Correction

IMMUNOLOGYCorrection for “Posttranscriptional regulation of interleukin-10 expression by hsa-miR-106a,” by Amit Sharma, Manish Kumar,Jyotirmoi Aich, Manoj Hariharan, Samir K. Brahmachari, AnuragAgrawal, and BalaramGhosh, which was first published March 23,2009; 10.1073/pnas.0808743106 (Proc. Natl. Acad. Sci. U.S.A. 106,5761–5766).

The authors note that Fig. 4C and Fig. 5D appeared incorrectlydue to inadvertent duplication errors. Fig. 5A (lane: Egr1) was in-advertently duplicated and presented as Fig. 4C. The U6 panel ofFig. 5B was inadvertently duplicated in Fig. 5D. The correctedfigures and their legends appear below. The corrected figures werederived from data collected at the same time as the original ex-periments. These errors do not affect the conclusions of the article.

Fig. 4. Transcriptional regulation of hsa-miR-106a. (A) Schematic representation of the region encompassing the cluster encoding hsa-miR-106a that codesfor 6 different miRNAs. TSS is indicated as a dark box is ≈2 kb upstream. Transcription factors, Egr1 (dark circle) and Sp1 (dark oval), were predicted to bind ata region between −305 to −317 base pair. (B) The hsa-miR-106a promoter construct (in pGL3 basic) or pGL3 basic vector was transfected in Jurkat cells, andrelative luciferase activity was plotted. Results are presented as mean ± SE of 3 independent experiments. (C) Western blotting was performed by usingunstimulated or PHA/PMA (combination) stimulated nuclear extracts from Jurkat cells to determine the expression of Egr1 (panel 1) and Sp1 (panel 2), andequal loading was determined by Lamin (panel 3). (D) EMSA were performed by using end-labeled hsa-miR-106a promoter oligonucleotide. Sp1 binds to thehsa-miR-106a promoter in unstimulated conditions. (E) Egr1 binds to the promoter region on stimulation with PHA/PMA combination at an overlappingbinding site. NE, nuclear extract; un, unstimulated; st, stimulated; ODN, oligonucleotide; 106a, hsa-miR-106a promoter oligonucleotide; Egr (M), Egr mutantoligonucleotide; Egr-1 SS, Egr1 super shifted band. (F) Northern blotting (panel 1) was performed by using the RNAs from Jurkat cells stimulated with PMA/PHA combination for various time points (in hours as H), as indicated. The membrane was stripped and Northern blotting was performed for U6 for nor-malization; (panel 2) the extreme left lane represents radiolabeled Decade Marker (Ambion). C–F are representative of 3 independent experiments.

12498–12499 | PNAS | June 2, 2020 | vol. 117 | no. 22 www.pnas.org

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

8, 2

020

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

8, 2

020

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

8, 2

020

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

8, 2

020

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

8, 2

020

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

8, 2

020

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

8, 2

020

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

8, 2

020

Published under the PNAS license.

First published May 26, 2020.

www.pnas.org/cgi/doi/10.1073/pnas.2006676117

Fig. 5. Activation of hsa-miR-106a promoter regulates IL-10 expression. (A) Western blotting was performed for Egr1 expression by using Jurkat cells nuclearextracts prepared after stimulation with or without PHA, PMA, or PMA/PHA combination for 2 h, as indicated. (B) Northern blotting was performed for hsa-miR-106a by using RNA from Jurkat cells treated with or without PHA, PMA, or PMA/PHA combination for 8 h. (C) Western blotting was performed forEgr1 expression by using Jurkat cells transfected with Egr1 siRNA or a random control oligonucleotide followed by stimulation without or with PMA/PHAcombination for 2 h, as indicated. (D) Northern blotting was performed for hsa-miR-106a by using RNA from Jurkat cells treated as in C, followed bytreatment with PMA/PHA combination for 8 h. For loading control, the membranes from B and Dwere stripped, and Northern blotting was performed for U6.(E) The Jurkat cells were treated with or without PHA, PMA, or PHA/PMA combination for 24 h, and IL-10 was measured in the culture supernatant by ELISA.The figure represents fold change in IL-10 levels with respect to the unstimulated control, and is presented as mean ± SE of 3 independent experiments. A–Dare representative of 3 independent experiments.

PNAS | June 2, 2020 | vol. 117 | no. 22 | 12499

CORR

ECTION

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

8, 2

020

Posttranscriptional regulation of interleukin-10expression by hsa-miR-106aAmit Sharmaa, Manish Kumara, Jyotirmoi Aicha, Manoj Hariharanb, Samir K. Brahmacharib, Anurag Agrawala,and Balaram Ghosha,1

aMolecular Immunogenetics Laboratory and bG. N. Ramachandran Knowledge Centre for Genome Informatics, Institute of Genomics and IntegrativeBiology, Mall Road, Delhi 110007, India

Edited by David W. Talmage, University of Colorado Health Sciences Center, Denver, CO, and approved February 11, 2009 (received for reviewSeptember 3, 2008)

IL-10 is a key regulator of the immune system that criticallydetermines health and disease. Its expression is finely tuned bothat the transcriptional and posttranscriptional levels. Although theimportance of posttranscriptional regulation of IL-10 has beenpreviously shown, understanding the underlying mechanisms isstill in its infancy. In this study, using a combination of bioinfor-matics and molecular approaches, we report that microRNA (hsa-miR-106a) regulates IL-10 expression. The hsa-miR-106a bindingsite in the 3� UTR of IL10 has been identified by site-directedmutagenesis studies. Also, the involvement of transcription fac-tors, Sp1 and Egr1, in the regulation of hsa-miR-106a expressionand concomitant decrease in the IL-10 expression, has also beendemonstrated. In summary, our results showed that IL-10 expres-sion may be regulated by miR-106a, which is in turn transcription-ally regulated by Egr1 and Sp1.

