EGFR Activation Increases Parathyroid Hyperplasia and Calcitriol Resistance in Kidney Disease

Abstract

Calcitriol, acting through vitamin D receptors (VDR) in the parathyroid, suppresses parathyroid hormone synthesis and cell proliferation. In secondary hyperparathyroidism (SH), VDR content is reduced as hyperplasia becomes more severe, limiting the efficacy of calcitriol. In a rat model of SH, activation of the EGF receptor (EGFR) by TGF-α is required for the development of parathyroid hyperplasia, but the relationship between EGFR activation and reduced VDR content is unknown. With the use of the same rat model, it was found that pharmacologic inhibition of EGFR activation with erlotinib prevented the upregulation of parathyroid TGF-α, the progression of growth, and the reduction of VDR. Increased TGF-α/EGFR activation induced the synthesis of liver-enriched inhibitory protein, a potent mitogen and the dominant negative isoform of the transcription factor CCAAT enhancer binding protein-β, in human hyperplastic parathyroid glands and in the human epidermoid carcinoma cell line A431, which mimics hyperplastic parathyroid cells. Increases in liver-enriched inhibitory protein directly correlated with proliferating activity and, in A431 cells, reduced VDR expression by antagonizing CCAAT enhancer binding protein-β transactivation of the VDR gene. Similarly, in nodular hyperplasia, which is the most severe form of SH and the most resistant to calcitriol therapy, higher TGF-α activation of the EGFR was associated with an 80% reduction in VDR mRNA levels. Thus, in SH, EGFR activation is the cause of both hyperplastic growth and VDR reduction and therefore influences the efficacy of therapy with calcitriol.

In patients with chronic kidney disease (CKD), parathyroid hyperplasia and increased synthesis and secretion of parathyroid hormone (PTH) are critical contributors to renal osteodystrophy and cardiovascular complications responsible for high morbidity and mortality rates.^1–4 The efficacy of 1,25-dihydroxyvitamin D (calcitriol, 1,25D), the hormonal form of vitamin D, to suppress PTH gene expression and parathyroid cell growth has rendered calcitriol or its less calcemic analogs the treatment of choice for secondary hyperparathyroidism (SH)⁵; however, as kidney disease progresses, parathyroid vitamin D receptor (VDR) levels decrease in parallel with the severity of parathyroid hyperplasia,^6,7 rendering nodular hyperplasia, the most aggressive form of SH,⁶ unresponsive to calcitriol (analog) therapy.⁷ The lack of appropriate experimental models has impeded earlier characterization of the pathogenesis of the association between the severity of parathyroid hyperplasia and the reduction of VDR. Primary cultures of parathyroid cells fail to mimic the suppression of growth by either calcitriol or high calcium (Ca) in vivo, as a result of a rapid loss of VDR⁸ and Ca-sensing receptor.⁹ Furthermore, despite Ca-sensing receptor reduction, high Ca stimulates rather than suppresses proliferation in parathyroid cell cultures¹⁰; therefore, the first goal of these studies was to design a valid model of SH. We showed that enhanced parathyroid levels of the potent growth promoter TGF-α and its receptor, the EGF receptor (EGFR), are the cause of the aggressive parathyroid growth that doubles gland size within 1 wk of the onset of kidney disease in rats.¹¹ Furthermore, parathyroid TGF-α induces its own expression,¹² thereby generating a positive feedforward autocrine growth loop similar to that found in aggressive EGFR-driven carcinomas.¹³ From these findings in early uremia and the higher TGF-α levels in hyperplastic parathyroid glands (PTG) from patients with CKD,¹⁴ we hypothesized that progressively higher levels of parathyroid TGF-α from week 2 to week 4 after five-sixths nephrectomy (NX) contribute to worsen EGFR-driven hyperplasia, causing a proportional reduction in VDR content. This hypothesis was tested using Erlotinib, a potent and highly specific inhibitor of EGFR activation. Erlotinib effectively inhibits EGFR-tyrosine phosphorylation upon ligand binding, a step mandatory for EGFR-driven growth, in vivo and in vitro.^15,16 Erlotinib treatment completely prevented not only further increases in parathyroid cell growth but also VDR reduction, raising a critical question: How does TGF-α activation of the EGFR reduce VDR expression?

