 672-678. Strewler, G.J., Stern, P.H., Jacobs, J.W. et al. (19871 J. Clin. Invest. 80, 1803-1807. Suva, L.J., Winslow, G.A., Wettenhall, R.E.H. et al. (1987) Science 237, 893-896. Suva, L.J., Mather, K.A., Gillespie, M.T. et al. (1989) Gene 77, 95-105. Suva, L.J., Gillespie, M.T., Center, R.J. et al. (1991) J. Bone Miner. Res. 6 (Suppl. 11, A450. Tashjian, A.H., Voelkel, E.F., Lazzaro, M. et al. (1985) Proc. Natl. Acad. Sci. USA 82, 4535-4538. Thiede, M.A. (1989) Mol. Endocrinol. 3, 1443-1447. Thiede, M.A., Strewler, G.J., Nissenson, R.A. et al. (1988) Proc. Natl. Acad. Sci. USA 85, 4605-4609. Thiede, M.A., Harm, SC., Hasson, D.M. and Gardner, R. (1991) Endocrinology 128, 2317-2323. Vasicek, T.J., McDevitt, B.E., Freeman, M.W. et al. (1983) Proc. Natl. Acad. Sci. USA 80, 2127-2131. Yasuda, T., Banville, D., Hendy, G.N. and Goltzman, D. (1989) J. Biol. Chem. 264, 7720-7725.