additive effects of micrornas and transcription factors on

Agné Kulyté,1 Yasmina Belarbi,1 Silvia Lorente-Cebrián,1 Clara Bambace,1 Erik Arner,1,2 Carsten O. Daub,3 Per Hedén,4

Mikael Rydén,1 Niklas Mejhert,1 and Peter Arner1

Additive Effects of MicroRNAsand Transcription Factors onCCL2 Production in HumanWhite Adipose Tissue

Adipose tissue inflammation is present in insulin-resistant conditions. We recently proposeda network of microRNAs (miRNAs) and transcriptionfactors (TFs) regulating the production of theproinflammatory chemokine (C-C motif) ligand-2(CCL2) in adipose tissue. We presently extended andfurther validated this network and investigated if thecircuits controlling CCL2 can interact in humanadipocytes and macrophages. The updatedsubnetwork predicted that miR-126/-193b/-92acontrol CCL2 production by several TFs, includingv-ets erythroblastosis virus E26 oncogene homolog 1(avian) (ETS1), MYC-associated factor X (MAX),and specificity protein 12 (SP1). This was confirmedin human adipocytes by the observation that genesilencing of ETS1, MAX, or SP1 attenuated CCL2production. Combined gene silencing of ETS1 andMAX resulted in an additive reduction in CCL2production. Moreover, overexpression of miR-126/-193b/-92a in different pairwise combinationsreduced CCL2 secretion more efficiently than eithermiRNA alone. However, although effects on CCL2secretion by co-overexpression of miR-92a/-193band miR-92a/-126 were additive in adipocytes, thecombination of miR-126/-193b was primarily additivein macrophages. Signals for miR-92a and -193b

converged on the nuclear factor-kB pathway. Inconclusion, TF and miRNA-mediated regulation ofCCL2 production is additive and partly relayed bycell-specific networks in human adipose tissue thatmay be important for the development of insulinresistance/type 2 diabetes.Diabetes 2014;63:1248–1258 | DOI: 10.2337/db13-0702

White adipose tissue (WAT) function plays an importantrole in the development of insulin resistance/type 2 di-abetes. Fat cells present in WAT secrete a number ofmolecules, collectively termed adipokines, which affectinsulin sensitivity by autocrine and/or paracrine mecha-nisms (1,2). In insulin-resistant obese subjects, WATdisplays a chronic low-grade inflammation, which isproposed to alter insulin signaling by increased fatty acidrelease and altered adipokine production (3).

Chemokine (C-C motif) ligand-2 (CCL2), also referredas monocyte chemoattractant protein-1, is an earlyproinflammatory signal in WAT that acts by promotingadipose infiltration of macrophages (4,5). After diet-induced obesity, Ccl2-/- mice are less insulin-resistant anddisplay a less pronounced inflammatory reaction in WAT(6). Adipocytes and macrophages both contribute sig-nificantly to CCL2 secretion from WAT.

1Department of Medicine, Huddinge, Lipid Laboratory, Karolinska Institutet,Stockholm, Sweden2RIKEN Center for Life Science Technologies, Division of Genomic Technologies,Yokohama, Japan3Department of Biosciences and Nutrition, Huddinge, Karolinska Institutet,Stockholm, Sweden4Akademikliniken, Stockholm, Sweden

Corresponding authors: Peter Arner, [email protected], and Agné Kulyté, [email protected].

Received 3 May 2013 and accepted 15 December 2013.

This article contains Supplementary Data online at http://diabetes.diabetesjournals.org/lookup/suppl/doi:10.2337/db13-0702/-/DC1.

© 2014 by the American Diabetes Association. See http://creativecommons.org/licenses/by-nc-nd/3.0/ for details.

1248 Diabetes Volume 63, April 2014

METABOLISM

Dow

nloaded from http://diabetesjournals.org/diabetes/article-pdf/63/4/1248/411822/1248.pdf by guest on 24 February 2022

http://crossmark.crossref.org/dialog/?doi=10.2337/db13-0702&domain=pdf&date_stamp=2014-03-07

mailto:[email protected]

mailto:agne<?show $6#?>[email protected]

mailto:agne<?show $6#?>[email protected]

http://diabetes.diabetesjournals.org/lookup/suppl/doi:10.2337/db13-0702/-/DC1


http://creativecommons.org/licenses/by-nc-nd/3.0/

http://creativecommons.org/licenses/by-nc-nd/3.0/

Little is known about the mechanisms regulating WATinflammation. Results in recent years have demonstratedthat microRNAs (miRNAs) play important roles in con-trolling inflammation in other conditions and non-adipose tissues, as reviewed (7). In mammals, miRNAsact by inhibiting the expression of target genes afterbinding to defined complementary sites in target 39-untranslated regions (UTRs) (8).

Cellular function is controlled by transcriptional(transcription factors [TFs]) and posttranscriptionalmechanisms (miRNAs). Using global expression profilingdata and mathematical models, the transcriptome can bemapped into transcriptional regulatory networks (TRNs),with regulatory elements (e.g., TFs and miRNAs) asnodes and their interactions as edges (9,10). We recentlydescribed and partly experimentally verified a miRNA-TFnetwork in the subcutaneous WAT of insulin-resistantobese women that controls the expression of CCL2 (11).This TRN suggests that several miRNAs and TFs interactin different combinations to regulate CCL2 production.Although the reason for this complex regulation and re-dundancy is not entirely clear, it may allow for more exactfine-tuning of gene expression and/or signal enhancementin response to specific environmental cues. In line withthis, recent studies in cancer cells have proposed thatpartly overlapping regulatory circuits allow signal ampli-fication in an additive or synergistic manner (12,13). Tothe best of our knowledge, it is not known whether thesetypes of mechanisms are present in WAT or whether theydiffer between specific cell types within the tissue.