Egr1 � IL-10 � microRNA � SP1 � miR-106a promoter

IL-10 is a key orchestrator of the immune system that has beenshown to be antiinflammatory in many model systems (1). It

is secreted by various cell types like different T cell subsets,macrophages, dendritic cells (DC), B-cells, mast cells, andeosinophils. It mediates a plethora of immunoregulatory eventssuch as maturation and activation of macrophages and DC,expression of MHC-II and B-7, activation of T cells, synthesis ofcytokines, and antibody production (2, 3). Also, dysregulation ofIL-10 leads to various immunological diseases, such as cancer,rheumatoid arthritis, asthma, infectious disorders, etc. (2).Therefore, it is likely that IL-10 expression is tightly regulated.

Regulation of IL-10 expression has been studied at thetranscriptional and posttranscriptional levels. Although Sp1 andSp3 have been found to regulate transcription, it has been shownthat IL10 mRNA is constitutively transcribed in many cells;however, the availability of its protein level is significantlydetermined by posttranscriptional mechanisms (4, 5). AU-richelements (ARE) in the 3� UTR of mouse IL10 that lead to thedegradation of its mRNA have been shown (6). Half life of IL10mRNA in normal melanocytes is much shorter than in mela-noma cell lines (7). Also, genetic variations in the 3� UTR of IL10have been shown to be associated with IL-10 levels that couldlead to disease pathogenesis (8, 9). Therefore, a pellucid under-standing of the posttranscriptional regulation of IL10 will be ofscientific and clinical significance.

A growing class of noncoding RNAs called microRNAs(miRNAs) is involved in posttranscriptional regulation of genes(10). MicroRNAs have been reported to modulate hematopoi-etic lineage differentiation (11), angiogenesis (12), cell adhesion(13), etc., indicating that they could have important roles innumerous biological processes (14). There is also a growing bodyof literature supporting the potential role of miRNAs in regu-lation of inflammatory processes (15). The importance of IL-10in the orchestration of the immune response, and the strongevidence for its posttranscriptional regulation, makes it an

attractive candidate for miRNA mediated regulation of inflam-mation.

Here, we report the identification of a miRNA, hsa-miR-106a,that regulates IL-10 expression. We also find that Sp1 and Egr1have an important role in hsa-miR-106a transcription and, thus,indirectly regulate the expression of IL-10 posttranscriptionally.

ResultsPrediction of miRNA Regulators of IL-10. To predict the miRNA thatmay regulate IL-10, we used a consensus approach by employing3 different miRNA target prediction software, as described inMaterials and Methods. We found that 8 miRNAs have potentialbinding site in the 3� UTR of IL10 (Table 1). Also, using miRex,we found that 5 miRNAs (namely hsa-miR-106a, miR-106b,miR-20a, miR-20b, and miR-93) were expressed in cells oflymphoid and myeloid origin, which are known to express IL-10.Because information regarding the expression of the remaining3 miRNAs was not available, we focused our study on hsa-miR-106a, miR-106b, miR-20a, miR-20b, and miR-93.

Identification of hsa-miR-106a As a Regulator of IL-10 Expression. Toidentify the miRNA(s) that may be involved in the regulation ofIL-10, initially Raji cell line was chosen, because it constitutivelyexpresses high level of IL-10 and, thus, presumably may havelower level of regulatory miRNA(s). When these cells weretransfected with the 5 predicted miRNA oligonucleotides, hsa-miR-106a oligonucleotide significantly down-regulated the ex-pression of IL-10 present in the culture supernatant (Fig. 1A).Interestingly, hsa-miR-106b, which differs from hsa-miR-106a inonly 2 nucleotide positions, also affected IL-10 expression, albeitat a lower level. In contrast, transfection with other 3 miRNA orcel-miR-67, a miRNA from Caenorhabditis elegans having nosequence identity with the predicted human miRNAs (used asnegative control), had no effect on the IL-10 expression (Fig.1A). Also, when increasing concentration of hsa-miR-106a wasused, IL-10 expression was found to be reduced in a dose-dependent manner (Fig. 1B).

To check whether hsa-miR-106a affects the stability of IL10mRNA, RT-PCR was performed by using RNA prepared fromRaji cells transfected with increasing concentration of hsa-miR-106a oligonucleotide. It was observed that it down-regulatedIL10 mRNA in a concentration-dependent manner (Fig. 1C),whereas cel-miR-67 had no effect. Thus, it is likely that hsa-miR-106a decreases IL-10 expression by degrading its mRNA.

Author contributions: A.S., S.K.B., and B.G. designed research; A.S., M.K., J.A., and M.H.performed research; A.S., M.K., A.A., and B.G. analyzed data; and A.S., A.A., andB.G. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Freely available online through the PNAS open access option.

1To whom correspondence should be addressed. E-mail: bghosh@igib.res.in.

