Abstract
Mutated alleles of the SLC26A2 (diastrophic dysplasia sulfate transporter or DTDST) gene cause each of the four recessive chondrodysplasias, i.e., diastrophic dysplasia (DTD), multiple epiphyseal dysplasia (MED), atelosteogenesis Type II (AO2), and achondrogenesis Type IB (ACG1B). SLC26A2 acts as an Na+-independent sulfate/chloride antiporter and belongs to the SLC26 anion transporter gene family, currently consisting of six homologous human members. Although Northern analysis has indicated some expression in all tissues studied, the only tissue known to be affected by SLC26A2 mutations is cartilage. Abundant SLC26A2 expression has previously been detected in normal human colon by in situ hybridization. We have used in situ hybridization and immunohistochemistry to examine multiple normal tissues for the expression of human SLC26A2. As expected, a strong signal for SLC26A2 mRNA and protein immunostaining were detected in developing fetal hyaline cartilage, while bronchial cartilage showed mRNA expression in adult tissues. SLC26A2 expression could also be detected in eccrine sweat glands, in bronchial glands, and in placental villi. In addition, immunoreactivity for the SLC26A2 protein was observed in exocrine pancreas. Our results suggest a more limited expression pattern for SLC26A2 than that found by Northern analysis. However, SLC26A2 expression is also detected in tissues not affected in chondrodysplasias caused by SLC26A2 mutations. (J Histochem Cytochem 49:973–982, 2001)
SLC26A2 (also known as diastrophic dysplasia sulfate transporter, or DTDST) is an anion transporter that is responsible for four recessively inherited chondrodysplasias of increasing severity, i.e., multiple epiphyseal dysplasia (MED; Superti–Furga et al. 1999), diastrophic dysplasia (DTD; Hästbacka et al. 1994), atelosteogenesis Type II (AO2; Hästbacka et al. 1996), and achondro-genesis Type IB (ACG1B; Superti–Furga et al. 1996). The severity of the phenotype correlates with the underlying combination of SLC26A2 mutations and depends on the residual activity retained by the defective protein. In patient fibroblasts and chondrocytes, sulfate uptake is impaired, leading to undersulfation of synthesized cartilage proteoglycans (Hästbacka et al. 1994; Rossi et al. 1996) and low overall sulfate content of the cartilage (Superti-Furga et al. 1996).
SLC26A2 acts as an Na+-independent sulfate/chloride antiporter (Hästbacka et al. 1994; Satoh et al. 1998) and is a member of the newly delineated SLC26 anion transporter gene family, currently consisting of six known homologous human genes. Other well-characterized members are the liver sulfate anion transporter SLC26A1 (also known as SAT-1) (Lohi et al. 2000), the major intestinal chloride/bicarbonate or chloride/hydroxide exchanger SLC26A3 (alias CLD or DRA) that causes congenital chloride diarrhea when defective (Höglund et al. 1996), SLC26A4 (also called PDS), a chloride/iodide or chloride/formate exchanger that causes Pendred syndrome when defective (Everett et al. 1997; Scott et al. 1999; Scott and Karniski 2000), and the recently cloned SLC26A5 (also known as prestin; Zheng et al. 2000), and a putative pancreatic anion exchanger SLC26A6 (Lohi et al. 2000). SLC26A3, SLC26A4, and SLC26A5 have very restricted expression profiles (Everett et al. 1999; Haila et al. 2000; Zheng et al. 2000).
Although the phenotype caused by SLC26A2 mutations suggests cartilage as the major expression site, Northern analysis has shown wide expression patterns, with some expression in all tissues studied (Hästbacka et al. 1994). However, the cellular localization of SLC26A2 in most of these tissues has never been determined. In situ hybridization (ISH) study of normal human colon has demonstrated abundant SLC26A2 expression in the surface epithelium of the upper one third of the colonic crypt (Haila et al. 2000).
In this study we investigated multiple normal tissues to determine the expression patterns of the human SLC26A2 gene and protein. Characterization of tissues and specific cell types that express SLC26A2 in vivo is important for elucidating the physiological function of the normal protein and the pathophysiology of MED, DTD, AO2, and ACG1B.
