Parvalbumin Is Expressed in Normal and Pathological Human Parathyroid Glands

Abstract

The parathyroid glands are of major importance in calcium homeostasis. Small changes in the plasma calcium (Ca²⁺) concentration induce rapid changes in parathyroid hormone (PTH) secretion to maintain the extracellular Ca²⁺ levels within the physiological range. Extracellular Ca²⁺ concentration is continuously measured by a G-protein-coupled Ca²⁺-sensing receptor, which influences the expression and secretion of PTH. The mechanism of signal transduction from receptor sensing to PTH secretion is not well understood, but changes in PTH secretion are tightly linked to changes in the cytosolic Ca²⁺ concentration. Using immunohistochemistry and Western blot analysis, we detected the EF Ca²⁺ binding protein parvalbumin (PV) in normal and in hyperplastic and adenomatous human parathyroid glands. The strongest PV signal was present in chief cells and water clear cells, whereas in oxyphilic cells only a weak signal was observed. Immunohistochemistry and in situ hybridization of the PTH indicated a co-localization of PV and PTH in the same cell types. Because changes in the cytosolic Ca²⁺ concentration are believed to influence the process of PTH secretion, a possible role of PV as a modulator of this Ca²⁺ signaling is envisaged.

Calcium (Ca²⁺) is implicated in the regulation of a variety of cellular processes, such as muscle contraction, cell division, cell fertilization, phagocytosis, hormone secretion, glucose metabolism, and even gene expression (Bading et al. 1993; Peunova and Enikolopov 1993). Because Ca²⁺ regulates such a multitude of processes, the maintenance of a steady concentration of Ca²⁺ both intra- and extracellularly is tightly regulated by several feedback control systems: Ca²⁺ absorption in the intestine, reabsorption/excretion in the kidney, and deposition/solubilization in the bones (Brown 1991; Hurwitz 1996). These three systems are regulated by three different hormones: parathyroid hormone (PTH), calcitonin (CT), and 1,25-dihydroxyvitamin D₃ [1,25 (OH)₂D₃], respectively.

The important role of the parathyroid gland as an overall regulator of Ca²⁺ homeostasis has been related to its exquisite capacity to sense and respond to minimal variations in the extracellular Ca²⁺ concentration (Brown 1991). Two different Ca²⁺-sensing receptors have been identified and cloned from parathyroid cells (Brown et al. 1993; Garrett et al. 1995). One of these receptors (the calcium receptor, CaR) belongs to the seven-transmembrane-helix G-protein-coupled receptor superfamily. Ca²⁺-dependent CaR activation results in the intracellular accumulation of inositol 1,4,5-triphosphate (IP₃) and diacylglycerol (DAG) (Brown 1991). An increase in intracellular Ca²⁺ on CaR stimulation has been described and linked to extracellular Ca²⁺-mediated alterations in PTH secretion (Brown 1991). A second, less-studied putative Ca²⁺ sensor (CAS) is a large protein (Mr 500,000) with a single transmembrane region, belonging to the low-density lipoprotein (LDL) receptor family (Lundgren et al. 1994).

The intracellular Ca²⁺ concentration is also tightly regulated by the ATP-dependent Ca²⁺ pump and by different types of Ca²⁺ exchangers located at the plasma membrane, endoplasmic/sarcoplasmic reticulum, and mitochondria (Pietrobon et al. 1990). In addition, levels of intracellular Ca²⁺ are modulated by cytosolic Ca²⁺ binding proteins, of which the super-family of EF-hand Ca²⁺ binding proteins is the largest and best characterized (Kretsinger 1980; Kawasaki and Kretsinger 1994). The best-known proteins of this superfamily are calmodulin, troponin C, calbindin D-28k, and PV.