In breast cancer cells, TGF-α and/or EGF activation of the EGFR induces the synthesis of liver-enriched inhibitory protein,¹⁷ the truncated, dominant negative isoform of the transcription factor CCAAT enhancer binding protein-β (C/EBPβ), one of six C/EBP family members, all transcriptional regulators with decisive roles in cell growth and differentiation.¹⁸ LIP is a potent mitogen that is responsible for the most aggressive forms of tumor progression^19,20 and that coexists in a cell with liver-enriched activating protein (LAP)2 and LAP1, full-length and amino-terminal extended C/EBPβ isoforms. In contrast to LIP, full-length isoforms are transcriptional activators and induce cell-cycle arrest.¹⁹ Thus, cellular LAP/LIP ratios determine C/EBPβ biologic effects. Increases in LIP reduce LAP/LIP ratios, increase proliferation, and reduce LAP transactivation of target genes.^18,21

The recent demonstration that LAP induces VDR gene transcription in osteoblasts and renal cells²² led us to propose that, in the PTG, TGF-α/EGFR induction of LIP could worsen proliferation rates and antagonize LAP induction of VDR gene expression. This hypothesis was tested in the human epidermoid carcinoma cell line A431, in which, similar to hyperplastic parathyroid cells, growth is driven by TGF-α/ EGFR, and confirmed in PTG from patients with SH.

RESULTS

Our rat model accurately reproduces the association between parathyroid hyperplasia and VDR reduction of human SH. Table 1 shows that PTG enlargement from week 2 to week 4 after five-sixths NX and 0.9% dietary phosphorus (P) intake were associated with a 25% reduction in VDR, sufficient to cause resistance to 1,25D suppression of serum PTH. A 4-ng dosage of 1,25D, which effectively normalized serum PTH when administered every other day during the first week after five-sixths NX,²³ was no longer effective by week 4 (control 237.6 ± 59.5 pg/ml [n = 10]; 1,25D_w4 163.7 ± 75.8 [n = 9]). Furthermore, whereas PTG enlargement from week 2 to week 4 correlated directly with enhanced parathyroid TGF-α (r = 0.64, P < 0.02; n = 12), EGFR content remained unchanged. Table 1 and Figure 1 show that inhibition of TGF-α/EGFR signaling using the EGFR-tyrosine kinase inhibitor Erlotinib during weeks 3 and 4 after five-sixths NX was sufficient to prevent further increases in PTG size, TGF-α levels, TGF-α activation of the EGFR as measured by phosphorylated extracellular signal–regulated kinase 1/2 (P-ERK1/2) content, and proliferating activity (proliferating cell nuclear antigen [PCNA]-positive cells per area at week 4 was 0.506 ± 0.053 [n = 12] in uremic controls versus 0.211 ± 0.025 [n = 14]; P < 0.001 in Erlotinib-treated rats). Erlotinib had no effect on parathyroid EGFR expression. Erlotinib inhibition of growth could not be attributed to changes in serum P or Ca levels or the degree of renal failure (Table 2). Thus, enhanced TGF-α self-induction and the resulting increased activation of unchanged EGFR levels determined the progression of parathyroid hyperplasia in experimental SH.

Table 1.

PTG weight and parathyroid TGF-α, VDR, and EGFR content^a

Parameter	2 Wk	4 Wk	4 Wk (2 Wk Erlotinib)
PTG weight/body wt (μg/g)	1.56 ± 0.08 (30)	2.16 ± 0.14 (12)b	1.29 ± 0.09 (15)c
TGF-α (IOD/area)	42.6 ± 5.3 (8)	60.5 ± 6.4 (12)d	36.4 ± 6.2 (15)c
VDR (IOD/area)	129.8 ± 12.3 (8)	98.2 ± 7.7 (12)d	129.1 ± 4.7 (15)c
EGFR (IOD/area)	97.8 ± 4.9 (7)	98.1 ± 4.8 (7)	95.6 ± 5.7 (9)

^aData are means ± SEM (number of rats) of PTG weight corrected per body weight or immunohistochemical quantification of TGF-α, VDR, and EGFR levels expressed as integrated optical density (IOD)/area at 2 and 4 wk after the onset of kidney disease by five-sixths NX in rats that were fed a high-P diet and received either vehicle (2 wk or 4 wk, respectively) or daily doses of 6 mg/kg body wt Erlotinib in 200 μ l of DMSO from week 2 to week 4 (4 wk [2 wk Erlotinib]). Numbers in parentheses represent number of rats from two independent experiments.

^bP < 0.01 versus 2 wk.

^cP < 0.01 versus 4 wk, by unpaired t test.

^dP < 0.05 versus 2 wk.

Figure 1.

Erlotinib inhibition of EGFR activation prevents TGF-α self-upregulation and VDR reduction in experimental SH. Representative immunohistochemical analysis of changes in parathyroid levels of TGF-α, EGFR, P-ERK1/2, and VDR in response to kidney disease and high dietary P (HP) in established experimental SH (week 2 to week 4 after the onset of kidney disease by five-sixths NX) and their regulation through inhibition of TGF-α activation of the EGFR with Erlotinib administration from week 2 to week 4. Magnifications: ×200; ×600 in inset.

Table 2.

