BASIC RESEARCH www.jasn.org

Abnormal miR-148b Expression Promotes AberrantGlycosylation of IgA1 in IgA Nephropathy

Grazia Serino,*† Fabio Sallustio,*† Sharon N. Cox,* Francesco Pesce,* andFrancesco P. Schena*†

*Nephrology, Dialysis and Transplantation Unit, Department of Emergency and Organ Transplantation, University ofBari, Bari, Italy; and †Centro Addestramento Ricerca Scientifica in Oncologia (C.A.R.S.O.) Consortium, Valenzano, Italy

ABSTRACTAberrant O-glycosylation in the hinge region of IgA1 characterizes IgA nephropathy. The mechanismsunderlying this abnormal glycosylation are not well understood, but reduced expression of the enzymecore 1, b1,3-galactosyltransferase 1 (C1GALT1) may contribute. In this study, high-throughput microRNA(miRNA) profiling identified 37 miRNAs differentially expressed in PBMCs of patients with IgA nephrop-athy compared with healthy persons. Among them, we observed upregulation of miR-148b, which poten-tially targets C1GALT1. Patients with IgA nephropathy exhibited lower C1GALT1 expression, whichnegatively correlated with miR-148b expression. Transfection of PBMCs from healthy persons with amiR-148b mimic reduced endogenous C1GALT1 mRNA levels threefold. Conversely, loss of miR-148bfunction in PBMCsof patientswith IgA nephropathy increasedC1GALT1mRNAandprotein levels to thoseobserved in healthy persons. Moreover, we found that upregulation of miR-148b directly correlated withlevels of galactose-deficient IgA1. In vitro, we used an IgA1-producing cell line to confirm that miR-148bmodulates IgA1 O-glycosylation and the levels of secreted galactose-deficient IgA1. Taken together,these data suggest a role for miRNAs in the pathogenesis of IgA nephropathy. Abnormal expression ofmiR-148b may explain the aberrant glycosylation of IgA1, providing a potential pharmacologic target forIgA nephropathy.

J Am Soc Nephrol 23: ccc–ccc, 2012. doi: 10.1681/ASN.2011060567

IgA nephropathy (IgAN) is considered the mostcommon form of primary GN throughout theworld, and about 40% of patients develop ESRD.1,2

The initial event in the pathogenesis of IgAN isthe mesangial deposition of IgA1, aberrantly gly-cosylated because the hinge-region O-linked gly-cans lack galactose.3–6 Different studies suggest thatthis alteration could lead to IgA1 self-aggregation,7

IgA1-IgG immune complex formation,8,9 and com-plex defective clearance, with sequential depositionin the glomeruli.10

In humans, each IgA1 heavy-chain hinge regionhas three to five 5 O-linked glycans that are builtby stepwise addition of monosaccharides, begin-ning with the addition of N-acetylgalactosamineby the enzyme N-acetylgalactosaminyltransferase2 and continuing with the addition of galactoseby the enzyme core 1, b1,3-galactosyltransferase 1(C1GALT1).11–14

IgAN is a complex multifactorial disease whosepathogenic mechanism is still unknown. Severalinvestigators have sought to identify specific geneticmarkers associated with the development and pro-gression of this disease;15 however, few studies havespecifically described intracellular mechanisms as-sociated with disease development.16 Recently, ourgroup identified new mechanisms associated with

Received June 13, 2011. Accepted December 19, 2011.

G.S. and F.S. contributed equally to this work.

Published online ahead of print. Publication date available atwww.jasn.org.

Correspondence: Dr. Francesco P. Schena, Nephrology Dialysisand Transplantation Unit, Department of Emergency and OrganTransplantation, University of Bari, Policlinico, Piazza G. Cesareno. 11, 70124, Bari, Italy. Email: fp.schena@nephro.uniba.it

Copyright © 2012 by the American Society of Nephrology

J Am Soc Nephrol 23: ccc–ccc, 2012 ISSN : 1046-6673/2305-ccc 1

the pathogenesis of IgAN and showed that WNT-b-cateninand PI3K/Akt pathways were highly activated in patients withIgAN.17 The basis for the abnormal glycosylation in IgAN isstill unknown, but some studies suggest that C1GALT1 couldbe involved because of its altered expression.18

On the other hand, the potential role of microRNAs(miRNAs) in the IgAN pathogenesis has been poorly in-vestigated. After the discovery of miRNAs, great effort hasfocused on determining their biologic functions and theirrelevance to diseases. In fact, deregulation of miRNAs hasbeen associated with several disease states, including kidneydiseases.19

To our knowledge, this study is the first to evaluate the globalmiRNA expression profile of IgAN patients9 PBMCs, whichare directly involved in the disease.20 We defined the miRNAsignature in patients with IgAN and showed that miR-148bregulating C1GALT1 explains the abnormal glycosylation pro-cess in IgAN. These results support an important and unre-ported role of this miRNA in the pathogenesis of IgAN andsuggest a pharmacologic rationale for the potential use of syn-thetic miRNA inhibitors to attenuate IgA1 deglycosylation inthe disease.

RESULTS

Identification of Differentially Expressed miRNAs inPatients with IgANThe role of miRNA expression in the pathogenesis of IgANhas not been well explored. To identify miRNAs differentiallyexpressed in IgAN, we analyzed their global expression profilein PBMCs of seven patients with IgAN and seven healthyparticipants. Among 723 human miRNAs represented on themicroarrays, 147were expressed in each sample. Unsupervisedhierarchical clustering analysis generated a tree with the IgANand healthy participants clearly separated into two groups(Figure 1). This separation was further confirmed by display-ing the relationships among miRNA expression patterns usingprincipal component analysis (Supplemental Figure 1). Afterwe applied a fold change threshold . 2 (false discovery rate[FDR], 0.01), 35 miRNAs were found to be significantly up-regulated and 2 were significantly downregulated in IgAN (Sup-plemental Table 1). To validate microarray results, weperformed quantitative real-time PCR (qRT-PCR) for miR-148b, miR-188-5p, miR-361-3p, miR-886-3p, let-7b, and let-7don miRNAs isolated from PBMCs of an independent set of 10patients with IgAN and 10 healthy persons with the same clini-cal and demographic characteristics as those in the populationused for microarray experiments (Table 1). The expression of allanalyzedmiRNAswas significantly higher in patients with IgAN,thereby confirming microarray results (Figure 2, A and B).

In silico Analysis of miRNA TargetsTo study the molecular mechanisms in which the miRNAs areinvolved, we performed a bioinformatic analysis to predict

target genes of the 35 upregulated miRNAs. To reduce thenumber of false-positive results, we used four differentalgorithms and listed only putative target genes predicted byat least two of them. On the basis of the results of bioinfor-matics analysis, we found that one of the potential targets ofmiR-148b was the gene C1GALT1, which plays an importantrole in the pathogenesis of IgAN. Of note, other miR-148bputative target genes were inversin (INVS) and phosphataseand tensin homologue (PTEN) (Supplemental Table 2), twogenes that we found downregulated in patients with IgAN.17

Ingenuity Pathway Analysis (IPA) software was then used toevaluate thebiologic interaction amongmiRNAs.Weuploadedthe 35 upregulated miRNAs in IPA, and two networks wereidentified. When we merged each of them with the networkresulting from the gene expression profile published in ourprevious work,17 we found that miRNAs were strongly inter-connected with the mRNA network (Supplemental Figure 2).In particular, let-7d directly regulated PTEN, miR-361 regu-lated INVS, miR-148b regulated both INVS and PTEN, andthree miRNAs (let-7d, let-7a, miR-98) indirectly regulate AKTthrough RET.

