Kidney International, Vol. 59 (2001), pp. 855–865

HORMONES – CYTOKINES – SIGNALING

p21WAF1 and transforming growth factor-a mediate dietaryphosphate regulation of parathyroid cell growth1

ADRIANA S. DUSSO, TRICIA PAVLOPOULOS, LECH NAUMOVICH, YAN LU, JANE FINCH,ALEX J. BROWN, JEREMIAH MORRISSEY, and EDUARDO SLATOPOLSKY

Renal Division, Department of Internal Medicine, Washington University School of Medicine, St. Louis, Missouri, USA

p21WAF1 and transforming growth factor-a mediate dietary Parathyroid (PT) hyperplasia, rather than hypertro-phosphate regulation of parathyroid cell growth. phy, is a major contributor to hyperparathyroidism in

Background. The parathyroid (PT) hyperplasia induced by patients with chronic renal failure [1–3]. Understandingrenal failure can be further enhanced by high dietary phosphate

the mechanisms regulating PT cell growth is crucial in(P) or completely abolished by P restriction. To identify poten-preventing PT gland enlargement, which results in in-tial mechanisms mediating these opposing effects of dietary P

on PT growth, this study first focused on p21WAF1 (p21) because creased synthesis and secretion of PT hormone (PTH)high P reduces while low P enhances serum 1,25-dihydroxyvita- and concomitant increases in bone resorption and oste-min D, whose potent antiproliferative properties result from itis fibrosa cystica, central features of renal osteodystro-the induction of p21. In addition to reducing p21, high phy [1, 3].P-induced PT growth could result from increased PT expres-

In normal conditions, PT cells rarely undergo mitosission of the growth promoter transforming growth factor-a[1, 3]. However, several factors, including uremia, hypo-(TGF-a), known to be elevated in hyperplastic and adenoma-

tous human PT glands. calcemia, hyperphosphatemia, and vitamin D deficiency,Methods. The time course for dietary P regulation of PT stimulate PT cell replication and, consequently, PTH

expression of TGF-a and p21 was assessed for seven days after secretion [4]. In rats, most of the PT growth induced by5/6 nephrectomy in rats and correlated with the degree of PT

high dietary P occurs within five days of the onset ofhyperplasia and secondary hyperparathyroidism.renal failure [5]. Conversely, dietary P restriction andResults. In P-restricted 5/6 nephrectomized rats, PT-p21

mRNA and protein increased by day 2, independent of changes 1,25-dihydroxyvitamin D [1,25(OH)2D3] administrationin serum 1,25-dihydroxyvitamin D, and remained higher than prevent both uremia-induced PT hyperplasia and en-in the high P counterparts for up to seven days. The PT hyper- hanced PTH secretion [1–8].plasia of the high P group could not be attributed to a reduction This study was designed to identify potential mecha-of PT-p21 expression from normal control values. Instead, PT–

nisms by which dietary phosphorus can either furtherTGF-a protein was higher in uremic rats compared with normalenhance or completely abolish PT cell growth in earlycontrols and increased further with high dietary P intake. PT

levels of proliferating cell nuclear antigen (PCNA), an index renal failure. We utilized 5/6 nephrectomized rats as anof cell mitoses, correlated inversely with p21 and directly with in vivo uremic model, because there are no PT cell linesTGF-a. Consistent with these findings, PT gland size and serum available. Moreover, in the most common in vitro model,PT hormone levels, similar in both dietary groups at day 2,

primary cultures of bovine PT cells, the concentration ofwere higher in the high P group by day 5. Induction of p21 byP in the incubation media has no effect on growth [2].low P and of TGF-a by high P was specific for the PT glands.

Dietary P had no effect either on intestinal growth or p21 or Abnormal PT growth could involve changes in eitherTGF-a protein content. the rate of cell proliferation or programmed cell death

Conclusions. These findings suggest that low P induction of (apoptosis). Our studies centered on P control of PT cellp21 could prevent PT hyperplasia in early uremia, whereasproliferation because no change in the number of cellshigh P enhancement of TGF-a may function as an autocrineundergoing apoptosis has been observed in the experi-signal to stimulate growth further.mental conditions affecting PT cell growth [1–3]. As formost eukaryotic cells, mitogens such as uremia and high1 See Editorial by Drueke, p. 1182.P could induce PT cells to abandon quiescence and divide

Key words: hyperparathyroidism, growth arrest, renal failure, uremia, by increasing the activity of cyclin/cyclin-dependent ki-hyperplasia, p21. nase (Cdk) complexes, decreasing the levels of Cdk in-

hibitors, or a combination of both [9–11]. Conversely,Received for publication June 16, 2000antimitogens, such as low dietary P, arrest cell growthand in revised form September 1, 2000

Accepted for publication September 13, 2000 by exerting the opposing effects on cell cycle regulators[9–11]. 2001 by the International Society of Nephrology

855

Dusso et al: p21WAF1 and TGF-a in PT growth856

Initial studies focused on dietary P regulation of PT Denmark). Plasma P and creatinine were determinedexpression of the Cdk inhibitor p21, based on the demon- using an autoanalyzer (COBAS-MIRA Plus, Branch-stration that low P increases whereas high P reduces burg, NJ, USA). Serum total calcium was measured byserum levels of 1,25(OH)2D3 [12, 13], whose potent anti- atomic absorption spectrophotometry using a Perkin-proliferative properties in several tissues result from di- Elmer 1100B spectrophotometer (Perkin-Elmer, Nor-rect transcriptional activation of p21 gene expression walk, CT, USA). Intact PTH levels were measured by[14–16]. The enhanced expression of the growth pro- an immunoradiometric assay specific for intact rat PTHmoter transforming growth factor-a (TGF-a) [17, 18] in (Immunotopics, San Clemente, CA, USA). 1,25(OH)2D3

hyperplastic and adenomatous human PT glands [19] led was extracted from plasma using the procedure devel-us to examine TGF-a as a potential mitogenic signal oped by Hollis [20] and measured with the radioreceptorinduced by high dietary P. assay of Reinhardt et al [21].

