Home | Looking for something? Sign In | New here? Sign Up | Log out


Subscribe Now: standard

Wednesday, May 11, 2011

Mechanism of Impaired NLRP3 Inflammasome Priming by Monophosphoryl Lipid A.

Wednesday, May 11, 2011

1Department of Microbiology and Immunology and Institute for Cellular Therapeutics, University of Louisville School of Medicine, 570 South Preston Street, Louisville, KY 40202, USA.

Monophosphoryl lipid A (MLA), a nontoxic derivative of the endotoxin lipopolysaccharide (LPS), has been approved in the United States for use as a vaccine adjuvant. LPS and MLA are ligands of Toll-like receptor 4 (TLR4), and it has been unclear why LPS triggers toxic inflammation, whereas MLA generates safe and effective immunostimulation. Signaling downstream of TLR4 is mediated by the adaptor proteins TRIF [Toll-interleukin-1 (IL-1) receptor (TIR) domain-containing adaptor-inducing interferon-ß], which is required for adaptive immune outcomes, and MyD88 (myeloid differentiation marker 88), which is responsible for many proinflammatory effects. Two models have provided nonexclusive explanations for the differential effects of LPS and MLA. According to the first model, MLA fails to induce maturation of the proinflammatory cytokine IL-1ß because it fails to activate caspase-1, which is required for the conversion of pro-IL-1ß into its bioactive form. The second model suggests that MLA triggers unequal engagement of both of the signaling adaptor pathways of TLR4, such that signaling mediated by TRIF is largely intact, whereas signaling mediated by MyD88 is incomplete. We show that the TRIF-biased signaling that is characteristic of low-toxicity MLA explains its failure to activate caspase-1. Defective induction of NLRP3, which depends on MyD88, led to decreased assembly of components of the IL-1ß-activating inflammasome required for the activation of preformed, inactive procaspase-1. In addition, we elucidated the contributions of MyD88 and TRIF to priming of the NLRP3 inflammasome and demonstrated that TRIF-biased TLR4 activation by MLA was responsible for the defective production of mature IL-1ß.

View the original article here

read more

Comparative transcriptomics of extreme phenotypes of human HIV-1 infection and SIV infection in sooty mangabey and rhesus macaque


J Clin Invest. doi:10.1172/JCI45235.
Copyright © 2011, The American Society for Clinical Investigation. Margalida Rotger1, Judith Dalmau2, Andri Rauch3, Paul McLaren4, Steve Bosinger5, Raquel Martinez1, Netanya G. Sandler6, Annelys Roque6, Julia Liebner6, Manuel Battegay7, Enos Bernasconi8, Patrick Descombes9, Itziar Erkizia2, Jacques Fellay1, Bernard Hirschel10, Jose M. Miró11, Eduard Palou12, Matthias Hoffmann13, Marta Massanella2, Julià Blanco2, Matthew Woods14, Huldrych F. Günthard15, Paul de Bakker4, Daniel C. Douek6, Guido Silvestri5, Javier Martinez-Picado2,16 and Amalio Telenti1,17

1Institute of Microbiology, University Hospital and University of Lausanne, Lausanne, Switzerland.
2AIDS Research Institute (IrsiCaixa), Hospital Universitari Germans Trias i Pujol, Badalona, Barcelona, Spain.
3Division of Infectious Diseases, University Hospital Bern, Bern, Switzerland.
4Division of Genetics, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts, USA.
5Yerkes National Primate Research Center and Emory Vaccine Center, Emory University, Atlanta, Georgia, USA.
6Human Immunology Section, Vaccine Research Center, National Institute of Allergy and Infectious Diseases, NIH, Bethesda, Maryland, USA.
7Division of Infectious Diseases and Hospital Epidemiology, University Hospital Basel, Basel, Switzerland.
8Ospedale Regionale, Lugano, Switzerland.
9Genomics Platform, University of Geneva, Geneva, Switzerland.
10Division of Infectious Diseases, University Hospital Geneva, Geneva, Switzerland.
11Hospital Clinic — Institut d’Investigacions Biomèdiques August Pi i Sunyer, University of Barcelona, Barcelona, Spain.
12Banc de Sang i Teixits, Barcelona, Spain.
13Kantonsspital St. Gall, Saint Gallen, Switzerland.
14Ragon Institute of MGH, MIT and Harvard, Boston, Massachusetts, USA.
15Division of Infectious Diseases and Hospital Epidemiology, University Hospital Zurich, University of Zurich, Zurich, Switzerland.
16Institució Catalana de Recerca i Estudis Avançats, Barcelona, Spain.
17Swiss HIV Cohort Study (SHCS), Lausanne, Switzerland.

