A plasma microRNA signature of acute lentiviral infection

A plasma microRNA sig
nature of acute lentiviralinfection: biomarkers of central nervous
system disease

Kenneth W. Witwera, Stephanie L. Sarbanesb,

Jonathan Liua and Janice E. Clementsa,c,d

Copyright © L

aDepartment of MbColumbia UniverUniversity School

Correspondence tUniversity School

Tel: +1 410 955 9Received: 29 Apri

DOI:10.1097/QAD

ISSN

Objective: Plasma microRNAs (miRNAs) are modulated during disease and are emer-ging biomarkers; they have not been characterized in HIV infection. Using ourmacaque/simian immunodeficiency virus (SIV) model of HIV, we sought to identifya plasma miRNA profile of acute lentiviral infection, evaluate its relationship withknown cellular and viral determinants of lentivirus-associated central nervous system(CNS) disease, and explore the potential of miRNAs to predict CNS disease.

Design: Plasma samples were obtained before inoculation and 10 days after inocu-lation from SIV-infected macaques.

Methods: Plasma miRNA expression profiles were determined by TaqMan low-densityarray for six individuals. miRNA expression was compared with levels of cytokines,virus, and plasma platelet count. miRNA results were confirmed by single miRNA-specific assays for 10 macaques. Nineteen individuals were used to validate a diseaseprediction test.

Results: A 45-miRNA signature of acute infection (differential expression with P<0.05after multiple comparison correction) classified plasma as infected or not. Severaldifferentially expressed miRNAs correlated with CNS disease-associated cytokinesinterleukin-6 and CCL2 and included predicted and/or validated regulators of thecorresponding mRNAs. miRNAs tracked with viral load and platelet count were alsopredictors of CNS disease. At least six miRNAs were significantly differentiallyexpressed in individuals with severe versus no CNS disease; in an unweightedexpression test, they predicted CNS disease.

Conclusion: Acute-phase differential expression of plasma miRNAs predicts CNSdisease and suggests that CNS damage or predisposition to disease progression beginsin the earliest phase of infection. Plasma miRNAs should be investigated further asleading indicators of HIV diseases as early as acute infection.

� 2011 Wolters Kluwer Health | Lippincott Williams & Wilkins

AIDS 2011, 25:2057–2067

Keywords: acute infection, biomarker, HIV, HIV-associated neurocognitivedisorders, microRNA, plasma, simian immunodeficiency virus

Introduction

Despite the availability of effective antiretroviral therapy(ART) [1], HIV-associated neurocognitive disorders

ippincott Williams & Wilkins. Unaut

olecular and Comparative Pathobiology, Johns Hsity, New York, New York, cDepartment of Neuof Medicine, Baltimore, Maryland, USA.

o Dr Kenneth W. Witwer, PhD, Department ofof Medicine, 733 North Broadway, BRB Suite 8

770; fax: +1 410 955 9823; e-mail: kwitwer1@l 2011; revised: 6 July 2011; accepted: 3 Augus

.0b013e32834b95bf

0269-9370 Q 2011 Wolters Kluwer Hea

(HANDs) – including HIV-associated dementia, minorcognitive/motor disorder, and asymptomatic neurocog-nitive impairment (HAND) [2–4] – are a growingburden as the HIV-infected population expands and ages.

horized reproduction of this article is prohibited.

opkins University School of Medicine, Baltimore, Maryland,rology, and dDepartment of Pathology, Johns Hopkins

Molecular and Comparative Pathobiology, Johns Hopkins31, Baltimore, MD 21205, USA.

jhmi.edut 2011.

lth | Lippincott Williams & Wilkins 2057

mailto:[email protected]

http://dx.doi.org/10.1097/QAD.0b013e32834b95bf

Co

2058 AIDS 2011, Vol 25 No 17

These disorders lie at an etiologic intersection of viral andpharmacologic causes, immune responses, comorbidities,and aging [5], with no established biological diagnostic orpredictive tools [6]. Reported biomarkers are measuredmostly in brain tissue or cerebrospinal fluid (CSF) [7].New, less invasive metrics of HAND risk and progressioncould guide decisions about treatment timing andregimen [8] and open novel therapeutic avenues.

Ideal HAND biomarkers would have prognostic valueeven during acute infection, especially because damagingeffects of early host responses may persist. The increasingprevalence of HAND despite viral suppression throughART [1] demonstrates that although retroviruses con-tribute directly to neuropathology [7], host responses toinfection are also etiologically important. Indeed, centralnervous system (CNS) innate responses persist withoutactive viral spread [9]. The extent to which acute phasehost responses contribute to or predict the developmentof HIV diseases such as HAND remains incompletelyunderstood and difficult to explore in patient cohorts.Our validated simian immunodeficiency virus (SIV)/macaque model of AIDS and HIV CNS disease, whichhas yielded several important CNS disease biomarkers[10–13], allows longitudinal monitoring even during thefirst days of infection [14,15]. Here, we report the resultsof our search for microRNA (miRNA) biomarkers ofHAND using this model.

miRNA signatures have emerged as promising diseasebiomarkers [16] and are particularly useful when found inblood, obviating the need for tissue biopsy or CSF. PlasmamiRNAs are remarkably stable [17], providing a windowinto largely inaccessible tissue compartments such as brain[18–21]. No studies have explored plasma miRNAs asbiomarkers of lentiviral disease. Research has focused onthe reciprocal influence of HIV and host miRNAs incultured cells [22,23]; miRNA profiles in cells [24] ortissue [25,26]; and proposed HIV-encoded miRNAs [27].

We describe a plasma miRNA signature of acute infectionin which at least 45 plasma miRNAs are differentiallyexpressed by 10 days postinoculation (dpi). This signatureincludes miRNAs that are involved in innate immuneregulatory networks and correlate with CNS antiviralresponses. Moreover, differential expression of at least sixmiRNAs predicts the development of CNS disease.

Methods

Ethics statementAnimal studies were approved by the Johns HopkinsUniversity Institutional Animal Care and Use Committeeand conducted in accordance with the WeatherallReport, the Guide for the Care and Use of LaboratoryAnimals, and the USDA Animal Welfare Act.

pyright © Lippincott Williams & Wilkins. Unautho

Animal studiesPlasma samples were from pigtailed macaque studies (seealso Table, Supplemental Digital Content 1, http://links.lww.com/QAD/A179, list of macaques) involvingdual inoculation with an immunosuppressive SIV swarmand a neurotropic clone [28]. Preinoculation and 10-dpiplasma from three macaques each that did or did notprogress to encephalitis were analyzed by array. Plasmafrom four additional macaques (two each with no orsevere disease) were added for quantitative PCR (qPCR)validation; three preinoculation samples per animal and a10-dpi plasma were used. Nine additional sample setswere incorporated to assess a six-miRNA test for CNSdisease.

microRNA isolationTotal RNA was isolated from 100 ml of plasma bymiRvana protocol for liquid samples (Ambion, Austin,Texas, USA). Synthetic ath-miR159a (UUUGGAUU-GAAGGGAGCUCUA, Integrated DNA Technologies,Coralville, Iowa, USA) was added to normalize RNA forprocessing variations and was measured with severalendogenous miRNAs for quality control.

qPCR arrays, analysis, and data availabilityPlasma RNA was amplified with TaqMan Array HumanMicroRNA AþB Cards v3.0 (Applied Biosystems/LifeTechnologies, Carlsbad, California, USA). Data wereanalyzed by the R/Bioconductor package Analysis ofHigh-Throughput PCR data (HTqPCR) [29], DataAssist(Applied Biosystems/Life Technologies), BRB-Array-Tools (http://linus.nci.nih.gov/BRB-ArrayTools.html)[30], and MultiExperiment Viewer (http://www.tm4.org/mev/) [31] (see also Text, Supplemental DigitalContent 2, http://links.lww.com/QAD/A179, supple-mental methods). Array data were deposited with GEO(ID GSE26057, http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE26057).