These studies demonstrate that in early renal failure,Reverse transcriptase-polymerase chain reactionphosphorus restriction may arrest uremia-induced PT

cell growth through a specific induction of the mRNA mRNA levels for the housekeeping gene glyceralde-and protein levels of the Cdk inhibitor p21, which occurs hyde 3-phosphate dehydrogenase (GAPDH) and theindependently of changes of serum 1,25(OH)2D3 concen- Cdk inhibitor p21 in PT glands and liver were determinedtrations. In addition, high P induction of TGF-a appears using semiquantitative reverse transcriptase-polymeraseto function as an autocrine signal to stimulate cell growth chain reaction (RT-PCR) followed by Southern blottingfurther, thus worsening PT hyperplasia and the resultant of the PCR products. Specifically, the pair of PT glandssecondary hyperparathyroidism. and a small portion of the liver from each rat were ho-

mogenized in 250 mL RNAzol (Tel-test Inc., Friends-wood, TX, USA). Total RNA from PT and liver tissuesMETHODSwas obtained according to the manufacturer’s protocols.

Male Sprague-Dawley rats, five to six weeks old,RNA content was quantitated spectrophotometrically at

weighing 220 to 240 g, and fed a standard normal chowa wavelength of 260 nm using a 10 mL microcuvette (Phar-

diet, were either sacrificed at day 0 or underwent 5/6macia Biotechnology AB, Uppsala, Sweden). Oligo dT-nephrectomy under chloral hydrate anesthesia (360 mg/kgprimed complementary DNA (cDNA) was prepared frombody weight). For 5/6 nephrectomy, several branches of1 mg of RNA using Superscript II RNAse H reversethe left renal artery were ligated, and the right kidneytranscriptase (GIBCO, Life Technologies, Grand Island,was excised [5]. Rats were fed starch based L-amino acidNY, USA). An amount of cDNA equivalent to 100 (fordefined diets (Dyets, Inc., Bethlehem, CA, USA). Dietsp21) and 25 ng (for GAPDH) of starting RNA was usedwere identical in carbohydrate, lipid, amino acid, andto ensure PCR amplification within the linear range for thefiber composition and contained a salt mix, which pro-PCR conditions specified later in this article. All primersvided either 0.2% P, 0.5% Ca (low P diet) or 0.8% P,were selected in different exons to ensure that genomic1% Ca (high P diet). Rats were sacrificed at time pointsDNA did not contribute to the measurements of mRNAranging from 12 hours to 38 days after the induction[22]. The 59 and 39 sequences for sense and antisenseof renal failure, as specified in individual experimentalprimers were ATG TCC AAT CCT GGT GAT GTCprotocols. On the day of sacrifice, PT glands, liver, andCGA CCT and TCA GGG TTT TCT CTT GCA GAA10 cm of the duodenal portion of the small intestineGAC CAA for p21 and AAT GCA TCC TGC ACCwere surgically removed. Blood was drawn for analyticalACC AA and GTA GCC ATA TTC ATT GTC ATAdeterminations. PT glands were weighed on a CAHN-31for GAPDH, respectively. Primers were purchased frommicrobalance (Cahn Instruments, Inc., Cerritos, CA,GIBCO BRL (Life Technology, Inc.). PCR conditionsUSA). Intestines were excised, washed with ice-cold sa-were denaturation at 948C for one minute, annealingline, and scraped to obtain the mucosal surface. Plasma558C for two minutes, and extension 728C for two min-was kept at 2208C, whereas all tissues were frozen inutes for 35 cycles. PCR products were separated in 1.2%liquid nitrogen and kept at 2708C until processing foragarose gels containing ethidium bromide in Tris-ace-RNA or protein. For immunohistochemistry, PT glandstate-ethylenediaminetetraacetic acid (TAE). The sizeswere kept in 10% formaldehyde overnight and thenof the expected PCR products were 496 bp for p21 andtransferred to 70% ethanol in water before mounting.515 bp for GAPDH. Gels were photographed on anAll experimental protocols were approved by the Ani-ultraviolet light box with Polaroid type 665 film to yieldmal Study Committee at Washington University Schoola negative. Gels were then denatured in NaCl/NaOHof Medicine (St. Louis, MO, USA).(1.5 mol/L/0.5 mol/L) for 10 minutes and neutralized in

Analytical determinations ammonium acetate/NaOH (1.125 mol/L/50 mmol/L) for10 minutes prior to transfer to a Zeta probe membraneIonized calcium (ICa) was measured by an ICa-specific

electrode (Model ICA-1; Radiometer, Copenhagen, in 10 3 standard saline citrate (SSC) overnight. After

Dusso et al: p21WAF1 and TGF-a in PT growth 857

cross-linking, membranes were prehybridized at 428C MA, USA) following the manufacturer’s protocol, priorto protein resolution by electrophoresis. After transferfor one hour and hybridized overnight at 428C with the

appropriate double-stranded DNA fragments of the se- and blocking, membranes were exposed to a mousemonoclonal anti–TGF-a (134 A-2B3; Oncogene Re-quence of the expected PCR products, labeled with the

Random Primed DNA labeling kit using 32P-dCTP and search Products, Cambridge, MA, USA). Blots werethen incubated with horseradish peroxidase-conjugatedrandom oligonucleotides as primers (Boehringer Mann-

heim, Mannheim, Germany). Blotted membranes were secondary antibody (Pierce, Rockford, IL, USA) for onehour at room temperature. Specific bands were detectedwashed twice in 1% sodium dodecyl sulfate (SDS), 0.2 3

SSC for 20 minutes. PCR products for GAPDH and using the Supersignal Substrate Working Solution byPierce and enhanced autoradiographic film (Amersham,p21 were quantitated by scanning densitometry using a

PhosphorImager (Molecular Dynamics, Sunnyvale, CA, Arlington Heights, IL, USA). Gels were scanned, andthe densitometric analysis was performed using Gel-ProUSA). The intra-assay coefficient of variation was 26%