This article contains supporting information online at www.pnas.org/cgi/content/full/0808743106/DCSupplemental.

www.pnas.org�cgi�doi�10.1073�pnas.0808743106 PNAS � April 7, 2009 � vol. 106 � no. 14 � 5761–5766

IMM

UN

OLO

GY

To confirm the involvement of miR-106a further, we choseJurkat cells that produce low basal level of IL-10, and, thus, mayproduce higher level of hsa-miR-106a. When Jurkat cells weretransfected with increasing concentration of anti-hsa-miR-106aoligonucleotide and IL-10 levels were measured in the culturesupernatant, a dose-dependent increase in the levels of IL-10 wasobserved (Fig. 1D). In contrast, a random oligonucleotide had noeffect on IL-10 expression. Thus, these experiments demonstratethat hsa-miR-106a regulates IL-10 expression.

Expression of hsa-miR-106a and Its Involvement in IL-10 Regulation inDifferent Cells. To specifically detect the expression of hsa-miR-106a, we standardized Northern blotting conditions so that wespecifically detect hsa-miR-106a, not hsa-miR-106b (Fig. 2A).Next, when blot was performed using RNA prepared from A549,Jurkat, Raji, HeLa, Hep-G2, and THP-1 cells, hsa-miR-106a wasfound to be expressed significantly in Jurkat, Raji, and THP-1cells (Fig. 2B). However, Hep-G2, HeLa and A549 cells ex-pressed lower amounts of hsa-miR-106a. To correlate the ex-pression of hsa-miR-106a with 3� UTR regulatory activity,

Jurkat, Raji, THP-1, and A549 cells were transfected withpMIR-REPORT-IL10 3� UTR (intact). A substantial fall inluciferase activity was observed in Jurkat, Raji, and THP-1 cells.In comparison, the decrease in A549 cells was not profound.Thus, the down-regulatory effect of IL10 3� UTR was found tobe correlated with the levels of mature hsa-miR-106a present inthese cells (Fig. 2C).

Confirmation of Target Site for hsa-miR-106a in IL10 3� UTR. Ourbioinformatics analysis predicted the binding of hsa-miR-106a tothe 3� UTR of IL10 in a region encompassing �4451 to �4478bases (Fig. 3A). When luciferase reporter vector containing theintact 3� UTR was transfected in Jurkat or Raji cells, a significantdown-regulation in the luciferase expression in both cells withrespect to control was observed (Fig. 3B). This down-regulationwas not seen in mutant 3� UTR construct lacking the predicted27 base hsa-miR-106a binding site. Also, when A549 cells, whichexpresses minimal level of hsa-miR-106a, were cotransfectedwith hsa-miR-106a oligonucleotide and pMIR-REPORT-IL103� UTR (intact), we observed �2-fold reduction in luciferase

Table 1. A list of predicted miRNAs (in the decreasing order of binding free energy) havingpotential binding sites in the 3� UTR of IL10

MicroRNAStartsite

Endsite

Binding freeenergy, Kcal/mol

Expression status inlymphoid and

myeloid origin cells

hsa-miR-20b�MIMAT0001413 625 647 �24.32 Yeshsa-miR-640� MIMAT0003310 654 670 �24.16 No datahsa-miR-93� MIMAT0000093 625 646 �23.88 Yeshsa-miR-20a� MIMAT0000075 625 647 �22.21 Yeshsa-miR-106b� MIMAT0003310 627 647 �20.77 Yeshsa-miR-597� MIMAT0003265 249 270 �20.52 No datahsa-miR-372� MIMAT0000724 619 646 �20.46 No datahsa-miR-106a� MIMAT0003310 621 647 �20.35 Yes

Fig. 1. Identification of hsa-miR-106a as a regulator of IL-10 expression. (A) Raji cells were transfected with or without the 5 candidate miRNA oligonucleotidesindividually, or with a negative control oligonucleotide, cel-miR-67 (having no sequence identity with any known human miRNAs); 24 h after transfection, IL-10was measured from the culture supernatants by ELISA, and plotted as fold change in IL-10 with respect to untransfected control. (B) Raji cells were transfectedwithout or with increasing concentrations of hsa-miR-106a or cel-miR-67, and IL-10 level was measured, as described in Materials and Methods. (C) Raji cells weretransfected with or without increasing concentrations of hsa-miR-106a or cel-miR-67 as described in B, and RNA was prepared followed by RT-PCR by using IL10specific primers. GAPDH RT-PCR was performed as loading control. Representative of 3 independent experiments. (D) Jurkat cells were transfected with orwithout increasing concentration of anti-hsa-miR-106a oligonucleotide or a negative control oligonucleotide followed by IL-10 measurement in the culturesupernatant by ELISA, 24 h after transfection. Results in A, B, and D are presented as mean � SE.

5762 � www.pnas.org�cgi�doi�10.1073�pnas.0808743106 Sharma et al.

expression (Fig. 3C). However, cotransfection with cel-miR-67did not have any appreciable effect on luciferase expression.Most importantly, when hsa-miR-106a was cotransfected alongwith pMIR-REPORT-IL10 3� UTR (mutant), there was mini-mal reduction in the luciferase expression. Also, cotransfectionof pMIR-REPORT-IL10 3� UTR (mutant) with cel-miR-67 hadno effect on luciferase expression (Fig. 3C). These experimentsindicate that mature hsa-miR-106a regulates IL-10 expression byinteracting with its 3� UTR.