Viral load and ELISAsViral load was measured by qPCR [32]. ELISAs forinterleukin (IL)-6 and CC motif chemokine ligand 2(CCL2) were from R&D Systems (Minneapolis,Minnesota, USA) [32].

Individual qPCR assaysTaqMan miRNA assay (Applied Biosystems/LifeTechnologies) reverse transcription and real-time stepsfollowed the manufacturer’s protocol (see also Text,Supplemental Digital Content 2, http://links.lww.com/QAD/A179, supplemental methods). Delta-deltathreshold cycle (DDCt) analysis included normalizationto ath-miR159a spike-in and comparison of each 10-dpisample to the mean of up to three preinfection samples.

Sequence analysesmiRNA sequences were obtained from miRBase(http://mirbase.org [33]). Human/macaque genomic

rized reproduction of this article is prohibited.

http://links.lww.com/QAD/A179



http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE26057

http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE26057



http://mirbase.org/

Plasma miRNA signature of SIV infection Witwer et al. 2059

comparisons were performed with UCSC genomebrowser (http://genome.ucsc.edu).

Target predictionsA Perl program combined RNAhybrid [34], TargetScan[35], and MiRanda [36] target prediction algorithms,assigning a score to candidate macaque miRNA/30 UTRpairs by the number of programs, predicting interactionand the number of calls per UTR.

Statistical methodsTwo-tailed t-tests (paired when appropriate) were donewith or without Welch’s correction; multiple comparisoncorrection was by Benjamini–Hochberg method [37].Values below 0.05 were considered significant; thosebelow 0.1 were considered to approach significance.Assessment of PCR results was based on DDCt. Viralloads were log-transformed.

Six-microRNA prediction testAn unweighted composite expression change score wasgenerated from six miRNAs with preinfection to post-infection expression changes in animals with no versusseveredisease.Animal-specificmiRNAexpressionchangeswere adjusted by the corresponding mean value of animalswith no CNS disease. Negative scores were zeroed. Valueswere summed to generate a composite score.

Results

Approximately 250 plasma microRNAs detectedby PCR arrayTaqMan low-density arrays identified differentiallyexpressed miRNAs in plasma from six pigtailed macaques(Macaca nemestrina) prior to and at 10 dpi. Almost 250miRNAs, on average, were detected in each sample, withover 120 miRNAs detected before a threshhold cycle (Ct)of 30 (see Figure, Supplemental Digital Content 3,http://links.lww.com/QAD/A179, Ct density curvesand the effects of normalization and filtering). There wasa significant decrease in the number of miRNAs detectedin acutely infected plasma (average¼ 241) compared withpreinfection plasma (average¼ 256; Fig. 1a, ‘All’). Adecrease was also apparent when the analysis wasrestricted to relatively abundant miRNAs (115 versus130; Fig. 1a, ‘Ct< 30’). Although these assays weredesigned to detect human miRNAs, 97% of miRNAsexpressed in macaque brain [38], as well as most predictedmacaque miRNAs [39], have 100% identity with humanorthologs. We have obtained concordant miRNA PCRresults with human and macaque samples [40].

Robust plasma microRNA profile changesaccompany acute retroviral infectionWe used three complementary analyses [hierarchicalclustering, principle component analysis (PCA), and

Copyright © Lippincott Williams & Wilkins. Unaut

random forest class prediction] to determine whetherexpression of miRNAs defines a signature of acuteinfection. Hierarchical clustering separated the samplesinto two distinct groups, representing preinfection andacute SIV infection (Fig. 1b; see also Figure, Supple-mental Digital Content 4, http://links.lww.com/QAD/A179, depicting the influence of normalization, filtering,and clustering method). PCA also revealed preinfectionand acute infection clusters (Fig. 1C). Ten-dpi plasmafrom one animal (CT64, the only female) falls betweenpreinfection and acute samples in a method-dependentmanner, often with greatest proximity to the preinfectionsample (see Text, Supplemental Digital Content 5,http://links.lww.com/QAD/A179, notes on sex, miR-NAs, and HIV). Random forest correctly classifiedsamples as infected or uninfected with 100% positive andnegative predictive power (see Table, SupplementalDigital Content 6, http://links.lww.com/QAD/A179,random forest results and classifying miRNAs).

Differentially expressed microRNAsTo identify differentially expressed miRNAs, we used twoanalysis programs (HTqPCR [29] and DataAssist) andmultiple normalization strategies (for details, see Supple-mental Methods; Text, Supplemental Digital Content 7,http://links.lww.com/QAD/A179, summary of normal-ization options; and Table, Supplemental Digital Content8, http://links.lww.com/QAD/A179, a comparison ofresults). Forty-five miRNAs were significantly differen-tially expressed in plasma at 10 dpi compared withpreinfection (Fig. 2, miRNAs with corrected P< 0.02;see also Table, Supplemental Digital Content 9, http://links.lww.com/QAD/A179, presenting all differentiallyexpressed miRNAs). This represents a minimum numberof differentially expressed miRNAs for three reasons.First, the assumptions of significance testing may not fullyreflect the extent of miRNA coregulation. Second,alternative analyses suggested several additional candidatedifferentially expressed miRNAs, such as miR-21 andmiR-34a. Third, quantile normalization may mask differ-ences among highly expressed miRNAs, for example,miR-21.