PV is found in lower and higher vertebrates, including humans (Wnuk et al. 1982; Pauls et al. 1996). Several PV isoforms exist, which have been placed in a subfamily of the superfamily of EF-hand Ca²⁺ binding proteins (Moncrief et al. 1990; Nakayama et al. 1992). Two different isoforms, a and β, can be distinguished by their different biophysical characteristics (Föhr et al. 1993). However, in adult human only a single α-PV form is expressed (Föhr et al. 1993) whereas a β-PV is found exclusively extraembryonically in the placenta and in human tumors (Heizmann and Berchtold 1987). PV has been localized in metabolically active cells such as fast-twitch muscle fibers, GABAergic neurons, and several endocrine glands (Berchtold 1989). PV can function as a cytosolic Ca²⁺ buffer, thereby acting as a fast relaxation factor in muscle contraction or protecting cells from cytotoxic Ca²⁺ overload (Pauls et al. 1996). A recent study on PV-knockout mice shows that PV may confer an advantage in the performance of rapid, phasic movements in fast-twitch muscle fibers (Schwaller et al. 1999). Furthermore, a possible indirect role of PV in signal transduction by modifying the Ca²⁺ signal has been proposed (Pauls et al. 1996).

In this study we looked for the presence of PV in parathyroid glands with different states of activity: normal, hyperplastic and adenomatous tissue. We report on the presence of PV in normal as well as diseased parathyroid glands and co-localization with PTH.

Materials and Methods

Isolation of Tissue

Human parathyroid tissue was obtained from surgical procedures. Normal parathyroid glands (10 specimens: six men, four women, age range 26-68) were isolated from thyroidectomy biopsies removed for goiter (four cases), follicular adenoma (five cases), and papillary carcinoma (one case). All eight cases of secondary hyperplasia (five men, three women, age range 29-69) were related to chronic renal failure. Ten specimens of dissected parathyroid adenoma (four men, six women, age range 46-76) were received fresh from the Department of Pathology of the Ospedale Civile di Belluno, Italy.

Immunohistochemistry

Tissues were routinely fixed in 10% neutral buffered formalin and embedded in paraffin. Immunohistochemical reactions were carried out according to the avidin-biotinylated peroxidase complex method (ABC) (Vector Laboratories; Burlingame, CA) as previously reported (Doglioni et al. 1987). Three different antibodies against PV were used: a polyclonal antiserum 4064 against rat muscle PV (SWant; Bellinzona, Switzerland), a monoclonal antibody PA-235 against carp muscle PV (Sigma; St Louis, MO), and a monoclonal antibody Parv-19 against frog muscle PV (Sigma). The antiserum 224-1 against human PTH was from Signet (Dedham, MA). Antisera against calbindin D-28k and calretinin were from SWant. Antibody dilutions were 1:4000 for PA-235, 1:2000 for Parv-19, 1:3000 for 224-1, and 1:10,000 for anti-calretinin. The specificity of these antibodies was extensively tested (Kägi et al. 1987; Celio et al. 1988). Immunoreactivity was semiquantitatively evaluated with regard to the percentage of immunostained cells.

In Situ Hybridization

Dewaxed and rehydrated tissue sections were digested using proteinase K (Sigma). Fluorescein-labeled parathormone oligonucleotide probe (Novocastra; New Castle, UK) was added to each section. A rabbit F(ab') anti-FITC antiserum conjugated with alkaline phosphatase (Novocastra) was subsequently applied and the reaction developed with nitroblue tetrazolium (NTB).

Tissue Extraction and Western Blot Analysis

Frozen tissue was extracted by sonication with 5 volumes of 4 mM EDTA, pH 7.0, supplemented with 1 μM pepstatin hemisulfate A, 0.4 mM phenylmethylsuphonyl fluoride, 150 μM L-1-(tosylamido)-2-phenylethyl chloromethyl ketone, 1 μM leupeptin, and 30 U/ml trasylol. Protein concentration of cleared supernatants was measured using the microbiuret assay (Itzhaki and Gill 1964).

Tissue extracts (60 μg/lane) were separated by 15% SDS-PAGE. Separated lysates were transferred onto nitrocellulose membrane (Hybond ECL; Amersham Life Science, Poole, UK). Membranes were blocked in 3% bovine serum albumin, 1% fetal calf serum in 200 mM NaCl, 1 mM CaCl₂, and 50 mM Tris-HCl, pH 7.4, and then incubated with primary polyclonal antisera diluted in blocking solution. The secondary antibody was peroxidase-conjugated goat anti-rabbit IgG (Sigma). Staining of secondary antibody was performed using the ECL technique (Super Signal Substrate; Pierce, Rockville, IL).

