The Extracellular Calcium-sensing Receptor Is Required for Calcium-induced Differentiation in Human Keratinocytes*

In cultured keratinocytes, the acute increase of the extracellular calcium concentration above 0.03 m M leads to a rapid increase in intracellular calcium concentration ([Ca 2 (cid:1) ] i ) and inositol trisphosphate production and, subsequently, to the expression of differentiation-re-lated genes. Previous studies demonstrated that human keratinocytes express the full-length extracellular calci-um-sensing receptor (CaR) and an alternatively spliced variant lacking exon 5 and suggested their involvement in calcium regulation of keratinocyte differentiation. To understand the role of the CaR, we transfected keratinocytes with an antisense human CaR cDNA construct and examined its impact on calcium signaling and cal-cium-induced differentiation. The antisense CaR cDNA significantly reduced the protein level of endogenous CaRs. These cells displayed a marked reduction in the rise in [Ca 2 (cid:1) ] i in response to extracellular calcium or to NPS R-467, a CaR activator, whereas the ATP-evoked rise in [Ca 2 (cid:1) ] i was not affected. Calcium-induced inhibition of cell proliferation and calcium-stimulated expression of the differentiation markers involucrin and transglutaminase were also blocked by the antisense CaR cDNA. When cotransfected with luciferase reporter vec-tors containing either the involucrin or transglutaminase promoter, the antisense CaR cDNA suppressed the calcium-stimulated promoter activities. These results indicate that CaR is required for mediating calcium signaling and calcium-induced differentiation in keratinocytes. basic vector (Promega, Madison, WI), linking them to the firefly luciferase gene to generate involucrin and transglutaminase promoter/reporter constructs (gifts from Dr. D. Ng), respectively. Transfection, Selection, and Luciferase Assay— First-passage kerati-* peroxidase-conjugated representative

Changes in the concentration of extracellular calcium affect the balance between proliferation and differentiation in epidermal keratinocytes (1,2). Elevation of the extracellular calcium concentration ([Ca 2ϩ ] o ) 1 above 0.03 mM (calcium switch) inhibits proliferation and induces the onset of terminal differentiation. One early response to the elevation of extracellular calcium is an increase in intracellular calcium concentration ([Ca 2ϩ ] i ) (3). Blocking the rise in [Ca 2ϩ ] i with an intracellular calcium chelator blocks the ability of extracellular calcium to induce differentiation (4). After the calcium switch, the levels of inositol 1,4,5-trisphosphate (IP 3 ) and diacylglycerol also increase rapidly (5,6). This is subsequently followed by elevated expression of differentiation-related genes such as involucrin (7) and transglutaminase (8,9), a substrate and enzyme, respectively, required for cornified envelope formation. Previous studies (10 -12) suggest the involvement of the extracellular calcium-sensing receptor (CaR) in mediating calcium signaling during keratinocyte differentiation. Activation of CaR with calcium or other polyvalent cations activates the phospholipase C signaling pathway, resulting in the generation of inositol 1,4,5trisphosphate and the release of calcium from intracellular stores (13,14). Human keratinocytes express the full-length CaR and an alternatively spliced variant of CaR lacking exon 5 (AltCaR) (15). Unlike the full-length CaR, AltCaR fails to mediate the acute IP 3 response to [Ca 2ϩ ] o . The full-length CaR message is maximally expressed in undifferentiated keratinocytes, but its level decreases as the cells differentiate. On the other hand, the message levels of AltCaR remain relatively unchanged throughout differentiation (15). These changes in CaR expression are consistent with the reduction in [Ca 2ϩ ] i and IP 3 response to [Ca 2ϩ ] o during differentiation (15) and further support a role for CaR in keratinocyte differentiation.
In the present study, we transfected an antisense CaR cDNA construct into undifferentiated keratinocytes to elucidate the role of CaRs in calcium sensing and calcium-induced differentiation. We report here that the full-length antisense CaR cDNA construct significantly reduced the protein level of endogenous CaRs. In these cells, the [Ca 2ϩ ] i response to an increase in [Ca 2ϩ ] o was inhibited, as were calcium-induced inhibition of proliferation and calcium-stimulated expression of the differentiation markers involucrin and transglutaminase. These results demonstrate that CaR-mediated calcium signaling is required for keratinocyte differentiation.