Address correspondence to: Javier Martinez-Picado, AIDS Research Institute — IrsiCaixa, Hospital Germans Trias i Pujol, 08916 Badalona, Spain. Phone: 34.93.4656374; Fax: 34.93.4653968; E-mail: jmpicado@irsicaixa.es. Or to: Amalio Telenti, Institute of Microbiology, Bugnon 48, CHUV, 1011 Lausanne, Switzerland. Phone: 41.79.556.0751; Fax: 41.21.314.4095; E-mail: Amalio.telenti@chuv.ch.

Authorship note: Margalida Rotger, Judith Dalmau, Andri Rauch, Paul McLaren, and Steve Bosinger contributed equally to this work.

Published May 9, 2011
Received for publication February 2, 2011, and accepted in revised form March 30, 2011.

High levels of HIV-1 replication during the chronic phase of infection usually correlate with rapid progression to severe immunodeficiency. However, a minority of highly viremic individuals remains asymptomatic and maintains high CD4+ T cell counts. This tolerant profile is poorly understood and reminiscent of the widely studied nonprogressive disease model of SIV infection in natural hosts. Here, we identify transcriptome differences between rapid progressors (RPs) and viremic nonprogressors (VNPs) and highlight several genes relevant for the understanding of HIV-1–induced immunosuppression. RPs were characterized by a specific transcriptome profile of CD4+ and CD8+ T cells similar to that observed in pathogenic SIV-infected rhesus macaques. In contrast, VNPs exhibited lower expression of interferon-stimulated genes and shared a common gene regulation profile with nonpathogenic SIV-infected sooty mangabeys. A short list of genes associated with VNP, including CASP1, CD38, LAG3, TNFSF13B, SOCS1, and EEF1D, showed significant correlation with time to disease progression when evaluated in an independent set of CD4+ T cell expression data. This work characterizes 2 minimally studied clinical patterns of progression to AIDS, whose analysis may inform our understanding of HIV pathogenesis.

HIV infection leads to severe immunodeficiency in most infected subjects, in an average of 10 years; however, there are marked departures from this estimate. Attention has been directed at understanding the determinants of nonprogressive disease, as exemplified by the clinical course of long-term nonprogressors and of elite controllers (1–3). The other extreme of the spectrum of disease — rapid progression — has been the subject of much less research. Rapid progressors (RPs) can be defined by a number of criteria — generally including progressive immunosuppression soon after seroconversion and, in many cases, high levels of viremia (4, 5). Limited data suggest that the concurrence of viral and host factors contributes to the severity of early disease (6). There are, however, few such individuals in clinical cohorts — the main limitations for prospective recruitment are the need to identify patients with a known date of infection (seroconverters), and the short window of clinical observation before antiretroviral treatment is initiated. These constrain the availability of relevant biological material for study.

There are also very rare individuals that can tolerate very high viral loads, comparable to those of RPs, while maintaining stable CD4+ T cell counts for many years in the absence of treatment. Choudhary et al. (7) described 3 HIV-infected individuals with long-term asymptomatic disease who maintained stable CD4+ T cell counts and low levels of immune activation, despite viral replication in the range of 104 to 105 HIV-1 RNA copies per ml of plasma. This profile of tolerance of viral replication is reminiscent of the pattern of SIV infection in the natural host. The importance of such model for the understanding of HIV/AIDS pathogenesis has been underscored by studies in sooty mangabeys and in African green monkeys (8–12). Sooty mangabeys have nonprogressive disease despite chronic virus replication that is characterized by low levels of immune activation, while pathogenic SIV infection of rhesus macaques is associated with chronic immune activation. The consequences of immune activation include increased cell turnover, the skewing of lymphocytes toward more activated and differentiated subpopulations, and the induction of cellular exhaustion, senescence, and low renewal potential (reviewed in ref. 13).

The first goal of the present study was to explore a set of standard criteria to identify HIV-infected individuals presenting those 2 distinct clinical patterns: rapid progression and the contrasting setting of nonprogressive disease, despite prolonged and very high levels of viremia (extreme viremic nonprogressors [VNPs]). We then used immunogenetic, genomic, and transcriptomic tools and biomarkers to identify differences between those extreme groups as well as exploring genomic patterns previously defined in comparative studies of SIV infection in the pathogenic and the nonpathogenic models of rhesus macaques and sooty mangabeys, respectively (8–10). The study revealed characteristic biomarker and transcriptome patterns and highlighted several genes of relevance for the understanding of pathogenesis of HIV-1–induced immunosuppression.