Transcriptional Regulation of hsa-miR-106a. Because the level ofhsa-miR-106a was found to vary in different cell lines (Fig. 2B),and this variation could be due to differential regulation of itsexpression, we sought to investigate the transcriptional regula-tion of hsa-miR-106a locus present on the X chromosome. Thus,a region encompassing 1 kilobase upstream of the putativetranscription start site (TSS) of hsa-miR-106a was analyzed byusing Transfac Alibaba prediction software (16). Two transcrip-tion factors, Egr1 and Sp1, were predicted to bind to this region(Fig. 4A). When this 1 kilobase putative promoter region wascloned upstream of the luciferase gene (pGL3 basic, a promoterless vector) and transfected in Jurkat cells, we observed a 4-foldinduction in the luciferase expression (Fig. 4B); thus, indicatingthat the putative promoter sequence possessed regulatoryactivity.

To establish the binding of putative transcription factors Sp1

and Egr1, we determined the levels of these proteins usingWestern blotting of nuclear extracts prepared from unstimulatedor stimulated Jurkat cells (PHA/PMA combination for 2 h).Although the basal level of Egr1 in Jurkat cells was undetectable,there was a significant increase after stimulation with PHA/PMA combination (panel 1 in Fig. 4C). However, the level of Sp1was unaffected by stimulation (panel 2). To demonstrate thebinding of Sp1 and Egr1, EMSAs were performed by usingend-labeled oligonucleotide probe from hsa-miR-106a potentialpromoter region (�285 to �317 bases) and nuclear extract fromunstimulated (for Sp1) Jurkat cells. Out of the shifted bands(Fig. 4D, lane 2), our cold chase experiments with miR-106apromoter oligonucleotide (lane 3), irrelevant oligonucleotidesNF-�B (lane 4), and Sp1 oligonucleotide (lane 5) indicated theuppermost band (with an arrow) could be Sp1. The specificity ofSp1 band was further confirmed by antibody experiments, whereanti-Sp1 antibody abolished this uppermost band (lane 6),whereas anti-Sp3 (lane 7) or normal goat IgG (lane 8) could not.

For Egr1, when nuclear extract from stimulated Jurkat cells wasincubated with labeled hsa-miR-106a promoter oligonucleotide,only 1 shifted band was observed (Fig. 4E, lane 3). However, theband was absent when unstimulated nuclear extract was used (Fig.4E, lane 2). Specificity of the shifted band was confirmed by coldchase with excess of Egr consensus oligonucleotide (lane 4), whichabolished the specific band. However, cold chase with Egr mutantoligonucleotide was not able to chase the specific band (lane 5). Thisobservation was further confirmed by antibody super shift exper-iments, where antibody against Egr1 was able to shift the band (lane6), whereas non specific IgG did not (lane 7). These experimentsindicate that both Sp1 and Egr1 bind at an overlapping sequence inthe promoter region, and could regulate the transcription ofhsa-miR-106a.

To determine the effect of induced Egr1 expression onhsa-miR-106a level, Jurkat cells were stimulated with PHA/PMA combination for different time intervals, and Northernblotting was performed. It showed that there was an increase inthe levels of hsa-miR-106a up to 8 h (Fig. 4F), and its levels

Fig. 2. Expression of hsa-miR-106a and its involvement in IL-10 regulation indifferent cells. (A) DNA oligonucleotides of hsa-miR-106a, hsa-miR-106b, anda random sequence of equal length were run on an acrylamide gel and probedwith hsa-miR-106a radiolabeled probe. The probe was able to specificallydiscriminate between 2 forms of the miRNA. (B) Northern blotting was per-formed with total RNA from A549, Jurkat, Raji, HeLa, Hep-G2, and THP-1 cellsby using radiolabeled probe specific for hsa-miR-106a. The membrane wasstripped, and Northern blotting was performed for U6, which was used fornormalization. Lower panel indicates U6 band. Representative of 3 indepen-dent experiments. (C) The pMIR-REPORT-IL10 3� UTR (intact) was transfectedin A549, Jurkat, Raji, and THP-1 cells, and luciferase expression was measured.Relative fold change in luciferase activity for each cell line was plotted withrespect to pMIR-REPORT. A dark bar indicates cells transfected with pMIR-REPORT vector, whereas a light bar indicates cells transfected with pMIR-REPORT-IL10 3� UTR (intact). Results are presented as mean � SE.

Fig. 3. Confirmation of target site for hsa-miR-106a in IL10 3� UTR. (A)Schematic representation of IL10 3� UTR indicating the binding site of hsa-miR-106a as predicted. This region encompassing �4451 to �4478 bases wasdeleted by PCR based site-directed mutagenesis. (B) Luciferase reporter vec-tors pMIR-REPORT -IL10 3� UTR (intact) or pMIR-REPORT-IL10 3� UTR (mutant)were transfected in Jurkat or in Raji cells. The fold change in relative luciferaseactivity was plotted. Dark and light bars indicate luciferase activity in Jurkatcells and Raji cells, respectively. (C) A549 cell line was cotransfected withhsa-miR-106a oligonucleotide or cel-miR-67 along with either pMIR-REPORT-IL10 3� UTR (intact) or pMIR-REPORT-IL10 3� UTR (mutant). The luciferaseactivity relative to pMIR-REPORT-IL10 3� UTR (intact) was plotted. Results in Band C are presented as mean � SE for 3 independent experiments.