Transblot ⁴⁵Ca²⁺ Overlay

The protocol of Maruyama et al. (1984) was followed. Proteins were first subjected to SDS-PAGE, then transferred to nitrocellulose membrane and finally incubated with ⁴⁵Ca²⁺ (specific activity 10-40 mCi/mg; Amersham). ⁴⁵Ca²⁺ binding proteins were visualized by autoradiography.

Results

Immunohistochemistry with Antibodies Against Different EF-hand Ca²⁺ Binding Proteins

To investigate whether PV, calbindin D-28k, and calretinin are present in normal and pathological parathyroid glands, 20 samples were analyzed by immunohistochemistry. In all cases strong PV-immunoreactive staining was found, indicating the presence of PV in normal (Figure 1A) and hyperplastic (Figure 1B) parathyroid, as well as adenoma (Figure 1C) and carcinoma (Figure 1D) of parathyroid glands. The same localization of PV-immunoreactive staining was found for three different antibodies directed against PV from different species. No immunoreactive signal was observed using antibodies directed against calbindin D-28k and calretinin (not shown).

PV immunoreactivity was localized in the chief cells (dark and light) and their morphological variants: the water clear cells and oxyphilic (oncocytic) cells (Figures 1A–1E). However, the intensity and intracellular localization of the immunoreactive staining was very different among the various cell types. Strongest PV signals were found in chief cells, with an even distribution throughout the cytoplasm and a variable staining in the nucleus. In water clear cells, strong PV staining was found in the cytoplasm close to the inner side of the plasma membrane and in the nucleus (Figure 1D). In oxyphilic cells, PV immunoreactivity was faint or absent (Figure 1E).

No immunostaining for any of the three Ca²⁺ binding proteins was found in thyroid gland, including C-cells (not shown).

Western Blot Analysis with Antibodies Against PV

Western Blot analysis was performed to demonstrate that the signals found in immunohistochemistry with different antibodies against PV were specifically due to the presence of human PV. Because normal human parathyroid glands are very small and difficult to isolate from the surrounding thyroid tissue, protein extracts of parathyroid tissue from hyperplasia and adenoma were used for this analysis (Figure 2).

A total of four hyperplastic parathyroids and six adenomas were investigated. In all protein extracts, the antibodies against rat PV detected a protein band with the same molecular weight (M_r of 12,000) as expected for human PV, indicating that the antibodies indeed detected human PV in immunohistochemistry (Figure 2). A strong signal for human PV was observed using the ECL technique, whereas only week signals were found using the less sensitive chloronaphthol staining.

Transblot ⁴⁵Ca²⁺ Overlay

⁴⁵Ca²⁺ overlay assays were performed to evaluate whether the human PV in the protein extracts from parathyroid bind Ca²⁺ ions. A single ⁴⁵Ca²⁺ signal for a protein band with an M_r of 10,000 was observed whereas at the position where PV is located no ⁴⁵Ca²⁺ signal was found. It is therefore assumed that PV is present at a rather low concentration in human parathyroid glands.

Immunohistochemistry with Antibodies Against Human PTH

Immunohistochemistry with antibodies against human PTH was performed to compare the localization of PV and PTH in different cells of the parathyroid glands.

For all tissues, the immunoreactivity for PV and PTH showed the same distribution and localization among the three different cell types. PTH immunoreactivity (Figure 1F) was strongest in chief and water clear cells, whereas it was absent in oxyphilic cells.

Western Blot Analysis with Antibodies Against Human PTH

Western blot analysis was performed to check whether the PTH-immunoreactive signal in immunohistochemistry is human PTH. The same protein extracts of hyperplastic and adenomatous parathyroid were used as in Western blots for PV identification.

The immunoreactive protein band reacting with antibodies directed against human PTH migrated at the correct molecular weight for human PTH (M_r 10,000), indicating that the antibodies indeed detected human PTH in immunohistochemistry (Figure 2C).

In Situ Hybridization of Parathyroid Tissue for PTH mRNA

PTH mRNA was found in all cells that showed positive immunoreactivity for PTH and PV, supporting the notion that these cells are actively producing PTH (Figure 1G).

