Department of Biomedical Sciences, Atlantic Veterinary College, University of Prince Edward Island, Charlottetown, Canada
Correspondence to: Sunny Hartwig, Department of Biomedical Sciences, Atlantic Veterinary College, University of Prince Edward Island, 550 University Avenue, Charlottetown, PE, C1A 4P3, Canada. E-mail: email@example.com
Kidney development begins at embryonic day (E)10.5 in mice, with the outgrowth of the ureteric bud from the Wolffian duct into the overlying metanephric mesenchyme (for a recent review, see Little and McMahon, 2012). In response to inducing signals from the ureteric bud, the metanephric mesenchyme condenses as a tight cap of nephron progenitor cells around ureteric bud tips, while the ureteric bud branches to form the collecting system. During nephrogenesis, nephron progenitors have the capacity to self-renew and to differentiate into pretubular aggregates, which in turn undergo mesenchymal-epithelial transition to sequentially form early renal vesicles, comma-shaped bodies, S-shaped bodies, and, ultimately, functional nephrons (Saxen and Sariola, 1987; Self et al., 2006; Kobayashi et al., 2008).
Defects in nephrogenesis and ureteric branching result in congenital anomalies of the kidney and urinary tract (CAKUT). With an incidence of 1:400 (Pope et al., 1999; Rumballe et al., 2010), CAKUT constitute one of the most frequent birth defects in humans, and the major cause of childhood renal failure. These pleiotropic malformations comprise a multitude of renal phenotypes including renal agenesis, hypoplasia, and renal dysplasia, associated with obstructive or refluxive anomalies of the ureter (Toka et al., 2010). CAKUT are predicted to have different genetic origins (Weber, 2012), yet the molecular pathogenesis of CAKUT is not well understood. Nephron deficiency is a hallmark feature of CAKUT (Puddu et al., 2009). In fact, nephron number varies greatly even within normal range, from 11,000 to 19,000 in mice (Merlet-Benichou et al., 1999), and from 200,000 to 1.8 million in humans (Hughson et al., 2003). Low nephron endowment within this normal range, though asymptomatic early in life, is associated with adult-onset hypertension (Mackenzie and Brenner, 1995; Walker et al., 2012), a leading cause of coronary heart disease, stroke, and renal failure in North America (Tedla et al., 2011). Despite the fundamental importance of nephron endowment to health and disease, only a minority of molecular mechanisms underlying nephron morphogenesis has been identified.
The Wilms' Tumour Suppressor-1 (WT1) gene encodes a DNA- and RNA-binding nuclear transcription factor that plays an essential role during nephrogenesis in part via its target gene effectors. However, endogenous gene targets that mediate WT1 function during renal development are largely unknown. We previously employed Wt1 ChIP-chip genome-wide location analysis in embryonic E18.5 mouse kidney tissue, as a means of identifying novel kidney development genes that might act down-stream of Wt1 to control nephrogenesis in vivo (Hartwig et al., 2010). Intriguingly, we identified all three members of the Sry-related high-mobility group (HMG) Box (Sox)-C subfamily—Sox4, Sox11, and Sox12—as candidate Wt1 gene targets by ChIP-chip. Members of the Sox gene family have been shown to play master regulatory roles in a multitude of developmental processes including dorso-ventral patterning (Chen et al., 2012), stemness (Avilion et al., 2003; Masui et al., 2007; Wang et al., 2012), male differentiation (Koopman et al., 1991; Wagner, 1994), neurogenesis (Bylund et al., 2003; Taranova et al., 2006; Bhattaram et al., 2010; Bergsland et al., 2011; Mu et al., 2012), cardiac outflow tract, B-lymphocyte (Schilham et al., 1996), and spinal cord development (Shim et al., 2012), and skeletogenesis (Akiyama and Lefebvre, 2011). However, SOX function in the developing kidney is largely undefined.