Through combinatorial gain-of-function/loss-of-function studies focusing on TFs and miRNAs, weinvestigated how transcriptional regulators (TFs andmiRNAs) can integrate effects on the expression ofinflammatory factors in WAT. First, we extended andfurther validated the proposed TRN (11). We sub-sequently silenced potentially important TFs in the net-work, alone or in combination, and investigated how thisaltered CCL2 production in human adipocytes. Finally,we determined whether concomitant overexpression ofmiRNAs in human adipocytes or macrophages resulted inincreased effects when compared with overexpressingeach miRNA alone.

RESEARCH DESIGN AND METHODS

Clinical Material

Cohort 1 comprised 30 obese (BMI .30 kg/m2) and26 nonobese (BMI ,30 kg/m2) women who had nochronic disease and were free of continuous medication.This cohort has been described in detail previously (11).The subjects came to the laboratory in the morning afteran overnight fast. Height, weight, and waist circumfer-ence were determined. An abdominal subcutaneous WATbiopsy specimen (;1.0–1.5 g) was obtained by needleaspiration as described (14). One part (300 mg) of thetissue was used for measurement of CCL2 release andexpressed per number of fat cells as described (15).

Another part of the tissue (at least 500 mg) was sub-jected to collagenase treatment, and mean adipocytevolume and weight were determined as described (16).Packed fat cells (200 mL) and intact WAT (400 mg) werefrozen at 270°C for future mRNA and miRNA mea-surements as described (11). The Karolinska UniversityHospital Ethical Committee (Stockholm, Sweden) ap-proved this study. All subjects were informed in detailabout the studies, and written informed consent wasobtained.

For in vitro studies, subcutaneous WAT was obtainedfrom healthy men and women undergoing cosmetic li-posuction. This experimental group was not selected forage, sex, or BMI. Isolation of human adipocyte progenitorcells from subcutaneous WAT was performed as de-scribed (17). The progenitor cells obtained from separateindividuals were not mixed.

Affymetrix GeneChip Human Gene 1.0 ST and miRNAArray Protocols

Data obtained from gene and miRNA arrays from humansubcutaneous WAT (Gene Expression Omnibus numberGSE25402) and the protocols have been described (11).

Cell Culture

Culture and in vitro differentiation of human adipocyteprogenitor cells were performed as described (17). 3T3-L1and THP1 cells were handled as recommended in the pro-tocols from American Type Culture Collection (Manassas,VA). For the induction of monocyte–macrophage differen-tiation, THP1 cells were stimulated with 50 ng/mL phorbol12-myristate 13-acetate (PMA).

miRNA and Small Interfering RNA Transfection

In vitro differentiated adipocytes (day 10–12 post-induction) were treated for 24–48 h with various con-centrations (5–40 nmol/L) of miRIDIAN miRNA mimicsfor overexpression of miRNA activity or 60 nmol/L ofmiRIDIAN miRNA Hairpin Inhibitors for native miRNAinhibition (Thermo Fisher Scientific, Lafayette, CO) andHiPerFect Transfection Reagent (Qiagen, Hilden,Germany), according to the manufacturers’ protocols.Optimal transfection conditions were determined inseparate titration experiments. Transfection efficiency wasassessed with real-time quantitative PCR (qPCR) usingmiRNA probes/assays (Applied Biosystems, Foster City,CA; and Qiagen, respectively). To rule out unspecificeffects, control cells were transfected with miRIDIANmiRNA Mimic or Hairpin Inhibitor Non-Targeting Nega-tive Controls (Thermo Fisher Scientific).

In some experiments, in vitro differentiated adipo-cytes were cotransfected with miScript Target Protectorfor the miR-92a binding site on specificity protein 12(SP1; 500 nmol/L; Qiagen) and miR-92a mimics (40nmol/L) or appropriate control reagents (miScript TargetProtector Negative Control; Qiagen) and miRIDIANmiRNA Mimic Non-Targeting Negative Control (ThermoFisher Scientific). THP1 cells were transfected with

diabetes.diabetesjournals.org Kulyté and Associates 1249

Dow


miRNA mimics, as described above, for adipocytes 48 hafter incubation with PMA.

For RNA interference experiments, adipocytes weretreated with various concentrations (10–40 nmol/L), asdescribed above, using ON-TARGETplus SMARTpoolsmall interfering RNA (siRNA) for v-ets erythroblastosisvirus E26 oncogene homolog 1 (avian) (ETS1), SP1, orMYC-associated factor X (MAX) and appropriate con-centrations of siRNA Non-Targeting Negative Control(Thermo Fisher Scientific) instead of miRNA reagents.

RNA Isolation, cDNA Synthesis, and Real-Time PCR

Total RNA was extracted from in vitro differentiatedadipocytes, 3T3-L1 cells, and THP1 macrophages, andpurity control was performed as described (11). Syn-thesis of coding-gene cDNA and real-time qPCR usingTaqman probes (Applied Biosystems) was performed asdescribed (11). Synthesis of miRNA cDNA and followingreal-time qPCR was performed using Applied Bio-systems reagents as described (11) or using the miScriptII RT Kit and miScript Primer Assays (Qiagen) accordingto the manufacturer’s instructions. Relative gene ex-pression was calculated using the comparative Ctmethod (i.e., 2DCt-target gene/2DCt-reference gene) with 18S asthe internal control. Expression of miRNA was normalizedto a reference gene RNU48 or SNORD68. Levels of 18S,RNU48, and SNORD68 did not differ between groups.