Figure 1. Unsupervised hierarchical clustering of miRNA ex-pression profile. miRNA expression pattern of PBMCs of sevenpatients with IgAN and seven healthy subjects (HSs) were ex-amined using Agilent array composed of 723 human and 76human viral miRNAs. A total of 147 miRNAs were expressed in allsamples, discriminating patients with IgAN from HSs (P,0.0001;FDR,0.01). Two principal clusters were identified on the basis ofdifferential miRNA expression.

2 Journal of the American Society of Nephrology J Am Soc Nephrol 23: ccc–ccc, 2012

BASIC RESEARCH www.jasn.org

Identification of miR-148b as a Regulator of C1GALT1ExpressionC1GALT1 is known as a directly involved in IgAN; however,the basis for its decreased function in the disease is unknown.Therefore,we evaluated theC1GALT1mRNAexpression levelsby qRT-PCR in the same set of RNA samples used in themicroarray validation. We found that C1GALT1 levels weresignificantly lower in patients with IgAN than in the healthyparticipants (P,0.0001; Figure 3).

The in silico analysis showed that the miR-148b increasecould be the cause of the C1GALT1 reduction. In fact, se-quence alignment of human miR-148b with 39-untranslatedregion (UTR) C1GALT1 identified a binding site (nucleotides1355–1361 in human C1GALT1, chr 12q13; Figure 4A) that iswell conserved among different species (Figure 4B).

To study whether the downregulation of C1GALT1 inpatients with IgAN was attributable to increased miR-148blevels, we tested the correlation between C1GALT1 mRNA and

miR-148b levels in patients with IgAN andthe healthy participants. A negative corre-lation was observed (R2=0.4; P,0.01; Fig-ure 4C): samples with higher miR-148bexpression levels showed lower levels ofC1GALT1 mRNA.

The inverse correlation observed be-tween the levels ofmiR-148b andC1GALT1mRNA suggests that this gene is probably atarget of miR-148b.

Moreover, we performed a bioinformaticanalysis to estimate the effect of some single-nucleotide polymorphisms (SNPs) on puta-tivemiRNAtargets interrogating the 39-UTRand predicting whether a SNP within thetarget site will disrupt/eliminate or enhance/create a miRNA binding site. Our results re-vealed that the 1365G/A polymorphism(rs1047763) in the 39-UTR of C1GALT1,which is associated with IgAN,21,22 affects

miR-148b binding sites in this gene. In fact, the biocomputa-tional analysis showed that the 1365G allele enhances miR-148bbinding (Figure 4D).

To study whether the upregulated expression of miR-148bwas specific to IgAN, we checked the miR-148b expression inPBMCs from controls with three additional diseases: 3 patientswithmembranoproliferativeGN type I , 5 with focal segmentalglomerulosclerosis, and 10 with Henoch–Schönlein purpuranephritis. miR-148b expression in these diseases has beencompared with that of IgAN and controls. We found that themiR-148b was again higher in patients with IgAN (P,0.0001)than in those with the other three diseases (Figure 5), con-firming that higher miR-148b levels are typical of IgAN.

miR-148b Downregulates C1GALT1 mRNA ExpressionThe nature of target of miR-148b and its mode of action inPBMCs have not been functionally identified; thus, we testedwhether miR-148b was able to modulate the expression of

Table 1. Demographic and clinical features of patients and healthy participants

Variable

Initial Sample Cohort Disease Controls Validation Sample Cohort

IgANHealthy

ParticipantsMPGN-I FSGS HSPN IgAN

HealthyParticipants

Participants (n) 25 25 3 5 10 50 50Men/women (n/n) 16∕9 15∕10 2/1 3/2 5/5 38/12 40/10Age (yr) 37.2610.2 36.1610.6 3569.6 42612.3 8.563 36612.3 43610.6Serum creatinine (mg/dl) 0.960.2 0.860.3 0.760.05 0.860.4 0.560.1 1.0060.5 0.960.3Estimate GFR (ml/min per 1.73 m2) 118622.3 106613.1 122610 116.8611 ND 120611 114612.3Proteinuria (g/24 h) 0.360.1 0.160.2 2.761.2 2.0560.9 ND 0.860.2 0.0660.02Systolic BP (mmHg) 121620.4 120610 133611.5 126615 118616 124613.3 11969Diastolic BP (mmHg) 7269 7468 8060.5 7966.5 6768 77.269.7 7469Total IgA (ng/ml) 0.1060.004 0.1260.01 ND ND 0.1560.01 0.1460.002 0.1360.001

Unless otherwise noted, values are expressed as mean 6 SD. IgAN, IgA nephropathy; MPGN-I, membranoproliferative GN type I; FSGS, focal segmentalglomerulosclerosis; HSPN, Henoch-Schönlein purpura nephritis; ND, not determined.

Figure 2. Differential expression validation. Validation of differential expression ofmiR-148b, miR-188-5p, miR-886-3p, let-7b, and let-7d (A) and miR-361-3p (B) in anindependent set of PBMCs from 10 patients with IgAN and 10 healthy subjects (HSs).Expression levels were quantified using qRT-PCR. ThemiRNA relative expressions werenormalized to the expression of U6. Expression levels of miR-148b, miR-188-5p, miR-886-3p, let-7b, let-7d, and miR-361-3p were found to be significantly higher in pa-tients with IgAN than in HSs. The histograms represent the mean 6 SEM. *P,0.03;**P,0.01.

J Am Soc Nephrol 23: ccc–ccc, 2012 miR-148b in IgA Nephropathy 3

www.jasn.org BASIC RESEARCH

C1GALT1. For this purpose, we performed some transienttransfection experiments ex vivo using PBMCs from an inde-pendent group of four patients with IgAN and four healthypersons.

We increased the amount of the endogenous miR-148bwithin PBMCs isolated from four healthy persons, transfectingshort RNA sequences that mimic the action of the miRNA, tosimulate the situation found in patients with IgAN. PBMCs ofthe four healthy persons, transfected with 25nM miR-148b

mimic, showed a three-fold reduction in endogenous C1GALT1mRNA levels (P,0.03; Figure 6A).

To provide evidence for a direct effect of miR-148b onC1GALT1expression,we transfected PBMCs isolated fromfiveother healthy persons with both a miR-148b mimic and amutated miR-148b mimic (the D20-21 miR-148b, with asubstitution from UG to AC in position 20-21 [59→39]).We found that the C1GALT1 mRNA expression decreasedonly after transfection with the wild-type miR-148b mimic(Figure 6B).

To further confirm that C1GALT1 mRNA was targeted bymiR-148b, we performed the reverse experiment, in which weused synthetic hairpinmiRNA inhibitors to silence the activityof miR-148b in PBMCs isolated from four patients with IgAN.The miRNA inhibitors were chemically modified acids de-signed to bind and to inhibit endogenous miR-148b. Thetransient transfection of IgAN PBMCs with 250 nM of miR-148b inhibitor led to a three-fold increase in endogenousC1GALT1 mRNA levels (P,0.01, Figure 6C). In line with ourcomputational analysis, we showed that miR-148b did notaffect Cosmc chaperone, which is involved in stability andactivity of C1GALT123 (Supplemental Figure 3).

miR-148b Downregulates C1GALT1 Protein ExpressionWe used the same strategy to functionally inhibit or enhancethemature form ofmiR-148b in PBMCs of patients with IgANand healthy persons, respectively, with the aim of establish-ing whether C1GALT1 protein expression was also effec-tively controlled by miR-148b. PBMCs of five healthy persons

Figure 3. C1GALT1 gene expression levels evaluated by real-time PCR in 10 patients with IgAN and 10 healthy subjects (HSs).C1GALT1 expression levels were significantly lower in patientswith IgAN than in HSs. C1GALT1 expression levels were nor-malized on the housekeeping gene b-actin. The scatter plotsrepresent the mean 6 SEM. #P,0.0001.