Mammalian Sox genes are categorized into eight subgroups (A–H) based on phylogenetic homology of their HMG box DNA-binding domain. SOX proteins from different subgroups share partial identity (<46%) in their DNA-binding domain and none outside this domain. In contrast, subgroup members share a high degree of identity both within and outside their DNA-binding domains, thus often exhibiting functional redundancy in tissues where they are co-expressed. Accordingly, all three members of the SoxC subfamily—Sox4, Sox11, and Sox12—exhibit overlapping expression patterns in multiple tissues during embryonic development, suggesting genetic redundancy within this subfamily (Dy et al., 2008; Hoser et al., 2008). In this study, we characterize the renal developmental expression pattern of the three SoxC genes, and show that strong Sox4 expression is detected in nephron progenitors and their derivative nephrogenic structures, while Sox11 appears preferentially expressed in differentiating nephrogenic structures during renal development. To begin our investigation of SoxC function during nephrogenesis, we ablated Sox4 function in nephron progenitors and their cellular descendants (Sox4nephron- mice) using an established Six2Cre-EGFP transgenic line (Kobayashi et al., 2005, 2008). Postnatal Sox4nephron- kidneys exhibit significantly reduced nephron number. Sox4nephron- mice exhibit reduced podocyte nephrin expression at P7, and develop severe proteinacious kidney injury within 2 weeks of birth, leading to end-stage renal disease. Collectively, our results demonstrate an essential requirement for Sox4 in normal renal development, and further support a role for Sox4 in glomerular development and/or adult glomerular maintenance in vivo.
RESULTS AND DISCUSSION
Dynamic Distribution of SoxC Transcripts During Renal Development
We performed quantitative real-time qPCR and RNA in situ hybridization to analyze the expression of SoxC transcripts at multiple stages of embryonic renal development, from the onset of renal development at E11.5 until E18.5, as well as at two postnatal time-points, namely, postnatal day (P) 7 when nephrogenesis is completed, and in adult P21 kidneys (Fig. 1). Absolute quantitative qPCR (normalized to 25 ng RNA) shows that the three SoxC transcripts are highly expressed throughout renal development (Fig. 1A–C). Similar to previous reports of overlapping SoxC expression in developing tissues (Dy et al., 2008), SoxC transcripts exhibit an overlapping but distinct pattern of expression in the early developing kidney: Sox4, Sox11, and Sox12 are all expressed in nephron progenitors and in the ureteric bud (Fig. 1D–F). Notably, strong Sox11 expression is also detected in epithelialized nephrogenic structures at E12.5. As development proceeds, strong Sox4 and Sox12 expression continues to be detected in nephron progenitors and ureteric bud lineages as well as in differentiating nephrogenic structures, as shown at E18.5 (Fig. 1G,G',I,I'). Similarly, Sox11 expression continues to be strongly detected in differentiating nephrogenic structures throughout renal development, including pretubular aggregates, comma-shaped and S-shaped bodies, with weak expression in nephron progenitors and ureteric bud (Fig. 1H,H').
Nephrogenesis is complete within the first week of postnatal life, coincident with the absence of expression of nephron progenitor markers including Six2, Cited1, Meox1, and Dpf3 (Mugford et al., 2009; Brown et al., 2011). At this stage, diffuse Sox4 and Sox12 expression are detected in multiple cell types of the kidney, and low levels of Sox11 are also detected. Focal SoxC expression is observed in epithelialized tubules primarily in the renal papilla, and declines to low levels by P21 (see Supp. Fig. S1, which is available online). Consistent with the established roles for Sox4 and Sox11 during embryonic development, our expression data suggest a primary role for SoxC genes in directing different stages of nephrogenesis during renal development in vivo.
SoxC proteins share a high degree of identity in their high-mobility group domain, and in the SoxC-specific transactivation domain. Of the three SoxC genes, Sox12 is the weakest transactivator in vitro (Dy et al., 2008). Sox12−/− mice develop normally, are fertile, and do not exhibit any obvious defects (Hoser et al., 2008). Genetic analyses of embryos with all combinations of SoxC -null alleles reveals that loss of Sox12 only slightly aggravates the phenotype of Sox4+/−;Sox11+/− and Sox4−/−; Sox11−/− embryos (Bhattaram et al., 2010). Collectively, these genetic studies demonstrate that Sox12 is dispensable, providing only minor contributions during embryonic development in vivo, possibly due to low transcriptional competence of its protein product.