Western Blot

Cells were lysed in radioimmunoprecipitation assaybuffer as described (18). SDS-PAGE was used to separate15–20 mg total protein, and Western blot wasperformed according to standard protocols. Themembranes were blocked in 3% enhanced chemilu-minescence Advance Blocking Agent (GE Healthcare,Buckinghamshire, U.K.). Primary antibodies againstETS1 were from Novocastra Leica Biosystems AB (Wet-zlar, Germany) and against SP1 were from Santa CruzBiotechnology (Santa Cruz, CA). b-Actin (Sigma-Aldrich,St. Louis, MO) was used as a loading control. Secondaryantibodies mouse/rabbit IgG-horseradish peroxidase werefrom Sigma-Aldrich. All tested primary antibodies detectingMAX were unspecific and results therefore could not bedemonstrated. Antibody-antigen complexes were detectedby chemiluminescence using ECL Select Western BlottingDetection Kit (GE Healthcare).

ELISA

CCL2 levels in conditioned media from in vitro differ-entiated adipocytes and macrophages were analyzed usingan ELISA assay from R&D Systems (Minneapolis, MN).

Luciferase Reporter Assay

Empty luciferase reporter vector and vector containinga part of 39UTR of SP1 (enclosing predicted binding sitefor hsa-miR-92a) were obtained from GeneCopoeia(Rockville, MD). The luciferase reporter assay in 3T3-L1cells was performed as described (11).

Global Gene Expression Analysis, Motif ActivityResponse Analysis, and Network Construction

Motif activity response analysis and construction of theadipocyte CCL2 TRN used in this study has been de-scribed in detail (11,19). Predicted targets of miRNAswere updated according to the latest release of TargetScan(March 2013).

Statistical Analyses

Data presented are mean 6 SEM. When appropriate, thedata were log-transformed to become normally distrib-uted. Results were analyzed with unpaired t test, linearregression (simple and multiple), or ANOVA (repeated-measures).

RESULTS

Strategy for Investigating Additive Effects of TFs andmiRNAs

To dissect the regulation of adipocyte CCL2 pro-duction by miRNAs and TFs, we extended and furthervalidated a recently described TRN present in humanfat cells and perturbed in obese insulin-resistant sub-jects (11). In brief, this in silico–generated TRN linksfor several miRNAs (miR-92a/-126/-193b/-652 andlet-7a), directly or indirectly, through one or two in-termediate TF steps, to CCL2 production. For thecurrent study, we focused on a subpart of the TRNcomprising two partly validated circuits, from miR-126and -193b to CCL2, and a third possible candidate,miR-92a, which was not fully characterized in ourprevious study (Fig. 1A).

miR-92a Regulates CCL2 Production by SP1 inAdipocytes

Because the regulatory edges linking miR-92a to CCL2have not been explored in detail, we first tested if miR-92a affected its predicted target genes SP1 and ETS1 inhuman adipocytes. We overexpressed or inhibited miR-92a using miRNA mimics/inhibitors and evaluated ex-pression of SP1 as well as mRNA expression of ETS1.Modulation of miR-92a expression levels affected themRNA and protein of SP1, whereas ETS1 expressionremained unaltered (Fig. 1B–D and Table 1), suggestingthat SP1, but not ETS1, is a direct target of miR-92a.The levels of CCL2 in conditioned media mirrored thechanges in SP1 mRNA/protein expression (Fig. 1B andD). Luciferase reporter assays confirmed that miR-92ainteracted directly with the 39UTR region of SP1(Fig. 1E).

To verify that SP1 could regulate CCL2 productionin human adipocytes independently of miR-92a, theexpression of SP1 was silenced using siRNA at twodifferent concentrations (10 and 20 nmol/L). SP1 geneknockdown significantly attenuated CCL2 secretion(Fig. 2A) and reduced SP1 mRNA/protein levels (Fig.2B and C). In addition, SP1 silencing affected mRNAexpression of two predicted direct targets in the TRN:

1250 Additive Regulation of CCL2 by miRNAs and TFs Diabetes Volume 63, April 2014

Dow


Figure 1—miR-92a regulates SP1 and CCL2 in human adipocytes. A: The proposed detailed CCL2 regulatory subnetwork in humanadipocytes depicting miR-193b, -126, and -92a (□) as well as motifs and cognate TFs (○) altered by obesity and targeting CCL2 (◇).T-bars represent inhibition; arrows represent stimulation. Bold edges represent physical interactions between nodes described pre-viously in adipocytes, and thin lines represent predictions by network analyses and/or interactions in other cell types. Under the NF-kBcomplex, the following TFs are merged: NF-kB1, RELA (v-rel reticuloendotheliosis viral oncogene homolog A [avian]), RELB, and REL.B: Predicted interaction between miR-92a and SP1 was investigated by overexpressing miR-92a and assessing the effects on SP1mRNA expression in human in vitro differentiated adipocytes. Changes of CCL2 secretion in the same experiments are shown in


Dow


CCL2 and v-rel reticuloendotheliosis viral oncogenehomolog [avian] (REL), a part of the nuclear factor(NF)-kB network (Table 1). To further investigate therole of miR-92a in CCL2 production by SP1, we con-comitantly inhibited expression of miR-92a and SP1.The combination of miR-92a inhibition and silencingof SP1 abolished their individual effects on CCL2production (Fig. 2D). Attenuation of miR-92a levelsincreased CCL2 secretion to the same extent as pre-viously shown in Fig. 1D. Moreover, simultaneoustransfection of miR-92a mimics and a target protector(corresponding to the miR-92a target sequence on SP139UTR) abolished the effect of miR-92a on CCL2 pro-duction (Fig. 2E).

Previous validations of the interactions in the CCL2TRN identified that miR-126 targeted CCL2 directly,whereas miR-193b affected its production indirectlythrough several TFs, including ETS1 and MAX (11). Inthe current study, we identified two additional TFs aspossible targets for miR-126 and miR-193b, REL andSP1, respectively (Fig. 1A). However, overexpression ofmiR-126 or miR-193b did not alter REL and SP1 mRNA

levels, suggesting that they are not true targets for thesemiRNAs in human adipocytes (Table 1).