(N 5 24). To minimize interassay variability, the follow- Analyzer (Media Cybernetics, Silver Spring, MD, USA).To compile the densitometric analysis from independenting precautions were taken to compile results from inde-

pendent experiments. In each independent experiment, experiments, for each time point, the average proteincontent of p21, TGF-a, or PCNA was obtained fromp21- and GAPDH-PCR products from samples from

uremic rats fed low and high P were quantitated in the three to four determinations per dietary group in thesame gel. For individual time points, the ratio LP/HPsame Southern blot for individual time points. For each

sample, mRNA levels of p21 were then normalized to (or HP/LP) was then calculated to express results asaverage increases in the low P group (or in the high PmRNA levels of GAPDH. To compile the results of

densitometric analysis from independent experiments, group) over that in the high P (or low P) counterparts.p21/GAPDH mRNA ratios for individual PT glands

Immunohistochemistry(N . 4) were averaged for each time point. The ratioslow P/high P and high P/normal control were then calcu- Immunohistochemical staining for p21, TGF-a, and

PCNA was performed on formalin-fixed, paraffin-embed-lated for every time point for each of the three indepen-dent experiments conducted. ded PT glands using mouse monoclonal anti-p21 anti-

body (SXM30; Pharmingen International), mouse mono-Western blot analysis clonal anti–TGF-a antibody (213-4.4; Oncogene), or

mouse monoclonal anti-PCNA (PC-10; Zymed Labora-Cell lysates were prepared from intestinal mucosaand PT glands (five to six PT glands from each experi- tories, Inc.), respectively, as primary antibodies, and a

commercial immunohistochemical staining kit (Histo-mental condition were pooled) in a modified RIPAbuffer [150 mmol/L NaCl, 1% NonidetP-40, 0.5% so- stain-Plus, mouse; Zymed Laboratories, Inc.). The speci-

ficity for the primary antibody was verified by immuno-dium deoxycholate, 0.1% sodium dodecyl sulfate (SDS),50 mmol/L, Tris, pH 8.0), and 1 3 protease inhibitor histochemical staining of rat PT tissue using mouse IgG1.

For quantitation, all precautions were taken to minimizecocktail; Boehringer Mannheim]. Total protein wasquantitated by the Bradford method (Bio-Rad Labora- variability. PT tissue was deparaffinized and rehydrated,

and endogenous peroxidase was quenched using 0.6%tories, Hercules, CA, USA). For p21 and proliferatingcell nuclear antigen (PCNA) analyses, equal aliquots of hydrogen peroxide in ethanol. When immunostaining for

either p21 or PCNA expression, tissue slices were micro-total protein (20 to 40 mg) were resolved by polyacryl-amide gel electrophoresis (PAGE) in 4 to 20% gradient waved at high power for 10 or 14 minutes, respectively.

Tissue was then blocked with 10% preimmune goat se-gels (Fisher Scientific, Pittsburgh, PA, USA) and wettransferred to 0.45 mm nitrocellulose membranes in 20% rum and incubated with primary antibody (5 mg/mL for

p21, 10 mg/mL for TGF-a, and 0.11 mg/mL for PCNA) formethanol in Tris-glycine at 100 V for one hour. Gelloading and transfer efficiencies were examined by stain- one hour at room temperature or overnight for PCNA.

Biotinylated secondary antibody was applied followeding total protein in the membranes with Ponceau Red(Sigma, St. Louis, MO, USA). After blocking, mem- by a straptavidin-horseradish peroxidase conjugate. The

immune complexes were visualized with aminoethyl car-branes were exposed to monoclonal antibodies for p21(SX118 mouse anti-p21; Pharmingen, San Diego, CA, bazole (AEC) substrate chromagen. Approximately 20

consecutive sections of tissue were cut for each PT gland.USA) and PCNA (PC10 mouse anti-PCNA; Zymed Lab-oratories, Inc., San Francisco, CA, USA) at 48C over- Immunohistochemical staining of TGF-a and PCNA pro-

teins was quantitated using a Nikon Diaphot-TMD micro-night. For TGF-a, equal aliquots of total protein (10 to40 mg) were resolved in 10 to 20% Tris-Tricine Gels (Bio- scope coupled to a camera and an image analysis system.

Images of stained tissue sections (magnification 3100)Rad Laboratories). To assess for glycosylated TGF-aisoforms, PT cell lysates were deglycosylated using pep- were acquired using a DAGE-330 color camera and cap-

tured with a Pentium P-166 IBM compatible computertide:N-glycosidase F (New England Biolabs, Beverly,

Dusso et al: p21WAF1 and TGF-a in PT growth858

Table 1. Serum chemistry

Day 2 Day 5 Day 7

Diet LP (10) HP (10) LP (9) HP (9) LP (10) HP (10)

Creatinine mg/dL 1.6560.05 1.8960.13 1.160.09 1.360.06 1.1860.08 1.1660.08ICa mg/dL 4.760.05 4.7260.06 4.4660.09 4.5260.06 4.860.07 4.7260.04Total Ca mg/dL 9.5060.10 9.9160.05a 10.060.13 9.4460.07a 10.260.16 9.860.13b

P mg/dL 6.1360.21 7.1560.18a 5.7860.29 7.460.31a 5.5860.31 7.7260.24a

Abbreviations are: LP, low phosphate; HP, high phosphate; ICa, ionized calcium; P, phosphate. Values represent the mean 6 SEM; (N) indicates number of rats.a and b indicate P , 0.01 and P , 0.05, respectively, for unpaired t test analysis between low and high dietary P for each time point

equipped with a Coreco frame grabber. The digitized RESULTSimages were converted to gray scale and analyzed using These studies were designed to assess the role for theImage-Pro Plus software (Media Cybernetics). All com- Cdk inhibitor p21 and TGF-a on the mitogenic effectsponents were obtained from Boyce Scientific, Inc. (St. of high dietary P and the antimitogenic properties of PLouis, MO, USA). Mize’s study provides a comprehen- restriction on uremia-induced PT growth.sive discussion of quantitative immunohistochemical