In contrast, independent, essential roles for Sox4 and Sox11 have been demonstrated in heart development in vivo. Sox4−/− mice die prior to E14 due to cardiac outflow tract malformation, exhibiting a primary failure of the endocardial ridge differentiation into semilunar valves (Schilham et al., 1996). Sox4 is also required for differentiation of T and B lymphocytes (Schilham et al., 1996, 1997), pancreatic β-cells (Wilson et al., 2005) and osteoblasts (Nissen-Meyer et al., 2007). Sox11−/− mice die shortly after birth, with similar but less severe heart malformations than Sox4−/− embryos, and exhibit other defects in organogenesis (Sock et al., 2004). Analyses of compound Sox4- and Sox11-null embryos demonstrate that Sox4 and Sox11 act partially redundantly to control key neural and mesenchymal progenitor cell fate decisions during organogenesis in vivo (Thein et al., 2010; Bhattaram et al., 2010; Bergsland et al., 2011; Shim et al., 2012). Compound Sox4−/−; Sox11−/− kidneys have not been studied in vivo, due to early embryonic lethality (E8.5) of compound Sox4−/−;Sox11−/− embryos. Notably, Sox4−/−;Sox11+/− embryos exhibit more severe phenotypic abnormalities than Sox4+/−;Sox11−/− embryos in early organogenesis, and are almost as severely affected as Sox4−/−; Sox11−/− embryos. These genetic data suggest that in tissues co-expressing both Sox4 and Sox11, Sox4 may play a greater role in early developmental processes, while Sox11 may function more strongly in later stages of development. These observations are consistent with the distinct expression pattern of Sox4 in early nephrogenic structures and Sox11 in more mature nephrogenic structures, collectively suggesting that Sox4 and Sox11 may function nonredundantly to control early and late aspects of nephrogenesis, respectively, in vivo.
Sox4 Mutant Mice Develop Postnatal Renal Disease
To begin our investigation of SoxC function during nephrogenesis, we ablated Sox4 function in nephron progenitors and their cellular descendants using a well-characterized Six2Cre-EGFP transgenic mouse strain (Kobayashi et al., 2008; Ho et al., 2011). Analysis of Sox4nephron- embryonic kidneys did not reveal any overt abnormalities; Sox4nephron- mutant newborns appeared normal, and were born at the expected Mendelian ratios. However, Sox4nephron- glomeruli appeared hypocellular at P7 (Fig. 2A vs. B; Fig. 3A vs. B), and nephron number was significantly (48%) reduced in Sox4nephron- kidneys compared to wild-type controls (Fig. 2C; P < 0.006; n=4). Within 2 weeks of birth, Sox4nephron- mice appeared smaller than wild-type littermates, reductions in weight that became significant within 5 weeks (11.7 ± 0.52%, P < 0.05, wild-type n=5; Sox4 mutant n=9). By P16, Sox4nephron- mutant kidneys had undergone a remarkable alteration, exhibiting a poorly organized renal medulla and cortex, and numerous dilated tubules containing PAS+ protein casts (Fig. 3B,D, vs. A,C). Abundant PAS+ protein reabsorption granules were observed in the cytoplasm of proximal tubules of Sox4nephron- mutant kidneys (Fig. 3D,D' vs. C,C'), evidencing substantial proximal tubular reabsorption of abnormally high levels of protein and lipoproteins in the urine, a hallmark of nephrotic range proteinuria and glomerular disease (Laurinavicius et al., 1999; Schwimmer et al., 2003). Renal corpuscles in Sox4nephron- mutants were markedly hypocellular compared to wild-type littermates and had poorly-defined glomerular tufts (Fig. 3B'vs. A'). Characteristically thin PAS+ basement membrane staining of the glomerular capillary loops and tubular epithelium in wild-type animals is lost in Sox4 mutants (Fig. 3D'vs.C'), indicating damage to glomerular capillary loops, together with increased mesangial matrix in these animals. Collectively, these findings are consistent with early-onset proteinuria and glomerular disease in Sox4 mutant animals.