ETS1 and MAX Regulate CCL2 Production in HumanAdipocytes by TRN in an Additive Fashion

According to Fig. 1A, ETS1 and MAX are entry TFs formiR-193b; moreover, they have a predicted interactionand have been validated as direct targets of miR-193b(11). To determine whether ETS1 and MAX could affectCCL2 secretion independently of miR-193b, we knockeddown ETS1 and MAX in adipocytes using various concen-trations of siRNAs, followed by measurements of mRNAlevels of their first predicted neighbors in the TRN (Fig.1A). We could confirm that silencing of the ETS1 genedecreased the mRNA levels of MAX, signal transducer andactivator of transcription 6, interleukin-4 induced (STAT6),nuclear factor of k light polypeptide gene enhancer inB-cells 1 (NFKB1), and CCL2, whereas silencing of MAXdecreased the mRNA levels of v-rel reticuloendotheliosisviral oncogene homolog B (RELB) (Table 1).

Silencing of ETS1 and MAX with 10 nmol/L of siRNAhad modest effects on CCL2 secretion, but higher con-centrations (i.e., 20 nmol/L) resulted in significantly

parallel (n = 8). C: Representative Western blots of SP1 protein after overexpression and inhibition of miR-92a. D: Quantification of SP1protein levels after overexpression and inhibition of miR-92a. Changes of CCL2 secretion are shown in parallel (n = 6–9). E: miR-92a wasoverexpressed together with SP1 39UTR luciferase reporter in 3T3-L1 cells, and luciferase activity was measured (n = 8). NegC, negativecontrol. ***P < 0.001; **P < 0.01; *P < 0.05.

Table 1—Interactions between miRNAs or TFs and first predicted neighbors

miRNA/TFsiRNA

Predicted firstneighbor

Fold changetreatment/NegC

t testP value Reference

miR-92a SP1 0.78 0.031 This study (Fig. 1B)ETS1 1.09 0.3108 This study

miR-193b ETS1 0.54 0.0001 (11)MAX 0.58 0.0001 (11)SP1 1.04 0.681 This study

miR-126 CCL2 0.67 0.0112 (11)REL 1.16 0.3714 This study

siSP1 CCL2 0.76 0.007 This studyREL 0.77 0.0365 This study and (28) (not adipocytes)NFKB1 0.85 0.3563 This studyETS2 NE — This study and (29) (not adipocytes)

siETS1 MAX 0.82 0.0178 This studyNFKB1 0.80 0.0242 This study and (27) (not adipocytes)STAT6 0.67 0.0023 This studyCCL2 0.56 0.0055 This study and (23) (not adipocytes)

siMAX RELB 0.84 0.0106 This study

NE, not evaluated due to low expression. Predicted interactions between miRNAs and first neighbors were investigated by over-expression of each miRNA and assessing mRNA expression of TFs in human in vitro differentiated adipocytes. Predicted TF inter-actions with first neighbors were investigated by transfecting human in vitro differentiated adipocytes with siRNA and assessing theeffects on mRNA expression of first neighbors. NegC, negative control.


Dow


decreased CCL2 production (Fig. 3A). Because CCL2protein levels were affected by both TFs, andMAX mRNAexpression was affected by silencing of ETS1, we assessedwhether ETS1 and MAX could interact with each other incontrolling CCL2 secretion. Indeed, concomitant silenc-ing of ETS1 and MAX using low concentrations (10nmol/L) of siRNA resulted in a more pronounced re-duction of CCL2 secretion (10+10 nmol/L of each siRNA)compared with single knockdown of either TF (Fig. 3A).Expression of CCL2 mRNA followed the same trend asCCL2 secretion (Supplementary Fig. 1A).

Knockdown efficiency was determined by quantifyingETS1 and MAX mRNA levels (Fig. 3B and C). There wasa marked decrease in mRNA levels of ETS1 and MAX

when silenced alone at 10/20 nmol/L or in combination(10+10 nmol/L). ETS1-siRNA treatment reduced proteinlevels of ETS1 (Fig. 3D). Owing to the poor specificity oftested MAX antibodies, we were not able to determineprotein levels of this TF.

As mentioned above, we observed that 20 nmol/LsiRNA (ETS1 or MAX) caused a more pronounced de-crease in CCL2 production than 10 nmol/L. Un-fortunately, we did not detect significant quantitativedifferences in the mRNA levels of ETS1 and MAX be-tween 10 and 20 nmol/L of siRNA at a given time point.To control for off-target effects, we treated in vitro adi-pocytes with transfection agent alone or transfected withvarious concentrations of siRNA nontargeting negative

Figure 2—SP1 regulates CCL2 in human adipocytes. A: TF SP1 was silenced with siRNA at 10 and 20 nmol/L final concentrations in invitro differentiated adipocytes, and secreted levels of CCL2 were assessed by ELISA (n = 12–18). B: TF SP1 was silenced with siRNA at 10and 20 nmol/L final concentration in in vitro differentiated adipocytes, and mRNA levels of SP1 were assessed (n = 6–9). C: Protein levelsof SP1 were measured after silencing of SP1 (n = 6). Representative blot is shown. D: Expression of SP1 and levels of miR-92a wereattenuated using siRNA and inhibitor of miRNA-92 individually or in pair with each other, and CCL2 secretion was assessed using ELISA(n = 6–14). E: Interaction of miR-92a-SP1 was studied by cotransfecting human in vitro differentiated adipocytes with miR-92a mimics andtarget protector (corresponds to the binding site of miR-92a on the 39UTR of SP1) after evaluation of CCL2 secretion by ELISA (n = 7). Torule out unspecific effects, cells in parallel were transfected with appropriate negative control (nontargeting negative control for miRNAmimics or inhibitor, nontargeting negative control for siRNA, or negative control for target protector). Levels of mRNA were calculatedusing a comparative Ct-method (i.e., 2DCt-target gene/2DCt-reference gene) with 18S as reference gene. SP1 protein levels were evaluated byWestern blotting. Levels of b-actin were used to control loading. Data were analyzed using the t test and are presented as fold change 6SEM relative to the negative control. NegC, negative control. ***P < 0.001; **P < 0.01; *P < 0.05.