Sharma et al. PNAS � April 7, 2009 � vol. 106 � no. 14 � 5763

IMM

UN

OLO

GY

declined gradually with time (16 and 24 h). It was also observedthat the levels of pre-hsa-miR-106a increased with time up to 4 h,and then slowly declines. Interestingly, the decline of pre-hsa-

miR-106a preceded the decline of mature hsa-miR-106a (com-pare 4 versus 8 h). These results indicate that induced Egr1increased the expression of hsa-miR-106a.

Activation of hsa-miR-106a Promoter Regulates IL-10 Expression. Tocorrelate Egr1 mediated activation of hsa-miR-106a promoterwith IL-10 expression, Jurkat cells were stimulated with PMA,PHA, or PMA/PHA combination for 2 h, and expression of Egr1was detected by Western blotting (Fig. 5A). It was observed thatalthough PMA (50 ng/mL) alone did not affect Egr1 levels, PHA(5 �g/mL) was able to induce it. However, it was the combinationof both PMA/PHA that was able to maximally induce Egr1 levels(Fig. 5A). Next, to determine the level of induced hsa-miR-106a,Jurkat cells were treated with PMA, PHA, or PMA/PHAcombination for 8 h. Stimulation with PMA/PHA combinationbut not PMA alone induced hsa-miR-106a expression (Fig. 5B).This upregulation could be inhibited by specifically knockingdown Egr1 expression by using Egr1 siRNA before stimulation(Fig. 5 C and D).

To correlate the level of Egr1 induction with the levels ofIL-10, Jurkat cells were treated with PHA, PMA, or PHA/PMAtogether for 24 h, and IL-10 was measured in the culturesupernatant by ELISA (Fig. 5E). It was observed that Jurkat cellsinduced with PMA alone resulted in 3.5-fold increase in IL-10expression (Fig. 5E, lane 3), whereas PHA alone or PMA/PHAtogether significantly reduced IL-10 levels (Fig. 5E, lanes 2 and4). Thus, the level of IL-10 was inversely correlated with the levelof Egr1-induced hsa-miR-106a expression.

DiscussionThe posttranscriptional regulation of IL-10 is a subject ofparamount clinical interest because of the crucial role of IL-10as a modulator of inflammation. In this study, we found thathsa-miR-106a has a key role in regulating IL-10 expression. Weprovide a combination of in silico and in vitro evidence includinga series of transfection experiments with various potential miRNAsin support of this conclusion.

Our finding of differential level of hsa-miR-106a (Fig. 2B) invarious cells supports miRNA mediated control of IL-10 is

Fig. 4. Transcriptional regulation of hsa-miR-106a. (A) Schematic representation of the region encompassing the cluster encoding hsa-miR-106a that codes for6 different miRNAs. TSS is indicated as a dark box is �2 kb upstream. Transcription factors, Egr1 (dark circle) and Sp1 (dark oval), were predicted to bind at a regionbetween �305 to �317 base pair. (B) The hsa-miR-106a promoter construct (in pGL3 basic) or pGL3 basic vector was transfected in Jurkat cells, and relativeluciferase activity was plotted. Results are presented as mean � SE of 3 independent experiments. (C) Western blotting was performed by using unstimulatedor PHA/PMA (combination) stimulated nuclear extracts from Jurkat cells to determine the expression of Egr1 (panel 1) and Sp1 (panel 2), and equal loading wasdetermined by Lamin (panel 3). (D) EMSA were performed by using end-labeled hsa-miR-106a promoter oligonucleotide. Sp1 binds to the hsa-miR-106a promoterin unstimulated conditions. (E) Egr1 binds to the promoter region on stimulation with PHA/PMA combination at an overlapping binding site. NE, nuclear extract;un, unstimulated; st, stimulated; ODN, oligonucleotide; 106a, hsa-miR-106a promoter oligonucleotide; Egr (M), Egr mutant oligonucleotide; Egr-1 SS, Egr1 supershifted band. (F) Northern blotting (panel 1) was performed by using the RNAs from Jurkat cells stimulated with PMA/ PHA combination for various time points(in hours as H), as indicated. The membrane was stripped and Northern blotting was performed for U6 for normalization; (panel 2) the extreme left lanerepresents radiolabeled Decade Marker (Ambion). C–F are representative of 3 independent experiments.

Fig. 5. Activation of hsa-miR-106a promoter regulates IL-10 expression. (A)Western blotting was performed for Egr1 expression by using Jurkat cells nuclearextracts prepared after stimulation with or without PHA, PMA, or PMA/PHAcombination for 2 h, as indicated. (B) Northern blotting was performed forhsa-miR-106a by using RNA from Jurkat cells treated with or without PHA, PMA,or PMA/PHA combination for 8 h. (C) Western blotting was performed for Egr1expression by using Jurkat cells transfected with Egr1 siRNA or a random controloligonucleotide followed by stimulation without or with PMA/PHA combinationfor 2 h, as indicated. (D) Northern blotting was performed for hsa-miR-106a byusing RNA from Jurkat cells treated as in C, followed by treatment with PMA/PHAcombination for 8 h. For loading control, the membranes from B and D werestripped, and Northern blotting was performed for U6. (E) The Jurkat cells weretreated with or without PHA, PMA, or PHA/PMA combination for 24 h, and IL-10was measured in the culture supernatant by ELISA. The figure represents foldchange in IL-10 levels with respect to the unstimulated control, and is presentedas mean � SE of 3 independent experiments. A–D are representative of 3independent experiments.