EXPERIMENTAL PROCEDURES
Cell Culture-Normal human keratinocytes were isolated from neonatal human foreskins and grown in serum-free keratinocyte growth medium (KGM; Clonetics, San Diego, CA) as described previously (9). Briefly, keratinocytes were isolated from newborn human foreskins by trypsinization (0.25% trypsin, 4°C, 16 h), and primary cultures were established in KGM containing 0.07 mM calcium. First-and secondpassage keratinocytes were plated in KGM containing 0.03 mM calcium and used in the experiments described below.
Vector Construction-To construct a vector expressing CaR antisense RNA, a 3.3-kb human CaR cDNA fragment (15) containing the entire open reading frame was subcloned in an antisense orientation into mammalian expression vector pcDNA3.1 (Invitrogen, Carlsbad, CA), which expresses a neomycin resistance gene. The 3.7-kb fragment of the human involucrin promoter (a gift from Dr. L. B. Taichman) (16)  . Finally, the cells were solubilized in 1 N NaOH, and the radioactivity in the washed acid precipitate was measured in a scintillation counter. Every experiment was done in triplicate and repeated twice.
Protein Analysis by Western Blot-Crude plasma membranes were isolated from cultured human foreskin keratinocytes and from human embryonic kidney HEK293 cells transfected with human cDNA for full-length or spliced variant CaR as described previously (15,19). Briefly, the cells were sonicated and centrifuged at 100,000 ϫ g for 30 min, and the membrane fractions were extracted with RIPA buffer (20 mM Hepes, pH 7.4, 150 mM NaCl, 1% deoxycholate, 1% Triton X-100, 0.1% SDS, 1 mM EDTA, 20 g/ml phenylmethylsulfonyl fluoride, 1 g/ml leupeptin, 1 g/ml pepstatin A, and 2 g/ml aprotinin). The protein concentration in these membrane preparations was determined by the BCA Protein Assay Kit (Pierce). Membrane protein samples were electrophoresed through 5% polyacrylamide gels and electroblotted onto polyvinylidene fluoride membranes (Immobilon-P; 0.45 m; Millipore, Bedford, MA). After blocking, the blot was incubated with a polyclonal anti-CaR antibody, either ADDR or BoCaR#3 (a gift from Dr. Dolores Shoback), which reacts with both full-length human CaR and its splice variant, AltCaR, or with another polyclonal anti-CaR antibody, hCaR4/6, which specifically reacts with AltCaR. The ADDR antibody was raised against the peptide corresponding to amino acids 215-236 of the human CaR (ADDDYGRPGIEKFREEAEERDI), and BoCaR#3 was raised against the peptide corresponding to amino acids 916 -936 of the bovine CaR (SSKSNSEDPFPQQQPKRQKQP) (20). The hCaR4/6 antibody was raised against a peptide corresponding to the fusion splice junction between exon 4 and exon 6 in human CaR (KKVEAWQVPFSNCSR). As a control for specificity, the antibody was preabsorbed with the specific synthetic peptide against which the antibody was raised. Subsequently, the samples were incubated with horseradish peroxidase-conjugated anti-rabbit IgG (Amersham Pharmacia Biotech) for 1 h at room temperature. The bound antibody was visualized using the SuperSignal West Dura Chemiluminescent Kit (Pierce) and subsequent exposure to x-ray film. The data presented are representative of three independent experiments.
Fluorescence Immunostaining of CaR in Cultured Keratinocytes-The localization of endogenous CaR proteins in cultured human keratinocytes was detected by immunostaining. Second-passage keratinocytes on coverslips were fixed with 4% paraformaldehyde for 30 min. After blocking, cells were incubated with a polyclonal anti-CaR antibody, BoCaR#2 (a gift from Dr. Dolores Shoback), which was raised against the peptide corresponding to amino acids 215-236 of the bovine parathyroid CaR (20) and subsequently incubated with fluoresceinconjugated anti-rabbit IgG. Finally, the cells on coverslips were mounted on glass slides using Gel-Mount (Biomeda, Foster City, CA) and examined with a Leica TCS NT/SP confocal microscope (Leica Microsystems, Heidelberg, Germany).