Materials and Methods
PCR Analysis of Expression
PCR analyses were done using PCR-ready human MTC panel I and II cDNAs (Clontech; Palo Alto, CA). Panels included cDNA samples from brain, heart, kidney, liver, lung, pancreas, placenta, skeletal muscle, colon, ovary, peripheral blood leukocyte, prostate, small intestine, spleen, testis, and thymus. Cartilage cDNA was used as a positive control. The ribs of a fetus at gestational age of approximately 20 weeks were first homogenized and total RNA was extracted using the RNeasy RNA Extraction Kit (Qiagen; Chatsworth, CA). First-strand cDNA was synthesized with TaqMan Gold RT-PCR Kit (Applied Biosystems; Foster City, CA) from 0.4 μg total RNA as a template in a 20-μl reaction with 1 × TaqMan RT buffer, 2.5 μM random hexamers, 5.5 mM magnesium chloride, 500 μM of each dNTP, 8 U RNase inhibitor, and 25 U MultiScribe reverse transcriptase. The reactions were incubated at 25C for 10 min and 48C for 30 min, followed by 95C for 5 min. The surrounding tissues of the ribs were scraped off and used for cDNA synthesis to detect possible DTDST expression in these structures and thus soft tissue contamination in cartilage cDNA PCR.
PCR assays were performed in 25-μl volumes using 1.0 μl cDNA as template, 1 μM of each primer, 1 × reaction buffer provided by the enzyme supplier, 0.28 mM of each nucleotide, and 0.5 U of AmpliTaq Gold DNA polymerase (Applied Biosystems). The following PCR conditions were used: 94C for 5 min, 37 cycles of 94C for 1 min, 57C for 1 min, and 72C for 1 min, followed by 72C for 8 min. The 542-bp product corresponds to nucleotides 543–1084 of the published cDNA sequence (GenBank#U14528).
Northern Blot Hybridization
Commercially available Human and Human IV Multiple Tissue Northern Blots (Clontech) were hybridized with a 32P-labeled 542-bp (nucleotides 543–1084) SLC26A2 cDNA fragment that was generated by PCR amplification using human lung cDNA as a template. After prehybridization overnight at 65C, the blots were hybridized for 18 hr at 65C in 10% dextran sulfate, 1 M NaCl, 1% SDS, and 160 μg/ml denatured salmon sperm DNA. After hybridization, the filters were washed at 65C in 3 × SSC, 0.1% SDS. Autoradiography was performed on X-ray films overnight at −20C. Equal loading was verified with control gene hybridization (data not shown).
Tissues
Formalin fixed, paraffin-embedded archival specimens from adult patients were obtained from the Department of Pathology, Haartman Institute, University of Helsinki (adult tissues) and Department of Pathology, University of Oulu (fetal tissues). ISH and IHC studies on fetal samples were approved by the ethics committees of the Departments of Medical Genetics and Dermatology. The following adult specimens were examined: endometrium at different menstrual phases (n = 5), ventricle (n = 4), duodenum (n = 3), jejunum (n = 2), ileum (n = 1), appendix (n = 2), colon (n = 7), sigmoid (n = 2), anal canal (n = 2), liver (n = 3), pancreas (n = 4), skin (n = 8), heart (n = 3), bronchus (n = 4), kidney (n = 3), prostate (n = 3), placenta (n = 6), testis (n = 4), thymus (n = 4), epididymis (n = 2), and articular cartilage (n = 3). Fetal rib cartilage (n = 2), complete fetuses at gestational ages of 6–7, 8, 10–12 weeks and selected tissues from older fetuses at 11 and 20 weeks were also studied.
In Situ Hybridization
A 549-bp fragment corresponding to bases 1422–1970 of the published human SLC26A2 cDNA (GenBank#U14528) was generated by PCR. This fragment was designed with a T7 RNA polymerase promoter at the 3′ end and SP6 RNA polymerase promoter at the 5′ end. Both sense and antisense probes were transcribed from the PCR product using the Riboprobe in vitro transcription system (Promega; Madison, WI). Antisense and sense RNA probes were labeled with α-[35S]-UTP and purified probes were used at 4 × 105 cpm/liter of hybridization solution.
As previously described (Prosser et al. 1989; Höglund et al. 1996), deparaffinized 5-μm tissue sections were digested with 1.0–10 μg/ml proteinase K for 30 min at 37C and treated with 0.25% acetic anhydride in 0.1 M triethanola-mine buffer for 10 min at room temperature. Hybridization was carried out overnight at 52C. After hybridization, the slides were washed under stringent conditions, including RNase A, and exposed to LM-1 emulsion (Amersham; Poole, England) for 21–50 days at 4C. The slides were developed and counterstained with hematoxylin and eosin.
Normal colon samples known to be positive were used as controls in each experiment and a sense RNA probe was used as a negative control (Haila et al. 2000).