Effects of dietary P on p21 mRNA levels in the PTstaining using computer image analysis [23]. The thresh-glands of uremic ratsold calibration was as follows. An average gray value

obtained from a blank area of the slide (no tissue) was Immediately after 5/6 nephrectomy, rats were fed dietsused as maximal incident light; minimal light (the black of identical caloric composition containing either low Pvalue) was set as 0; and an average gray value (minus two (0.2%) or a high P (0.8%), for seven days. Ca content wasstandard deviations) obtained from the negative control increased to 1% in the high P diet to prevent the devel-tissue was used to set the level of the lowest detectable opment of hypocalcemia and was reduced to 0.5% in theintensity of staining. This threshold calibration sets the low P diet to avoid hypercalcemia. There was no signifi-lowest level of detectable staining at a point slightly cant difference in growth rate between dietary groups asdarker than the staining of the negative control. The demonstrated by a mild (2.1 6 1.0%, N 5 24) elevationintensity of staining of TGF-a was then quantitated using in body weight in P restricted animals by day 7.the optical density function of the software. The average The time course for changes in serum levels of creati-optical density per tissue section was calculated by divid- nine, ICa, total Ca, and P in both dietary groups is showning the sum integrated optical density by the sum area. in Table 1. In the uremic rats fed a low P diet, serum PPixels containing an intensity of staining outside the set concentrations were lower than in the high P group bythreshold were automatically excluded from calculations day 2 after 5/6 nephrectomy. Serum 1,25(OH)2D3 levels,by the software program. Excluded areas include tissue however, did not increase in the low P group (33.4 6with the same staining intensity as negative controls or 3.6 pg/mL, N 5 10) compared with those in uremic ratsnonstaining areas such as fibrous tissue or holes in the fed a high-P diet (30.2 6 3.6 pg/mL, N 5 10) and re-section. To eliminate variation, the microscope light mained similar in both dietary groups throughout thesource intensity used during image capture was kept seven-day period examined. The adjustments in dietaryconstant for all sections stained on a given day. Ca intake kept serum ICa similar in both dietary groups

Proliferating cell nuclear antigen expression was mea- throughout the length of the study.sured as number of nuclei stained positively for PCNA Figure 1 shows a representative time course for p21per PT gland section divided by the sum area of the mRNA expression in rat-PT glands up to day 7 after thesection. onset of renal failure. By day 2 after the onset of uremia,

For p21, no immunohistochemical quantitation was P restriction enhanced p21 mRNA twofold to threefoldperformed due to the high cytoplasmic p21 staining in above the levels expressed in the PT glands of rats fedrat PT tissue, which diminishes the accuracy of this the high P diet. PT-p21 mRNA remained higher in themethod of analysis in assessing nuclear p21 content. low P group compared with their high P counterparts

up to day 7. In contrast, in the PT glands of rats fed theStatistics high P diet, p21 mRNA levels remained unchanged with

Nonparametric unpaired t test or analysis of variance time. The compilation of densitometric analysis from(ANOVA) was employed to assess statistical significance three independent experiments confirmed the results de-of the differences between high and low dietary P groups picted in Figure 1. From day 2 after the onset of uremiafor individual time points or sequential variation within up to day 7, an average 1.89 6 0.12-fold higher p21

mRNA expression was observed in P-restricted rats com-experimental groups, respectively.

Dusso et al: p21WAF1 and TGF-a in PT growth 859

Fig. 1. Time course for the effect of dietaryphosphate (P) on parathyroid (PT) mRNAlevels of the cyclin-dependent kinase (Cdk)inhibitor p21WAF1 (p21). (Upper panel) A rep-resentative Southern blot of polymerase chainreaction (PCR) products for p21 and thehousekeeping gene GAPDH from PT glandsof rats fed a low phosphorus (LP) or highphosphorus (HP) diet at days 2 and 7 after5/6 nephrectomy. (Lower panel) p21/GAPDHmRNA ratios obtained after densitometricanalysis. Symbols are: (m) high phosphorus(HP), (j) low phosphorus (LP). Results rep-resent the mean 6 SEM, N 5 4.

pared with the high P group (P , 0.05). In contrast, nochanges in PT-p21 mRNA levels were observed in theuremic rats fed a high P diet compared with normalcontrols (1.03 6 0.17). The induction of p21 mRNA byP restriction at day 2 of renal insufficiency appears tobe specific for the PT glands since no change in p21/GAPDH mRNA levels was observed in the liver (LP,0.524 6 0.088, N 5 6; HP, 0.632 6 0.026, N 5 6) of theseuremic rats.

Effects of dietary P on p21 and PCNA proteincontent in rat PT glands in early uremia

Figure 2A depicts the results of Western blot analysesfor the effects of low and high dietary P on PT-p21protein content at days 2, 5, and 7 after 5/6 nephrectomy.Higher p21 levels were evident in whole cell extractsfrom a pool of PT glands from rats fed the low P dietat days 5 and 7. The barely detectable p21 protein contentin the PT glands of rats fed the high P diet, at days 2 and5, when 20 mg of total protein were used for Westernblots, precluded accurate densitometric analysis of PT-p21expression with time. Immunohistochemical staining forp21 at days 5 and 7 (Fig. 3) supports the increase inPT-p21 protein in P-restricted uremic rats suggested byWestern blot analysis.

Western blot analyses of PT expression of the markerof mitoses PCNA (Fig. 2B) show a higher PCNA contentFig. 2. Effect of dietary phosphate (P) on parathyroid (PT) p21 and

proliferating cell nuclear antigen (PCNA) protein content. (A) Repre- in the PT glands of rats fed the high P diet comparedsentative Western blot analyses of p21 protein expression in cell lysates with the levels expressed in the low P group at all time(20 mg of total protein for days 2 and 5 and 40 mg for day 7) from PT

points. Compilation of results from densitometric analy-glands of rats fed a low P (LP) or high P (HP) diet for two, five, orseven days after 5/6 nephrectomy. (B) Representative Western blot of sis of PT-PCNA protein content, shown in the bottomPCNA expression in cell lysates (40 mg of total protein) of PT glands left panel of Figure 6, demonstrated that the increase inof uremic (U) rats fed a low P (LP) or high P (HP) diet for two, five,or seven days after 5/6 nephrectomy. PCNA protein in uremic rats fed high P reaches statisti-

Dusso et al: p21WAF1 and TGF-a in PT growth860

Fig. 3. p21 protein expression in rat parathy-roid (PT) glands. Representative photomicro-graphs of immunohistochemical staining ofp21 in PT tissue of uremic rats fed a low Pdiet (LP) or a high P diet (HP) at days 5 and7 after the onset of renal failure. Magnification3100.