Figure 4. miR-148b targets C1GALT1. (A) Genomic localization of miR-148b in chromosome 12q13. (B) Sequence alignment of the miR-148b base-pairing sites in the 39-UTR of C1GALT1 mRNA showing that the regions complementary to miR-148b are highly conservedamong human, chimp, mouse, rat, dog, and chicken. The “seed” sequences of miR-148b complementary to C1GALT1 are shown ingray. (C) Linear correlation between the expression of C1GALT1 and the expression of miR-148b in PBMCs of 10 patients with IgANand 10 healthy subjects (HSs). C1GALT1 mRNA levels inversely correlated with miR-148b expression levels (R2=0.4; P,0.01). (D)Output of bioinformatic analysis performed by MicroSNiPer algorithms showing that the 1365G/A polymorphism (rs1047763) affectsmiR-148b binding sites in the 39-UTR of C1GALT1. The 1365G allele (arrow) enhances miR-148b binding, whereas the 1365A alleledoes not appear in the output as a possible binding site for miR-148b.

4 Journal of the American Society of Nephrology J Am Soc Nephrol 23: ccc–ccc, 2012

BASIC RESEARCH www.jasn.org

transiently transfectedwith 25 nMofmiR-148bmimic resultedin a remarkable reduction inC1GALT1protein expression (2.2-fold; P,0.002) as revealed by Western blot. Of note, the in-crease in the miR-148b level in PBMCs of healthy personsresulted in the same C1GALT1 protein expression levels as inPBMCs of patients with IgAN (Figure 7A). On the contrary, inthe reverse experiment, PBMCs of five patients with IgANtransfected with 250 nM of miR-148b inhibitor showed a sig-nificant increase in C1GALT1 protein production (1.7-fold in-crease; P,0.01). Surprisingly, C1GALT1 protein levels inpatients with IgAN increased to the same levels as in healthypersons (Figure 7B). Altogether our results demonstrate thatmiR-148b targets C1GALT1, which is responsible for a de-creased activity of the gene in patients with IgAN.

miR-148b Expression and Altered IgA1 GlycosylationRelationshipTo further validate the relationship between miR-148b ex-pression and altered IgA1 glycosylation, we enrolled an in-dependent cohort of 50 patients with IgAN and 50 healthypersons. Again, miR-148b expression levels, evaluated byqRT-PCR, were significantly higher in PBMCs of patientswith IgAN than in those of healthy persons (P,0.001;

Figure 8A).We next evaluated the galactose-deficient

(Gal-deficient) IgA1 in serum of the samepatients with IgAN and healthy personsusing helix aspersa agglutinin (HAA) lec-tin binding assay. We found that the Gal-deficient IgA1 was significantly higher inpatients with IgAN (mean OD 6 SD,0.9860.1) than in healthy persons (meanOD, 0.5360.02; P,0.0001) (Figure 8B),whereas total IgA levels were similar inthe two groups (0.13760.002 and 0.13260.002 mg/ml, respectively) (Figure 8C). Pear-son correlation analysis between miR-148bexpression level and Gal-deficient IgA1level showed a significant positive correla-tion (r = 0.4; P,0.0001; Figure 8D), sup-porting our data on C1GALT1 regulationby miR-148b and explaining the abnormalincrease in Gal-deficient IgA1 consequentto higher miR-148b levels in patients withIgAN. To confirm the direct cause-and-effect relationship between miR-148b ex-pression and Gal-deficient IgA production,we used as amodel the human B lymphomaDAKIKI cells producing IgA1.24 We trans-fected DAKIKI cells with miR-148b mimicand inhibitor and evaluated changes in theGal-deficient IgA1 levels by means of HAAlectin binding assay. DAKIKI cells transfec-ted with 50 nMofmiR-148bmimic resultedin a remarkable increase in Gal-deficient

Figure 5. miR-148b expression levels evaluated by real-time PCRin PBMCs from 3 patients with membranoproliferative GN type I(MPGN-I), 5 with focal segmental glomerulosclerosis (FSGS), and10 with Henoch-Schönlein purpura nephritis (HSPN). miR-148bexpression in these diseases and in healthy subjects (HSs) hasbeen compared with that in IgAN. We found that miR-148b levelswere higher in patients with IgAN than in those with MPGN-I,FSGS, and HSPN and in HSs. The miRNA relative expressionswere normalized to the expression of U6. The histograms repre-sent the mean 6 SEM. **P,0.01; #P,0.0001.

Figure 6. miR-148b regulates C1GALT1 mRNA expression. (A) C1GALT1 expressionlevels were analyzed by real-time PCR in PBMCs of healthy subjects (HSs) aftertransfection with miR-148b mimic. Increasing the amount of miR-148b within thePBMCs of HSs resulted in a three-fold reduction in endogenous C1GALT1 mRNAlevels. (B) C1GALT1 expression levels were analyzed by real-time PCR in PBMCs of fiveother HSs after transfection with the mutated miR-148b mimic, D20-21, and miR-148bmimic wild-type. C1GALT1 mRNA expression decreased only after transfection withthe wild-type miR-148b mimic. (C) C1GALT1 expression levels were analyzed by real-time PCR in PBMCs of patients with IgAN after transfection with miR-148b inhibitor.Silencing the activity of miR-148b within IgAN PBMCs led to a three-fold increase inendogenous C1GALT1 mRNA levels. “Mock” indicates mock-transfected cells goingthrough the transfection processes without addition of mimic/inhibitor miRNA. Expres-sion data were normalized on the housekeeping gene b-actin. Data are representativeof four independent experiments for A and C and five independent experiments for B(mean 6 SEM). **P,0.01.

J Am Soc Nephrol 23: ccc–ccc, 2012 miR-148b in IgA Nephropathy 5

www.jasn.org BASIC RESEARCH

IgA1 levels (1.5-fold increase; P,0.002; Figure 8E). In con-trast, in the reverse experiment, DAKIKI cells transfected with500 nMof miR-148b inhibitor showed a significant reductionin Gal-deficient IgA1 levels (1.8-fold; P,0.01; Figure 8E).These data further confirmed the role of miR-148b in IgA1O-glycosylation regulation.