Immunohistochemistry
Antisera were raised in rabbits against two synthetic peptides ERQEKSDTNFKEFVIK and TVRDSLTNGEYCKKEEEN, corresponding to bases 196–243 and 2092–2145, respectively, of the published cDNA sequence (GenBank#U14528). Peptide synthesis and antibody production were purchased from Research Genetics (Huntsville, AL).
The specificity of antibodies was demonstrated by Western blotting (Figure 2) using homogenized osteosarcoma tissue. After centrifugation at 12,000 × g for 10 min, the supernatant was collected and diluted 1:4 in Laemmli sample buffer (Pharmacia; Uppsala, Sweden) containing 5% of β-mercapto-ethanol. Denatured proteins were separated on a 7.5% polyacrylamide gel and the gel was blotted onto Hybond C-extra (Amersham) membrane using standard protocols. Affinity-purified primary antibodies at 2 μg/ml were used. Normal rabbit IgG (Dako; Glostrup, Denmark) was used as a negative control antibody. Peroxidase-conjugated anti-rabbit IgG was diluted 1:10,000 in 0.1% Tween-20/PBS containing 2.5% non-fat milk and was used as the secondary antibody. The protein bands were visualized by chemiluminescence according to standard protocols.
MTC panel PCR revealed the intense 542-bp band for SLC26A2 expression in lung, placenta, and colon, and a weaker band in pancreas, kidney, and testis cDNAs. Faint bands could also be detected in brain, heart, liver, peripheral blood leukocyte, small intestine, spleen, and thymus. Positive control reaction with fetal cartilage cDNA demonstrates a band with similar size. In Northern blot hybridization using the same 542-bp fragment as a probe, an approximately 8.4-kb transcript is present in all lanes. Strong signal is detected with placenta, prostate, testis, and colon mRNA.
Serial sections to those used for ISH were used for immunohistochemistry. The peroxidase–antiperoxidase technique was performed using the Vectastain Elite ABC Kit (Vector Laboratories; Burlingame, CA). For pretreatment, the depar-affinized slides were boiled in a microwave oven for 5 min in 10 mM citrate buffer (pH 6.0) or 0.01 M EDTA buffer (pH 8.0). One percent SDS for 5 min was used as a pretreatment for skin sections containing eccrine sweat glands. Anti-SLC26A2 sera were diluted 1:2500–1:4000. Diaminobenzidine (DAB) was used as the chromogenic substrate and the slides were counterstained with hematoxylin. Preimmune serum was used as a negative control on parallel sections.
Results
cDNA Panel PCR
Intense bands were amplified from lung, placenta, and colon cDNAs and weaker bands from pancreas, kidney, and testis cDNAs of the multiple-tissue cDNA panel (MTC panel, Clontech; Figure 1). Faint bands could be detected in brain, heart, liver, peripheral blood leukocyte, small intestine, spleen, and thymus. Fetal cartilage cDNA was used as a positive control and demonstrated a single band with identical electrophoretic mobility (Figure 1).
Northern Hybridization
Northern analysis showed that the 542-bp SLC26A2 probe detects a single approximately 8.4-kb transcript, as expected (Figure 1). Strong signal was detected in placenta, prostate, testis, and colon, and weak signal in all the other tissues. The result corresponds to earlier Northern analysis (Hästbacka et al. 1994). The intensity of the signal among tissues, however, is dissimilar.
Specificity of Anti-SLC26A2 Antiserum
Western blotting of osteosarcoma lysate using affinity-purified antibodies demonstrated a band with mobility corresponding approximately to the size of 90 kD (Figure 2). Anti-rabbit IgG used as a negative control did not detect this band. In addition, both SLC26A2-specific antibodies as well as normal anti-rabbit IgG recognized two smaller bands.
Western blotting analysis of osteosarcoma tissue lysate with affinity purified carboxy terminal (Lane 1) and amino terminal (Lane 3) SLC26A2 antibodies under denaturing conditions. Normal rabbit IgG (Lanes 2 and 4) was used as a negative control. Positions of the molecular weight markers are indicated in the middle.
Human SLC26A2 mRNA (
Immunohistochemistry
Cartilage. Intense SLC26A2 mRNA expression was seen in hyaline cartilage of developing long bone of the limb in an 11-week-old fetus. The hypertrophic chondrocytes in the diaphysic maturation zones of limb long bones demonstrated strong SLC26A2 mRNA expression (Figures 3A and 3B). The SLC26A2 protein expression was detected in the hypertrophic and proliferative chondrocytes in developing bones of an 11-week-old fetus (Figures 3C and 3D). Fetal periosteum also demonstrated focal immunoreactivity for the SLC26A2 protein (Figure 3C). In addition, chondrocytes in adult bronchial cartilage showed mRNA expression (Figures 3F–3I).