Northern Analysis-Total RNA was isolated using RNA-STAT 60 reagent (TEL-TEST "B", Inc., Friendswood, TX) according to the manufacturer's instructions. Thirty g of total RNA was electrophoresed on a 1% agarose-formaldehyde gel and blotted onto a nylon filter (Hybond-N ϩ ; Amersham Pharmacia Biotech). The blots were hybridized with the appropriate 32 P-labeled cDNA probe (P1-2 for involucrin (a gift from Dr. Howard Green), hTG for transglutaminase (a gift from Dr. Robert Rice), and 18 S ribosomal RNA probe for normalization). To detect the antisense CaR transcript, the RNA blots were hybridized with a 32 Plabeled 320-nucleotide sense strand RNA probe. The blots were then washed twice in 1ϫ SSC/0.1% SDS and twice in 0.1ϫ SSC/0.1% SDS at 65°C. Bound radioactivity was detected by exposing the blot to x-ray film. The data presented are representative of three independent experiments.

Measurement of [Ca 2ϩ ] i -The [Ca 2ϩ
] i response to elevated [Ca 2ϩ ] o was measured using a Dual-wavelength Fluorescence Imaging System (Intracellular Imaging Inc., Cincinnati, OH) as described previously (22). Briefly, preconfluent keratinocytes were transfected with the antisense human CaR cDNA construct or pcDNA3.1 vector on a coverslip in KGM containing 0.03 mM calcium. Three days after 300 M G418 selection, transfected cells were loaded with 5 M Fura-2 AM (Molecular Probes, Eugene, OR) in 0.1% Pluronic F127 in buffer A (20 mM Hepes, 120 mM NaCl, 5 mM KCl, 1 mM MgCl 2 , 1 mg/ml sodium pyruvate, and 1 mg/ml glucose) containing 0.07 mM calcium. Cells were then washed and measured in buffer A containing 0.03 mM calcium before exposure to 1.2 or 5 mM calcium or to the calcimimetic compound NPS R-467 (provided by Dr. Edward F. Nemeth; NPS Pharmaceuticals, Salt Lake City, UT), a selective CaR activator (24). The cells were alternately illuminated with 340 and 380 nm light, and the fluorescence at emission wavelength 510 nm was recorded. The signals from 15-40 single cells for each measurement were recorded. Each sample was calibrated by the addition of 20 M ionomycin (R max ) followed by 20 mM EGTA/ Tris, pH 8.3 (R min ). Intracellular calcium concentration was calculated from the ratio of emission at the two excitation wavelengths based on the formula (23) [Ca 2ϩ ] i ϭ K d Q(R Ϫ R min )/(R max Ϫ R), R ϭ F 340 /F 380 , Q ϭ F min /F max at 380 nm, and K d for Fura-2 for Ca 2ϩ is 224 nM.
[Ca 2ϩ ] i Measurement in Suspended Keratinocytes-[Ca 2ϩ ] i measurement of suspended keratinocytes was carried out as described previously (11). Briefly, undifferentiated keratinocytes were transfected with the antisense human CaR cDNA construct or pcDNA3.1 vector and selected by G418 as described above. Transfected cells were trypsinized and loaded with Fura-2 AM in buffer A at room temperature for 1 h. Cells were then washed and resuspended in buffer A at a cell density of 1 ϫ 10 6 cells/ml. Cells were placed in a quartz cuvette, and [Ca 2ϩ ] i was measured in buffer A containing 0.03 mM Ca 2ϩ , after which the response to 1 M NPS R-467 was recorded. Then 1.2 mM Ca 2ϩ was added to record the response to extracellular calcium. Fluorescence was recorded with a fluorimeter (650-40; PerkinElmer Life Sciences) using 340 and 510 nm for excitation and emission wavelengths, respectively. Each sample was calibrated by the addition of 20 M ionomycin (F max ) followed by 0.1% Triton X-100 and 20 mM EGTA/Tris, pH 8.3 (F min ).
where K d for Fura-2 for Ca 2ϩ is 224 nM. Each data point shows the mean Ϯ SD of triplicate determinations.