Body weight and serum chemistry^a

Parameter	Body Weight	Creatinine (mg/dl)	Total Ca (mg/dl)	Ionized Ca (mg/dl)	P (mg/dl)
UHP + vehicle (7)	269.0 ± 3.51	1.39 ± 0.12	9.83 ± 0.16	4.51 ± 0.10	8.43 ± 0.16
UHP + Erlotinib (9)	272.4 ± 3.68	1.20 ± 0.12	10.32 ± 0.20	4.68 ± 0.12	8.06 ± 0.64

^aData are means ± SEM. UHP, uremic (five-sixths NX) rats fed a high-P diet and receiving daily doses of either vehicle or 6 mg/kg body wt of Erlotinib in 200 μ l of DMSO. Numbers in parentheses represent number of rats from one of two independent experiments.

Erlotinib suppression of EGFR activation from week 2 to week 4 was also sufficient to prevent further reductions in parathyroid VDR content. Parathyroid VDR was 34% higher in Erlotinib-treated animals than in uremic controls (Table 1, Figure 1). Furthermore, 1 wk of Erlotinib treatment was sufficient to reverse the resistance to 1,25D suppression of PTH observed by week 4 after five-sixths NX. Whereas Erlotinib administration per se had no effect on serum PTH (Figure 2), when 1,25D (4 ng) administration during week 4 followed 1 wk of Erlotinib treatment (6 mg/kg, week 3), serum PTH decreased by 80% (Erlotinib_w3+1,25D_w4 68.7 ± 12.5 [n = 9]; Erlotinib_w3 225.0 ± 53.2 [n = 10]), an effect unrelated to changes in serum Ca or P levels (Table 3). These results establish a direct cause–effect relationship between EGFR activation and the reduction in parathyroid VDR responsible for calcitriol resistance.

Figure 2.

Erlotinib inhibition of EGFR-induced reduction of parathyroid VDR reverses resistance to 1,25D suppression of PTH in experimental SH. (Top) Experimental protocol (see Table 2 for the number of animals and serum chemistries per experimental condition). Bars and error bars represent means ± SEM of serum PTH levels at week 4 from rats that were subjected to five-sixths NX, were fed a high-P diet, and received daily doses of vehicle (4wHP); 1,25D, 4 ng in 200 μl of propylene glycol, every other day during week 4 (1,25D); Erlotinib, 6 mg/kg body wt in 200 μl of DMSO, daily during week 3 (Erlotinib); or Erlotinib during week 3 followed by 1,25D during week 4 (Erlotinib + 1,25D). **P < 0.01 for the statistical significance of the difference between the uremic control group (4wHP) and rats receiving Erlotinib + 1,25D by ANOVA.

Table 3.

Serum chemistry^a

Parameter	Creatinine (mg/dl)	Total Ca (mg/dl)	P (mg/dl)
Control (n = 10)	1.42 ± 0.08	9.44 ± 0.14	7.07 ± 0.76
1,25D (w4) (n = 9)	1.40 ± 0.08	9.98 ± 0.26	8.27 ± 0.70
Erlotinib (w3) (n = 10)	1.29 ± 0.04	9.56 ± 0.18	7.46 ± 0.53
Erlotinib (w3)-1,25D (w4) (n = 9)	1.35 ± 0.06	10.02 ± 0.12	7.23 ± 0.32

^aData are means ± SEM of serum levels at week 4 after five-sixths NX, from rats fed a high-P diet for 4 wk and receiving vehicle (control); 1,25D at a dose of 4 ng in 200 μ l of propylene glycol, every other day during week 4 [1,25D (w4)]; 6 mg/kg body wt of Erlotinib in 200 μ l of DMSO during week 3 [E (w3)], or Erlotinib during week 3 followed by 1,25D during week 4 [Erlotinib (w3)-1,25D (w4)].

Next, our hypothesis that TGF-α/EGFR-induced increases in LIP content could simultaneously induce growth and reduce VDR expression was tested in A431 cells. Figure 3 shows that Erlotinib inhibited TGF-α activation of the EGFR dosage dependently, as measured by parallel decreases in P-ERK levels, causing progressive reductions of LIP, PCNA, and growth rates (3-(4,5 dimethylthiazol-2-yl)2-5-diphenyl tetrasodium bromide assay). In fact, Erlotinib-induced increases in LAP/LIP ratios inversely correlated with PCNA levels (r = 0.90, P < 0.01). More significant, the highest Erlotinib concentration (0.25 μM) caused a three-fold increase in VDR content (Figure 3D). For direct evaluation of whether increases in LIP downregulate VDR gene expression, A431 cells were transfected with a luciferase reporter driven by the human VDR promoter. As expected for a C/EBPβ-driven promoter, ectopic LAP expression enhanced basal promoter activity by 104.8 ± 8.5-fold (n = 5 independent experiments, each with triplicate determinations per experimental condition). Figure 4, top, shows that the induction of promoter activity by ectopic LAP was impaired by 75.7% (n = 3) by simultaneous LIP coexpression and by 36.6% when endogenous LIP synthesis was induced through further activation of the EGFR with exogenous ligand (EGF, 17 nM). Erlotinib (0.25 μM) treatment of A431 cells exposed to EGF prevented EGF/EGFR induction of endogenous LIP and consequently EGF inhibition of VDR promoter activity. Western blot analysis of cell lysates from transfected A431 cells (Figure 4, bottom) corroborated that the expected changes in cellular LIP content were achieved upon transfection and/or treatment. Thus, TGF-α/EGFR-induced increases in LIP reduced LAP/LIP ratios, thereby increasing proliferation rates and reducing VDR gene transcription and VDR protein levels.