Validation of array results by qPCRArray results were validated by individual quantitativestem-loop real-time PCR for selected differentiallyexpressed and unchanged miRNAs. Seven differentiallyexpressed RNAs (miR-1274A, miR-1233, miR-190b,miR-660, miR-146a, miR-155, and miR-34a�)were drawn from HTqPCR analysis; two (miR-21 andmiR-34a) were from DataAssist analysis, and the tenthwas RNU48. The five unchanged miRNAs were let-7band let-130b, let-186, let-411, and let-424. Plasmasamples were from the six macaques analyzed by PCRarray and four additional macaques. We included up tothree preinfection samples per animal to address thepossibility of preinfection miRNA variations. Eight of 10differentially expressed RNAs were confirmed (Fig. 3a).


http://genome.ucsc.edu/










Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

2060 AIDS 2011, Vol 25 No 17

280

(a)

(b)

(c)

270

260

250

240

230

220Pre

All Ct < 30

# m

iRN

As

# m

iRN

As

10 dpi

160

150

140

130

120

100

110

90Pre 10 dpi

0.18

0.14

0.10

Hei

ght

0.06

01P

006

PR

d2

PLc

2

PLb

2

0.4

0.2

0.0

–0.2

–0.4

–0.34 –0.32 –0.30

PC1

PLc2

PLc2

PWf2

PWf2

01P006

01P006

CT64

CT64

PRd2

PRd2

PLb2

PLb2

PC

2

–0.28 –0.26 –0.24

PW

f2

CT

64

CT

64

PLc

2

01P

016

PR

d2

PLb

2

PW

f2

Fig. 1. MicroRNAs detected in macaque plasma distinguish infected from uninfected samples. (a) Compared with preinfectionplasma, at 10 days post infection (dpi), there was a significant decrease (P¼0.015, paired t-test) in the total number of detectedmicroRNAs (miRNAs) (‘All’). In all but one individual, the decline (P¼0.026) also occurred among relatively abundant miRNAsthat amplified before cycle 30 (’Ct< 30’). (b) Hierarchical clustering of infected (red) and uninfected (blue) samples. (c) Principlecomponent analysis identifies clusters based on infection status. The input for analyses in (b) and (c) was the set of all miRNAsdetected by qPCR array.


RNA1274Blet-7d1274A

1233

190b1303720

let-7g1267192

103

RNU48

191

215

let-7e660

30b301a

26b26a146a

1290625*

150142-3p -4.2

-4.3

-4.75.1

4.6-4.7

-4.7

-4.9

-4.9

5.0

5.7-6.1

-6.7

-7.76.2

9.8-6.6

6.0

12.7

7.08.6

9.8-8.2

-5.2

11.9 0.0002

0.0009

0.0009

0.0016

0.0020

0.0023

0.00510.0054

0.0055

0.0059

0.00600.0060

0.0060

0.0077

0.00840.0094

0.0097

0.0110

0.01100.01100.0113

0.01170.0123

0.0172

0.0172

-4.3

-4.4-9.3

-5.1-9.5

-2.4

3.0-8.2

-3.2

10.96.1

2.5

-4.4

-2.1-4.2

1.9

2.0

1.4

2.1-2.6

-3.0

-5.2-1.5

1.6

20 25 30 35 40

2.420.24

0.19

20.71+

34.29

+

+

5.25

0.13

9.46

-0.01

0.18

21.390.23

18.46

0.26

0.260.38

0.23

6.017.96

36.57

0.360.33

t adj. p ddCt FC Normalized Ct

Fig. 2. MicroRNAs differentially expressed from preinfec-tion to 10 days postinfection. Quantile-normalized array datawere assessed for significance, with Benjamini–Hochbergcorrection for multiple comparisons. Differentially expressedmiRNAs with P<0.02 (paired, corrected t-test) are shownhere. The right column depicts the normalized thresholdcycles of differentially expressed miRNAs (P<0.02, pairedt-test) for the six macaques before infection (circles) andduring acute infection (arrowheads). adj. p, adjusted P value;Ct, ; FC, fold change; t, T statistic.

For miR-155, results were confirmed for the original sixanimals, but differential expression was not significant inthe expanded group. All five miRNAs with no significantchanges according to array analysis were confirmed asunchanged (Fig. 3b). A discordant outcome was found foronly one miRNA, miR-190b (Fig. 3b, box).

Coregulation of differentially expressedmicroRNAs during acute lentiviral infectionTo assess whether plasma miRNAs are related by genomicproximity or transcriptional environment, we performedall pairwise correlations of miRNAs followed byhierarchical clustering (see Figure, Supplemental DigitalContent 10, http://links.lww.com/QAD/A179, den-drogram depicting related miRNAs). Several miRNAsgrouped by genomic proximity. Five of six members ofthe miR-17–92 cluster [41] were detected. miR-17,miR-19b, and miR-20a grouped together, as didmiR-19a and miR-92a. miR-1 and miR-133a areencoded antisense to the E3 ubiquitin protein ligaseMIB1 on chromosome 18 (human and macaque); their


levels correlated significantly (P< 0.005). miR-193b andmiR-365 were also significantly correlated (P< 0.0002)and located within a 5-kB region on chromosome 16(human) or 20 (macaque). Other miRNAs may be relatedby proximity to transcription factor binding sites.miR-155 and miR-223, downregulated during acuteinfection as analyzed by array, correlated significantly(P< 0.0001). Although located on different chromo-somes – 21 and X (human), 3 and X (macaque) – they areencoded near C/EBPb and POU domain transcriptionfactor binding sites. C/EBPb has activating andinhibitory forms. The latter is preferentially induced byacute innate immune responses [42]. miR-374a andmiR-374b are encoded near similar binding sites. Theirexpression levels correlated (P< 0.002), and miR-374bclustered with miR-155 and miR-223.

MicroRNAs correlate with innate immuneresponses that predict central nervous systemdiseaseTo investigate the link between plasma miRNAs andinnate immune responses with implications for CNSdisease, we calculated correlations of plasma miRNAlevels with the abundance of two cytokines that predictCNS disease in our model, IL-6 and CCL2. We alsowrote a program that integrates results of three miRNAtarget prediction algorithms to assess the potential fordirect miRNA regulation of the macaque IL-6 andCCL-2 30 UTRs. Negatively correlated miRNAs withpredicted target sites are strong candidates for directregulation of the corresponding transcript. SeveralmiRNAs with negative correlations to CCL2 proteinwere predicted to interact with the CCL2 30 UTR(Table 1). However, no miRNAs with positive corre-lations were predicted. For IL-6, members of the let-7family, along with miR-210, miR-26a, miR-30b, andmiR-30c, have predicted binding sites in the 30 UTR andare negatively correlated (Table 1). Only one positivelycorrelated miRNA has a predicted target. Importantly,human IL-6 is directly regulated by let-7 family members[43], and the human and macaque IL-6 miRNArecognition elements in the 30 UTR are identical. Here,let-7 members have the strongest correlations with IL-6(see Figure, Supplemental Digital Content 11, http://links.lww.com/QAD/A179, let-7e and IL-6 expression).

MicroRNAs correlate with viral load andplateletsWe next sought miRNAs that correlate with twoadditional biomarkers of lentiviral CNS disease: CSFviral load [28] and plasma platelet count [13]. Severalplasma miRNAs correlated positively with viral load (seeTable, Supplemental Digital Content 12, http://links.lww.com/QAD/A179, showing miRNAs correlatedwith viral load). These miRNAs may be upregulatedby the virus or antiviral responses, target the transcripts ofantiviral proteins, or directly enhance virus production.We did not find plasma miRNAs with strong negative







Co

2062 AIDS 2011, Vol 25 No 17

10

(a) (b)

5

0

2112

3312

74A

34a*

346a

RNU48 34a

660

155

let-7

b

130b 186

411

424

190b

–5

10 d

pi-p

re-in

fect

ion

Ct c

hang

eno

rmal

ized

to s

pike

-in c

ontr

ol–10

Fig. 3. Validation of array results by individual qPCR. Individual quantitative PCR assays confirmed differential expression (a) orlack thereof (b) for 14 selected RNAs. The box (b) indicates a discordant result for miR-190b, which was differentially expressedaccording to array analysis. qPCR results were normalized to a spiked-in synthetic control RNA. Significance (two-tailed, one-sample t-tests for difference from a theoretical mean of 0) is indicated as follows: �P�0.01; ��P<0.005; ��P<0.001.

correlations with virus. In contrast, many miRNAs werefound to correlate positively or negatively with plateletcount (see Table, Supplemental Digital Content 13,http://links.lww.com/QAD/A179, platelet–miRNAcorrelations).