Dow



control, followed by mRNA measurements. Indeed, basalexpression levels of CCL2, ETS1, or MAX remained stable(Supplementary Fig. 2).

Taken together, experiments in Figs. 1–3 demonstratethat miR-193b and -92a attenuate CCL2 production inhuman adipocytes independently by TFs nodes com-prising ETS1/MAX and SP1, respectively.

miRNA-193b, -126, and -92a Cooperatively AffectCCL2 Production in Human Adipocytes

Because our data presented above indicate that TFs areable to additively control CCL2 secretion, we assessed ifthis mode of action is also present when combiningmiRNAs. We have previously shown that single

overexpression of miR-193b, -126, or -92a at high con-centrations (i.e., 40 nmol/L) downregulate CCL2 pro-duction to a substantial degree (30–50%) (11). Here weassessed how co-overexpression of miRNAs affectedCCL2 secretion. A caveat is that overexpression of mul-tiple miRNA mimics, particularly at high concentrations,increases the risk for off-target effects. Therefore, lowerconcentrations of miRNA mimics were used in thecombined transfections. At 5 nmol/L, single over-expression of miR-193b, -126, and -92a had significantlyweaker effects on CCL2 secretion compared with con-comitant overexpression of miR-92a with miR-193b or-126 (Fig. 4A). However, combining miR-193b and -126did not cause a more pronounced effect than miR-126

Figure 3—ETS1 and MAX regulate CCL2 production in human adipocytes. A: TFs ETS1 and MAX were silenced with siRNA alone at 10 or20 nmol/L final concentration in in vitro differentiated adipocytes, and secreted levels of CCL2 were assessed by ELISA. ETS1 and MAXwere also cosilenced in pair with each other (10+10 nmol/L each miRNA), and secreted levels of CCL2 were assessed (n = 11–17). B: ETS1and MAX were silenced alone (10 and 20 nmol/L) or cosilenced in pair with each other (10+10 nmol/L each miRNA), and mRNA of ETS1was assessed (n = 11–17). C: ETS1 and MAX were silenced alone (10 and 20 nmol/L) or cosilenced in pair with each other (10+10 nmol/Leach miRNA), and mRNA of MAX was assessed (n = 11–17). D: Protein levels of ETS1 were measured after silencing (n = 6). Repre-sentative blot is shown. To rule out unspecific effects, cells in parallel were transfected with siRNA nontargeting negative control. Levels ofmRNA were calculated using a comparative Ct-method (i.e., 2DCt-target gene/2DCt-reference gene) with 18S as the reference gene. ETS1 proteinlevels were evaluated by Western blotting. Levels of b-actin were used to control loading. Data were analyzed using the t test and arepresented as fold change 6 SEM relative to the negative control. NegC, negative control. ***P < 0.001; **P < 0.01; *P < 0.05.


Dow



alone. The effects on CCL2 secretion were paralleled bysimilar changes in mRNA expression (SupplementaryFig. 1B).

To rule out the possibility that the additive effects onCCL2 production by combined overexpression of miRNAswere not due to alterations in transfection efficiency, wequantified miRNA levels after transfection. Indeed, therewere no differences in miRNA abundance irrespective ofwhether they were overexpressed alone or in pairs(Supplementary Fig. 3).

As suggested in Fig. 1A, the signaling circuits for miR-193b and -92a converge on TFs in the NF-kB pathway.We therefore quantified mRNA levels of NFKB1 and

RELB in the miRNA co-overexpression experiments. Al-though no changes occurred in the expression levels ofNFKB1 (Fig. 4B), RELB mRNA expression was attenuatedby the combined overexpression of miR-92a with miR-193b or miR-126 (Fig. 4C).

Cooperative Effects of miRNA 193b, -126, and -92a onCCL2 Production in Human Macrophages

That CCL2 is secreted by several other cell types presentwithin WAT, among which macrophages are probably themost important (20), is well known. Therefore, weassessed the effects of miR-193b, -126, and -92a on CCL2secretion in the THP1 human monocyte/macrophage cell

Figure 4—miRNAs regulate CCL2 in an additive manner in human adipose tissue. A: miR-193b, -126, and -92a were overexpressed aloneat 5 nmol/L final concentration or co-overexpressed in pairs (5+5 nmol/L of each miRNA) in in vitro differentiated adipocytes, and secretedlevels of CCL2 were assessed by ELISA (n = 16–24). B: miR-193b, -126, and -92a were overexpressed alone at 5 nmol/L final con-centration or co-overexpressed in pairs (5+5 nmol/L each miRNA) in in vitro differentiated adipocytes, and mRNA levels of NFKB1 wereassessed (n = 14–19). C: miR-193b, -126, and -92a were overexpressed alone at 5 or 10 nmol/L final concentration or co-overexpressed inpairs (5+5 nmol/L of each miRNA) in in vitro differentiated adipocytes, and mRNA levels of RELB were assessed (n = 14–19). D: miR-193b,-126, and -92a were overexpressed alone at 10 or 20 nmol/L final concentration or co-overexpressed in pairs (10+10 nmol/L each miRNA)in in vitro differentiated THP1 cells, and secreted levels of CCL2 were assessed by ELISA (n = 9). To rule out unspecific effects, cells inparallel were transfected with an equal concentration of nontargeting negative control for miRNA mimics. Levels of mRNA were calculatedusing a comparative Ct-method (i.e., 2DCt-target gene/2DCt-reference gene) with 18S as the reference gene. Results were analyzed using thet test and are presented as fold change 6 SEM relative to the negative control. NegC, negative control. ***P < 0.001; **P < 0.01; *P <0.05.