5764 � www.pnas.org�cgi�doi�10.1073�pnas.0808743106 Sharma et al.

operational in cells that actively produce it. Recently, it has beendemonstrated that regulatory T (Treg) cells have a miRNAprofile distinct from conventional CD4 T cells, and some miR-NAs, including hsa-miR-106a, were found to be down-regulated(17). In our study, the identification of hsa-miR-106a thatregulates IL-10 expression further supports this observation. Itcan also be seen that there is a thin line between regulation anddysregulation of IL-10 expression. It is observed that in 46% ofhuman T cell leukemias, pri-miR-106–363 cluster is overex-pressed (18), presumably modulating IL-10 expression that couldpromote clonal expansion and leukemic cell survival. Our cur-rent results with the promoter revealed the presence of over-lapping binding sites for Sp1 and Egr1 that may have potentialregulatory function for hsa-miR-106a, a member of this cluster.

Sp1 is a well characterized ubiquitous transcription factor thatregulates a vast array of genes, including IL-10 (4, 5). The earlygrowth response gene product (Egr1) is a zinc finger transcrip-tion factor with a characteristic brisk kinetics of induction. It caneither positively or negatively regulate gene transcription inresponse to stimulation, because Egr1 contains both transacti-vation and repression domains (19). It has been proposed thatEgr1 might be up-regulated in response to environmental chal-lenge. Overlapping binding sites for Egr1 and Sp1 have beenpreviously seen in the promoter of several genes. For example,it has been previously demonstrated that under quiescent con-ditions the promoters of PDGF-A chain (A) and �-(1)-adrenergic receptor are occupied by Sp1, which regulates basalexpression of these genes (20, 21). However, on stimulation,levels of Egr1 rise, allowing Egr1 to displace Sp1 from thisregion; thus, inducing the expression of PDGF-A as a responseto the induction (21). Physical and functional cooperation be-tween Sp1 and Egr1 has also been seen at IL-2R� promoter inT cells (22). Here, constitutive level of IL-2R� is maintained bySp1, whereas on induction, Egr1 physically interacts with Sp1 tomaximize transcription. Similarly, in our study, Egr1 expressionin Jurkat cells resulted in the induction of pre-hsa-miR-106afollowed by mature hsa-miR-106a. This finding seems to be inconsonance with the observation that Sp1 is present in Jurkatcells at basal levels and may be responsible for basal transcriptionof this miRNA. Perturbation of this equilibrium by Egr1 wouldthen up-regulate hsa-miR-106a; thus, negatively regulating IL-10. This notion was further supported by our observation thatEgr1 stimulated hsa-miR-106a expression was negatively corre-lated with IL-10 levels. However, the mechanism by which Sp1and Egr1 cooperate for regulation of hsa-miR-106a expressionremains to be elucidated.

PHA and PMA have been known to induce vast array of genesby activating p38 (23), p42/44 ERK (24), and JNK MAP kinases(25). These 3 types of MAP kinases can be activated individually orsimultaneously; thus, suggesting their independent signaling roles.Although PMA alone is known to induce the activation of Sp1 inT cells (4), both PMA and PHA have been shown to activate Egr1in CEM cells (T cell line) (26). IL-10 has been previously shown tonegatively regulate LPS-stimulated Egr1 expression in mouse mac-rophages (27). Thus, it would be interesting to study the effect IL-10on Egr1 induced expression of hsa-miR-106a.

The regulatory loops and feedback mechanism for miRNAmediated posttranscriptional regulation of IL10 may be criticalin maintaining immune homeostasis, and its dysregulation maylead to disease. It seems plausible that 3 regulatory mechanismslike transcriptional regulation, ARE mediated 3� UTR regula-tion, and hsa-miR-106a mediated 3� UTR regulation may crit-ically regulate IL-10 expression. In fact, recently, it has beenshown that inhibitory function of NKp46�CD49b�CD3� NKcells recruited in the hepatic granulomas in experimental modelof visceral leishmaniasis was correlated with high IL-10 produc-tion as a result of increased stability of IL10 mRNA (28). It isvery likely that hsa-miR-106a mediated posttranscriptional con-

trol could potentially be involved in fine tuning the critical levelof IL-10 expression in a context-dependent manner, and hencedictates the outcome of immune response to specific externalstimuli.

In summary, our results suggest that IL-10 expression is regulatedby hsa-miR-106a, which is in turn transcriptionally regulated byEgr1 and Sp1. Further studies examining the relevance of thisregulatory mechanism to human disease are warranted. It wouldalso be interesting to investigate the effect of pathological condi-tions (like rheumatoid arthritis, cancer, and viral infections, wherean altered IL-10 expression is an important aspect of diseasepathogenesis) on this regulatory pathway involving IL-10, hsa-miR-106a, and Egr1. Also, identifying other targets of hsa-miR-106a andthe effect of activation of Egr1 on other members of the miRNAcluster would be of considerable interest.