DISCUSSION

IgAN is the most common primary form of GN worldwide,25

leading to ESRD in 40% of patients. There is a consensusopinion that aberrant glycosylation of IgA1 is directly involvedin the pathogenesis of the disease.3–5,10,26 This abnormality ismanifested by a deficiency of galactose in the hinge-region O-linked glycans of IgA1. Biosynthesis of these glycans beginswith the addition of N-acetylgalactosamine by the enzyme N-acetylgalactosaminyltransferase 2 and continues with the ad-dition of galactose by the enzyme C1GALT1.11–14

Themechanisms leading to this aberrant glycosylation havebeen studied extensively; however, to date, the nature ofabnormally O-glycosylated IgA1 is still obscure. Kudo et al.demonstrated an important role for C1GALT1 in the molec-ular basis for the aberrant IgA1 glycosylation in IgAN.18 How-ever, it is controversial whether IgA1 is undergalactosylatedas a consequence of functional changes in C1GALT1 activity oras a consequence of its reduced expression.14,27 Moreover, themolecular mechanisms behind the impaired functioning of

C1GALT1 are again unknown. Accumulat-ing evidence suggests the involvement ofmiRNAs in the pathogenesis of some kid-ney diseases, such as acute kidney rejection,lupus nephritis, and renal ischemia reper-fusion injury.28–30 However, miRNAs po-tentially involved in the pathogenesis ofIgAN have been partially studied in renaltissue.31

Starting from a microarray miRNA ex-pression profile in PBMCs of patients withIgAN and using several bioinformatic ap-proaches, we identified 37 miRNAs that aredifferentially regulated in IgAN patientscompared with healthy persons, along withtheir candidate target genes.

Furthermore, we overlappedmodulatedmiRNAs with previously identified genepathways involved in IgAN17 to identifycandidate target genes involved in the path-ogenesis. This analysis revealed that themiRNAs and genes differentially expressedin IgAN were strongly interconnected, par-ticularly the six miRNAs whose expressionwe validated by qRT-PCR.

Among 37 miRNAs significantly dereg-ulated in patients with IgAN, miR-148b

aroused our interest because its putative target genes wereC1GALT1, INVS, and PTEN, three genes downregulated inthese patients.11,17 We found a reduced expression ofC1GALT1 in patients with IgAN, and we showed that miR-148b expression negatively correlated with the C1GALT1 ex-pression levels in the same patients.

In the last years, two large-scale case-control associationstudies pointed out the role of C1GALT1 in the pathogenesisof IgAN, reporting some C1GALT1 genetic variants associatedwith the genetic susceptibility to IgAN in the Italian andChinese populations.21,22 Both studies identified a positiveassociation between the susceptibility to IgAN and aC1GALT1 SNP at the position 1365 G/A (rs1047763) in the39-UTR regulatory region. Furthermore, a correlation be-tween this SNP and a reduced C1GALT1 expression in homo-zygous 1365G/G individuals was shown.22

Our bioinformatic analysis revealed that rs1047763 SNPis located within the miR-148b binding site on the C1GALT139-UTR region, and the binding affinity of miR-148b is greaterwhen the 1365G-variant allele is present. These results couldexplain the lower C1GALT1 expression levels in patients withIgAN carrying the 1365G/G genotype.22

Supporting our bioinformatic analysis, we found thatC1GALT1 mRNA levels were significantly lower in patientswith IgAN. Moreover, using the strategy to functionally en-hance or inhibit the mature form of miR-148b, we confirmedbiologically, in an ex vivo experiment, that miR-148b targetsC1GALT1 and can modulate its mRNA and protein levels. Of

Figure 7. miR-148b regulates C1GALT1 expression at protein level. (A) Transfection ofPBMCs of healthy participant with 25 nM miR-148b mimic resulted in a 2.2-fold re-duction in C1GALT1 protein expression. Increasing the miR-148b levels within PBMCsof healthy participants resulted in the same C1GALT1 protein levels as in IgAN PBMCs,as shown by Western blot. (B) Western blot representing protein levels of C1GALT1 inIgAN PBMCs after transfection with 250 nMmiR-148b inhibitor. A significant increase inC1GALT1 protein production was shown in IgAN PBMCs transfected with miR-148binhibitor (1.7-fold increase). miR-148b loss of function led to the same C1GALT1protein levels as in PBMCs of healthy participants. “Mock” indicates mock-transfectedcells going through the transfection processes without addition of mimic/inhibitormiRNA. In both experiments, b-actin was used as endogenous control. Data arerepresentative of five independent experiments (mean 6 SEM). **P,0.01, §P# 0.001.

6 Journal of the American Society of Nephrology J Am Soc Nephrol 23: ccc–ccc, 2012

BASIC RESEARCH www.jasn.org

note, the loss of function of miR-148b in PBMCs isolated frompatients with IgAN led to C1GALT1 protein levels similar tothose observed generally in healthy persons. In contrast, miR-148b overexpression in PBMCs isolated from healthy personsled to a reduction in C1GALT1 protein to levels similar to thoseexpressed in patients with IgAN.

Moreover, patients with IgAN who had higher serum levelsofGal-deficient IgA1 also showed higher levels ofmiR-148b. Incontrast, lower levels of serum Gal-deficient IgA1 in healthy

personswere associatedwith lowmiR-148blevels. This finding supports our data onC1GALT1 regulation by miR-148b andexplains the abnormal increase in Gal-deficient IgA1due to highmiR-148b levels inpatients with IgAN. The role of miR-148b inIgA1 O-glycosylation regulation was alsoconfirmed by transfection experiments inDAKIKI human B lymphoma cells showingthat miR-148b modulated levels of secretedGal-deficient IgA1.

On the basis of our results, we canhypothesize that in IgAN, lower levels ofC1GALT1 due to the inhibiting effect ofoverexpressed miR-148b may lead to anincrease in circulating deglycosylatedIgA1.8,9,32–34 These galactose-deficientIgA1 molecules are predisposed to aggre-gation and formation of immune com-plexes;8 they are recognized by naturallyoccurring circulating antibodies.9 Conse-quently, these complexes are less effectivelycleared from the circulation35 and effi-ciently bind to mesangial cells, inducingcellular proliferation, overproduction ofextracellular matrix, and synthesis of in-flammatory cytokines that can initiate orperpetuate the course of GN.36–38

Our study has one important limitation.We did not study the effect of miR-148bon the process of IgA1 glycosylation in anexperimental model of IgAN because IgA1is present exclusively in humans and hom-inoid primates.39,40 However, our studypoints overall to a new regulatory mecha-nism of IgAN that can explain the aberrantglycosylation of IgA1 responsible for thepathogenesis of the disease. Moreover, themiR-148b expression levels may be used todevelop a method to diagnose the diseasethrough a noninvasive technique, whichmay replace, in the future, the renal biopsyas a diagnostic approach.

In conclusion, our study shows, for thefirst time, a global miRNA expression pro-file in PBMCs of patients with IgAN and a

miRNA signature. We also provide evidence for a previouslyunknown key regulatory mechanism of IgAN by discover-ing the deregulation of miR-148b, which could explain theaberrant glycosylation of IgA1 in the disease. These findingscould also have therapeutic implications because the inhibitionof miR-148b expression reverses the lower levels of C1GALT1typical of IgAN. Therefore, miR-148b levels may be manip-ulated to provide useful new therapeutic approaches to thedisease.

Figure 8. miR-148b regulates Gal-deficient IgA1 levels. (A) The miR-148b expressionlevels evaluated by real-time PCR in 50 patients with IgAN and 50 healthy subjects(HSs). miR-148b expression levels were significantly higher in patients with IgAN than inHSs. miR-148b expression levels were normalized to the expression of U6. The his-tograms represent the mean 6 SEM. §P,0.001. (B) shows the serum levels of Gal-deficient IgA1 in patients with IgAN and HSs. The Gal-deficient IgA1 was significantlyhigher in sera obtained from the patients with IgAN than in sera from HSs. The relativelectin binding per unit IgA1 was calculated as the OD value of lectin over the OD valueof total IgA. (C) The serum level of total IgA in patients with IgAN and HSs determinedby ELISA. IgA levels in sera obtained from 50 patients with IgAN and 50 HSs weresimilar. The histograms represent the mean 6SEM. #P,0.0001. (D) Linear correlationbetween the expression of miR-148b and the Gal-deficient IgA1 levels in 50 patientswith IgAN and 50 HSs. miR-148b levels directly correlated with Gal-deficient IgA1levels (r=0.4,; P,0.0001). (E) The Gal-deficient IgA1 levels in supernatants obtainedfrom DAKIKI cells after transfection with miR-148b mimic and inhibitor. A significantincrease in Gal-deficient IgA1 production was shown in DAKIKI cells transfected with50 nM miR-148b mimic (1.5-fold increase). On the contrary, the transfection with 500nM miR-148b inhibitor led to a significant reduction in Gal-deficient IgA1 levels (1.8-fold). The relative lectin binding per unit IgA1 was calculated as the OD value of lectinover the OD value of total IgA. The histograms represent the mean 6 SEM. §P,0.001.