MicroRNAs associated with central nervoussystem disease severityqPCR assays also showed that at least six miRNAs –miR-125b, miR-21, miR-34a, miR-1233, miR-130b,and miR-146a – were differentially upregulated inanimals that developed severe versus no CNS disease(Fig. 4a). Change in expression was significant for miR-125b, miR-21, miR-4a, and miR-1233 and approachedsignificance for miR-130b and miR-146a. For com-parison, there were no significantly different changes formiR-34a� (the hairpin partner of miR-34a), miR-155,miR-1274A, and let-7b.


Table 1. MicroRNA–cytokine correlations and target predictions.

CCL2-negative CCL2-positive

miRNA r TP miRNA r TP

let-7e �0.84 222 0.82let-7d �0.80 1274B 0.81let-7g �0.80 885–5p 0.80

142–3p �0.78 192 0.7930b �0.76 See 30c 1303 0.7626a �0.74 1233 0.7430c �0.73 MM 1274A 0.7426b �0.66 1267 0.71103 �0.64 215 0.69301 �0.61 720 0.64150 �0.58 625M 0.61

409–3p �0.57 T 375 0.56142–5p �0.56 MM 660 0.56

15b �0.55 125b 0.56191 �0.55 34aM 0.55210 �0.55 132 0.53

331–3p �0.54 146a 0.53374a �0.50 T 1260 0.51

Differentially expressed plasma microRNAs (miRNA) correlate (r¼ correlati(IL)-6 as determined by ELISA. miRNA target algorithms TargetScan (T), miRathe 30 UTR of macaque CCL2 and IL-6 for negatively correlated miRNAs. Rein a single UTR.

A six-microRNA expression test predicts centralnervous system diseaseA robust predictive tool for HAND or its constituentdisorders is likely to consist of multiple components [44].Therefore, we combined results for the six validated,upregulated miRNAs into an unweighted composite testof expression change. Samples from nine additionalanimals were added to the analysis for a total of 19.Thirteen of 14 animals that developed CNS diseasescored higher than all five no disease controls (Fig. 4b,P< 0.0005). With values of 0, 1, 2, 3, and 4 assigned tonone, mild, moderate, and severe CNS disease, respect-ively, composite expression scores correlated with out-come severity (r¼ 0.68, P< 0.002). However, omittingthe no disease samples, the severity/score correlation wasnot significant (see also Figure, Supplemental DigitalContent 14, http://links.lww.com/QAD/A179, graphof individual components of the prediction tool, stressing


IL-6-negative IL-6-positive

miRNA r TP miRNA r TP

let-7e �0.77 T 885–5p 0.71409–3p �0.74 1274B 0.71

let-7g �0.73 T 1274A 0.69let-7d �0.67 TMR 532–5p 0.68

210 �0.66 M 215 0.67155 �0.66 660 0.66103 �0.66 192 0.66

24–2M �0.65 146a 0.6330c �0.59 MMRR 511 0.62454 �0.56 132 0.62191 �0.56 1233 0.60 R26a �0.55 T 125b 0.5930b �0.54 See 30c 222 0.59

142–3p �0.52 1303 0.58331–3p �0.51 378 0.56

1267 0.55

on coefficient) with cerebrospinal fluid levels of CCL2 and interleukinnda/microRNA.org (M), and RNAHybrid (R) predict target sites (TP) inpeated letters in the ‘TP’ column denote multiple predicted target sites




6

4

2

-ddC

tC

ompo

site

test

sco

re

0

–2

12

8

4

0

125b

None

Mild

Mod

erat

e

Sever

e

34a 21

1233

130b

146a

let-7

b34

-a*

155

1274

A

None

p < 0.0005

p < 0.002r = 0.68

Severe

(a)

(b)

Fig. 4. A subset of microRNA changes during acute infectionare associated with central nervous system disease outcome.(a) qPCR results (�DDCt, with normalization to ath-miR159aspike-in control) from preinfection to 10 days postinoculationfor samples from all 10 animals used in the validation of arrayresults were re-analyzed after grouping by known CNS dis-ease outcome: ‘none’ or ‘severe’ pathology. Significance isindicated by ��P<0.01; �P<0.05; ^P<0.1. All P valueswere determined by Student’s t-test; error bars are standarderror of the mean. (b) An unweighted composite expressionscore based on the sum of expression changes of microRNA(miRNA)-125b, miRNA-34a, miRNA-21, miRNA-1233,miRNA-130b, and miRNA-146a was calculated for eachanimal. Expression score means for the group with no appar-ent central nervous system (CNS) disease and the group of allanimals with CNS disease were significantly different, asdetermined by an unpaired, two-tailed t-test with Welch’scorrection. Values of 1, 2, 3, and 4 were assigned to none,mild, moderate, and severe CNS disease, respectively, forcalculation of the correlation coefficient (r).

the necessity of a multicomponent predictive tool).Increased expression of these six miRNAs, thus,associates with CNS disease development.

Discussion

We demonstrate for the first time that acute retroviralinfection is accompanied by changes in circulating hostmiRNA levels, defining a signature that successfully


classifies plasma as infected or uninfected. Further, wepropose the first circulating miRNA biomarkers forlentivirus-associated CNS disease: six plasma miRNAs, ascomponents of a composite test, predict development ofCNS disease in our macaque/SIV model. The conserva-tion of primate miRNAs [38,39] and the similarities ofimmunologic and virologic parameters in our model andHIV infection [15,45,46] suggest implications for humandisease, justifying intensified investigation of plasmamiRNAs as biomarkers for HIV-associated diseases.

miRNAs are not simply incidental biomarkers; they canalso illuminate disease pathogenesis and accelerate thesearch for therapeutics. Our results show that let-7 familymembers, direct regulators of IL-6 [43], inverselycorrelate with this known predictor of CNS disease inthe macaque model. In addition, at the height of IL-10production [15], miR-146a is upregulated and miR-155downregulated, mirroring recent results in mouse andhuman [47,48]. miRNAs may be important targets forimmunotherapy of retrovirus-associated conditions.

miRNAs also correlate with platelet count, anothervalidated predictor of lentiviral CNS disease in macaqueand human. Platelets contain an abundance of functionalmiRNAs [49–52], which play an important regulatoryrole because only posttranscriptional regulation is possiblein these anucleate cell fragments. Platelet decline duringretroviral disease can be caused by progenitor deficiency,abnormal activation, immunological destruction, andexposure to high levels of cytokines [53,54], but reasonsfor the ties between platelet decline and CNS diseaseremain unclear. Our data suggest that miRNAs mayprovide insight into the underlying molecular events, andthat platelets are an important source of plasma miRNAs.