Fig. 4. Transforming growth factor-a (TGF-a)protein expression in the parathyroid (PT)glands of normal and uremic rats. (Upperpanel) Representative Western blot of a gly-cosylated TGF-a (pre-TGF-a) precursor inPT glands of normal and 5/6 nephrectomizedrats. Each lane represents equal amount oftotal protein (20 mg) from cell lysates froma pool of 10 glands from each experimentalcondition. (Lower panel) Representative West-ern blot of TGF-a expression in PT glands ofuremic (U) rats fed a low P (LP) or high P(HP) diet for two, five, or seven days after5/6 nephrectomy. Each lane represents equalamount of total protein (20 mg) from cell ly-sates from a pool of 10 glands for each experi-mental condition.

cal significance only by day 5. Immunohistochemical as- sion observed in secondary hyperparathyroidism in hu-sessment of PT PCNA expression in sections from at mans. Figure 4 (top panel) shows a Western blot analysisleast three PT glands demonstrated a uniform distribu- of TGF-a protein content in cell lysates from three pools,tion of mitotic activity and confirmed the results from each of 10 PT glands from normal and 5/6 nephrecto-Westerns. The number of nuclei staining positively for mized rats obtained 10 days after the onset of renalPCNA per PT gland area increased in the high dietary failure. Densitometric analysis demonstrated 87.6%P group from 0.62 6 0.293 at day 2 to 2.14 6 0.44 and higher levels of a putative TGF-a precursor of 22 to4.40 6 0.34 at days 5 and 7, respectively. In the P-restricted 25 kD in uremic rats compared with normal controlsrats PT-PCNA remained unchanged at days 2 (0.35 6 (P , 0.05). Treatment of PT-cell lysates with peptide N0.033) or 5 (0.43 6 0.11) and increased modestly by

glycosidase [24] identified this band as a glycosylatedday 7 (1.32 6 0.26). Differences in PT PCNA contentTGF-a precursor. We next examined whether dietary Pbetween dietary groups reached statistical significanceaffected the expression of the glycosilated TGF-a pre-at days 5 and 7.cursor in uremic rat PT glands. Figure 4 (bottom panel)

Effects of dietary P on TGF-a expression in rat PT shows that in uremic rats fed the high P diet, PT–TGF-aglands in uremia precursor was slightly higher (24% 6 4.3) at day 2,

peaked by day 5 (61% 6 2.0, P , 0.05) after the onsetInitial studies examined whether uremic rats repro-duced the pattern of the enhanced TGF-a protein expres- of uremia, and remained higher than in the P restricted

Dusso et al: p21WAF1 and TGF-a in PT growth 861

Fig. 5. TGF-a protein expression in rat parathyroid glands. Representative photomicrographs of immunohistochemical staining of TGF-a in PTtissue of uremic rats fed a low P diet (LP) or a high P diet (HP) at days 2, 5, and 7 after the onset of renal failure. Magnification 3100.

rats up to day 7 (52% 6 12, P , 0.05). The time course for normal controls to 0.063 6 0.008, (N 5 4) one monthafter 5/6 nephrectomy. Three days after reducing thethe enhancement of PT content of the TGF-a precursor

paralleled the increases in PT-PCNA protein levels, both dietary P content, the PT–TGF-a protein content de-creased to an average IOD/area of 0.0380 6 0.007 (N 5reaching statistical significance by day 5 after nephrec-

tomy. Immunohistochemical studies using a monoclonal 4, P , 0.05), indistinguishable from PT–TGF-a levels innormal controls. Conversely, the increase in PT-p21 wasantibody that recognizes all forms of TGF-a confirmed

the results of Western blots. Higher levels of TGF-a evident only by day 7 after switching dietary P intakefrom high to low (data not shown).protein are expressed in PT glands from uremic rats fed

a high P diet compared with their low P counterpartsEffects of dietary P on PT gland enlargement andthroughout the seven-day period examined (Fig. 5). Nosecondary hyperparathyroidism in early uremiaimmunostaining for TGF-a was apparent in sections of

PT tissue in which mouse IgG1 replaced the primary The time course for the effects of high and low dietary Pon serum PTH levels and PT gland weight after 5/6 ne-antibody for TGF-a (data not shown). Immunohisto-

chemical quantitation of PT–TGF-a expression with phrectomy is depicted in Figure 6. At day 2 after theinduction of uremia, there was no significant increase intime is depicted in the top panel of Figure 6. In uremic

rats fed a high P diet, TGF-a expression was only slightly PT gland size, and serum PTH levels are similar betweenthe two groups. However, by day 5, both PT gland weighthigher than in the P-restricted group by day 2. Maximal

expression of PT–TGF-a protein content was observed and PTH levels increased significantly in the high-P groupand remained higher than in P-restricted rats up to day 7.at day 5, with TGF-a levels persisting elevated in the

high P for up to seven days after the onset of renal Conversely, P restriction prevented both the enlarge-ment of the PT gland and the increases in serum PTH.failure. To validate further an association between high

dietary P and induction of PT–TGF-a expression, theEffects of dietary P on intestinal cell growthfollowing protocol was conducted: 5/6 nephrectomized

rats were fed a 0.9% P diet for one month and then Dietary P content had no effect on the rate of prolifer-ation of cells from intestinal mucosa, as shown in Figureswitched to a low P diet (0.2%) for one week. Immuno-

histochemical assessment of PT–TGF-a expression was 7. Intestinal PCNA protein content was similar in bothdietary groups at days 2 and 5. Furthermore, in contrastperformed at days 0, 3, and 7 after switching dietary P

from 0.9 to 0.2%. PT–TGF-a content increased from a to the elevation in PT-PCNA in uremic rats fed a high Pdiet, intestinal PCNA decreased by 36.3 6 12.5% (N 5basal average IOD/area of 0.038 6 0.005 (N 5 4) in

Dusso et al: p21WAF1 and TGF-a in PT growth862

Fig. 6. Effect of dietary P on PT expression of TGF-a and PCNA, PT gland weight, and serum PTH in early uremia. PT content of TGF-a andPCNA protein, serum PTH levels, and PT gland weight (PTGw) in rats fed a low-phosphorus (LP; m) or high-phosphorus (HP; j) diet for sevendays after 5/6 nephrectomy. The results represent the mean from at least eight rats per group (*P , 0.05).