Figure 3.

Erlotinib inhibition of LIP synthesis arrests growth and increases VDR content in A431 cells. (A) Representative Western blot analysis of the changes in LAP, LIP, and PCNA expression induced by Erlotinib inhibition of TGF-α activation of EGFR (P-ERK1/2) in whole-cell extracts (40 μg of total protein) from A431 cells treated with increasing dosages of Erlotinib for 84 h. (B) Regression analysis of Erlotinib-induced changes in LAP/LIP ratios and A431 growth (PCNA/glyceraldehyde-3-phosphate dehydrogenase [GAPDH]). Results from two independent experiments are expressed as percentage of untreated cells. (C) Erlotinib dosage-dependent reduction of A431 growth. Bars and error bars represent means ± SEM of 3-(4,5 dimethylthiazol-2-yl)2-5-diphenyl tetrasodium bromide assays (see the Concise Methods section) from four independent experiments with 10 replicates per experimental condition. (D) Representative Western blot analysis of the increases in VDR expression in A431 cells in response to the inhibition of LIP synthesis by the highest (0.25 μM) dosage of Erlotinib (top) in whole-cell extracts from A431 cells treated as indicated in A. Bars and error bars represent means ± SEM of the densitometric analysis of VDR expression, corrected by protein loading (HDAC2) and expressed as percentage of control levels from at least four independent experiments.

Figure 4.

Increases in LIP antagonize LAP induction of VDR gene expression in A431 cells. Promoter reporter analysis in A431 cells transfected with a Luciferase reporter driven by the human VDR promoter containing two C/EBP binding sites (top) and a β-galactosidase expression vector to correct for transfection efficiency, with and without simultaneous coexpression of expression vectors for LAP (ectopic LAP) and/or LIP (LIP; see the Concise Methods section for details). Twenty-four hours after transfection, cells were treated for 24 h with vehicle, EGF (17 μM) to induce endogenous LIP synthesis, or Erlotinib (0.25 μM) to suppress EGFR-induction of LIP synthesis. Bars and error bars represent means ± SEM of the fold induction of VDR promoter activity from their respective controls from at least three independent experiments, each with triplicate determinations per experimental condition. *P < 0.05, **P < 0.01 versus control (ectopic LAP). (Bottom) Representative Western blot analysis of changes in cellular LAP (endogenous: human LAP [hLAP]; ectopic: rodent LAP [rLAP]) and LIP upon the transfection (empty vector, rodent LAP and LIP expression vectors [eLAP and eLIP, respectively]) and/or treatment described previously. LIP/GAPDH indicates cellular LIP levels induced by either ectopic expression or treatment.

Next, the contribution of TGF-α/EGFR induction of LIP to the association between high proliferation rates and reduced VDR was examined in hyperplastic PTG from patients with SH. Figure 5 confirms the reduced VDR levels reported for nodular hyperplasia,⁷ which was associated with enhanced TGF-α expression and EGFR activation, as measured by increased P-ERK1/2 content; however, high cross-reactivity of the antibody directed to the domain of the C/EBPβ molecule common for LAP and LIP impedes accurate immunohistochemical analysis of LIP expression in diffuse and nodular areas. Therefore, Western blot analysis was conducted in whole-cell extracts from hyperplastic human glands, which accurately measures LAP and LIP levels but fails to discriminate diffuse from nodular areas. Parathyroid PCNA correlated directly with LIP expression (r = 0.97, P < 0.001; n = 9; Figure 6B), suggesting higher LIP content in the highly proliferating nodular areas. Assessment of parathyroid LAP content showed variability in LAP1/LAP2 ratios (e.g., compare lanes 3 and 4 in Figure 6A) with higher LAP2 in six of nine glands. Because LAP1 seems to inhibit cyclin D1 directly,²⁴ increases in LIP may not be the only impairment to parathyroid LAP actions in SH. The higher TGF-α content and TGF-α activation of EGFR in nodular areas adjacent to diffuse areas of the same PTG coincided with lower VDR mRNA levels by in situ hybridization (Figure 7). Furthermore, real-time PCR quantification of VDR mRNA levels was 80% lower (P < 0.05) in three exclusively nodular glands compared with the two diffuse glands studied. The statistical significance of the differences was calculated from the 95% confidence interval for Cyde threshold VDR (lower limit 12.4; upper limit 20.02) for nodular glands (n = 3), which does not include the average Cyde threshold VDR mRNA (11.69) in diffuse glands. Thus, in human SH, enhanced TGF-α/EGFR induction of LIP is associated directly with the severity of growth and could contribute to reduce VDR mRNA and protein levels.