Sox4nephron- mutants progress to end-stage renal disease between 5–9 months of age. Histologic analysis of Sox4nephron- mutant kidneys at 5 months revealed an abundance of grossly dilated tubules containing protein casts (Fig. 4B,D vs. A,C). Glomeruli of Sox4nephron- mutants lack clear vascular tufts, and exhibit widespread sclerosis with reduced urinary space between the parietal and visceral epithelial layers of Bowman's capsule (Fig. 4B' vs. A'). PAS staining showed significant thickening of capillary basement membranes and marked mesangial matrix accumulation in Sox4nephron- mutant glomeruli, together with focal loss of proximal tubular brush border indicative of proximal tubule injury (Fig. 4D' vs. C'). Significant renal pathology was also observed by electron microscopy in these same Sox4nephron- mutant mice. Assessment of wild-type kidneys revealed glomeruli with a normal, uniform glomerular basement membrane (GBM), and regular, interdigitating podocyte foot processes (Fig. 5A,A'). In contrast, Sox4nephron- mutant glomeruli display extensive podocyte foot process effacement and apparent fusion characteristic of proteinuric glomerular disease (Fig. 5B vs. A). The GBM in these mice is markedly thickened and numerous subepithelial aggregates of electron-dense deposits are present (Fig. 5C,C'), a hallmark of membranous nephropathy (Pavelka and Roth, 2005). Together, histologic and ultrastructural analyses demonstrate that the structure and function of the glomerular filtration barrier is abrogated in Sox4nephron- mutant mice.
Our observation of reduced nephron endowment in Sox4nephron- mice demonstrates an essential role for Sox4 in nephrogenesis in vivo. However, the severity of the early-onset proteinuric glomerular injury observed in postnatal Sox4nephron- mice is not easily explained by a reduction in nephron number, and suggests a requirement for Sox4 in podocyte development and/or adult podocyte maintenance. Given the complex and poorly-understood interplay between podocytes, endothelium, mesangium, and glomerular basement membrane, it is possible that alterations in podocyte gene expression or morphology could cause secondary responses in these cell types, which would further exacerbate glomerular injury. In order to assess whether Sox4 may regulate glomerular development and/or maintenance, we performed indirect immunofluorescence for key glomerular markers in Sox4nephron- kidneys at E17.5 and in P7 kidneys (Fig. 6). At E17.5, Sox4nephron- glomeruli did not exhibit alterations in expression of early glomerular markers including glomerular basement proteins nidogen (Fig. 6A vs. B) and β1 integrin (data not shown), glomerular mesangial markers including NG2 chondroitan sulfate (Fig. 6E vs. F), desmin, and PDGF-R (data not shown), endothelial protein PECAM-1 (Fig. 6I vs. J), podocyte slit diaphragm proteins nephrin (Fig. 6M vs. N) and podocin (Fig. 6Q vs. R), or in the podocyte nuclear marker Wt1 (Fig. 6U vs. V). However, by P7, Sox4nephron- glomeruli exhibited a specific reduction in nephrin expression, while expression of podocin as well as other markers of the endothelium, mesangium, and basement membrane continued to be normally expressed. Nephrin is a large transmembrane protein of the immunoglobulin (Ig) superfamily, and a key structural component of the podocyte slit diaphragm (Tryggvason et al., 2006). The regular spacing between adjacent podocyte foot processes is maintained through adhesive interactions between nephrin and other Ig superfamily proteins expressed on opposing foot processes. Loss or abnormal function of nephrin leads to proteinuria due to loss of the slit diaphragm and foot process effacement (Jones et al., 2009; New et al., 2013). Our results demonstrate reduced nephrin expression in P7 Sox4nephron- glomeruli, suggesting that loss of nephrin expression may partly underlie the observed glomerular phenotype observed in P16 Sox4nephron- mice, and indicating a role for Sox4 in regulation of nephrin expression.
The number of Wt1+ podocyte nuclei is also reduced in Sox4nephron- kidneys at P7 (Fig. 6X vs. W). Nephrin is a transcriptional WT1 target (Wagner et al., 2004). Alternatively, loss of nephrin expression in Sox4nephron- kidneys may be secondary to reduced WT1 expression in Sox4nephron- podocytes, suggesting a role for Sox4 in regulation of WT1 in podocytes. Importantly, reduced numbers of Wt1+ podocytes in Sox4nephron- renal corpuscles may reflect reduced podocyte numbers in these mice, suggesting that aberrant nephrin expression and glomerular injury may be partly attributable to podocyte loss in Sox4nephron- kidneys, supporting a role for Sox4 in podocyte development in vivo.