Colon. We have previously shown that in normal colon SLC26A2 mRNA is detected in the upper one third of the crypt epithelium, mainly in the absorptive epithelial and goblet cells, whereas the signal was absent both at the bottom of the crypts and in luminal surface epithelium (Haila et al. 2000). Immunohistochemical studies revealed a similar distribution of the SLC26A2 protein in normal colon epithelium (Figures 4A and 4E). With microwave boiling as a pretreatment, colon surface epithelial cells demonstrated constant immunoreactivity in the cytoplasm on the apical side of the nucleus (Figures 4A and 4C). When epitopes were unmasked using 1% SDS instead of microwave boiling, immunoreactivity was detected focally in the surface epithelium and especially at the apical edge of immunoreactive cells (Figures 4D and 4E).
Placenta. Trophoblasts and syncytiotrophoblasts covering the surface of the chorionic villi demonstrated abundant SLC26A2 mRNA (Figures 5A–5C) and protein expression (Figure 5D). Both strong cytoplasmic and apical plasma membrane immunoreactivity for the SLC26A2 protein could be detected, whereas the villous stromal tissue was negative (Figure 5D).
Sweat Gland. Eccrine sweat glands showed both SLC26A2 protein (Figure 6A) and mRNA (Figures 6D and 6E) expression. Abundant immunoreactivity was detected in epithelial cells in the coiled secretory portion of the eccrine sweat gland (Figure 6C).
Pancreas and Liver. In the pancreas, acinar cells and duct epithelium of large ducts demonstrated strong immunoreactivity (Figures 7A, 7B, 7D, and 7E). By contrast, cells of the islets of Langerhans were negative (Figure 7D). In addition, protein or cell debris in acinar lumen secretion stained with the SLC26A2 antiserum (Figure 7D). In the liver, signal for the SLC26A2 mRNA was detected only in occasional cells and no specific immunostaining for the protein could be seen (data not shown).
Immunostaining of human colon with antiserum against the carboxy terminal peptide. When microwave pretreatment is used, constant perinuclear cytoplasmic immunostaining is detected (
Bronchial Sections. High expression levels of both SLC26A2 mRNA and protein were detected in submucosal seromucous glands in the airway section of human bronchi (Figures 7F-7H). SLC26A2 expression was also observed in the tracheal surface epithelium (Figures 7I and 7J).
ISH and immunostaining of human placenta. Trophoblastic cells on the surface of the chorionic villi demonstrate both mRNA (
Discussion
In this work we present the expression patterns of human SLC26A2 (also known as DTDST) by ISH and IHC in vivo in different tissues and cell types. This is the first study demonstrating the expression at the cellular level in morphologically preserved tissues. SLC26A2 is responsible for four chondrodysplasias if mutated, and its function as a sulfate and chloride transporter has been established (Satoh et al. 1998). At present, the exact biological role of SLC26A2 in most tissues in vivo is uncertain, and only cartilage is known to be affected when the protein is defective. Therefore, detailed knowledge of the cellular distribution of the human SLC26A2 in different tissues and cell types is necessary.
As expected, SLC26A2 expression was detected in several tissues. Cartilage, colon, placenta, bronchial glands, tracheal epithelium, pancreas, and eccrine sweat glands demonstrated immunoreactivity for the protein. Expression was observed in many different cell types but was confined mostly to secretory structures. Tissues such as testis, thymus, and prostate remained negative, and these results probably indicate that SLC26A2 expression level is so low that it escapes the ISH and IHC methods, although Northern analysis suggested expression. Expression can also be localized in a small, restricted area or structure. It is possible that the expression is variable and is regulated by certain events, whereas basal mRNA expression under normal physiological conditions is weak and the protein is not translated.
Interestingly, two antisera targeted against distinct parts of the SLC26A2 protein recognized expression in different tissues. Antiserum against the carboxy terminal peptide recognized expression in colon, placenta, bronchial glands, and tracheal epithelium, whereas antiserum against the amino terminal peptide recognized SLC26A2 expression in fetal cartilage, pancreas, and eccrine sweat glands. Both antisera demonstrated SLC26A2 protein expression in pancreatic ducts. However, under denaturing conditions in Western blotting analysis, both antisera recognized the same ∼90-kD mobility fragment in homogenized osteosarcoma tissue. In addition, no discrepancy between SLC26A2 mRNA expression by ISH and protein expression could be detected in this study.