The plasma miRNA signature of acute infection likelyreflects both virus-specific and general inflammation-associated responses. A subset of signature miRNAscorrelates with virus levels. miR-221, reduced in 10-dpiplasma, is similarly suppressed in HIV-transfected cells[55]. In contrast, miR-146a upregulation may be ageneral response, modulating potentially damaginginflammation. Although several studies have examinedplasma miRNA profiles during chronic viral infection(e.g., [56,57]), comparison with acute infections isneeded to pinpoint retrovirus-specific responses withinour results.

Curiously, no miRNAs had significant negative corre-lations with viremia, as one might expect of miRNAswith direct anti-HIV properties, including miR-29family members [58,59] and those that inhibit HIVreplication through the viral UTR, for example, inresting CD4þ T lymphocytes [22]. This finding shouldnot, however, be interpreted as inconsistent with antiviralroles for miRNAs. There is no reason to assume thatcells with miRNA-mediated virus resistance would


Co

2064 AIDS 2011, Vol 25 No 17

specifically export protective miRNAs. Because thesecells are only a fraction of those contributing miRNA tothe blood, their miRNAs may not be released in amountssufficient to cause significant change in plasma levels.Furthermore, antiviral miRNAs could be an intrinsichost defense: rising concentrations of viral RNA wouldshift the balance of miRNA targets, making virusregulation more favorable stoichiometrically or kineti-cally even without miRNA expression changes.

In contrast, the miRNAs described here are important inthe developing, normal, and diseased brain, includingmiR-132 (Fragile X [60], Alzheimer’s [61]); miR-155(Down syndrome [62,63]); the let-7 family (neuronaldifferentiation [64], synaptic development [65], andopioid receptor regulation [66]); miR-18a (glucocorti-coid receptor/stress [67]); and miR-103, miR-191,miR-30b/c, and miR-495, which decrease in plasmaby 10 dpi and regulate the neuroprotective brain-derivedneurotrophic factor [68]. miR-155, miR-26a, andmiR-34a are associated with interferon-b [40,69], themost important type I interferon in brain [70]. Of the sixCNS disease-predictive miRNAs (miR-21, miR-34a,miR-125b, miR-130b, and miR-146a), at least five areexpressed in brain [38,71]. miR-21 is upregulated byexcitotoxic processes, targeting the neuroprotectivetranscription factor MEF2C in HAND [72]; expressionchanges are reported in glioblastoma [73] and braintrauma [21]. miR-34a is involved in neuronal apoptosisand was targeted for knockdown in a murine Alzheimer’sdisease model [74]. miR-125b is implicated in Downsyndrome [63,75], Fragile X [60], and Alzheimer’s disease[62]. Also Alzheimer’s disease-associated [76] miR-146ais upregulated in HIVencephalitis [77] and is tied to Rettsyndrome [78] and epilepsy [79]. miR-130b expressionmarks glioma progression [71].

It is also noteworthy that miR-125b targets p53 [80],whereas miR-34a is stimulated by p53 [81,82], targetsthe silent information regulator [82], and contributes toreplicative senescence [83] and aging in primate brain[84]. Upregulation of miR-125b and miR-34a suggeststhat aging-related regulatory processes are engaged in anescalating conflict during acute infection. The outcomeof this arms race could affect long-term predisposition tolentivirus-associated neurologic disease and prematureaging.

How and why would plasma biomarkers reflect processesin the CNS? Possible answers are not necessarily mutuallyexclusive. First, many but not all [14] host responses toinfection are shared by periphery and CNS [15,85]. Asecond link is the blood–brain barrier (qua, ‘relaystation’): plasma or brain miRNAs and cytokinesinfluence barrier signaling or physically cross it. Theendothelium is an important source of CNS cytokines[86], and release of miRNAs into circulation wouldcouple plasma miRNAs and CNS cytokines by reciprocal


regulation in a common compartment. Further researchis needed to show whether and how plasma miRNAsmove from blood to brain; that brain-derived factorsaffect peripheral miRNA synthesis; and that the blood–brain barrier endothelium is an important source ofplasma miRNAs.

Third, miRNAs within the CNS may be exported. Thephysical disruption of brain trauma, in which plasmamiRNA profiles change substantially [19], presents anobvious mechanism for miRNA transfer, and vascular-ization explains miRNA-containing glioblastoma vesiclesin blood [18]. However, changes in plasma miRNAduring therapy for mental disorders [20] suggest an exportprocess. We posit that miRNAs are released from braincells, perhaps for signaling purposes [87], and entercirculation through regions with minimal blood–brainbarrier (e.g., choroid plexus, hypothalamus) or collect inCSF. CSF is renewed several times each day, its contentsfiltered into the blood. Concentration in CSF and plasmahalf-life of theoretical CSF-derived miRNA-containingprotein complexes [88], lipoprotein particles [89], ormicrovesicles would determine the relative concen-trations of CSF and plasma miRNAs. Longitudinal CSFand brain tissue profiling, including measurements ofacute phase samples, is uniquely possible with the SIVmacaque model and will elucidate links between brain,CSF, and blood miRNAs. It will show how early thesemiRNAs (and their targets) are affected in brain, andthe effect of miRNA regulatory networks on CNSoutcomes.

We note that expression change in the individualmacaque, not absolute miRNA levels at one time point,was predictive of disease. Ongoing work will establishwhether single-time point levels suffice at later stages ofinfection or whether a preinfection baseline is a universalrequirement. However, this finding has implications forthe sensitivity of miRNA testing in general. miRNAbiomarkers have been proposed for many diseases andmiRNA-based diagnostics are currently in development.Samples from individual patients, stored prior to infectionor disease, would provide ideal controls, increasingsensitivity and multiplying the potential applications ofmiRNA profiling. It may, thus, be worthwhile toexamine the feasibility of banking plasma samples fromhealthy individuals.

In summary, our observations suggest that physiologicalchanges during acute retroviral infection predict and/orcontribute to the pathogenesis of incompletely under-stood HIV-associated disorders such as HAND and HIV-associated premature aging. Human samples should bestudied to confirm that miRNAs, including thosereported here, are associated with human disease. Furtherhuman and animal studies are needed to determine theextent to which plasma miRNAs affect disease processes,and whether altering their expression therapeutically will



influence disease progression. As demonstrated by miR-122 and hepatitis C virus, for which a miRNA-targetingtreatment is effective in nonhuman primates [90] and hasentered phase II trials, miRNA biomarkers provide cluesto disease pathogenesis and furnish novel therapeutictargets. These are urgently needed as the prevalence ofHAND continues to rise.

Acknowledgements

Study conception and design was by K.W.W. and J.E.C.;data acquisition was performed by K.W.W. and S.L.S.;contribution of analysis tools was by J.L.; analysis andinterpretation of the results was performed by K.W.W.,S.L.S., and J.E.C.; drafting and/or revision of manuscriptwas done by K.W.W., S.L.S., and J.E.C.; and J.E.C.obtained funding.