Dow





line. We previously used high concentrations of mimicreagents (40 nmol/L) to show that miR-193b and -126affected CCL2 production in these cells (40 nmol/L)(11). To observe possible additive effects of miRNAson CCL2 production, we overexpressed miRNAs atlower concentrations (10 nmol/L). In this experimen-tal setting, overexpression of miR-92a alone signifi-cantly decreased CCL2 secretion, whereasoverexpressing miR-193b caused somewhat increased(albeit not statistically significant) CCL2 secretion.In contrast to the results obtained in adipocytes,co-overexpression miR-193b and -126 decreased thesecretion of CCL2 significantly, whereas the combina-tions of miR-193b/-92a and miR-126/-92a showed lessprominent effects (Fig. 4D).

Similar to the experiments in adipocytes, we quanti-fied the expression of miRNAs after transfection in THP1macrophages. There were no differences in the abundanceof miR-126 or -92a irrespective of whether they wereoverexpressed alone or in pairs (Supplementary Fig. 4).Owing technical problems of tested probes for miR-193b,we were not able to determine levels of this miRNA inTHP1 cells.

Association Between Expression of Adipose TissuemiR-193b, -126, and -92a and CCL2 Secretion

The possible in vivo relevance was assessed by corre-lating expression levels of miRNAs (miR-193b, -126,and -92a) with CCL2 secretion in adipose tissue ofcohort 1 (results in Table 2). In simple regression, onlymiR-193b correlated significantly (and negatively)with CCL2 secretion. However, when miRNA levelswere combined in a stepwise regression, miR-193b/-92a and miR-193b/-126 interacted significantly inthe negative correlation with CCL2. The combinationsexplained 21–24% of the interindividual variations inCCL2 secretion (adjusted r2). There was no significantcorrelation with CCL2 release when miR-92a and -126were combined as regressors (values not shown).miRNA-193b always entered as the first step in themodels.

As shown in Fig. 1A, miR-193b targets two TFs in theTRN: ETS1 and MAX. We applied regression analysis tofurther investigate if the activity of these TFs covaried intheir relationship with adipose CCL2 secretion (Table 2).In simple regression, the motif activity of either TF wassignificantly and positively correlated with adipose CCL2secretion (Table 2). In a stepwise regression, their motifactivities acted together in a significant manner withregard to their correlation with CCL2 secretion. Togetherthey explained 40% of the CCL2 variation (i.e., adjustedr2). ETS1 entered as the first step in the relationship. Wealso investigated by regression analysis if the activity ofSP1 (target of miR-92a) covaried with adipose CCL2 se-cretion. A simple regression showed the SP1 motif ac-tivity was significantly and positively correlated withadipose CCL2 secretion (r = 0.42, P = 0.003).

DISCUSSION

In this study, we tested whether combined silencing ofTFs (ETS1 and MAX) and overexpression of miRNAs(miR-193b/-126/-92a) resulted in augmented effects inCCL2 production. Our main conclusion is that the threeinvestigated miRNAs (miR-193b, -126, and -92a) regu-lated CCL2 production by distinct pathways involvingspecific TFs. Concomitant overexpression of miRNA pairsor paired silencing of their cognate target TFs (ETS1 andMAX) resulted in additive effects on CCL2 production.The enhanced effects of miRNAs pairs were qualitativelydifferent in human adipocytes compared with humanmacrophages. This suggests that TRNs regulating CCL2production are cell-specific and that the combined effectsof miRNAs influence cellular phenotypes.

Several biological processes, including diseases such ascancer, have been causatively associated with dis-turbances in the interplay between miRNAs and TF invitro and in vivo, as reviewed (21). MiRNAs and TFs aretransactivating factors that interact with cis-regulatoryelements, potentially generating complex combinatorialeffects. Here, we extensively characterized the firstdownstream targets of miR-193b and -92a in the pro-posed CCL2 TRN. By knocking down ETS1, MAX, andSP1, we demonstrate that these TFs downregulate CCL2production in fat cells independently of the targetingmiRNAs. This suggests that these three TFs are upstreamregulators of CCL2 in human adipose tissue. The resultswith ETS1 and SP1 are in accordance with findings innonadipose tissue demonstrating that both TFs control

Table 2—Correlation between CCL2 secretion, expressionof miRNAs in adipose tissue and activity of ETS1 and MAXdetermined by motif activity response analysis

Regressor

Single regression Stepwise regression

rvalue

Pvalue

Fvalue

Pvalue

miR-193b alone 20.39 0.007 — —

miR-126 alone 20.21 0.15 — —

miR-92a alone 20.28 0.053 — —

miR-193b + -92a — —

For miR-193b 10.0For miR-92a — — 5.8 0.002

miR-193b + -126 — —

For miR-193b 13.8For miR-126 — — 7.2 0.0009

ETS1 alone 0.57 ,0.0001 — —

MAX alone 0.41 0.004 — —

ETS1 + MAX — —

For ETS1 19.8For MAX — — 7.7 ,0.001

An F value .4 indicates that the regressors have significantlyentered the model. CCL2 secretion was expressed per numberof fat cells.


Dow



CCL2 (22,23). Although the MYC/MAX/MAD family arewell-established regulators of key processes in basic cellphysiology, including cell metabolism (24), their in-volvement in the regulation of CCL2 has not been de-scribed before. Our study therefore suggests a novel rolefor MAX in the regulation of adipose inflammation. Be-cause ETS1 and MAX were both direct targets for miR-193b, we explored the effects on CCL2 production ofcombined knockdown of these two genes. This resultedin more pronounced inhibition of CCL2 productioncompared with single knockdown of either TF. Thus,miR-193b may amplify its signal to CCL2 throughinteractions within the TF network.