SLC26A2 expression in eccrine sweat glands. Antiserum against the amino terminal peptide demonstrates strong immunoreactivity in the secretory portion of eccrine sweat glands (
Whereas SLC26A2 expression levels in adult cartilage were almost undetectable, strong signal for mRNA was observed in most mature hypertrophic chondro-cytes at gestational Week 11 in developing fetal cartilage. In addition to hypertrophic chondrocytes, the protein immunostaining was detected in proliferative chondrocytes, whereas the primitive reserve chondro-cytes were negative. The SLC26A2 protein expression in proliferative chondrocytes corresponds well with their active biosynthesis of sulfated proteoglycans and thus with the undersulfation of proteoglycans detected in patients with the defective SLC26A2 transporter. The sulfation of the chondroitin sulfate increases constantly with gestational age. At gestational Week 11, 6-sulfation, which is known to be more sensitive to extracellular inorganic sulfate depletion (Ito et al. 1982; Rossi et al. 1996), predominates in chondroitin sulfate chains (Roughley et al. 1987). Proper 6-sulfation has also been demonstrated to be more essential for the growth response of the cells to fibroblast growth factor (Pye et al. 1998). The strong signal for mRNA in hypertrophic chondrocytes is, however, interesting. During hypertrophy, synthesis of matrix proteins, such as collagens, gradually replaces proteoglycan synthesis. However, SLC26A2 expression in hypertrophic chondrocytes suggests that it has a role during chondrocyte hypertrophy and in the events that lead to matrix mineralization.
In addition to cartilage, SLC26A2 expression was detected in many different tissues and cell types that are not known to be affected in the four chondrodys-plasias. In our earlier work, SLC26A2 mRNA was shown to be expressed in the upper one third of colonic crypt epithelium (Haila et al. 2000). Immunohistochemistry confirmed the protein expression in colon as well. However, two staining patterns that depended on pretreatment were detected. Heavy microwave pre-treatment appears to destroy the apical, most likely functional, form of the protein, whereas the cytoplasmic form embedded inside the cell, probably corresponding to the protein under construction and to terminal maturation in Golgi apparatus, is revealed.
Other tissues that showed SLC26A2 mRNA and protein expression included placental villi, eccrine sweat glands, airway submucosal glands, and tracheal epithelial cells. In addition, protein expression was detected in exocrine pancreas. The presence of an active sulfate transport mechanism or even an SO4 2–/Cl− exchanger, as well as sulfated macromolecules in these tissues, has been reported (Seutter et al. 1970; Parmley et al. 1984; Elgavish et al. 1987; Cole and Rastogi 1991; Elgavish and Meezan 1992). Expression of SLC26A2 makes it as an excellent candidate responsible for sulfate or chloride transport processes in these tissues, but the clinical significance and the possible compensatory mechanisms of defective transport caused by SLC26A2 mutations remain subjects for further study.
In pancreas, (
The results reported here represent important steps towards defining the role of SLC26A2 (formerly DTDST) in anion transport processes of multiple tissues. SLC26A2 expression in cartilage and the pheno-type caused by its defects establish its major role in providing enough inorganic sulfate for abundant biosynthesis of sulfated macromolecules in chondrocytes. However, it is still surprising that the phenotype of diastrophic dysplasia is restricted to cartilage and bone, although in vivo protein expression can be detected in many tissues. Many possible explanations exist. The residual activity of the mutated SLC26A2 protein combined with the use of alternative sulfate sources (Esko et al. 1986; Rossi et al. 1996) in diastrophic dysplasia is most probably able to supply enough sulfate for the modest needs of most cell types other than chondrocytes. Most structures and cell types that express SLC26A2 are responsible for secretion that is usually performed by a complex cooperation of multiple transporters. It is possible that secretory epithelium is able to compensate for defective SLC26A2 function by regulating the expression of other transporters.
Footnotes
Acknowledgments
Supported by the Helsinki University Research Fund, the Academy of Finland, The Finnish Medical Foundation, the Duodecim Foundation, the Research and Science Foundation of Farmos, the Sigrid Juselius Foundation, and the Finnish Pediatric Foundation, Ulla Hjelt Fund.
We thank Dr Riitta Herva and Dr Juha-Pekka Turunen for their pathology expertise. The skillful technical assistance of Ms Alli Tallqvist and Ms Ranja Eklund is gratefully acknowledged.