Fig. 7. Effect of dietary P on intestinal ex-pression of PCNA, p21, and TGF-a in earlyuremia. Compilation of Western blot analysisof intestinal content of PCNA (m), p21 (d),and TGF-a (j) in rats fed a low-phosphorus(LP) or high-phosphorus (HP) diet for sevendays after 5/6 nephrectomy. The results areexpressed as HP/LP percentage and representthe mean from two independent experiments,each with at least four rats per experimentalgroup.

8) at day 7 in this dietary group. This reduction, however, differences in p21 and TGF-a between dietary groupsat days 2 and 5. However, the mild reduction in mitoticdid not reach statistical significance. Intestinal p21 and

TGF-a content changed as expected from the time activity, shown by lower PCNA at day 7 in the high Pgroup, occurred simultaneously with an increase in p21course for PCNA expression. There were no significant

Dusso et al: p21WAF1 and TGF-a in PT growth 863

(46.4% 6 19.7, N 5 7) and a reduction (28.1% 6 23.8, in both a slow-growing tissue, the liver at day two afterthe onset of uremia, and in a rapidly growing tissue, theN 5 4) in TGF-a. Neither change reached statistical

significance. intestinal mucosa, at days 5 and 7. No induction of p21mRNA by low dietary P occurred in the liver. The un-changed intestinal content of both p21 and PCNA with

DISCUSSIONdietary P manipulations supports the specificity for the

These studies demonstrate a role for enhanced expres- PT glands of the antimitogenic effects P restriction insion of the Cdk inhibitor p21 and TGF-a in the opposing these uremic rats.effects of dietary P on PT cell growth in early renal failure In contrast to the antiproliferative effects of increasesin rats. Initial studies examined dietary P regulation of in p21, a reduction of p21 mRNA levels of 50%, achievedp21 expression in rat PT glands. The rationale for the by transfecting antisense p21, prevented epidermal growthstudies was the well-known effects of high dietary P to factor (EGF)-induced growth arrest in A431 cells [30].reduce and of P restriction to increase serum 1,25(OH)2 Moreover, down-regulation of p21 accounts for growthD3 in normal individuals and in patients with mild and stimulation by TGF-b1 in human embryonic lung fibro-moderate renal failure [12, 13]. 1,25(OH)2D3, in turn, blasts [31]. A reduction in p21 appears to regulate thedirectly activates p21 gene transcription [14]. 1,25(OH)2 proliferative response of glomerular epithelial cells toD3 induction of p21 is responsible for the potent antipro- injury, and to underlie the development of progressiveliferative properties of 1,25(OH)2D3 in cells of the mono- glomerulonephritis [32]. In our studies, however, highcyte-macrophage lineage [14], keratinocytes [15], and dietary P had no detectable effect on p21 mRNA orprostate cancer cell lines [16]. The results from these protein levels, suggesting that reduction of intracellularinitial studies suggested that low P induction of p21 p21 was not the main mechanism mediating the mito-mRNA may mediate the antiproliferative effects of P genic properties of high P in the PT glands.restriction on uremia-induced PT cell growth. Two days In the search for mitogenic signals triggered by highafter 5/6 nephrectomy, P restriction decreased serum P dietary P, we focused on TGF-a. TGF-a, known to pro-and induced a twofold increase in PT p21 mRNA levels, mote growth not only in malignant transformation butwhich remained twofold higher than in the high P group in normal tissues [17, 18], is enhanced in hyperplasticup to day 7. This modest increase in p21 mRNA resulted and adenomatous human PT glands [19]. The mature,in a substantially higher expression of p21 protein in soluble form of TGF-a is produced from double proteo-P-restricted rats compared with the group on a high P lytic cleavage of a transmembrane precursor [17]. Bothdiet. These findings suggest post-transcriptional regula- membrane-anchored and soluble forms of TGF-a signaltion of p21 expression in PT glands, similar to that re- through the EGF receptor [17, 18], which is normallyported in fibroblasts and murine erythroleukemia cells expressed in the PT glands [19]. Our studies demonstrate[25]. Low P induction of PT p21 was not mediated by that induction of TGF-a by high dietary P could mediate1,25(OH)2D3. The degree of dietary P restriction utilized the enhancement of the rate of PT cell proliferationin these protocols was insufficient to increase serum already increased by uremia. Similar to the results of1,25(OH)2D3 even one month after 5/6 nephrectomy [5]. immunohistochemistry for TGF-a in human PT glands,This was an expected finding. A milder reduction of uremic rats expressed higher levels of a glycosilatedrenal function is required for a 0.2% P diet to induce TGF-a precursor. Moreover, time course studies using1,25(OH)2D3 synthesis [26] or a dietary P intake below both Western blot analysis of the glycosylated TGF-a0.2% in rats with the degree of renal failure employed precursor and immunohistochemical quantitation of allin the current protocols [27]. Clearly, pathways other TGF-a isoforms demonstrate that in uremic rats fed athan the 1,25(OH)2D3-vitamin D receptor axis mediate high-phosphorus diet, the levels of TGF-a were slightlyboth low P activation of p21 gene transcription and the higher compared with their low-phosphorus counter-post-transcriptional induction of p21 protein expression. parts by day 2 after 5/6 nephrectomy, peaked by day 5,

Induction of p21 was shown to be sufficient to induce and remained higher than in the low P group up to daygrowth arrest in monocyte-macrophages [14], keratino- 7. The increases in TGF-a induced by high P followed acytes [15] and human cancer cells [28], as well as to pattern identical to that of PT-PCNA protein expressionsuppress tumorigenicity in vivo [29]. In uremic rat PT and were also specific for the PT-glands. High dietary Pglands, a role for low P induction of p21 expression in did not increase intestinal TGF-a protein expression.arresting PT cell growth is supported by the finding that In the PT glands, increases in TGF-a content couldthe increases in p21 protein with time correlate inversely induce PT cell growth through an autocrine mechanismwith PT levels of PCNA, a marker of mitotic activity. or through less characterized juxtacrine and reverse jux-Taking into account the temporal changes in the rates tacrine pathways involving the transmembrane TGF-aof PT cell proliferation in early uremia [5], the specificity isoform from an adjacent PT cell, as demonstrated in

other tissues [33, 34]. Whichever the downstream mecha-for the effects of dietary P on p21 expression was tested