To study the combined effects of miRNAs in humanadipocytes, we overexpressed miR-193b/-126/-92a in-dividually and in pairs. In fat cells, the combination ofmiR-92a with miR-193b or -126 caused a more markeddownregulation of CCL2 production compared with ei-ther miRNA alone. This demonstrates two types of signalamplifications, which are summarized in Fig. 5. For miR-92a and -193b, there are interactions through the TFnetwork that in an additive manner amplify the effectson CCL2 secretion (interaction A1 and A3). For miR-92a

and -126, the effects are most probably obtained due toa combination of the TF network (miR-92a acts throughSP1) and the direct interaction of miR-126 with CCL2(interaction A2).

We also investigated if combinations of miRNAs couldalter signaling further downstream in the TF network.Co-overexpression of miR-92a with miR-193b or miR-126downregulated RELBmRNA expression. This suggests thatin adipocytes, an additive effect of miRNAs on CCL2production may be due to signals converging onto the NF-kB/REL pathway. These results are in concordance withpreviously published data on other cell types demon-strating that NF-kB/REL are downstream of SP1, ETS1,and MYC (an interaction partner with MAX) (25–27).

To study cell-specific miRNA effects on CCL2 regula-tion, we performed similar experiments in the humanmacrophage THP1 cell line. In contrast to fat cells, onlythe combination of miR-193b and miR-126 had an ad-ditive effect on CCL2 production (Fig. 5; interaction M1).Whether the miRNA-TF regulatory circuits established inadipocytes is also acting in macrophages remains to bedetermined. Nevertheless, our data demonstrate thatmiRNAs regulate CCL2 differentially in human macro-phages compared with fat cells.

Are additive effects of miRNA signaling of clinicalrelevance? Unfortunately, determining this in vivo is notpossible. To shed some light on this relevant issue, weperformed extensive correlation analyses betweenmiRNA expression, TF activities, and CCL2 secretion inhuman subcutaneous adipose tissue. The combined ex-pression of miR-193b/-92a or miR-193b/-126 explainedinterindividual variations in CCL2 secretion to a largerextent than either miRNA alone. The same was true forthe combined expression of MAX and ETS1. This sup-ports the hypothesis that additive effects of miRNA andTFs may be clinically relevant although, admittedly,caution should be taken when extrapolating these sta-tistical correlations into an in vivo situation. It is alsoimportant to stress that the TRN constructed using thepresent approach is not complete and most likelyincludes other factors not evaluated in the current andprevious analysis (11).

Most of the work characterizing miRNAs associatedwith obesity and/or insulin resistance concerns singlemiRNAs rather than combinations or clusters of miRNAs.We propose the following model for how miRNAs affectCCL2 production in human adipose tissue, where the sig-nal may be relayed in at least three different ways (Fig. 5):1) as the combined effects of direct interactions with CCL2and indirect interactions with specific TFs, 2) as the in-teraction of several miRNAs in a TF network, and 3) as theinteraction of a single miRNA with several TFs. Thesecombined effects are specific for different cell types withinadipose tissue (e.g., fat cells and macrophages). In addition,there are probably other signaling pathways that could beimportant. For example, the expression of other cytoche-mokines is probably under the control of other TRNs, and

Figure 5—Suggested model of integrated regulation of CCL2 bymiRNAs and TFs in human adipocytes and macrophages. The vali-dated transcriptional regulatory subnetwork including three miRNAs(miR-126, -193b, and -92a; □), three TFs (MAX, ETS1, and SP1; ○),and an NF-kB subnetwork all directly or indirectly regulating the ex-pression and secretion of CCL2 (◆) from the adipose tissue in obesity(upper part). In the lower part of the model, miRNAs (miR-126 and-193b;□) and possible TF network (○) regulating CCL2 production inmacrophages are presented. T-bars represent inhibition, and arrowsrepresent stimulation. Bold edges represent interactions betweennodes described in this and previous studies in adipocytes, and thinlines are predicted by network analyses or shown in other cells types.Dashed edges/lines in the lower part represent possible interactions.


Dow


other WAT regions besides the subcutaneous fat could becontrolled by miRNAs in a different way.

In summary, human adipose CCL2 production is reg-ulated by a local miRNA-TF network that allows diversesignal amplification and is specific for different cell typespresent in the tissue. This may be an important factorlinking adipose tissue inflammation, insulin resistance,and type 2 diabetes.

Acknowledgments. The authors thank Eva Sjölin, Gaby Åström,Elisabeth Dungner, Kerstin Wåhlén, Britt-Marie Leijonhufvud, Katarina Hertel,and Yvonne Widlund (Department of Medicine, Lipid Laboratory, KarolinskaInstitutet) for excellent technical assistance.

Funding. This work was supported by several grants from the SwedishResearch Council, the Swedish Diabetes Foundation, Diabetes Wellness, theDiabetes Program at Karolinska Institutet, the Swedish Society of Medicine,Tore Nilsson Foundation, Foundation for Gamla Tjänarinnor, Åke WibergFoundation, European Association for the Study of Diabetes/Lilly program, NovoNordisk Foundation, and a research grant from the Ministry of Education,Culture, Sports, and Technology in Japan to the RIKEN Center for Life ScienceTechnologies.

Duality of Interest. No potential conflicts of interests relevant to thisarticle were reported.