Figure 5.

Reduced VDR is associated with enhanced TGF-α and EGFR activation in human SH. Representative immunohistochemical (VDR and P-ERK1/2) and immunofluorescence (TGF-α) analysis of parathyroid VDR, TGF-α, and P-ERK1/2 content in diffuse and nodular areas in human PTG (see the Concise Methods section for details). Magnification, ×400.

Figure 6.

Increases in parathyroid LIP content correlate directly with growth rates in human SH. (A) Representative Western blot analysis of LAP, LIP, PCNA, and HDAC2 content from sections of parathyroid tissue with no histologic characterization of the type of hyperplasia (see the Concise Methods section for detail). (B) Regression analysis of LIP versus PCNA (corrected by protein loading). Each point represents LIP and PCNA levels per parathyroid tissue section obtained by densitometric quantification from Western blot analysis in whole-cell extracts (40 μg of total protein) from nine hyperplastic PTG from patients with SH.

Figure 7.

Parathyroid LIP content inversely correlates with VDR mRNA expression. (Top) Representative in situ hybridization for VDR mRNA levels using the tyramide amplification signal in the nodular and diffuse areas of the same PTG. (Bottom) VDR mRNA in nodular (n = 3) compared with diffuse (n = 2) human PTG as assessed by quantitative real-time PCR. *P < 0.05, calculated using the 95% confidence interval for the nodular glands (see the Concise Methods section). Magnification, ×400.

DISCUSSION

This work identified EGFR activation as the cause of the strong association between severe hyperplastic growth and the reduction in VDR that triggers the onset of calcitriol resistance in SH. More significant, it delineated a mechanism by which EGFR activation reduces VDR expression (Figure 8). Our rat model of established SH reproduced the increased proliferating activity, TGF-α expression, and decreased VDR, with unchanged EGFR levels of human SH. In human hyperplastic glands, TGF-α content parallels growth rates, whereas EGFR content is similar in normal tissue, parathyroid adenomas, and the parathyroid hyperplasia secondary to kidney disease.¹⁴ With the use of this validated model of SH, a clear mechanistic relationship was established between the degree of parathyroid EGFR activation and the severity of growth and VDR reduction. Erlotinib inhibition of parathyroid TGF-α/EGFR activation was sufficient to prevent not only TGF-α self-upregulation and, consequently, further increases in TGF-α, cell proliferation, and gland enlargement, but also VDR reduction. Indeed, Erlotinib treatment reversed the response of hyperplastic rat PTG to calcitriol from resistance to an 80% reduction of serum PTH. The pathophysiologic relevance of our finding that EGFR activation drives the reduction in VDR responsible for calcitriol resistance extends beyond SH. In the kidney, EGFR activation occurs upon nephron reduction, prolonged renal ischemia, or prolonged exposure to angiotensin II (AngII) and is the cause of accelerated progression of renal lesions, which include not only tubular hyperplasia but also mononuclear cell infiltration, tubulointerstitial inflammation, and interstitial fibrosis.^25,26 Similar to hyperplastic rat PTG, increases in TGF-α mediate AngII-induced activation of renal EGFR.²⁷ This underscores the importance of identifying the mechanisms enhancing parathyroid TGF-α expression in CKD; however, TGF-α is one of six EGFR-activating ligands, including EGF, amphiregulin, heparin-binding EGF, betacellulin, and epiregulin.²⁸ In CKD, increases in parathyroid or renal expression of EGFR ligands and/or G protein–coupled receptor (GPCR) ligands, known to transactivate the EGFR, could further aggravate the severity of EGFR-driven growth and VDR reduction induced by TGF-α.^14,29 Indeed, EGFR transactivation upon binding to their cognate G protein–coupled receptor mediates the mitogenic properties of molecules that circulate at high levels in patients with CKD, including endothelin I,^30–32 prostaglandin E2,³³ and AngII,^34,35 in the vasculature and in the hypertrophic growth of cardiac myocytes^30,33,34,36; therefore, in CKD, progressive decreases in VDR levels induced by increased EGFR activation/transactivation, at multiple sites in addition to the PTG, could partially account for the resistance to calcitriol antiproliferative and anti-inflammatory actions that accelerate the progression of SH and renal and cardiovascular disorders.

Figure 8.