Dusso et al: p21WAF1 and TGF-a in PT growth864

nism for TGF-a signaling in the PT glands, it is clear nal triggered by uremia, further enhanced by high dietarythat high P induction of TGF-a is specific for the PT P, and prevented and/or totally counteracted by P restric-glands, occurs within five days of the onset of uremia, tion through induction of p21.and correlates with enhanced expression of PCNA, en-largement of the PT glands and increased serum PTH ACKNOWLEDGMENTSlevels. Further support for induction of TGF-a to medi- This work was supported in part by National Institute of Diabetesate the mitogenic properties of high dietary P came from and Digestive and Kidney diseases grants DK-09976, DK-30178, DK-

07126, and DK53774, and by GelTex Pharmaceuticals, Inc. (Waltham,a reduction in PT–TGF-a to levels found in normal con-MA, USA). The authors wish to express their appreciation to Dr.trols, which occurred three days after reducing dietaryChristine Sorenson (Department of Biochemistry) for valuable sugges-

P intake from 0.9 to 0.2% in uremic rats one month after tions for the experimental approach; Dr. Maria Isabel Colombo andCourtney Gelvermann (Department of Cell Biology and Physiology)5/6 nephrectomy. Although these powerful opposing ef-for their assistance in the initial settings of Western blots for rat PT-fects of high and low dietary P on PT growth and second-p21; Cindy Ritter (Renal Division) for her assistance in immunohisto-

ary hyperparathyroidism, it is still unclear how the PT chemical quantitation; and Dr. Gustavo Blanco (Department of Cellgland senses extracellular inorganic P levels. One candi- Biology and Physiology) and Marcos Vidal, M.S. (Renal Division) for

their help in the densitometric analysis and graphs of Western blotsdate molecule is the sodium-dependent P cotransporterand immunohistochemistry.PiT-1. PiT-1 expression in PT-glands is enhanced by

1,25-dihydroxyvitamin D and P restriction, two maneu- Reprint requests to Adriana S. Dusso, Ph.D., Renal Division, Box8126, Department of Internal Medicine, 660 South Euclid Avenue, St.vers that lead to arrest PT cell growth in uremia [35, 36].Louis, Missouri 63110, USA.Interestingly, similar to the specificity for the PT glandE-mail: adusso@im.wustl.edu

of dietary P regulation of growth, P restriction has noeffect on intestinal expression of the P cotransporter REFERENCES[35]. However, further research is mandatory to identify

1. Parfitt AM: The hyperparathyroidism of chronic renal failure: Asignal transduction pathways activated by dietary P todisorder of growth. Kidney Int 52:3–9, 1997regulate PT growth downstream from the sodium- 2. Slatopolsky E, Finch J, Denda M, et al: Phosphorus restriction

dependent/P cotransporter and upstream from the induc- prevents parathyroid gland growth: High phosphorus directly stim-ulates PTH secretion in vitro. J Clin Invest 97:2534–2540, 1996tion of PT-p21 or TGF-a. A post-transcriptional mecha-

3. Silver J, Bar Sela S, Naveh-Many T: Regulation of parathyroidnism, similar to that operating for P control of PTH cell proliferation. Curr Opin Nephrol Hypertens 6:321–332, 1997synthesis, namely, RNA-binding proteins that regulate 4. Naveh-Many T, Rahaminow R, Livni N, et al: Parathyroid cell

proliferation in normal and chronic renal failure rats: Effects ofthe stability of PTH gene transcripts [37, 38], could medi-calcium, phosphate and vitamin D. J Clin Invest 96:1786–1793, 1995ate dietary P induction of p21 and TGF-a. RNA-binding

5. Denda M, Finch J, Slatopolsky E: Phosphorus accelerates theproteins induced by hypophosphatemia may reduce the development of parathyroid hyperplasia and secondary hyperpara-

thyroidism in rats with renal failure. Am J Kidney Dis 28:596–602,stability of transcripts for growth promoter genes includ-1996ing TGF-a while increasing that of inducers of growth

6. Wang Q, Palnitkar S, Parfitt AM: Parathyroid cell proliferationarrest such as p21. in the rat: Effect of age, and of phosphate administration and

In summary, we have demonstrated that the mitogenic recovery. Endocrinology 137:4558–4562, 19967. Wang Q, Palnitkar S, Parfitt AM: The basal rate of cell prolifera-properties of high dietary P in rat PT glands in early

tion in normal human parathyroid tissue: Implications for theuremia cannot be attributed to down-regulation of the pathogenesis of hyperparathyroidism. Clin Endocrinol 46:343–349,expression of the Cdk-inhibitor p21, but to enhanced 1997

8. Szabo A, Merke J, Beier E, et al: 1,25(OH)2 vitamin D3 inhibitsexpression of TGF-a. High P induction of TGF-a in theparathyroid cell proliferation in experimental uremia. Kidney Inturemic rat PT gland may constitute an autocrine signal 35:1049–1056, 1989

that further stimulates uremia-induced PT cell prolifera- 9. Nasmyth K: Viewpoint: Putting the cell cycle in order. Science274:1643–1645, 1996tion. In contrast, the induction of PT-p21 mRNA and

10. Hatakeyama M, Herrera RA, Makela T, et al: The cancer cellprotein levels by P restriction, which occurs indepen-and the cell cycle clock. Cold Spring Harbor Symp Quant Biol 59:

dently of increases in serum 1,25(OH)2D3, may constitute 1–10, 199411. Roberts JM, Koff A, Polyak K, et al: Cyclins, cdks and cyclinan antimitogenic signal that counteracts an earlier mito-

kinase inhibitors. Cold Spring Harbor Symp Quant Biol 59:31–38,genic signal triggered by the onset of renal failure, thus1994arresting PT gland growth. The rapid return of PT–TGF-a 12. Portale AA, Halloran BP, Morris RC Jr: Physiologic regulation

content to normal levels by P restriction also suggests of serum concentration of 1,25-dihydroxyvitamin D by phosphorusin normal men. J Clin Invest 83:1494–1499, 1989that low P may counteract uremia-induced PT cell