Author Contributions. A.K. and Y.B. designed the study, performed thefunctional analyses, and wrote the manuscript. S.L.-C. and C.B. performed thefunctional analyses and wrote the manuscript. E.A. and C.O.D. conductedmicroarray data, motif activity response, and network analyses, and wrote themanuscript. P.H. collected tissue and wrote the manuscript. M.R. designed thestudy, collected tissue, and wrote the manuscript. N.M. designed the study;conducted microarray data, motif activity response, and network analyses; per-formed the functional analyses; and wrote the manuscript. P.A. designed thestudy and wrote the manuscript. All authors contributed to data interpretation andreviewed and approved the final manuscript. A.K. and P.A. are the guarantors ofthis work and, as such, had full access to all the data in the study and takeresponsibility for the integrity of the data and the accuracy of the data analysis.

References1. Fain JN. Release of inflammatory mediators by human adipose tissue is

enhanced in obesity and primarily by the nonfat cells: a review. MediatorsInflam 2010;2010:513948.

2. Ouchi N, Parker JL, Lugus JJ, Walsh K. Adipokines in inflammation andmetabolic disease. Nat Rev Immunol 2011;11:85–97

3. Cancello R, Clément K. Is obesity an inflammatory illness? Role of low-grade inflammation and macrophage infiltration in human white adiposetissue. BJOG 2006;113:1141–1147

4. Anderson EK, Gutierrez DA, Hasty AH. Adipose tissue recruitment of leu-kocytes. Curr Opin Lipidol 2010;21:172–177

5. Deshmane SL, Kremlev S, Amini S, Sawaya BE. Monocyte chemoattractantprotein-1 (MCP-1): an overview. J Interferon Cytokine Res 2009;29:313–326.

6. Kanda H, Tateya S, Tamori Y, et al. MCP-1 contributes to macrophageinfiltration into adipose tissue, insulin resistance, and hepatic steatosis inobesity. J Clin Invest 2006;116:1494–1505

7. Sayed D, Abdellatif M. MicroRNAs in development and disease. Physiol Rev2011;91:827–887

8. Bartel DP. MicroRNAs: genomics, biogenesis, mechanism, and function.Cell 2004;116:281–297

9. Martinez NJ, Walhout AJ. The interplay between transcription factors andmicroRNAs in genome-scale regulatory networks. Bioessays 2009;31:435–445

10. Schadt EE. Molecular networks as sensors and drivers of common humandiseases. Nature 2009;461:218–223

11. Arner E, Mejhert N, Kulyté A, et al. Adipose tissue microRNAs as regulatorsof CCL2 production in human obesity. Diabetes 2012;61:1986–1993

12. Hobert O. Common logic of transcription factor and microRNA action.Trends Biochem Sci 2004;29:462–468

13. Ivanovska I, Cleary MA. Combinatorial microRNAs: working together tomake a difference. Cell Cycle 2008;7:3137–3142

14. Kolaczynski JW, Morales LM, Moore JH Jr, et al. A new technique forbiopsy of human abdominal fat under local anaesthesia with Lidocaine. IntJ Obes Relat Metab Disord 1994;18:161–166

15. Arvidsson E, Viguerie N, Andersson I, Verdich C, Langin D, Arner P. Effectsof different hypocaloric diets on protein secretion from adipose tissue ofobese women. Diabetes 2004;53:1966–1971

16. Löfgren P, Hoffstedt J, Näslund E, Wirén M, Arner P. Prospective andcontrolled studies of the actions of insulin and catecholamine in fat cells ofobese women following weight reduction. Diabetologia 2005;48:2334–2342

17. van Harmelen V, Skurk T, Hauner H. Primary culture and differentiation ofhuman adipocyte precursor cells. Methods Mol Med 2005;107:125–135

18. Stenson BM, Rydén M, Venteclef N, et al. Liver X receptor (LXR) regulateshuman adipocyte lipolysis. J Biol Chem 2011;286:370–379

19. FANTOM Consortium; Suzuki H, Forrest AR, van Nimwegen E, et al. Thetranscriptional network that controls growth arrest and differentiation ina human myeloid leukemia cell line. Nat Genet 2009;41:553–562

20. Fain JN. Release of interleukins and other inflammatory cytokines byhuman adipose tissue is enhanced in obesity and primarily due to thenonfat cells. Vitam Horm 2006;74:443–477

21. Pitto L, Ripoli A, Cremisi F, Simili M, Rainaldi G. microRNA(interference)networks are embedded in the gene regulatory networks. Cell Cycle 2008;7:2458–2461

22. Ping D, Boekhoudt G, Zhang F, et al. Sp1 binding is critical for promoterassembly and activation of the MCP-1 gene by tumor necrosis factor. J BiolChem 2000;275:1708–1714

23. Zhan Y, Brown C, Maynard E, et al. Ets-1 is a critical regulator of Ang II-mediated vascular inflammation and remodeling. J Clin Invest 2005;115:2508–2516

24. Dang CV, Resar LM, Emison E, et al. Function of the c-Myc oncogenictranscription factor. Exp Cell Res 1999;253:63–77

25. Hirano F, Tanaka H, Hirano Y, et al. Functional interference of Sp1 andNF-kappaB through the same DNA binding site. Mol Cell Biol 1998;18:1266–1274

26. Keller U, Nilsson JA, Maclean KH, Old JB, Cleveland JL. Nfkb 1 is dispensablefor Myc-induced lymphomagenesis. Oncogene 2005;24:6231–6240

27. Lambert PF, Ludford-Menting MJ, Deacon NJ, Kola I, Doherty RR. Thenfkb1 promoter is controlled by proteins of the Ets family. Mol Biol Cell1997;8:313–323

28. Viswanathan M, Yu M, Mendoza L, Yunis JJ. Cloning and transcriptionfactor-binding sites of the human c-rel proto-oncogene promoter. Gene1996;170:271–276

29. Carbone GM, McGuffie EM, Collier A, Catapano CV. Selective inhibition oftranscription of the Ets2 gene in prostate cancer cells by a triplex-formingoligonucleotide. Nucleic Acids Res 2003;31:833–843


Dow


additive effects of micrornas and transcription factors on

Documents