EGFR activation is the cause of the association between hyperplastic growth and VDR reduction in SH. Enhanced parathyroid TGF-α expression, TGF-α self-upregulation, and the resulting EGFR activation determine the severity of growth and VDR reduction. Erlotinib inhibition of EGFR activation is sufficient to prevent not only further increases in parathyroid TGF-α and cell growth but also VDR reduction. The higher TGF-α/EGFR induction of LIP, a potent mitogen and an antagonist of VDR gene transcription, in nodular hyperplasia supports a role for parathyroid LIP as a molecular link between growth exacerbation and VDR reduction.

TGF-α/EGFR induction of LIP synthesis contributes to the link between severe EGFR-driven growth and VDR reduction. In human cancer, whereas LAP acts as a constitutive suppressor of cyclin D1,³⁷ LIP overexpression is associated with aggressive growth and invasive carcinomas,¹⁹ suggesting that LIP antagonism on LAP function acts as an oncogenic event. Also in normal adipocytes, increases in LIP disrupt terminal differentiation and induce a transformed phenotype.²¹ The strong correlation between parathyroid LIP expression and proliferating activity suggests that high TGF-α/EGFR induction of parathyroid LIP could enhance the propensity to nodularity. Indeed, higher parathyroid TGF-α levels and EGFR activation occur in nodular hyperplasia; however, the lack of nodule formation in rat SH impedes direct assessment of the contribution of enhanced LIP to the switch from diffuse to nodular hyperplasia. Reductions in parathyroid LAP1, known to suppress cyclin D1 directly,²⁴ could further aggravate the hyperplastic growth driven by increased LIP. These findings and calcitriol efficacy in suppressing parathyroid²³ and A431 cell growth³⁸ suggest novel calcitriol antiproliferative properties. Calcitriol could either downregulate TGF-α/EGFR induction of LIP synthesis or prevent LIP-induced decreases in cellular LAP/LIP ratio through induction of LAP, as demonstrated in osteoblasts,^22,39 renal cells,²² and monocyte-macrophages,^40,41 the last through a VDR-mediated mechanism.⁴⁰

In SH, LIP antagonism on LAP transactivation of the VDR gene could mediate the reduction in VDR levels upon EGFR activation. In A431 cells, Erlotinib inhibition of TGF-α/EGFR induction of LIP enhanced VDR protein levels and VDR promoter activity. In contrast, increases in cellular LIP, either by ectopic expression or by inducing endogenous synthesis through EGF treatment, markedly reduced LAP transactivation of the VDR gene. In support of a contribution of increases in LIP to EGFR-driven VDR reduction, the lower VDR mRNA levels occurred in the highly proliferative nodular areas, eliciting higher TGF-α activation of the EGFR. Therapeutic strategies that enhance LAP/LIP ratio should simultaneously control EGFR-driven disorders and calcitriol resistance.

In conclusion, our study highlights a key role for the degree of TGF-α expression, TGF-α self-upregulation, and TGF-α/EGFR induction of LIP synthesis, in PTG and A431 cells, in determining not only growth rates but also VDR gene expression and, consequently, cell-specific responsiveness to the 1,25D/VDR complex (Figure 8). Identification of the relative contribution of increases in TGF-α, EGFR-activating ligands, and EGFR transactivation in reducing VDR expression in CKD should help in the design of better therapies to overcome calcitriol resistance.

CONCISE METHODS

Experimental Protocols

Female Sprague-Dawley rats that were 5 to 6 wk or age and weighed 200 to 225 g were subjected to five-sixths NX as described previously²³ and fed a high-P diet for 4 wk (0.9% P, 0.6% Ca; Dyets, Bethlehem, PA). Two independent experiments were conducted per experimental protocol. In each, basal levels of parathyroid hyperplasia and VDR content were assessed 2 wk after five-sixths NX. In protocol 1, rats received either vehicle (200 μl of DMSO) or daily doses of 6 mg/kg body wt OSI-774 in 200 μl of DMSO (Erlotinib; provided by Genentech, San Francisco, CA), from week 2 to week 4, as described previously.¹¹ In protocol 2, rats received either vehicle or daily doses of 6 mg/kg body wt Erlotinib in 200 μl of DMSO only during week 3, followed by intraperitoneal administration of either vehicle or 4 ng of 1,25D in 200 μl of polypropylene glycol during week 4, as described previously.²³ Rats were killed at the end of week 4, and blood was drawn for analytical determinations. PTG were surgically removed and weighed on a CAHN-31 microbalance (Cahn Instruments, Cerritos, CA). PTG volume was measured indirectly, as PTG area after compressing the PTG to a constant height, as described previously.⁴² For immunohistochemistry, PTG were kept in 10% formaldehyde overnight and then transferred to 70% ethanol in water before mounting. Plasma P, creatinine, ionized Ca, and intact PTH levels were measured as described previously.¹¹ Protocols were approved by the Animal Study Committee at Washington University.