J Am Soc Nephrol 23: ccc–ccc, 2012 miR-148b in IgA Nephropathy 7

www.jasn.org BASIC RESEARCH

CONCISE METHODS

Sample CollectionA total of 75 patients with IgAN, 75 healthy participants, 10 patients

with Henoch-Schönlein purpura nephritis, 3 patients with mem-

branoproliferative GN type I, and 5 patients with focal segmental

glomerulosclerosis were included in this study. Written informed

consent was obtained from each patient and healthy participant.

The study was initially conducted on a cohort of 25 patients with

IgAN and 25 healthy participants. Seven in each group were included

in the microarray experiment, 10 participants from each group were

used for microarray validation, and 8 participants from each group

were used for the in vitro transfections. Patients with IgAN had nor-

mal renal function; patients with severe renal damage were excluded

from the study. Each patient was selected according to the following

demographic and clinical features: age (18–50 years), moderate histo-

logic lesions (G2 according to the Schena classification41), creatinine

clearance .90 ml/min per 1.73 m2 (evaluated by the Cockcroft-Gault

formula), and clinical follow-up (at least 7 years). The healthy partic-

ipants were selected on the basis of their demographic characteristics

and overlapped completely with IgAN group. All healthy participants

had a negative result on a urine test for blood and proteins. All patients

with diabetes, chronic lung diseases, neoplasm, or inflammatory dis-

eases and patients receiving antibiotics, corticosteroids, and nonste-

roidal anti-inflammatory agents were excluded from the study. No

patients had symptomatic coronary artery diseases or a family his-

tory of premature cardiovascular diseases.

To validate some of the results obtained, we collected blood

samples from an independent cohort of 50 patients with IgAN and 50

healthy persons matched in their demographic and clinical features

with the participants included in the first group.

Ten patients with Henoch-Schönlein purpura nephritis, three

with membranoproliferative GN type I, and five with focal segmental

glomerulosclerosis were selected as disease controls.

All participants were from the Italian population. The study was

carried out according to the principles of the Declaration of Helsinki

and was approved by our institutional ethics review board.

The main demographic and clinical features of our patients and

controls included in the study are summarized in Table 1. There were

no statistically significant differences between the IgAN and healthy

groups for any measure considered.

PBMCs were isolated by density separation over a Ficoll-Hypaque

(Ficoll-Paque Plus, GE Healthcare) gradient (460 g for 30 minutes).

PBMCs were washed three times with PBS (pH 7.4)/1 mM EDTA

(Sigma). Cells were then counted, and their viability was determined

by trypan blue exclusion.

Serum was collected at the same clinical time point, and whole

blood was centrifuged at 3000 rpm for 10 minutes at room tem-

perature.

miRNA IsolationIsolated PBMCs were used for total RNA extraction by means of

miRNeasy Mini kit (Qiagen) according to the manufacturer’s pro-

tocol. Deoxyribonuclease treatment was carried out to remove any

contaminating DNA (RNase-Free DNase Set, Qiagen).

Total RNA, including small RNA fractions, was then eluted in

ribonuclease-freewater. TheRNAconcentrationwasdeterminedwith

NanoDrop Spectrophotometer (Nanodrop Technologies). The

miRNA quality was assessed using Agilent 2100 Bioanalyzer (Agilent

Technologies), and only samples with RNA integrity number . 8.5

were used.

miRNA MicroarraymiRNAmicroarray analysis was performed using the Agilent Human

miRNAMicroarrays V2, whichwere based on SangermiRBase release

10.1, according to the manufacturer’s protocol. Briefly, 600 ng of

total RNA isolated from PBMCs of seven patients with IgAN and

seven healthy participants were first dephosphorylated with a calf in-

testine alkaline phosphatase treatment for 30 minutes at 37°C before

labeling. Samples were diluted with DMSO, denatured for 10 minutes

at 100°C, and labeled using pCp-Cy3 in T4 RNA ligation buffer.

The labeled RNA was hybridized, washed, stained, and scanned

with an Agilent microarray scanner (G2565BA, Agilent). Microarray

data analysiswasperformedusingAgilent FeatureExtractionSoftware

9.5.1.1.

Microarray data are available under accession number GSE25590

at the Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/).

Statistical Analyses and BioinformaticsFor microarray analysis, the raw expression signals were log-

transformed, normalized, and filtered according to the median cor-

rected signal of all the miRNAs with an intensity .100 (which is

considered as expressed), which resulted in the selection of 147

miRNAs out of the original set of 723 miRNAs. The preprocessed

microarray data were imported into the R language for statistical

analysis computing (http://www.r-project.org). miRNAs displaying

differential expression between patients with IgAN and healthy par-

ticipants were detected using a two-sample t test. Probe sets were

sorted after a P value showed a statistically significant difference

and were adjusted to account for multiple testing using the FDR

method of Storey and Tibshirani.42 Only genes that were significantly

(FDR,0.01) modulated in patients with IgAN compared with healthy

participants were considered for further analysis. Significance of differ-

ential expression, as determined by the enrichment analysis, was per-

mutated 1000 times. To determine the features that are most closely

correlated with IgANphenotype, we applied amost stringent filter with

FDR,0.01 and a fold change. 2. Two-dimensional hierarchical clus-

tering was performed using Spotfire decision site 8.0 (http://spotfire.

tibco.com). To assess biologic relationships among genes, we used In-

genuity Pathway Analysis software (IPA, Ingenuity System, Redwood

City, CA). IPA computes a score for each network according to the fit of

the set of supplied focusmiRNAs/genes. These scores indicate the likeli-

hood that focus genes would belong to a network versus those ob-

tained by chance. miRNA targets were predicted by means of miR-

Base 13.0 (http://microrna.sanger.ac.uk),43 TargetScan 5.1 (http://

www.targetscan.org),44 PicTar (http://pictar.org),45 and miRWalk

(http://www.ma.uni-heidelberg.de) algorithms. Potential targets were

chosen overlapping results from the four algorithms and selecting tar-

gets of genes predicted by at least two of the algorithms; the targets were

selected by a score cutoff computed by a weighted sum of a number of

8 Journal of the American Society of Nephrology J Am Soc Nephrol 23: ccc–ccc, 2012

BASIC RESEARCH www.jasn.org

sequence and context features of the predicted miRNA:mRNA duplex.

Bioinformatic analysis to estimate the effect of SNPs on putative

miRNA targets was carried out by microSNiPer software (Clinical

Brain Disorders Branch, National Institutes of Health, Bethesda,

MD; http://cbdb.nimh.nih.gov/microsniper)46 interrogating the

39-UTR and predicting whether a SNP within the target site will

disrupt/eliminate or enhance/create a microRNA binding site. A

banded Smith-Waterman dynamic program was used along the

best-matched regions. The algorithm output only those miRNAs/

seeds in which the alignment was shifted because of a change in

the alleles.