13. Portale AA, Booth BE, Halloran BP, et al: Effect of dietarygrowth not only through the induction of p21 expression, phosphorus on circulating concentrations of 1,25-dihydroxyvitaminbut also by preventing the enhancement of PT–TGF-a. D and immunoreactive parathyroid hormone in children with mod-

erate renal insufficiency. J Clin Invest 73:1580–1589, 1984Based on these results and the evidence in the litera-14. Liu M, Lee MH, Cohen M, et al: Transcriptional activation of theture on the mechanisms of action of p21 [39, 40] and

Cdk inhibitor p21 by vitamin D3 leads to the induced differentiationTGF-a [17, 18], we conclude that the induction of PT– of the myelomonocytic cell line U937. Gene Dev 10:142–153, 1996

15. DiCunto F, Topley G, Calautti E, et al: Inhibitory function ofTGF-a expression may constitute an early mitogenic sig-

Dusso et al: p21WAF1 and TGF-a in PT growth 865

p21Cip1/WAF1 in differentiation of primary mouse keratinocytes tides linked to an internalization peptide inhibit human cancer cellgrowth. Cancer Res 57:1442–1446, 1997independent of cell cycle control. Science 280:1069–1072, 1998

16. Zhuang SH, Burnstein KL: Antiproliferative effect of 1alpha, 29. Yang ZY, Perkins ND, Ohno T, et al: The p21 cyclin-dependentkinase inhibitor suppresses tumorigenicy in vivo. Nat Med 1:1052–25-dihydroxyvitamin D3 in human prostate cancer cell line LNCaP

involves reduction of cyclin-dependent kinase 2 activity and persis- 1056, 199530. Ohtsubo M, Gamou S, Shimizu N: Antisense oligonucleotide oftent G1 accumulation. Endocrinology 139:1197–1207, 1998

17. Kumar V, Bustin SA, McKay IA: Transforming growth factor WAF1 gene prevents EGF-induced cell cycle arrest in A431 cells.Oncogene 16:797–802, 1998alpha. Cell Biol Int 19:373–388, 1995

18. Driman DK, Kobrin MS, Kudlow JE, et al: Transforming growth 31. Miyazaki M, Ohashi R, Tsuji T, et al: Transforming growth factor-beta1 stimulates or inhibits cell growth via down- or up-regulationfactor-a in normal and neoplastic human endocrine tissues. Hum

Pathol 23:1360–1365, 1992 of p21/Waf1. Biochem Biophys Res Commun 246:873–880, 199832. Shankland SJ, Floege J, Thomas SE, et al: Cyclin kinase inhibitors19. Gogusev J, Duchambon P, Soermann-Chopard C, et al: De novo

expression of transforming growth factor-a in parathyroid gland are increased during experimental membranous nephropathy: Po-tential role in limiting glomerular epithelial cell proliferation intissue of patients with primary or secondary uraemic hyperparathy-

roidism. Nephrol Dial Transplant 11:2155–2162, 1996 vivo. Kidney Int 52:404–413, 199733. Wong ST, Winchell LF, McCune BK, et al: The TGF-alpha pre-20. Hollis BW: Assay of circulating 1,25-dihydroxyvitamin D metabo-

lites using a novel single extraction and purification procedure. cursor expressed on the cell surface binds to the EGF receptor onadjacent cells, leading to signal transduction. Cell 56:495–450, 1989Clin Chem 32:2060–2063, 1996

21. Reinhardt TA, Horst RL, Orf W, et al: A microassay for 34. Shum L, Reeves SA, Kuo AC, et al: Association of the transmem-brane TGFa precursor with a protein kinase complex. J Cell Biol1,25-dihydroxyvitamin D3 not requiring high performance liquid

chromatography: Application to clinical studies. J Clin Endocrinol 125:903–916, 199435. Tatsumi S, Segawa H, Morita K, et al: Molecular cloning andMetab 58:91–98, 1984

22. Morrissey JJ, Ishidoya S, McCracken R, et al: Control of p53 hormonal regulation of PiT-1, a sodium-dependent phosphate co-transporter from rat parathyroid glands. Endocrinology 139:1692–and p21 (WAF1) expression during unilateral ureteral obstruction.

Kidney Int 50(Suppl 57):84–92, 1996 1699, 199836. Miyamoto K, Tatsumi S, Morita K, et al: Does the parathyroid23. Mize RR: Quantitative image analysis for immunohistochemistry

and in situ hybridization. J Neurosci Methods 54:219–237, 1994 see phosphate? Nephrol Dial Transplant 13:2727–2729, 199837. Silver J, Yalcindag C, Sela-Brown A, et al: Regulation of the24. Bringman TS, Linqquist PB, Derynck R: Different transforming

growth factor-a species are derived from a glycosylated and palmi- parathyroid hormone gene by vitamin D, calcium and phosphate.Kidney Int 56(Suppl 73):S2–S8, 1999toylated transmembrane precursor. Cell 48:429–440, 1987

25. Macleod KF, Sherry N, Hannon G, et al: p53-dependent and 38. Sela-Brown A, Silver J, Brewer G: Identification of AUF1 asa parathyroid hormone mRNA-39-untranslated region-bindingindependent expression of p21 during cell growth, differentiation,

and DNA damage. Gene Dev 9:935–944, 1995 protein that determines parathyroid hormone mRNA stability.J Biol Chem 275:7424–7429, 200026. Rodriguez M, Martin-Malo A, Martinez ME, et al: Calcemic

response to parathyroid hormone in renal failure: Role of phospho- 39. Waga S, Hannon GJ, Beach D, et al: The p21 inhibitor of cyclin-dependent kinases controls DNA replication by interaction withrus and its effect on calcitriol. Kidney Int 40:1055–1062, 1991

27. Wu S, Finch J, Zhong M, et al: Expression of the renal 25-hydro- PCNA. Nature 369:574–578, 199440. Chuang LSH, Ian HI, Koh TW, et al: Human DNA-(cytosine-5)xyvitamin D-24-hydroxylase gene: Regulation by dietary phos-

phate. Am J Physiol 271(1 Pt 2):F203–F208, 1996 methyltransferase-PCNA complex as a target for p21WAF1. Science277:1996–2000, 199728. Bonfanti M, Taverna S, Salmona M, et al: P21WAF1-derived pep-