Materials and MethodsIn Silico Identification of miRNA Binding Sites. Based on experimental obser-vations of miRNA binding to the 3� UTR, several target prediction softwarehave been developed. Although each of them has been used extensively fortarget site detection, they are associated with many false positive results dueto lack of specificity of prediction. To avoid the overprediction, we used aconsensus approach employing 3 widely-used software to perform the targetprediction. Only those miRNA target pairs were identified that were detectedby all 3 software to bind to the same location in the target 3� UTR sequence(29). All known 470 human miRNAs (miRBase v9) were queried against humanIL10 3� UTR using miRanda (30), RNAhybrid (31), and TargetScan (32).

Analysis of miRNA Expression Profiles. To check which among the 8 miRNAs,predicted to target IL10, were expressed in cells of lymphoid and myeloidorigin, we collected data from the miRNA expression resource (miRex; http://miracle.igib.res.in/mirex). This database allows cross-comparison of miRNAexpression profile data across various platforms collected from public repos-itories like Gene Expression Omnibus (National Center for BiotechnologyInformation-GEO) and ArrayExpress. The prenormalized data retrieved fromthese repositories are internally normalized to a standard score (z-score) byusing an in-house developed platform (miRex; http://miracle.igib.res.in/mirex). Expression of each miRNA in a particular tissue is represented as a foldchange. The database currently contains a total set of 1,507 experimentsspanning 18 datasets.

Plasmids. Plasmids pMIR-REPORT-IL10 3� UTR (intact), pMIR-REPORT-IL10 3�

UTR (mutant) and hsa-miR-106a promoter in pGL3 (basic) constructs wereprepared as described in detail in SI Materials and Methods.

Cell Culture and Transfection. A549, Raji, Jurkat, and THP-1 cells were grownand transfected as described in SI Materials and Methods.

IL-10 ELISA. Because IL-10 is a secretory cytokine, its levels were detected fromculture supernatant of transfected cells by performing ELISA (R&D Systems)following the manufacturer’s manual.

RNA Extraction and RT-PCR. RNA was prepared from cells treated with orwithout PHA (5 �g/mL), PMA (50 ng/mL), or combination of PHA/PMA (5 �g/mLand 50 ng/mL) for different time intervals by using TRIzol reagent (Invitrogen)following the product manual. RNA was stored at �70 °C until further use.RT-PCR was performed by using specific primers for IL10 and GAPDH followingthe standard protocol.

The hsa-miR-106a Northern Blotting. RNA isolated from A549, Jurkat, Raji,THP-1, HeLa, and HepG2 cells were resolved on acrylamide gel, transferred onnylon membrane (Gene screen plus; PerkinElmer), and hybridized withGamma 32P-labeled hsa-miR-106a oligonucleotide probe, as described in SIMaterials and Methods.

EMSA. Nuclear extracts were prepared from Jurkat cells as previously described(33). These extracts were incubated with radio-labeled double stranded oli-gonucleotide probes for hsa-miR-106a potential promoter region. Protein-DNA complexes were separated on acrylamide gels. Super shift experimentswere performed as described in SI Materials and Methods.

Sharma et al. PNAS � April 7, 2009 � vol. 106 � no. 14 � 5765

IMM

UN

OLO

GY

Western Blotting. Nuclear protein was prepared as described previously (33),and Western blotting was performed for Egr1, Sp1, and Lamin, as described inSI Materials and Methods.

Statistical Analysis. Results are given as mean of 3 independent experiments �SEM. An independent 2-tailed Student’s t test was performed. Differenceswere considered statistically significant for P � 0.05. For normalization of

transfection efficiency, cells were cotransfected with vector pMIR-REPORTbeta-Gal as indicated.

ACKNOWLEDGMENTS. We thank Mr. Nomesh Goplani for technical help. A.S.,M.K., and M.H. are supported by fellowships from the Council of Scientific andIndustrial Research (CSIR). This work is supported by a grant from CSIR (TaskForce Project Grant NWP0033) India.

1. Asadullah K, Sterry W, Volk HD (2003) Interleukin-10 therapy-review of a new ap-proach. Pharmacol Rev 55:241–269.

2. O’Garra A, Barrat FJ, Castro AG, Vicari A, Hawrylowicz C (2008) Strategies for use ofIL-10 or its antagonists in human disease. Immunol Rev 223:114–131.

3. O’Garra A, Vieira P (2007) T(H)1 cells control themselves by producing interleukin-10.Nat Rev Immunol 7:425–428.

4. Tone M, Powell MJ, Tone Y, Thompson SA, Waldmann H (2000) IL-10 gene expressionis controlled by the transcription factors Sp1 and Sp3. J Immunol 165:286–291.

5. Ma W, et al. (2001) The p38 mitogen-activated kinase pathway regulates the humaninterleukin-10 promoter via the activation of Sp1 transcription factor in lipopolysac-charide-stimulated human macrophages. J Biol Chem 276:13664–13674.

6. Powell MJ, Thompson SA, Tone Y, Waldmann H, Tone M (2000) Posttranscriptionalregulation of IL-10 gene expression through sequences in the 3�-untranslated region.J Immunol 165:292–296.

7. Brewer G, Saccani S, Sarkar S, Lewis A, Pestka S (2003) Increased interleukin-10 mRNAstability in melanoma cells is associated with decreased levels of A � U-rich elementbinding factor AUF1. J Interferon Cytokine Res 23:553–564.