Human PTG

Sections of human PTG with no histologic diagnosis of diffuse or nodular hyperplasia were provided by Dr. Alex Brown (Washington University, St. Louis, MO). PTG with clear diagnosis of nodular or diffuse hyperplasia were received in Oviedo, Spain, from who had CKD and had undergone parathyroidectomy (Vall d'Hebron Hospital, Barcelona, Spain, and Santo Antonio Hospital, Oporto, Portugal). Immediately after parathyroidectomy, PTG were divided in two, with special care to minimize contamination with thyroid tissue, and either processed for histologic diagnosis or submerged in an RNAse inhibitor (RNAse later; Ambion, Austin, TX).

Immunohistochemistry

Rat PTG.

Immunohistochemical staining for PCNA, TGF-α and EGFR, and P-ERK1/2 (1:25; Cell Signaling, Danvers, MA) was performed as described previously.^11,23 For VDR (MAB 1360; Chemicon International, Temecula, CA; 1:30 dilution) immunostaining, tissue sections were microwaved in 10 mM citric acid buffer (pH 6.0) for 10 min, blocked with 2% gelatin, and incubated with primary antibody for 1 h at room temperature, followed by biotinylated secondary antibody (Sigma, St. Louis, MO; 1:200) and streptavidin-horseradish peroxidase conjugate. Immune complex formation was visualized with aminoethyl carbazole substrate-chromagen and quantified as described previously.^23,43

Human PTG.

For antigen retrieval, tissue sections were microwaved in 10 mM citric acid buffer (pH 6.0) for 15 min. Serial sections were blocked with 10% goat serum and incubated with primary antibodies overnight at 4°C followed by identical procedures as for rats except for a 1:100 dilution of the anti-VDR. For TGF-α immunofluorescence, secondary antibody (Alexa 594) was applied for 1 h at room temperature.

In Situ Hybridization for Human VDR

Paraffin sections of human PTG were processed according to the Tyramide Signal Amplification System. Briefly, sections were deparaffinized and rehydrated, and proteins were denatured with 200 mM HCl, followed by permeabilization with proteinase K, acetylation with 0.5% acetic anhydride in 1.3% triethanolamine in DEPC water, quenching with 0.3% hydrogen peroxide in methanol, prehybridization at 55°C for 2 h, and hybridization with 300 ng of Dig-RNA probe (600 bp from ATG) per slide at 72°C overnight. After a brief washing, samples were treated with RNase A/T1 at 37°C for 30 min. Slides were then washed at 72°C with 2 × SSC/50% formamide and subjected to the tyramide signal amplification system (GenPoint kit; K0620; DakoCytomation, Carpenteria, CA) according to the manufacturer's instructions. After DAB chromogen reaction, reactions were stopped by water and samples were mounted with glycerol vinyl alcohol mounting solution (Invitrogen, Carlsbad, CA).

Cell Culture and Proliferation Assay

A431 (ATCC, Manassas, VA) cells were grown in 10% FBS DMEM (Invitrogen, Carlsbad, CA) containing 4 mM l-glutamine, 4.5 g/L glucose, and 1.5 g/L sodium bicarbonate at 37°C in 5% CO₂. Cells plated at a concentration of either 10⁶ cells in a 10-cm plate or 10⁵ cells in 96-well plates were synchronized at G₀ by serum deprivation (serum-free DMEM) for 8 h and treated with Erlotinib (0, 0.025, 0.1, and 0.25 μM) in 2% FBS DMEM for 60 h followed by fresh treatment in 1% BSA DMEM up to 84 h. The colorimetric (3-(4,5 dimethylthiazol-2-yl)2-5-diphenyl tetrasodium bromide) assay kit (Chemicon International) measured A431 proliferation.

Promoter-Reporter Analysis

The luciferase reporter driven by the human VDR promoter containing the two reported C/EBP-binding sites²² was generated in our laboratory. Briefly, the fragment (−1554, +101) of the human VDR promoter was PCR-amplified with forward primer 5′-TCTTCCCACAAGAATCACCG-3′ and reverse primer 5′-TTCCCGCTGCTCCCGGGTTC-3′. The amplified DNA fragment was inserted into pGL2-Basic vector (between KpnI and Hind III). The β-galactosidase expression plasmid was obtained from Dr. Michael Rauchman (St. Louis University, St. Louis, MO). Wild-type rat C/EBPβ expression plasmids and their dominant negative isoform, cloned in a pSCT vector, originally described by Descombes and Schibler,⁴⁴ were provided by Dr. Kilberg (University of Florida, Gainesville, FL). A431 cells, plated in six-well plates at a concentration of 3 × 10⁵ cells/ml of media per well, were transfected using Myrus Transfection reagents following the manufacturer's protocol using 4 μl of transfection reagent per 1 μg of DNA. A total of 1 μg of the human VDR luciferase reporter and 0.1 μg of the β-galactosidase expression plasmid were transfected per well, together with 0.5 μg of the expression vectors for C/EBPβ (LAP) or its dominant negative isoform (LIP), when indicated. The total amount of transfected DNA was kept constant by addition of pGEM DNA when required. Upon an overnight incubation, cells were treated with vehicle, EGF (17 nM; Sigma), Erlotinib (0.25 μM), or the combination for 24 h. Cells were then lysed and luciferase and β-galactosidase activities were measured using Luciferase reporter system (Promega, Madison, WI) and Galacto-Light (Applied Biosystems, Foster City, CA), respectively. Western Blot analysis of lysates from 1.2 × 10⁶ transfected A431 cells (10-cm plates) estimated the efficacy of expression vectors and treatment in modulating LAP/LIP ratios.