MedicalEditor Michael Linde contributed to the editing ofthis manuscript. We thank Joseph L. Mankowski, M.Christine Zink, and Lucio Gama for advice; SuzanneQueen and Brandon Bullock for expert technicalassistance; Gerald Vandergrift (Life Technologies) andthe JHU DNA Analysis Facility for microarray guidance;and all members of the Molecular and ComparativePathobiologyRetrovirusLaboratory for helpful discussion.

Funding was provided by the National Institutes ofHealth (MH070306 to J.E.C.) and by the Johns HopkinsBrain Science Institute.

Conflicts of interestThere are no conflicts of interest.

Portions of this work were presented by K.W.W. at the1st Annual Workshop on HIV and Aging, September2010, Baltimore, Maryland, USA; at the KeystoneSymposium on microRNAs and Human Disease,February 2011, Banff, Alberta, Canada; and at the18th Conference on Retroviruses and OpportunisticInfections, March 2011, Boston, Massachusetts, USA.

References

1. Simioni S, Cavassini M, Annoni JM, Rimbault Abraham A,Bourquin I, Schiffer V, et al. Cognitive dysfunction in HIVpatients despite long-standing suppression of viremia. AIDS2010; 24:1243–1250.

2. Rockstroh J, Guaraldi G, Deray G. HIV and the body: a reviewof multidisciplinary management. HIV Med 2010; 11 (Suppl2):1–8.

3. Antinori A, Arendt G, Becker JT, Brew BJ, Byrd DA, Cherner M,et al. Updated research nosology for HIV-associated neuro-cognitive disorders. Neurology 2007; 69:1789–1799.

4. Robertson KR, Smurzynski M, Parsons TD, Wu K, Bosch RJ,Wu J, et al. The prevalence and incidence of neurocognitiveimpairment in the HAART era. AIDS 2007; 21:1915–1921.

5. McArthur JC, Steiner J, Sacktor N, Nath A. Human immuno-deficiency virus-associated neurocognitive disorders: mind thegap. Ann Neurol 2010; 67:699–714.


6. Kaul M. HIV-1 associated dementia: update on pathologicalmechanisms and therapeutic approaches. Curr Opin Neurol2009; 22:315–320.

7. Letendre SL, Ellis RJ, Ances BM, McCutchan JA. Neurologiccomplications of HIV disease and their treatment. Top HIVMed 2010; 18:45–55.

8. Letendre S, Marquie-Beck J, Capparelli E, Best B, Clifford D,Collier AC, et al. Validation of the CNS penetration-effective-ness rank for quantifying antiretroviral penetration into thecentral nervous system. Arch Neurol 2008; 65:65–70.

9. Zink MC, Brice AK, Kelly KM, Queen SE, Gama L, Li M, et al.Simian immunodeficiency virus-infected macaques treatedwith highly active antiretroviral therapy have reduced centralnervous system viral replication and inflammation but persis-tence of viral DNA. J Infect Dis 2010; 202:161–170.

10. Mankowski JL, Queen SE, Clements JE, Zink MC. Cerebrospinalfluid markers that predict SIV CNS disease. J Neuroimmunol2004; 157:66–70.

11. Zink MC, Coleman GD, Mankowski JL, Adams RJ, Tarwater PM,Fox K, et al. Increased macrophage chemoattractant protein-1in cerebrospinal fluid precedes and predicts simian immuno-deficiency virus encephalitis. J Infect Dis 2001; 184:1015–1021.

12. Wachtman LM, Skolasky RL, Tarwater PM, Esposito D, SchifittoG, Marder K, et al. Platelet decline: an avenue for investigationinto the pathogenesis of human immunodeficiency virus-associated dementia. Arch Neurol 2007; 64:1264–1272.

13. Wachtman LM, Tarwater PM, Queen SE, Adams RJ, MankowskiJL. Platelet decline: an early predictive hematologic marker ofsimian immunodeficiency virus central nervous system disease.J Neurovirol 2006; 12:25–33.

14. Alammar L, Gama L, Clements JE. Simian immunodeficiencyvirus infection in the brain and lung leads to differential type IIFN signaling during acute infection. J Immunol 2011;186:4008–4018.

15. Witwer KW, Gama L, Li M, Bartizal CM, Queen SE, Varrone JJ,et al. Coordinated regulation of SIV replication and immuneresponses in the CNS. PLoS One 2009; 4:e8129.

16. Edwards JK, Pasqualini R, Arap W, Calin GA. MicroRNAs andultraconserved genes as diagnostic markers and therapeutictargets in cancer and cardiovascular diseases. J CardiovascTransl Res 2010; 3:271–279.

17. Cortez MA, Calin GA. MicroRNA identification in plasma andserum: a new tool to diagnose and monitor diseases. ExpertOpin Biol Ther 2009; 9:703–711.

18. Skog J, Wurdinger T, van Rijn S, Meijer DH, Gainche L, Sena-Esteves M, et al. Glioblastoma microvesicles transport RNA andproteins that promote tumour growth and provide diagnosticbiomarkers. Nat Cell Biol 2008; 10:1470–1476.

19. Redell JB, Moore AN, Ward Iii NH, Hergenroeder GW, DashPK. Human traumatic brain injury alters plasma microRNAlevels. J Neurotrauma 2010; 27:2147–2156.

20. Rong H, Liu TB, Yang KJ, Yang HC, Wu DH, Liao CP, et al.MicroRNA-134 plasma levels before and after treatment forbipolar mania. J Psychiatr Res 2010; 45:92–95.

21. Lei P, Li Y, Chen X, Yang S, Zhang J. Microarray based analysisof microRNA expression in rat cerebral cortex after traumaticbrain injury. Brain Res 2009; 1284:191–201.

22. Huang J, Wang F, Argyris E, Chen K, Liang Z, Tian H, et al.Cellular microRNAs contribute to HIV-1 latency in restingprimary CD4R T lymphocytes. Nat Med 2007; 13:1241–1247.

23. Triboulet R, Mari B, Lin YL, Chable-Bessia C, Bennasser Y,Lebrigand K, et al. Suppression of microRNA-silencing pathwayby HIV-1 during virus replication. Science 2007; 315:1579–1582.

24. Houzet L, Yeung ML, de Lame V, Desai D, Smith SM, Jeang KT.MicroRNA profile changes in human immunodeficiency virustype 1 (HIV-1) seropositive individuals. Retrovirology 2008;5:118.

25. Tatro ET, Scott ER, Nguyen TB, Salaria S, Banerjee S, Moore DJ,et al. Evidence for alteration of gene regulatory networksthrough microRNAs of the HIV-infected brain: novel analysisof retrospective cases. PLoS One 2010; 5:e10337.

26. Noorbakhsh F, Ramachandran R, Barsby N, Ellestad KK, Le-Blanc A, Dickie P, et al. MicroRNA profiling reveals newaspects of HIV neurodegeneration: caspase-6 regulates astro-cyte survival. FASEB J 2010; 24:1799–1812.