Discussion

Our results show for the first time the presence of the EF-hand Ca²⁺ binding protein PV in normal and hyperplastic parathyroid glands as well as in adenoma and carcinoma of the human parathyroid. Calbindin D-28k and calretinin, two other EF-hand proteins of the same superfamily, were absent in these glands. PV was found by immunohistochemistry mainly in chief cells and water clear cells, whereas oxyphilic cells have a much lower concentration. In particular, the distribution of PV reactivity in nucleus and cytoplasm in normal and pathalogical parathyroid is different. It is predominantly cytoplasmic in normal (Figure 1A) and hyperplastic (Figure 1B) parathyroid and predominantly nuclear in adenomas and carcinomas of parathyroid. This is partially due to the fact that the nuclear reaction in normal gland is masked in part by a strong coloration with hematoxylin. However, no connection between the nuclear reactivity and the normal course of the pathology can be drawn. In the adenomas (Figures 1C and 1E), PV reactivity is also present (with a lower intensity compared to the adenoma) in the remaining suppressed gland.

Figure 1

Immunohistochemical localization of PV and PTH in normal and pathological parathyroid glands. Sections of normal parathyroid (A), hyperplasia (B), adenoma water clear cell with marked PV nuclear reactivity (C), carcinoma (D), adenoma oxyphilic cell with PV faint reactivity (E) were immunostained with primary antibodies against rat PV (A-E) and human PTH (F). Sections were counterstained with hematoxylin and antibodies were visualized by the ABC method. (G) In situ hybridization with a PTH-probe. Original magnification × 200.

Figure 2

Western blot analysis of human PV and human PTH. Proteins of soluble extracts from hyperplastic parathyroid and adenoma were separated by one-dimensional SDS-PAGE and stained with Coomassie brilliant blue (a) or transferred onto nitrocellulose membrane (b,c). PV- or PTH-immunoreactive protein bands were identified using antibodies against rat PV (b) or human PTH (c) and subsequent visualization by enhanced chemiluminescence. Rat PV (1 μg) (Lanes 1, 5), prestained low molecular weight protein standard (BioRad; Hercules, CA) (Lanes 2, 6, 10), protein extracts from hyperplasia (Lanes 3, 7, 11) and adenoma (Lanes 4, 8, 12), and human PTH (100 ng) (Lane 9).

Nuclear reactivity was also observed with other Ca²⁺ binding proteins (S-100, calretinin), but the biological implication is yet unknown. The PV signal in immunohistochemistry was further confirmed by Western blot analysis. The Western blot signal was strong using ECL visualization, whereas only faint signal was found using the conventional chloronaphthol method, indicating that the PV concentration in parathyroid tissue is rather low. Accordingly, ⁴⁵Ca²⁺ transblot overlay assays with parathyroid extracts did not identify a Ca²⁺ binding protein at the molecular range of PV. Because different dilutions of PV were tested and the detection limit for ⁴⁵Ca²⁺ was found to be below 1 μg (not shown), this indicated that the loaded samples (60 μg/lane) contained less than 1 μg of PV. Nevertheless, in all cases PV immunohistochemistry resulted in a clear and reproducible staining of parathyroid but not of thyroid cells. Therefore, PV immunohistochemistry could be used as a marker to identify and distinguish parathyroid from thyroid tissue in particular cases, such as when parathyroid gland has a follicular aspect.

In humans, the immunohistochemical localization of PV has been extensively used to map many areas of the brain (Pauls et al. 1996). In addition, PV has been immunohistochemically detected in human muscle (Föhr et al. 1993) and Leydig cells (Davidoff et al. 1993). On Western blot analysis, the presence of PV was also reported in human cerebellum and kidney, and with the more sensitive RT-PCR method PV mRNA was detected in muscle spindles, thymus, lung, diaphragm, and heart (Föhr et al. 1993). The parathyroid glands mediate their endocrine function through the production of PTH (Brown 1991). The three different cell types are assigned to different activities in PTH expression. Whereas chief and water clear cells actively express high amounts of PTH, oxyphilic cells are functionally inactive (de Lellis 1993). Interestingly, alterations in PTH secretion caused by changes in extracellular Ca²⁺ concentrations are tightly linked to changes of intracellular Ca²⁺ levels (Brown 1991). A close inverse correlation has been found between PTH secretion and intracellular Ca²⁺ concentrations (Brown 1991). High extracellular Ca²⁺ causes a decrease in PTH secretion and elicits an initial spike of cytosolic Ca²⁺, which is followed by a sustained increase (Brown 1991) or by oscillatory elevations of the cytosolic Ca²⁺ concentration (Miki et al. 1995; Ridefelt et al. 1995). On the basis of these data, it has been proposed that temporal or spatial characteristics of changes in the intracellular Ca²⁺ concentration may play a mediatory role in the regulation of PTH secretion (Brown 1991).