Expression of CaR Proteins in Human Keratinocytes-We
previously reported that human keratinocytes express the transcripts of the CaR, both the full-length form and its alternately spliced variant (AltCaR) lacking exon 5 (15). To detect the presence of CaR proteins, we analyzed membrane extracts from preconfluent human keratinocytes by Western blotting. The CaR was detected using a peptide affinity-purified polyclonal antibody against human CaR, ADDR (Fig. 1A). Although this antibody could detect both forms of CaR, the reaction with full-length CaR is stronger than that with AltCaR (Fig. 1A). ADDR specifically detected three bands of 120, 160, and 185 kDa in the human keratinocytes grown in serum-free medium containing either 0.03 or 1.2 mM calcium (Fig. 1A, left panel). ADDR also detected a minor band of 130 kDa in keratinocyte membranes, which was best observed on longer-exposed film (data not shown). The 160-kDa band corresponds to one of the two major glycosylated forms of CaR (with sizes of 140 and 160 kDa), whereas the 120-kDa band corresponds to the nonglycosylated form of CaR expressed in HEK293 cells transfected with cDNA for the full-length human CaR (Fig. 1A, left panel). The 130-kDa band corresponds to the single band of CaR expressed in HEK293 cells transfected with cDNA for AltCaR. The size of the 130-kDa band is consistent with the spliced variant lacking 77 amino acids of exon 5 that we have previously shown has altered glycosylation. The identity of the 185-kDa band remains unclear, although it specifically reacted with ADDR antibody. Preincubation of ADDR antibody with the specific peptide to which it was raised prevented detection of these bands (Fig. 1A, right panel). These bands were also detected by another anti-CaR antibody, BoCaR#3, which was raised against a peptide derived from the intracellular domain of CaR protein (data not shown). To confirm that the 130-kDa protein present in keratinocyte membranes is the spliced variant of human CaR, the same blot was incubated with hCaR4/6 antibody, which specifically reacts with the CaR splice variant (Fig. 1B). This antibody detected the 130-kDa band in keratinocytes and in HEK293 cells transfected with cDNA for the spliced variant of CaR (Fig. 1B, left) but not the full-length CaR. No immunoreactive band was observed when the blot was incubated with the same antibody preabsorbed with the specific peptide against which it was raised (Fig. 1B, right panel).
We next performed fluorescence immunostaining to detect CaR protein in fixed preconfluent keratinocytes cultured in 1.2 mM calcium using an anti-CaR antibody, BoCaR#2, which recognizes the full-length and spliced forms of human CaR. Examination by confocal microscopy confirmed that a low level of CaR protein is localized on the plasma membrane in some of the cells, whereas the majority of the CaR protein is expressed throughout the cytoplasm of keratinocytes (Fig. 2).
Transfection Northern analysis using a sense strand CaR RNA as probe showed that the keratinocytes transfected with the CaR antisense cDNA construct efficiently expressed the full-length CaR antisense strand transcript (Fig. 3A). Moreover, Western analysis of the membrane proteins by BoCaR#3 (Fig. 3B) or hCaR4/6 ( Fig. 3C) antibody showed that the expression of both the 160-kDa full-length CaR (Fig. 3B) and its 130-kDa spliced variant (Fig. 3C) was decreased in the cells transfected with the antisense CaR construct as compared with the cells transfected with empty vector. In this experiment, hCaR4/6 also detected a 115-kDa band in HEK293 cells but not in keratinocytes (Fig. 3C); however, this band is not detected by BoCaR#3 antibody. We next evaluated the effects of the CaR antisense cDNA construct on calcium sensing by examining the [Ca 2ϩ ] i response to raised [Ca 2ϩ ] o in the transfected keratinocytes. As shown in Fig. 4A (Fig. 4B). In this experiment, 2) Ca 2ϩ and from HEK293 cells transfected with the cDNA for full-length human CaR (CaR), AltCaR, or pcDNA3.1 vector (vector). Ten g of HEK293 cell membrane protein or 150 g of normal human keratinocyte membrane protein was loaded in each lane and separated on a 5% SDS-polyacrylamide gel electrophoresis gel. After blotting to a nylon membrane, the CaR was detected using (A) ADDR, a polyclonal antibody against human CaR that recognizes both full-length human CaR and its alternatively spliced variant, or (B) hCaR4/6, an antibody that specifically reacts with the alternatively spliced CaR variant. In addition to the three bands (120, 160, and 185 kDa) shown in A, a 130-kDa band (observed on longer-exposed film; data not shown) was specifically detected by ADDR in keratinocytes. Detection by hCaR4/6 confirmed that the 130-kDa protein present in keratinocyte membranes is the spliced variant of human CaR. The specificity of the immunostaining of these bands was confirmed by incubation with the antibodies preabsorbed with the specific peptides against which they were raised.