Co

2066 AIDS 2011, Vol 25 No 17

27. Narayanan A, Kehn-Hall K, Bailey C, Kashanchi F. Analysis ofthe roles of HIV-derived microRNAs. Expert Opin Biol Ther2011; 11:17–29.

28. Zink MC, Suryanarayana K, Mankowski JL, Shen A, Piatak M Jr,Spelman JP, et al. High viral load in the cerebrospinal fluid andbrain correlates with severity of simian immunodeficiencyvirus encephalitis. J Virol 1999; 73:10480–10488.

29. Dvinge H, Bertone P. HTqPCR: high-throughput analysis andvisualization of quantitative real-time PCR data in R. Bioinfor-matics 2009; 25:3325–3326.

30. Simon R, Lam A, Li MC, Ngan M, Menenzes S, Zhao Y. Analysisof gene expression data using BRB-array tools. Cancer Inform2007; 3:11–17.

31. Saeed AI, Bhagabati NK, Braisted JC, Liang W, Sharov V, HoweEA, et al. TM4 microarray software suite. Methods Enzymol2006; 411:134–193.

32. Barber SA, Herbst DS, Bullock BT, Gama L, Clements JE. Innateimmune responses and control of acute simian immuno-deficiency virus replication in the central nervous system.J Neurovirol 2004; 10 (Suppl 1):15–20.

33. Griffiths-Jones S, Grocock RJ, van Dongen S, Bateman A, EnrightAJ. miRBase: microRNA sequences, targets and gene nomen-clature. Nucleic Acids Res 2006; 34:D140–D144.

34. Rehmsmeier M, Steffen P, Hochsmann M, Giegerich R. Fast andeffective prediction of microRNA/target duplexes. RNA 2004;10:1507–1517.

35. Lewis BP, Burge CB, Bartel DP. Conserved seed pairing, oftenflanked by adenosines, indicates that thousands of humangenes are microRNA targets. Cell 2005; 120:15–20.

36. John B, Enright AJ, Aravin A, Tuschl T, Sander C, Marks DS.Human microRNA targets. PLoS Biol 2004; 2:e363.

37. Benjamini Y, Hochberg Y. Controlling the false discovery rate:a practical and powerful approach to multiple testing. J RoyalStat Soc 1995; 57:289–300.

38. Shao NY, Hu HY, Yan Z, Xu Y, Hu H, Menzel C, et al.Comprehensive survey of human brain microRNA by deepsequencing. BMC Genomics 2010; 11:409.

39. Brameier M. Genome-wide comparative analysis of micro-RNAs in three nonhuman primates. BMC Res Notes 2010; 3:64.

40. Witwer KW, Sisk JM, Gama L, Clements JE. MicroRNA regula-tion of IFN-beta protein expression: rapid and sensitive mod-ulation of the innate immune response. J Immunol 2010;184:2369–2376.

41. Mendell JT. miRiad roles for the miR-17-92 cluster in devel-opment and disease. Cell 2008; 133:217–222.

42. Dudaronek JM, Barber SA, Clements JE. CUGBP1 is required forIFNbeta-mediated induction of dominant-negative CEBPbetaand suppression of SIV replication in macrophages. J Immunol2007; 179:7262–7269.

43. Iliopoulos D, Hirsch HA, Struhl K. An epigenetic switch invol-ving NF-kappaB, Lin28, Let-7 microRNA, and IL6 links inflam-mation to cell transformation. Cell 2009; 139:693–706.

44. McGuire D. CSF biomarkers in HIV dementia: through a glassdarkly. Neurology 2009; 73:1942–1944.

45. Clements JE, Gama L, Graham DR, Mankowski JL, Zink MC. Asimian immunodeficiency virus macaque model of highlyactive antiretroviral treatment: viral latency in the peripheryand the central nervous system. Curr Opin HIV AIDS 2011;6:37–42.

46. Clements JE, Mankowski JL, Gama L, Zink MC. The acceleratedsimian immunodeficiency virus macaque model of humanimmunodeficiency virus-associated neurological disease: frommechanism to treatment. J Neurovirol 2008; 14:309–317.

47. McCoy CE, Sheedy FJ, Qualls JE, Doyle SL, Quinn SR, Murray PJ,et al. IL-10 inhibits miR-155 induction by toll-like receptors.J Biol Chem 2010; 285:20492–20498.

48. Monk CE, Hutvagner G, Arthur JS. Regulation of miRNA trans-cription in macrophages in response to Candida albicans. PLoSOne 2010; 5:e13669.

49. Landry P, Plante I, Ouellet DL, Perron MP, Rousseau G, ProvostP. Existence of a microRNA pathway in anucleate platelets. NatStruct Mol Biol 2009; 16:961–966.

50. Bruchova H, Yoon D, Agarwal AM, Mendell J, Prchal JT.Regulated expression of microRNAs in normal and polycythe-mia vera erythropoiesis. Exp Hematol 2007; 35:1657–1667.

51. Kannan M, Atreya C. Differential profiling of human red bloodcells during storage for 52 selected microRNAs. Transfusion2010; 50:1581–1588.


52. Nagalla S, Shaw C, Kong X, Kondkar AA, Edelstein LC, Ma L,et al. Platelet microRNA-mRNA co-expression profiles corre-late with platelet reactivity. Blood 2011; 117:5189–5197.

53. Flaujac C, Boukour S, Cramer-Borde E. Platelets and viruses: anambivalent relationship. Cell Mol Life Sci 2010; 67:545–556.

54. White MJ, Kile BT. Apoptotic processes in megakaryocytes andplatelets. Semin Hematol 2010; 47:227–234.

55. Yeung ML, Bennasser Y, Myers TG, Jiang G, Benkirane M, JeangKT. Changes in microRNA expression profiles in HIV-1-trans-fected human cells. Retrovirology 2005; 2:81.

56. Li LM, Hu ZB, Zhou ZX, Chen X, Liu FY, Zhang JF, et al. SerummicroRNA profiles serve as novel biomarkers for HBV infec-tion and diagnosis of HBV-positive hepatocarcinoma. CancerRes 2010; 70:9798–9807.

57. Ji F, Yang B, Peng X, Ding H, You H, Tien P. CirculatingmicroRNAs in hepatitis B virus-infected patients. J Viral Hepat2011; 18:e242–e251.

58. Ahluwalia JK, Khan SZ, Soni K, Rawat P, Gupta A, Hariharan M,et al. Human cellular microRNA hsa-miR-29a interferes withviral nef protein expression and HIV-1 replication. Retrovirol-ogy 2008; 5:117.

59. Nathans R, Chu CY, Serquina AK, Lu CC, Cao H, Rana TM.Cellular microRNA and P bodies modulate host-HIV-1 inter-actions. Mol Cell 2009; 34:696–709.

60. Edbauer D, Neilson JR, Foster KA, Wang CF, Seeburg DP,Batterton MN, et al. Regulation of synaptic structure andfunction by FMRP-associated microRNAs miR-125b andmiR-132. Neuron 2010; 65:373–384.