In summary, we have defined the relative abundance and subcellular distribution of the three SoxC genes during renal development. Postnatal analysis of Sox4 mutant mice demonstrates an essential requirement for at least one member of the SoxC subfamily in nephrogenesis in vivo, and suggests potential roles for Sox4 in podocyte development and/or adult podocyte maintenance. Future studies will assess whether Sox4 transcriptionally regulates nephrin expression in vitro. Conditional gene deletion studies will be aimed at ablating Sox4 in podocytes to determine the specific role of Sox4 in the podocyte lineage, and to evaluate the independent role of Sox11 as well as the combined role of Sox4/Sox11 signaling during renal development in vivo.
Histology, Immunofluorescence, and RNA In Situ Hybridization
For histology, tissues were fixed in 10% neutral buffered formalin and processed to form paraffin blocks. Tissues were sectioned at a thickness of 5 μm and subsequently stained with hematoxylin and eosin (H&E) and periodic acid-Schiff (PAS). For indirect immunofluorescence, kidneys were frozen in Tissue-TEK 4583 OCT (Sakura Finetek) in a liquid nitrogen–cooled bath. To visualize nephrin expression, the kidneys were fixed in 4% paraformaldehyde at 4°C overnight before embedding in OCT. For all other antibodies, cryosections were fixed in methanol at −20°C for 10 minutes. Immunofluorescence staining on 4-μm cryosections was performed as described (Hartwig et al., 2010). The primary antibodies used were as follow: rabbit anti-Wt1 (Santa Cruz Biotechnology, Santa Cruz, CA), mouse anti-E-Cadherin (610181, BD Bioscience, San Jose, CA), goat anti-nephrin (N-20, Santa Cruz), rat anti-β1 integrin (MAB1997, Chemicon, Temecula, CA), rabbit anti-podocin (P0372, Sigma, St. Louis, MO), rabbit anti-NG2 chondroitin sulfate (AB5320, Millipore, Billerica, MA), rabbit anti-PDGFRβ (3169, Cell Signaling, Danvers, MA), rat anti-nidogen (MAB 1883, Chemicon), rat anti-PECAM1 (553370, BD Biosciences), and mouse anti-desmin (M0760, Dako, Carpinteria, CA). Fluorescence was detected using secondary antibodies conjugated to Texas Red and FITC fluorochromes (Jackson ImmunoResearch, West Grove, PA). RNA in situ hybridization was performed on frozen sections at a thickness of 8 μm using antisense probes. Corresponding sense negative controls did not produce background hybridization signals. SoxC probes were generated by PCR amplification using the following primers: Sox4 forward, 5′GAACGCCTTTATGGTGTGGT3′;
Sox4 reverse, 5′GGTAGACGCGCTT CACTTTC3′;
Sox11 forward, 5′GAGCCTGTACGA CGAAGTGC3′;
Sox11 reverse, 5′TCAAAGAGCCAC AAGCTTCA3′;
Sox12 forward, 5′AGAGCGGGTTC TCTTTAGGC3′;
Sox12 reverse, 5′TCAGCATGGGA CAACACATT3′.
PCR products were cloned into the pCRII- TOPO vector (Invitrogen, Carlsbad, CA) and verified by sequencing.
Glomeruli were counted as described using the acid maceration method (MacKay et al., 1987; Godley et al., 1996; Hoy et al., 2011). Briefly, kidneys were decapsulated, cut into 2 mm3 pieces and incubated in 5 ml of 6 N HCL at 37°C for 90 min, and tissue fragments were gently disrupted by aspirating up and down using a 10-mL pipette. The suspension was diluted to a total volume of 30 ml with water and incubated at 4°C overnight with rotation. The preparation was resuspended and a 500-μL aliquot removed to a counting chamber. Glomerular number was counted under phase microscopy. Each sample was counted in triplicate. The number of glomeruli per kidney was determined by correcting the mean of the glomerular counts for dilution.