FIG. 2. Fluorescence immunostaining of the CaR protein endogenously expressed in human keratinocytes.
Keratinocytes were stained with an anti-CaR antibody, BoCaR#2, followed by fluorescein-conjugated anti-rabbit IgG and examined with a confocal microscope. A low level of CaR was detected on the plasma membrane in some cells, whereas most of the CaR was detected within the cytoplasm. To test the specificity of the antisense CaR cDNA transfection, we examined its impact on the [Ca 2ϩ ] i response induced by ATP through the activation of P2 purinergic receptors and by the calcimimetic NPS R-467, a positive modulator of the CaR (24). We have previously shown that extracellular ATP evoked a rapid and transient mobilization of calcium from intracellular sources in keratinocytes (25). In the current study, ATP (100 M) induced a rapid increase in [Ca 2ϩ ] i (from 223 Ϯ 12 to 295 Ϯ 21 nM; n ϭ 21) in keratinocytes transfected with empty vector in the presence of 1.2 mM calcium (Fig. 5A). ATP also evoked a rapid rise in [Ca 2ϩ ] i to the same magnitude (from 106 Ϯ 10 nM to 172 Ϯ 18 nM; n ϭ 15) in the cells transfected with the antisense CaR cDNA, whereas extracellular calcium failed to induce a [Ca 2ϩ ] i response (Fig. 5A). We have previously shown that NPS R-467 raised [Ca 2ϩ ] i and stimulated differentiation in keratinocytes (12). As shown in Transfection of the Human Antisense CaR cDNA Suppressed Calcium-induced Keratinocyte Differentiation-One of the key characteristics of calcium-induced keratinocyte differentiation is an inhibition of cell proliferation (1). To investigate the role of CaR in calcium-induced differentiation, we first examined the effect of antisense CaR cDNA transfection on proliferation by measuring the rate of DNA synthesis. After transfection with the CaR antisense cDNA construct or empty vector and selection by G418 in KGM containing 0.03 mM calcium for 3 days, keratinocytes were either maintained in 0.03 mM calcium or switched to 1.2 mM calcium for 24 h, and the rate of DNA synthesis was determined by [ 3 H]thymidine incorporation. In the keratinocytes transfected with vector, 24 h of incubation in 1.2 mM calcium resulted in a reduction of DNA synthesis to 46.6 Ϯ 0.1% as compared with the control cells maintained in 0.03 mM calcium (Fig. 6). In contrast, DNA synthesis was minimally affected by 1.2 mM calcium (99 Ϯ 19% of control) in the keratinocytes transfected with antisense CaR cDNA (Fig. 6).
We next examined the ability of the antisense CaR construct to block calcium induction of the differentiation markers involucrin and transglutaminase. As shown in Fig. 7, 48 h of incubation in 1.2 mM calcium significantly increased the levels of involucrin and transglutaminase mRNA in keratinocytes transfected with vector. However, the stimulation of involucrin and transglutaminase expression by calcium was blocked by the antisense CaR cDNA construct. Basal levels of involucrin and transglutaminase mRNA at 0.03 mM calcium were little affected by the antisense CaR cDNA construct.
To determine whether CaR mediates the calcium-induced transcriptional activation of the involucrin and transglutaminase genes, involucrin or transglutaminase promoter/reporter constructs were transiently cotransfected with the antisense CaR cDNA construct, and their response to calcium was then evaluated (Fig. 8). Raising [Ca 2ϩ ] o from 0.03 to 1.2 mM induced a nearly 5-fold increase in involucrin promoter activity. Antisense CaR cDNA suppressed this increase (Fig. 8A). Similarly, transglutaminase promoter activity was increased 3-fold in response to 1.2 mM extracellular calcium, an increase that was completely blocked by the antisense CaR cDNA construct (Fig.  8B). These data indicate that the induction of both involucrin and transglutaminase by calcium requires CaR.