Clinical and immunogenetic profiles. We identified 6 individuals that fulfilled strict clinical criteria of VNPs and had material available for analysis; plots of the infection course for each VNP individual are shown in Figure 1. We further identified 66 individuals who fulfilled the criteria of rapid progression and had materials available for study; the collective plot is shown in Figure 2. Notably, at the time of analysis, VNPs had higher levels of viral replication (set point HIV RNA, 5.4 log10 cp/ml; interquartile range [IQR], 5.1–5.5 log10 cp/ml]) compared with those of RPs (set point HIV RNA, 4.7 log10 cp/ml [IQR, 4.3–5.2 log10 cp/ml]). Transcriptome analysis also included 9 elite/viremic controllers (ECs) and 5 chronic progressors, as previously defined (5). Patient characteristics are detailed in Supplemental Table 1 (supplemental material available online with this article; doi: 10.1172/JCI45235DS1).

Figure 1 Individual viral loads and CD4+ T cell profiles of VNPs. Viremia is shown in red, and CD4+ T cell count is shown in blue.

Figure 2 Evolution of T cell count in individuals with a profile of rapid progression. The first CD4+ T cell count determination (black symbols) and the last CD4+ T cell count determination (red symbols) (connected by dashed lines) in individuals, defined by the progression to fewer than 350 CD4+ T cells (denoted by the gray area) in fewer than 3 years after seroconversion. Only values beyond the 3-month window after seroconversion are considered.

The HLA and KIR alleles were determined in all individuals, compared across clinical groups and compared to the allele frequencies of 1,609 participants of the SHCS (Supplemental Figure 1). Protective alleles were underrepresented, and risk alleles were more common in RPs compared with the general population. In contrast to HLA alleles, there was no depletion of protective KIR alleles or KIR/HLA combinations in RPs (Supplemental Figure 2). HLA alleles of VNPs are shown in Supplemental Table 2. To determine whether any common variants of very large effect could be implicated in mediating rapid progression, the study was completed with a genome-wide association across an approximately 500,000-loci study that included 66 RPs and 757 participants of the SHCS. No SNPs reached genome-wide significance (Supplemental Figure 3 and Supplemental Table 3), likely due to the limited power to detect anything other than very large effect sizes. A previous genome-wide association study of rapid progression (4) identified 8 SNPs that passed the study-wide false discovery rate (FDR) cutoff of 25%. These failed confirmation in our study (Supplemental Table 4).

Transcriptome analysis in CD4+ T cells. To investigate differences at the transcriptome level between RPs and VNPs, we performed microarray analysis on purified CD4+ cells from 27 RPs, 5 VNPs, 5 chronic progressors, and 9 ECs (Supplemental Table 1B). RPs, with and without transcriptome analysis, were similar with regard to CD4+ T cell counts and HIV viral load. The median CD4+ T cell counts at baseline were 440 cells/µl (IQR, 350–506 cells/µl) and 382 cells/µl (IQR, 315–497 cells/µl) for those with and without transcriptome analysis, respectively; the median baseline HIV viral loads were 4.8 cp/ml (IQR, 4.1–5.5 cp/ml) and 4.9 cp/ml (IQR, 4.3–5.1 cp/ml). During follow-up, the median CD4+ T cell counts were 263 cells/µl (IQR, 197–313 cells/µl) and 223 cells/µl (IQR, 186–299 cells/µl), and median HIV viral loads were 4.8 cp/ml (IQR, 4.3–5.4 cp/ml) and 5.0 cp/ml (IQR, 4.4–5.2 cp/ml) (P > 0.4 for all comparisons). Thirteen (20%) individuals had an AIDS-defining event within 3 years of seroconversion.