A two-tailed t test was used to assess differences in biologic fea-

tures between patients with IgANand healthy participants. A Pearson

correlation test was used to study continuous variables. All values

were expressed as the mean 6 SEM of data obtained from at least

three independent experiments. Results were considered statistically

significant at P,0.05.

qRT-PCRTotal RNA, including small RNA fractions, was reverse transcribed

with miScript Reverse Transcription Kit (Qiagen) following the

manufacturer’s instructions.

Real-time RT-PCR for the quantification of a subset of miRNAs

(miR-148b, miR-188-5p, let-7d, let-7b, miR-361-3p, and miR-886-3p,

plus an endogenous control) was carried out with miScript Primer

Assays andmiScript SYBRGreen PCR Kit fromQiagen. Real-time PCR

amplification reactions were performed in triplicate in 25 ml of final

volume via SYBR Green chemistry on iCycler (Bio-Rad).

Normalization was performed with a small nucleolar RNA U6

endogenous control. Comparative real-time PCR was performed in

triplicate, including no-template controls. Relative expression was

calculated using the 2-DCt method.

For C1GALT1 and Cosmc expression analysis, total RNA was

reverse-transcribed with QuantiTect Reverse Transcription Kit

(Qiagen) following the manufacturer’s instructions. qRT-PCR

amplification reactions were performed in triplicate in 25-ml final

volumes using SYBR Green chemistry on an iCycler. qRT-PCR was

performed using the QuantiTect Primer Assay and the QuantiFast

SYBR Green PCR mix (Qiagen). Genes were amplified according to

the manufacturer’s directions. The b-actin gene amplification was

used as a reference standard to normalize the target signal.

miR-148b Mimic and Inhibitor TransfectionIsolated PBMCs were cultured at 2 3 106 cells per well in a 12-well

plate with RPMI-1640 supplemented with 2 mM L-glutamine, 1 mM

sodium pyruvate, 1 mM nonessential amino acids, 25 mM HEPES

buffer, and 10% heat-inactivated FBS. The transfection of miRNAs

mimic and inhibitor was carried out using TransIT-TKOTransfection

Reagent (Mirus) in accordance with manufacturer’s procedure. For

each transfection, 25 nM of miR-148b mimic (Qiagen), 250 nM of

miR-148b inhibitor (Qiagen), and 25 nM of a mutated miR-148b

mimic (D20-21 miR-148b, with a substitution from UG to AC in

position 20-21 [59→39]; Qiagen) were used. In transfection

experiments, a mock-transfection control was performed by putting

cells through the transfection procedure without adding miRNA. The

validated nonsilencing siRNA sequence AllStars Negative Control

siRNA (50 nM, Qiagen) was used as negative control and Syn-hsa-

miR-1 miScript miRNAmimic (25 nM, Qiagen) and anti–hsa-miR-1

miScript miRNA inhibitor (250 nM, Qiagen) were used as positive

controls for miR-148b mimic and miR-148b inhibitor transfection,

respectively. Each transfection experiment was done in triplicate.

After transfection, cells were incubated for 24 hours or 48 hours at

37°C in 5% CO2 and used, respectively, for total RNA extraction and

protein extraction.

For analysis of IgA1 glycosylation, the surface IgA1-positive

human B lymphoma cell line, DAKIKI, was purchased from

ATCC.24,47 Cells were cultured at 1 3 105 cells per well in a 12-well

plate with RPMI 1640 medium containing 10% heat-inactivated FBS,

1 mM sodium pyruvate, and 2 mM glutamine. The transfection of

miRNAs mimic and inhibitor was carried out using TransIT-TKO

Transfection Reagent (Mirus) in accordance with the manufacturer’s

procedure. For each transfection, 50 nM of miR-148b mimic (Qiagen)

and 500 nM of miR-148b inhibitor (Qiagen) were used. Each trans-

fection experiment was done in triplicate. After transfection, cells

were incubated for 72 hours at 37°C in 5% CO2.

Western Blot AnalysisThe amount of C1GALT1, after transfection, was determined by

Western blotting analysis. Total protein extract were prepared

with lysis buffer containing 150mMNaCl, 50mMTris-HCl (pH 8),

1% NonidetP-40, 0.1% sodium deoxycholate, and 0.1% SDS, plus

proteinase inhibitors. The protein concentration was determined

by the Bradford assay (BioRad). Eightymicrograms of each protein

lysate was separated on a 10% SDS-PAGE and transferred to

polyvinylidene difluoride membrane (Millipore). The membranes

were incubated in 5% nonfat milk powder diluted in PBS

containing 0.1% Tween-20 for 2 hours at room temperature and

probed with a mouse monoclonal anti-C1GALT1 antibody (Santa

Cruz Biotechnology) in blocking buffer overnight at 4°C. Finally,

membranes were incubated with secondary antibody of horserad-

ish peroxidase conjugated goat anti-mouse IgG (BioRad). Immu-

nocomplexes were detected with enhanced chemiluminescence

method (GE Healthcare). The same membranes were stripped

and reprobed with anti–b-actin monoclonal antibody (Sigma).

Images of autoradiography were acquired using a scanner EPSON

Perfection 2580 Photo (EPSON) and quantified by Image J 1.34

Software (http://rsb.info.nih.gov/ij). The ratio between intensities of

C1GALT1 and b-actin bands was used to normalize C1GALT1 ex-

pression in each sample.

Determination of Serum IgAIgAcontent in serum fromeachparticipantwasmeasured in duplicate

using ELISA. Briefly, high-adsorption polystyrene 96-microwell

plates (Corning Inc.) were coated overnight with 5 mg of F(ab9)2fragment goat anti-human IgA antibody per ml (Jackson Immuno-

Research Laboratories) in PBS at 4°C. Plates were blocked with 1%

BSA in PBS containing 0.05% Tween-20 for 90 minutes at room

temperature. Samples diluted in blocking buffer and standard human

IgA (Calbiochem) were added to each well and then incubated for 90

minutes at room temperature. The captured IgA was then detected

J Am Soc Nephrol 23: ccc–ccc, 2012 miR-148b in IgA Nephropathy 9

www.jasn.org BASIC RESEARCH

with biotin-labeled F(ab9)2 fragment of goat IgG anti-human IgA

(Biosource). The binding was measured after addition of avidin–

horseradish peroxidase conjugate (ExtrAvidin; Sigma-Aldrich), and

the reaction was developed with the peroxidase chromogenic sub-

strate o-phenylenediamine–H2O2 (Sigma-Aldrich). The color re-

action was stopped with 2 N H2SO4, and the OD at 490 nm was

determined in a microplate reader (GDV, Italy; model DV 990 B/V6).

IgA1 concentration in unknown samples was determined by interpo-

lation of the respective OD into the appropriate standard curve.