Quantitative Real-Time RT PCR (qPCR)
For embryonic stages, mouse embryos were excised from pregnant CD1 mice. Total RNA was isolated from three pooled biological kidney samples dissected at E11.5, E12.5, E14.5, E16.5, or E18.5. P7 and P21 kidneys were harvested from 3 animals. Freshly harvested mouse kidneys were stabilized in RNA Later (Ambion, Austin, TX) at 4°C overnight with rotation and then stored at −80°C according to manufacturer's instructions. Tissues were homogenized using a Tissue-Tearor (Biospecifics Products, Lynbrook, NY) and total RNA was extracted using the RNAqueous 4-PCR kit (Ambion). RNA was reverse-transcribed using the Superscript III First-Strand cDNA Synthesis System (Invitrogen). For quantification of SoxC mRNA levels, real-time PCR was performed using PerfeCTa SYBR Green Master Mix (Quanta Biosciences, Gaithersburg, MD) and the Rotor-Gene 3000 detection system (Qiagen, Valencia, CA). The following SoxC primer sequences were used:
Sox4 forward, 5′GATCTCCAAGC GGCTAGGCAAA3′;
Sox4 reverse, 5′GTAGTCAGCCA TGTGCTTGAGG3′';
Sox11 forward, 5′GGACCTGGAT TCCTTCAGTGAG3′;
Sox11 reverse, 5′GTGAACACC AGGTCGGAGAAGT3′;
Sox12 forward, 5′CATTTCGAATT CCCGGACTA3′';
Sox12 reverse, 5′GGTCGGCGATAC TAGACGAG3′.
For absolute copy number determination, standard plots were constructed from known concentrations of serially diluted full-length SoxC cDNA plasmids (TrueORF Gold cDNA in pCMV6-Entry, OriGene Technologies, Rockville, MD) over the concentration range of 101−107 copies/μL, as previously described (Pfaffl, 2004; Dhanasekaran et al., 2010). Each standard dilution point was performed in triplicate over the complete standard-curve range to achieve a reliable standard curve for each measured parameter (R2 > 0.99; Efficiency (E) ≥ 1.95). Standard curves were performed for both target and plasmid cDNA in 3 or more replicates on different days to validate reproducibility (in all cases, R2 > 0.99, E ≥1.95). qPCR experiments and analyses were performed in adherence to the Minimum Information for Publication of Quantitative Real-Time PCR Experiments (MIQE) guidelines (Bustin et al., 2010). Supplemental Table S1 contains a qPCR checklist listing relevant technical information.
The Sox4 null (Sox4−) allele (Schilham et al., 1996), conditional Sox4 (Sox4C) allele (Penzo-Mendez et al., 2007) and Six2-Cre BAC transgenic line (Kobayashi et al., 2008) have been previously characterized. All animal experiments were carried out in accordance with the policies of the Animal Care Committee at the Atlantic Veterinary College, University of Prince Edward Island.
Transmission Electron Microscopy
Kidneys were cut into 3–4-mm-thick sections, immediately immersed in 2% glutaraldehyde in 0.1 M phosphate buffer, pH 7.2, and refrigerated for 30 min at 4°C. The cortical region was trimmed from each section and cut into smaller pieces and stored in 2% glutaraldehyde in phosphate buffer at 4°C overnight. Tissues were washed twice with phosphate buffer, post-fixed in 1% osmium tetroxide in phosphate buffer for 1 hr at room temperature, then washed for 10 min in 0.1 M phosphate buffer. Tissue was dehydrated through a graded series of ethanols, cleared in propylene oxide, infiltrated and embedded in Epon. Semi-thin sections (0.5 μm) were cut and stained with 1% toluidine blue in 1% sodium tetraborate solution and examined with a light microscope. All samples that contained at least two glomeruli were re-cut to generate ultrathin sections (90 nm). Ultrathin sections were stained with uranyl acetate and Sato's lead stain. Sections were examined using a Hitachi H7500 transmission electron microscope operated at 80 kV and digital images were captured using an AMT XR40 side-mounted camera.
Funding from the Kidney Research Scientist Core Education and National Training (KRESCENT) Program and the Kidney Foundation of Canada (S.H.) and the German Research Foundation (KA 3217/2-1, M.K.) is gratefully acknowledged. We thank Dr. Veronique Lefebvre for the Sox4C mouse line (NIH/NIAMS R01 grant AR54153), Zoe Grutzner, Kathleen Jones, Joy Knight, and Ashley Patriquen for their outstanding technical support. We gratefully acknowledge Drs. Norman Rosenblum, Tarek Saleh, Darcy Shaw, Kevin Burns, Adeera Levin, and Mr. Wim Wolfs for their mentorship to S.H. S.H. gratefully acknowledges Daniel Hartwig, Sharon MacKenzie, David MacKenzie, Florence Harris-Eze, Cheagun Pak and Ansoon Pak for their outstanding support sine qua non. Soli Deo Gloria.