Western Blots

For protein extraction, for A431 whole-cell extracts, cells washed with PBS were added to 500 μl of RIPA Buffer 1% (2% RIPA Buffer:Tris-HCl 50 mM [pH 7.4], 1% NP-40, 0.25% Na-deoxycholate, 150 mM NaCl 1 mM, EDTA, 1 mM PMSF, 1 μg/ml aprotinin, 1 μg/ml leupeptin, 1 μg/ml pepstatin, 1 mM Na₃VO₄, and 1 mM NaF) with 40 μl/ml fresh protease inhibitor cocktail (Roche, Indianapolis, IN), sonicated on ice for 40 s, and centrifuged for 20 min at 14,000 rpm at 4°C to collect supernatants. Human PTG were homogenized in hypotonic conditions (300 mM sucrose, 25 mM Tris ([pH 7.4], 0.4 mM EDTA, 0.4 mM EGTA, 0.4 mM PMSF, 1 mM benzamidine, 1 mM NaN₃), and total protein was obtained using Active Motif Kit. Protein concentration was quantified by Bio-Rad Protein Assay (Bio-Rad Laboratories Hercules, CA). A total of 30 to 40 μg of total protein was resolved by 12% SDS-PAGE and electroblotted onto nitrocellulose (Immobilon-P transfer membrane; Millipore, Waltham, MA) or polyvinylidene difluoride (Hybond-P; Amersham Bioscience, Buckinghamshire, UK) membranes. After probing overnight at 4°C with primary antibodies (VDR [Santa Cruz Biotechnology; Santa Cruz, CA; 1:500], P-ERK [Cell Signaling; 1:500], LAP [Biolegend, San Diego, CA; 1:300], C/EBPβ [LAP+LIP; Santa Cruz; 1:1000], PCNA [Zymed; 1:500], and loading controls HDAC2 [Santa Cruz; 1:2000] or glyceraldehyde-3-phosphate dehydrogenase [Calbiochem, 1:2000]), blots were visualized by enhanced chemiluminescence (SuperSignal West Pico; Pierce, Rockford, IL) and analyzed by Gel Pro Analyzer (Media Cybernetics, Silver Spring, MD).

RNA Isolation and Quantitative Real-Time PCR

Total RNA was extracted using TRI Reagent from Sigma (Sigma-Aldrich, St. Louis, MO) following the manufacturer's instructions and quantified by UV-VIS spectrophotometer (Nanodrop Technologies, Wilmington, DE) measuring absorbance at 260 and 280 nm. Samples with A₂₆₀/A₂₈₀ ratios under 1.8 were discarded. Quantitative reverse transcriptase–PCR was performed on an ABI Prism 7000 Sequence Detection System (PE Applied Biosystems; Foster City, CA) using a Taqman Universal PCR Master Mix for human VDR and 18s genes, the former used as target gene and the latter as endogenous control. The cDNA was obtained from 2 μg of total RNA using Taqman Reverse Transcription reagents following manufacturer's instructions. Both genes were analyzed by using predeveloped assays: Taqman Gene Expression Assays-On-Demand Hs00172113_m1 (VDR) and Eukaryotic 18s rRNA Endogenous Control reagent. All reactions (in quadruplicate) were performed by amplification of both the target and the endogenous genes in the same plate. Preliminary experiments checked for reaction sensitivity using serial dilutions of the cDNA (1:1, 1:10, 1:100, and 1:1000) amplified for target and endogenous genes. Relative quantitative evaluation (User bulletin #2) of the VDR gene was performed by comparison of threshold cycles between diffuse and nodular tissue, considering the former as the calibrator and the latter as the test sample, as described previously.⁴⁵

Statistical Analyses

ANOVA was used to assess statistical differences among all experimental groups under study. Multiple comparisons using the stringent Bonferroni test (or unpaired t test analysis when indicated) measured the statistical significance of the differences between two experimental groups.

DISCLOSURES

E.S. is a consultant/speaker for Genzyme and Abbott Laboratories. All other authors have no conflicts of interest to disclose.