Detection of Gal-deficient IgA1Gal-deficient IgA1was detected by binding of the lectin,HAA (Sigma-

Aldrich), which is specific for terminal GalNAc, as reported else-

where.48,49 High-adsorption polystyrene 96-microwell plates (Corning

Inc.) were coated overnight with 3 mg of F(ab9)2 fragment goat anti-

human IgA antibody per ml (Jackson ImmunoResearch Laboratories)

in PBS at 4°C. Plates were blocked for 3 hours at room temperature with

1% BSA/PBS containing 0.1% Tween-20. Samples diluted in blocking

buffer were added to each well and incubated overnight at 4°C. The

captured IgAwas subsequently desialylated by treatment for 3 hours at

37°C with 20 mU/ml neuraminidase from Vibrio cholerae (Sigma-

Aldrich) in 10 mM sodium acetate buffer (pH 5).7 Samples were then

incubated for 3 hours at 37°Cwith 2mg of GalNAc-specific biotinylated

HAA lectin per ml (Sigma-Aldrich) diluted in blocking buffer. The

lectin binding was detected with avidin–horseradish peroxidase con-

jugate (ExtrAvidin) diluted in blocking buffer, and the reaction was

developed with the peroxidase chromogenic substrate o-phenylenedi-

amine-H2O2 (Sigma-Aldrich). The color reactionwas stopped with 2N

H2SO4, and the OD at 490 nm was determined in a microplate reader

(GDV, Italy; model DV 990 B/V6).

The relative lectin binding per unit IgA1 was calculated as the OD

value of lectin over the OD density value of IgA concentration.

ACKNOWLEDGMENTS

We are grateful to the patients with IgAN and healthy blood donor

participants for their cooperation in this study. We express our ap-

preciation toDr. Gabriella Aceto andDr. Anna Rita Cappiello for help

with collecting clinical samples frompatients withHenoch-Schönlein

purpura nephritis.

This work was supported by grants from the MiUR (COFIN-

PRIN 2006069815) and Regione Puglia (BISIMANE project,

H31D08000030007).

DISCLOSURESAn Italian patent entitled “Method and kit for the diagnosis of IgA ne-

phropathy” (patent no. MI2010A002007) was issued on October 28, 2010, to

the University of Bari.

REFERENCES

1. Radford MG Jr, Donadio JV Jr, Bergstralh EJ, Grande JP: Predictingrenal outcome in IgA nephropathy. JAmSocNephrol 8: 199–207, 1997

2. Schena FP, Coppo R: IgA nephropathies. In: Oxford Textbook of

Clinical Nephrology, 3rd ed., edited by Davison AM, Cameron JS,Grünfeld JP, Ponticelli C, Ritz E, Winearls CG, van T Persele C, Oxford,Oxford University Press, 2005, pp 469–501

3. Conley ME, Cooper MD, Michael AF: Selective deposition of immu-noglobulin A1 in immunoglobulin A nephropathy, anaphylactoid pur-pura nephritis, and systemic lupus erythematosus. J Clin Invest 66:1432–1436, 1980

4. Barratt J, Feehally J: IgA nephropathy. J Am Soc Nephrol 16: 2088–2097, 2005

5. Coppo R, Amore A: Aberrant glycosylation in IgA nephropathy (IgAN).Kidney Int 65: 1544–1547, 2004

6. Novak J, Julian BA, Tomana M, Mesteck J: Progress in molecular andgenetic studies of IgA nephropathy. J Clin Immunol 21: 310–327, 2001

7. Kokubo T, Hiki Y, Iwase H, Tanaka A, Toma K, Hotta K, Kobayashi Y:Protective role of IgA1 glycans against IgA1 self-aggregation and adhe-sion to extracellular matrix proteins. J Am Soc Nephrol 9: 2048–2054,1998

8. Tomana M, Matousovic K, Julian BA, Radl J, Konecny K, Mestecky J:Galactose-deficient IgA1 in sera of IgA nephropathy patients is presentin complexes with IgG. Kidney Int 52: 509–516, 1997

9. Tomana M, Novak J, Julian BA, Matousovic K, Konecny K, Mestecky J:Circulating immune complexes in IgA nephropathy consist of IgA1 withgalactose-deficient hinge region and antiglycan antibodies. J Clin In-

vest 104: 73–81, 199910. Mestecky J, TomanaM, Crowley-Nowick PA, Moldoveanu Z, Julian BA,

Jackson S: Defective galactosylation and clearance of IgA1 moleculesas a possible etiopathogenic factor in IgA nephropathy. Contrib

Nephrol 104: 172–182, 199311. Suzuki H, Moldoveanu Z, Hall S, Brown R, Vu HL, Novak L, Julian BA,

Tomana M, Wyatt RJ, Edberg JC, Alarcón GS, Kimberly RP, Tomino Y,Mestecky J, Novak J: IgA1-secreting cell lines from patients with IgAnephropathy produce aberrantly glycosylated IgA1. J Clin Invest 118:629–639, 2008

12. Eijgenraam JW, van Kooten C: IgA1 glycosylation in IgA nephropathy:As sweet as it can be. Kidney Int 73: 1106–1108, 2008

13. Buck KS, Smith AC,Molyneux K, El-Barbary H, Feehally J, Barratt J: B-cellO-galactosyltransferase activity, and expressionofO-glycosylationgenesin bone marrow in IgA nephropathy. Kidney Int 73: 1128–1136, 2008

14. Qin W, Zhou Q, Yang LC, Li Z, Su BH, Luo H, Fan JM: Peripheral Blymphocyte beta1,3-galactosyltransferase and chaperone expressionin immunoglobulin A nephropathy. J Intern Med 258: 467–477, 2005

15. Beerman I, Novak J, Wyatt RJ, Julian BA, Gharavi AG: The genetics ofIgA nephropathy. Nat Clin Pract Nephrol 3: 325–338, 2007

16. Lai AS, Lai KN: Molecular basis of IgA nephropathy. Curr Mol Med 5:475–487, 2005

17. Cox SN, Sallustio F, Serino G, Pontrelli P, Verrienti R, Pesce F, TorresDD, Ancona N, Stifanelli P, Zaza G, Schena FP: Altered modulation ofWNT-beta-catenin and PI3K/Akt pathways in IgA nephropathy. KidneyInt 78: 396–407, 2010

18. Kudo T, Iwai T, Kubota T, Iwasaki H, Takayma Y, Hiruma T, Inaba N,Zhang Y, Gotoh M, Togayachi A, Narimatsu H: Molecular cloning andcharacterization of a novel UDP-Gal:GalNAc(alpha) peptide beta 1,3-galactosyltransferase (C1Gal-T2), an enzyme synthesizing a core 1structure of O-glycan. J Biol Chem 277: 47724–47731, 2002

19. Kato M, Arce L, Natarajan R: MicroRNAs and their role in progressivekidney diseases. Clin J Am Soc Nephrol 4: 1255–1266, 2009

20. Barratt J, Smith AC, Molyneux K, Feehally J: Immunopathogenesis ofIgAN. Semin Immunopathol 29: 427–443, 2007

21. Pirulli D, Crovella S, Ulivi S, Zadro C, Bertok S, Rendine S, Scolari F,Foramitti M, Ravani P, Roccatello D, Savoldi S, Cerullo G, Lanzilotta SG,Bisceglia L, Zelante L, Floege J, Alexopoulos E, Kirmizis D, GhiggeriGM, Frascà G, Schena FP, Amoroso A European IgAN Consortium:Genetic variant of C1GalT1 contributes to the susceptibility to IgAnephropathy. J Nephrol 22: 152–159, 2009

10 Journal of the American Society of Nephrology J Am Soc Nephrol 23: ccc–ccc, 2012

BASIC RESEARCH www.jasn.org

22. Li GS, ZhangH, Lv JC, Shen Y,WangHY: Variants of C1GALT1 gene areassociated with the genetic susceptibility to IgA nephropathy. KidneyInt 71: 448–453, 2007