61. Lukiw WJ. Micro-RNA speciation in fetal, adult andAlzheimer’s disease hippocampus. Neuroreport 2007; 18:297–300.

62. Pogue AI, Cui JG, Li YY, Zhao Y, Culicchia F, Lukiw WJ. MicroRNA-125b (miRNA-125b) function in astrogliosis and glial cellproliferation. Neurosci Lett 2010; 476:18–22.

63. Kuhn DE, Nuovo GJ, Martin MM, Malana GE, Pleister AP, JiangJ, et al. Human chromosome 21-derived miRNAs are over-expressed in down syndrome brains and hearts. BiochemBiophys Res Commun 2008; 370:473–477.

64. Sempere LF, Freemantle S, Pitha-Rowe I, Moss E, Dmitrovsky E,Ambros V. Expression profiling of mammalian microRNAsuncovers a subset of brain-expressed microRNAs with possibleroles in murine and human neuronal differentiation. GenomeBiol 2004; 5:R13.

65. Corbin R, Olsson-Carter K, Slack F. The role of microRNAs insynaptic development and function. BMB Rep 2009; 42:131–135.

66. He Y, Yang C, Kirkmire CM, Wang ZJ. Regulation of opioidtolerance by let-7 family microRNA targeting the mu opioidreceptor. J Neurosci 2010; 30:10251–10258.

67. Uchida S, Nishida A, Hara K, Kamemoto T, Suetsugi M, Fuji-moto M, et al. Characterization of the vulnerability to repeatedstress in Fischer 344 rats: possible involvement of microRNA-mediated down-regulation of the glucocorticoid receptor. Eur JNeurosci 2008; 27:2250–2261.

68. Mellios N, Huang HS, Grigorenko A, Rogaev E, Akbarian S. Aset of differentially expressed miRNAs, including miR-30a-5p,act as posttranscriptional inhibitors of BDNF in prefrontalcortex. Hum Mol Genet 2008; 17:3030–3042.

69. O’Connell RM, Taganov KD, Boldin MP, Cheng G, Baltimore D.MicroRNA-155 is induced during the macrophage inflamma-tory response. Proc Natl Acad Sci U S A 2007; 104:1604–1609.

70. Griffin DE. Immune responses to RNA-virus infections of theCNS. Nat Rev Immunol 2003; 3:493–502.

71. Malzkorn B, Wolter M, Liesenberg F, Grzendowski M, Stuhler K,Meyer HE, et al. Identification and functional characterizationof microRNAs involved in the malignant progression of glio-mas. Brain Pathol 2010; 20:539–550.

72. Yelamanchili SV, Datta Chaudhuri A, Chen L-N, Xiong H, FoxHS. MicroRNA-21 dysregulates the expression of MEF2C inneurons in monkey and human SIV/HIV neurological disease.Cell Death Dis 2010; 1:e77.

73. Chan JA, Krichevsky AM, Kosik KS. MicroRNA-21 is an anti-apoptotic factor in human glioblastoma cells. Cancer Res 2005;65:6029–6033.

74. Wang X, Liu P, Zhu H, Xu Y, Ma C, Dai X, et al. miR-34a, amicroRNA up-regulated in a double transgenic mouse model ofAlzheimer’s disease, inhibits bcl2 translation. Brain Res Bull2009; 80:268–273.



75. Kuhn DE, Nuovo GJ, Terry AV Jr, Martin MM, Malana GE,Sansom SE, et al. Chromosome 21-derived microRNAs providean etiological basis for aberrant protein expression in humanDown syndrome brains. J Biol Chem 2010; 285:1529–1543.

76. Cui JG, Li YY, Zhao Y, Bhattacharjee S, Lukiw WJ. Differentialregulation of interleukin-1-receptor-associated kinase-1(IRAK-1) and IRAK-2 by micro RNA-146a and NF-{kappa}Bin stressed human astroglial cells and in Alzheimer’s disease.J Biol Chem 2010; 285:5189–5197.

77. Rom S, Rom I, Passiatore G, Pacifici M, Radhakrishnan S, DelValle L, et al. CCL8/MCP-2 is a target for mir-146a in HIV-1-infected human microglial cells. FASEB J 2010; 24:2292–2300.

78. Urdinguio RG, Fernandez AF, Lopez-Nieva P, Rossi S, HuertasD, Kulis M, et al. Disrupted microRNA expression caused byMecp2 loss in a mouse model of Rett syndrome. Epigenetics2010; 5:656–663.

79. Aronica E, Fluiter K, Iyer A, Zurolo E, Vreijling J, van Vliet EA,et al. Expression pattern of miR-146a, an inflammation-associated microRNA, in experimental and human temporallobe epilepsy. Eur J Neurosci 2010; 31:1100–1107.

80. Le MT, Teh C, Shyh-Chang N, Xie H, Zhou B, Korzh V, et al.MicroRNA-125b is a novel negative regulator of p53. GenesDev 2009; 23:862–876.

81. Chang TC, Wentzel EA, Kent OA, Ramachandran K, MullendoreM, Lee KH, et al. Transactivation of miR-34a by p53 broadlyinfluences gene expression and promotes apoptosis. Mol Cell2007; 26:745–752.

82. Yamakuchi M, Ferlito M, Lowenstein CJ. miR-34a repression ofSIRT1 regulates apoptosis. Proc Natl Acad Sci U S A 2008;105:13421–13426.


83. Fujita K, Mondal AM, Horikawa I, Nguyen GH, Kumamoto K,Sohn JJ, et al. p53 isoforms Delta133p53 and p53beta areendogenous regulators of replicative cellular senescence.Nat Cell Biol 2009; 11:1135–1142.

84. Somel M, Guo S, Fu N, Yan Z, Hu HY, Xu Y, et al. MicroRNA,mRNA, and protein expression link development and agingin human and macaque brain. Genome Res 2010; 20:1207–1218.

85. Kaul M, Lipton SA. Mechanisms of neuroimmunity and neuro-degeneration associated with HIV-1 infection and AIDS.J Neuroimmune Pharmacol 2006; 1:138–151.

86. Langford D, Masliah E. Crosstalk between components of theblood brain barrier and cells of the CNS in microglial activationin AIDS. Brain Pathol 2001; 11:306–312.

87. Camussi G, Deregibus MC, Bruno S, Cantaluppi V, Biancone L.Exosomes/microvesicles as a mechanism of cell-to-cell com-munication. Kidney Int 2010; 78:838–848.

88. Arroyo JD, Chevillet JR, Kroh EM, Ruf IK, Pritchard CC, GibsonDF, et al. Argonaute2 complexes carry a population of circu-lating microRNAs independent of vesicles in human plasma.Proc Natl Acad Sci U S A 2011; 108:5003–5008.

89. Vickers KC, Palmisano BT, Shoucri BM, Shamburek RD, Re-maley AT. MicroRNAs are transported in plasma and deliveredto recipient cells by high-density lipoproteins. Nat Cell Biol2011; 13:423–433.

90. Lanford RE, Hildebrandt-Eriksen ES, Petri A, Persson R, LindowM, Munk ME, et al. Therapeutic silencing of microRNA-122 inprimates with chronic hepatitis C virus infection. Science 2010;327:198–201.


Documents

A plasma microRNA signature of acute lentiviral infection