PV is believed to play an important role in intracellular Ca²⁺ buffering and in Ca²⁺/Mg²⁺ exchange (Pauls et al. 1996). Most PVs contain two functional metal ion binding sites that can bind Ca²⁺ with high affinity (K_Ca = 10⁷-10⁹ M^-1) and Mg²⁺ with moderate affinity (K_Mg = 10³-10⁵ M^-1) in a competitive fashion (Pauls et al. 1996). Under resting conditions, these sites are both occupied by two Mg²⁺ ions. An increase in the cytosolic free Ca²⁺ level to 1 μM or higher results in dissociation of Mg²⁺ and in subsequent binding of Ca²⁺ to both sites. Accordingly, Ca²⁺ spikes induced by brief depolarization pulses (Chard et al. 1993; Dreessen et al. 1996), ionophore, or KCl (Müller et al. 1996) are reduced in neurons microinjected with or overexpressing PV, demonstrating that PV buffers Ca²⁺ in a cytoplasmic environment. Therefore, it has been hypothesized that in neurons PV may act as a Ca²⁺ buffer that protects cells from cytotoxic Ca²⁺ overload (Nitsch et al. 1989; Yoshida et al. 1993; Pauls et al. 1996). In the parathyroid glands, PV may influence PTH expression or secretion by modifying intracellular Ca²⁺ signals. The same calcium receptor (CaR) of parathyroid glands is found in other tissues involved in Ca²⁺ homeostasis, e.g., kidney (Riccardi et al. 1995), and in C-cells of the thyroid gland (Garrett et al. 1995). Elevations of extracellular Ca²⁺ in both C-cells and chief cells of the parathyroid induce an increase in intracellular Ca²⁺. However, the consequences of this cytosolic Ca²⁺ increase are opposed. In chief cells it inhibits PTH secretion and in C-cells it stimulates calcitonin secretion (Fried and Tashjian 1986). Therefore, despite the fact that the two cell types are sensing extracellular Ca²⁺ through the same Ca²⁺-sensing receptor, a fundamental different mechanism for processing of the Ca²⁺ signal in both cells must be postulated. Whereas in parathyroid cells extracellular Ca²⁺-induced elevations of cytosolic Ca²⁺ are due to Ca²⁺ release from intracellular stores and subsequently to influx by voltage-independent Ca²⁺ channels, in C-cells the release from internal stores appears to play only a small role. Cytosolic Ca²⁺ elevations in the C-cells appear mainly to be due to Ca²⁺ influx from voltage-dependent Ca²⁺ channels (Scherubl et al. 1991). It is interesting that C-cells of the human thyroid gland are devoid of PV, whereas the distribution of PV in the kidney (Schneeberger and Heizmann 1986) matches fairly well with the distribution of CaR (Riccardi et al. 1995) in the distal tubule and proximal collecting duct, where the fine regulation of Ca²⁺ readsorption takes place. CaR is also present in the cortical thick ascending limb, the proximal tubule, and the medullar ascending limb, where PV has not been found.

Finally, the presence of CaR is not limited only to cells involved in mineral ion metabolism but has also been found in brain (Ruat et al. 1995). A comparison of PV expression and CaR expression in brain has yet to be established.

In conclusion, the presence of the Ca²⁺ binding protein PV in PTH-expressing parathyroid cells raises interesting questions regarding its potential role as a Ca²⁺ buffer and/or as a modulator of Ca²⁺ signals within these cells.