23. Ju T, Cummings RD: A unique molecular chaperone Cosmc requiredfor activity of the mammalian core 1 b 3-galactosyltransferase. ProcNatl Acad Sci USA 99: 16613–16618, 2002

24. Yamada K, Kobayashi N, Ikeda T, Suzuki Y, Tsuge T, Horikoshi S,Emancipator SN, Tomino Y: Down-regulation of core 1 beta1,3-galactosyltransferase and Cosmc by Th2 cytokine alters O-glycosylationof IgA1. Nephrol Dial Transplant 25: 3890–3897, 2010

25. Julian BA, Waldo FB, Rifai A, Mestecky J: IgA nephropathy, the mostcommon glomerulonephritis worldwide. A neglected disease in theUnited States? Am J Med 84: 129–132, 1988

26. Julian BA, Novak J: IgA nephropathy: An update. Curr Opin NephrolHypertens 13: 171–179, 2004

27. Allen AC, Topham PS, Harper SJ, Feehally J: Leucocyte beta 1,3 galac-tosyltransferase activity in IgA nephropathy.Nephrol Dial Transplant 12:701–706, 1997

28. AnglicheauD, Sharma VK, Ding R, Hummel A, Snopkowski C, DadhaniaD, Seshan SV, SuthanthiranM:MicroRNAexpression profiles predictiveof human renal allograft status. Proc Natl Acad Sci USA 106: 5330–5335, 2009

29. Te JL, Dozmorov IM, Guthridge JM, Nguyen KL, Cavett JW, Kelly JA,Bruner GR, Harley JB, Ojwang JO: Identification of unique microRNAsignature associated with lupus nephritis. PLoS ONE 5: e10344, 2010

30. Godwin JG, Ge X, Stephan K, Jurisch A, Tullius SG, Iacomini J: Iden-tification of a microRNA signature of renal ischemia reperfusion injury.Proc Natl Acad Sci USA 107: 14339–14344, 2010

31. Wang G, Kwan BC, Lai FM, Choi PC, Chow KM, Li PK, Szeto CC: In-trarenal expression of microRNAs in patients with IgA nephropathy.Lab Invest 90: 98–103, 2010

32. Allen AC, Harper SJ, Feehally J: Galactosylation of N- and O-linkedcarbohydrate moieties of IgA1 and IgG in IgA nephropathy. Clin ExpImmunol 100: 470–474, 1995

33. Allen AC, Bailey EM, Brenchley PE, Buck KS, Barratt J, Feehally J:Mesangial IgA1 in IgA nephropathy exhibits aberrantO-glycosylation:observations in three patients. Kidney Int 60: 969–973, 2001

34. Hiki Y,OdaniH, TakahashiM,YasudaY,NishimotoA, IwaseH, ShinzatoT,Kobayashi Y,Maeda K:Mass spectrometry proves under-O-glycosylationof glomerular IgA1 in IgA nephropathy. Kidney Int 59: 1077–1085,2001

35. Roccatello D, Picciotto G, Torchio M, Ropolo R, Ferro M, FranceschiniR, Quattrocchio G, Cacace G, Coppo R, Sena LM, et al: Removal sys-tems of immunoglobulin A and immunoglobulin A containing com-plexes in IgA nephropathy and cirrhosis patients. The role ofasialoglycoprotein receptors. Lab Invest 69: 714–723, 1993

36. Novak J, TomanaM,Matousovic K, Brown R, Hall S, Novak L, Julian BA,Wyatt RJ, Mestecky J: IgA1-containing immune complexes in IgA ne-phropathy differentially affect proliferation of mesangial cells. KidneyInt 67: 504–513, 2005

37. Moura IC, Arcos-Fajardo M, Sadaka C, Leroy V, BenhamouM, Novak J,Vrtovsnik F, Haddad E, Chintalacharuvu KR, Monteiro RC: Glycosylationand size of IgA1 are essential for interaction with mesangial transferrinreceptor in IgA nephropathy. J Am Soc Nephrol 15: 622–634, 2004

38. Moura IC, Arcos-Fajardo M, Gdoura A, Leroy V, Sadaka C, Mahlaoui N,Lepelletier Y, Vrtovsnik F, Haddad E, Benhamou M, Monteiro RC: En-gagement of transferrin receptor by polymeric IgA1: Evidence for apositive feedback loop involving increased receptor expression andmesangial cell proliferation in IgA nephropathy. J Am Soc Nephrol 16:2667–2676, 2005

39. Peppard JV, Kaetzel CS, Russell MW: Phylogeny and comparativephysiology of IgA. In: Mucosal Immunology, 3rd ed., edited byMesteckyJ, Bienenstock J, Lamm,ME, Mayer L, McGhee JR, Strober W, AcademicPress, Burlington, MA, 2005, pp 195–210

40. Mestecky J, Moro I, Kerr MA, Woof JM: Mucosal immunoglobulins. In:Mucosal Immunology, 3rd ed., edited by Mestecky J, Bienenstock J,Lamm,ME,Mayer L,McGhee JR, StroberW,Academic Press, Burlington,MA, 2005, pp 153–181

41. Manno C, Strippoli GF, D’Altri C, Torres D, Rossini M, Schena FP: Anovel simpler histological classification for renal survival in IgA ne-phropathy: A retrospective study. Am J Kidney Dis 49: 763–775, 2007

42. Storey JD, Tibshirani R: Statistical significance for genomewide studies.Proc Natl Acad Sci USA 100: 9440–9445, 2003

43. Griffiths-Jones S, Grocock RJ, van Dongen S, Bateman A, Enright AJ:miRBase: MicroRNA sequences, targets and gene nomenclature.Nucleic Acids Res 34[Database issue]: D140–D144, 2006

44. Friedman RC, Farh KK, Burge CB, Bartel DP: Most mammalian mRNAsare conserved targets of microRNAs. Genome Res 19: 92–105, 2009

45. Krek A, Grün D, Poy MN, Wolf R, Rosenberg L, Epstein EJ,MacMenamin P, da Piedade I, Gunsalus KC, Stoffel M, Rajewsky N:Combinatorial microRNA target predictions. Nat Genet 37: 495–500,2005

46. Barenboim M, Zoltick BJ, Guo Y, Weinberger DR: MicroSNiPer: A webtool for prediction of SNP effects on putative microRNA targets. HumMutat 31: 1223–1232, 2010

47. Steinitz M, Klein G: EBV-transformation of surface IgA-positive humanlymphocytes. J Immunol 125: 194–196, 1980

48. Moore JS, Kulhavy R, Tomana M, Moldoveanu Z, Suzuki H, Brown R,Hall S, Kilian M, Poulsen K, Mestecky J, Julian BA, Novak J: Reactivitiesof N-acetylgalactosamine-specific lectins with human IgA1 proteins.Mol Immunol 44: 2598–2604, 2007

49. Moldoveanu Z, Wyatt RJ, Lee JY, Tomana M, Julian BA, Mestecky J,HuangWQ, Anreddy SR, Hall S, HastingsMC, Lau KK, CookWJ, NovakJ: Patients with IgA nephropathy have increased serum galactose-deficient IgA1 levels. Kidney Int 71: 1148–1154, 2007

This article contains supplemental material online at http://jasn.asnjournals.org/lookup/suppl/doi:10.1681/ASN.2011060567/-/DCSupplemental.

J Am Soc Nephrol 23: ccc–ccc, 2012 miR-148b in IgA Nephropathy 11

www.jasn.org BASIC RESEARCH