Conceivably the neomycin analogue G418 could affect the [Ca 2ϩ ] i response and cell differentiation in keratinocytes by stimulating the CaR. However, our pilot experiments had shown that 300 M G418, which was used to enrich transfected cells, did not raise [Ca 2ϩ ] i in keratinocytes in the presence of 0.03 mM [Ca 2ϩ ] o (data not shown). Furthermore, G418 at this concentration did not drive cultured keratinocytes toward differentiation because only very low levels of the differentiation markers involucrin and keratinocyte transglutaminase were detected in the transfected cells before exposure to 1.2 mM [Ca 2ϩ ] o (Fig. 7). Moreover, similar results were obtained in keratinocytes transfected with a CaR antisense construct carrying a hygromycin-resistant gene and subsequently selected by 100 g/ml hygromycin (data not shown).

DISCUSSION
Extracellular calcium tightly controls the balance between proliferation and differentiation in a number of different cell types. In mesenchymal fibroblasts (26,27) and human ovarian surface epithelial cells (28), increased extracellular calcium promotes proliferation. It has been shown that the CaR modulates this proliferative response to extracellular calcium through activation of Src kinase and mitogen-activated protein kinase (27). However, in keratinocytes (1), breast epithelial cells (29), and intestinal epithelial cells (30), differentiation is promoted, whereas proliferation is inhibited by elevated extracellular calcium. The question this study addresses is whether the CaR is also involved in that cellular response to calcium.
We reported previously that keratinocytes express the fulllength CaR and its alternatively spliced variant (15). The transcript of the full-length CaR is predominantly expressed in undifferentiated keratinocytes, and its level decreases as the cells differentiate (15). The changes in CaR expression are consistent with the finding that the [Ca 2ϩ ] i response to [Ca 2ϩ ] o is maximal in undifferentiated cells and is attenuated as the cells differentiate (15). The notion that CaR modulates the cellular response to extracellular calcium is further supported by the ability of a selective CaR activator, NPS R-467, to potentiate the [Ca 2ϩ ] i response to [Ca 2ϩ ] o and to activate the genes required for cornified envelope formation (12). In the present study, we confirmed the expression of CaR proteins in keratinocytes. Because the blockage of CaR production specifically reduced the [Ca 2ϩ ] i response to [Ca 2ϩ ] o and to NPS R-467, our results indicate that CaR mediates calcium sensing directly or indirectly in keratinocytes. Transfection of the antisense CaR cDNA also suppressed the calcium-induced inhibition of proliferation and calcium-stimulated expression of involucrin and transglutaminase genes. These results are complementary to our findings (22) in a mouse model in which the synthesis of full-length CaR was disrupted (31). The loss of the full-length CaR altered the morphologic appearance of the epidermis and resulted in a reduction in the expression of the differentiation related gene, loricrin (22). Furthermore, epidermal keratinocytes isolated from these animals displayed a reduced [Ca 2ϩ ] i response to [Ca 2ϩ ] o (22). These studies indicate that CaR is one of the critical factors that modulate keratinocyte differentiation in vitro and in vivo by mediating calcium sensing and/or by coupling to other calcium signaling molecules important for proliferation and differentiation of these cells.
Immunocytochemical localization studies revealed that in keratinocytes, a low level of CaR protein is expressed on the plasma membrane, whereas its cellular localization is mostly cytoplasmic. Nonetheless, substantial cytoplasmic localization of CaR, often a perinuclear distribution, is commonly observed in other cell types, i.e. rat chondrogenic RCJ.C5.18 cells (32), mouse osteoblastic MC3T3-E1 cells (33), osteoblasts and articular and growth plate chondrocytes (34), and pancreatic acinar cells (35). Currently, it is unclear whether intracellular CaR represents nascent receptor protein being processed through the biosynthetic pathway or whether it has distinct biological functions. Indeed, the possibility that CaR mediates calcium sensing not only on the plasma membrane but also within the endoplasmic reticulum has been suggested (36). Although our observations that the antisense CaR specifically reduced the [Ca 2ϩ ] i response to [Ca 2ϩ ] o can be understood as indicating a role for calcium sensing of extracellular calcium by the CaR, it is conceivable that antisense CaR transfection could disrupt a CaR-mediated release of calcium from intracellular stores or a CaR-mediated increase in calcium influx through calcium channels. In either case, the [Ca 2ϩ ] o -mediated increase in [Ca 2ϩ ] i would be blocked by inhibition of CaR production, and differentiation would subsequently be aborted. Thus, the role of the CaR in calcium-mediated keratinocyte differentiation remains open for further investigation, but this study indicates that it has such a role.