8. Lyon H, et al. (2004) IL10 gene polymorphisms are associated with asthma phenotypesin children. Genet Epidemiol 26:155–165.

9. Kurreeman FA, Schonkeren JJ, Heijmans BT, Toes RE, Huizinga TW (2004) Transcriptionof the IL10 gene reveals allele-specific regulation at the mRNA level. Hum Mol Genet13:1755–1762.

10. Bartel DP, et al. (2004) MicroRNAs: Genomics, biogenesis, mechanism, and function.Cell 116:281–297.

11. Chen CZ, Li L, Lodish HF, Bartel DP (2004) MicroRNAs modulate hematopoietic lineagedifferentiation. Science 303:83–86.

12. Chen Y, Gorski DH (2008) Regulation of angiogenesis through a microRNA (miR-130a)that down-regulates antiangiogenic homeobox genes GAX and HOXA5. Blood111:1217–1226.

13. Harris TA, Yamakuchi M, Ferlito M, Mendell JT, Lowenstein CJ (2008) MicroRNA-126regulates endothelial expression of vascular cell adhesion molecule 1. Proc Natl AcadSci USA 105:1516–1521.

14. varez-Garcia I, Miska EA (2005) MicroRNA functions in animal development and humandisease. Development 132:4653–4662.

15. Taganov KD, Boldin MP, Baltimore D (2007) MicroRNAs and immunity: Tiny players ina big field. Immunity 26:133–137.

16. Grabe N (2002) AliBaba2: Context specific identification of transcription factor bindingsites. In Silico Biol 2:S1–S15.

17. Cobb BS, et al. (2006) A role for Dicer in immune regulation. J Exp Med 203:2519–2527.18. Landais S, Landry S, Legault P, Rassart E (2007) Oncogenic potential of the miR-106–363

cluster and its implication in human T-cell leukemia. Cancer Res 67:5699–5707.

19. Thiel G, Cibelli G (2002) Regulation of life and death by the zinc finger transcriptionfactor Egr-1. J Cell Physiol 193:287–292.

20. Bahouth SW, Beauchamp MJ, Vu KN (2002) Reciprocal regulation of beta(1)-adrenergicreceptor gene transcription by Sp1 and early growth response gene 1: Induction ofEGR-1 inhibits the expression of the beta(1)-adrenergic receptor gene. Mol Pharmacol61:379–390.

21. Khachigian LM, Williams AJ, Collins T (1995) Interplay of Sp1 and Egr-1 in the proximalplatelet-derived growth factor A-chain promoter in cultured vascular endothelial cells.J Biol Chem 270:27679–27686.

22. Lin JX, Leonard WJ (1997) The immediate-early gene product Egr-1 regulates thehuman interleukin-2 receptor beta-chain promoter through non canonical Egr and Sp1binding sites. Mol Cell Biol 17:3714–3722.

23. Loza MJ, Foster S, Peters SP, Penn RB (2006) Beta-agonists modulate T-cell functions viadirect actions on type 1 and type 2 cells. Blood 107:2052–2060.

24. Li YQ, Hii CS, Der CJ, Ferrante A (1999) Direct evidence that ERK regulates theproduction/secretion of interleukin-2 in PHA/PMA-stimulated T lymphocytes. Immu-nology 96:524–528.

25. Kesherwani V, Sodhi A (2007) Involvement of tyrosine kinases and MAP kinases in theproduction of TNF-alpha and IL-1beta by macrophages in vitro on treatment withphytohemagglutinin. J Interferon Cytokine Res 27:497–505.

26. Cogswell PC, Mayo MW, Baldwin AS, Jr (1997) Involvement of Egr-1/RelA synergy indistinguishing T cell activation from tumor necrosis factor-alpha-induced NF-kappa B1transcription. J Exp Med 185:491–497.

27. Kamimura M, et al. (2005) Interleukin-10 suppresses tissue factor expression in lipopo-lysaccharide-stimulated macrophages via inhibition of Egr-1 and a serum responseelement/MEK-ERK1/2 pathway. Circ Res 97:305–313.

28. Maroof A, et al. (2008) Posttranscriptional regulation of IL10 gene expression allowsnatural killer cells to express immunoregulatory function. Immunity 29:295–305.

29. Hariharan M, Scaria V, Pillai B, Brahmachari SK (2005) Targets for human encodedmicroRNAs in HIV genes. Biochem Biophys Res Commun 337:1214–1218.

30. John B, et al. (2004) Human MicroRNA targets. PLoS Biol 2:e363.31. Rehmsmeier M, Steffen P, Hochsmann M, Giegerich R (2004) Fast and effective pre-

diction of microRNA/target duplexes. RNA 10:1507–1517.32. Lewis BP, Shih IH, Jones-Rhoades MW, Bartel DP, Burge CB (2003) Prediction of

mammalian microRNA targets. Cell 115:787–798.33. Kumar S, Sharma A, Madan B, Singhal V, Ghosh B (2007) Isoliquiritigenin inhibits

IkappaB kinase activity and ROS generation to block TNF-alpha induced expression ofcell adhesion molecules on human endothelial cells. Biochem Pharmacol 73:1602–1612.

5766 � www.pnas.org�cgi�doi�10.1073�pnas.0808743106 Sharma et al.