Principal component analysis identified 4 outliers that were excluded from further analysis. Various parameters were assessed as covariates (clinical center, gender, age, CD4+ T cell viability and laboratory date, and microarray chip batch); we retained chip batch as a statistically significant covariate. To contrast specific patient profiles, we applied a Bayesian approach to the analysis of gene expression (14). Analysis of RPs versus ECs identified 14 differentially expressed genes at a FDR-adjusted P value of less than 0.05. Interferon-stimulated genes (ISGs) are well known to be upregulated in patients with progressive HIV disease. Consistent with this knowledge, 6 ISGs, IFI44 (and its ligand IFI44L), MX1, EIF2AK2, IFI6, LY6E, TRIM22, were upregulated in RPs. Other upregulated genes included SYNCRIP that encodes a nuclear ribonucleoprotein (hnRNP-Q) associated with the APOB mRNA editosome complex that may modulate the posttranscriptional C to U RNA-editing PRIC285 that encodes a helicase acting as a transcriptional coactivator for a number of nuclear receptors, EPSTI1 and MRPS18B. Genes downregulated in RPs included TRK1, which encodes a kinase, and FOXJ2, a transcriptional activator. Next, we specifically searched genes uniquely associated with the VNP profile by contrasting this profile with that of RPs or chronic progressors. This analysis failed to identify FDR-adjusted differentially expressed genes.

Transcriptome analysis in CD8+ T cells. We also performed microarray analysis on purified CD8+ T cells derived from the same PBMC samples used for CD4+ T cell transcriptome analysis. Expression analysis was successfully completed for 25 RPs and 5 VNPs as well as 5 chronic progressors and 8 elite and viremic controllers (Supplemental Table 1B). No outliers were identified, and all samples progressed to further analysis. As above, we retained microarray chip batch as covariate in all definitive analyses. Using the same sensitive Bayesian approach as for the CD4+ T cell analysis (14), contrasting of RPs and ECs yielded 317 differentially expressed genes at a FDR-adjusted P value less than or equal to 0.05 (Supplemental Table 5). Among the 180 genes upregulated in RPs, prominent groups of genes included multiple members of the proteasome and interferon-induced immunoproteasome, ISGs, and cell cycle, cell division, and metabolic genes indicating cell proliferation (Supplemental Figure 4). No apparent mechanisms were deduced from the collective analysis of 137 genes downregulated in RPs by using EMBL Search Tool for the Retrieval of Interacting Genes/Proteins (STRING), Ingenuity Pathway Analysis 7.0 (IPA), and KEGG pathway analysis (see Methods). As for the CD4+ T cells, we specifically searched genes uniquely associated with the VNP profile by contrasting this profile with that of RPs or chronic progressors. Given power limitations, this analysis failed to identify FDR-adjusted differentially expressed genes. Thus, we proceeded to the analysis of specific pathways and of the genes identified in primate studies of nonpathogenic SIV infection (9, 10).

Analysis of genes of the interferon response. Recent publications (8–12) highlight a distinctive downregulation of the interferon response after SIV infection of natural host species, such as sooty mangabeys and African green monkeys. In contrast, SIV infection of the pathogenic models of rhesus or pig-tailed macaque is characterized by persistence of deregulated interferon responses. Consistent with the primate model of natural infection, we observed a lower level of expression of ISGs (see Methods for the specific ISGs) in CD8+ T cells of individuals with a VNP profile in comparison with that of individuals with a RP profile (Figure 3) (difference of the means, median –0.21 [IQR, –0.05 to –0.40]; paired t test, P = 0.014). However, these differences were not observed in CD4+ T cells, (difference of the means, 0.01 [IQR, 0.13 to –0.04]; P = 0.59). As expected, more profound differences in expression of ISGs were found in the comparison between ECs and RPs (median –0.36 [IQR, –0.13 to –0.59], P = 2.5 × 10–5 in CD4+ T cells, and median –0.33 [IQR, –0.21 to –0.59], P = 3.8 × 10–6 in CD8+ T cells) (Figure 3). The expression of SOCS1, involved in a negative feedback loop in the regulation of signal transduction through the JAK/STAT5 pathway, was higher in CD4+ T and CD8+ T cells of VNPs and ECs compared with that of RPs; the differences were statistically significant for the comparison of ECs and RPs in CD4+ T cells (P = 0.02) (Figure 3). This trend was not observed for a second regulator, ADAR.

Figure 3 Analysis of differential expression of ISGs. Consistent with the primate natural infection model, a representative set of ISGs (n = 29) had lower expression levels in CD8+ T cells from VNPs than in those from RPs. The box-and-whisker plot indicates that the differences are more pronounced for the comparisons of ECs and RPs. The horizontal bars indicate the median values, the boxes indicate the 25th to 75th percentiles, and whiskers indicate extremes. Each dot represents the difference in expression value for a given gene across groups. The profiles of the inhibitor of interferon response, SOCS1, and of ADAR are highlighted (blue and red, respectively).

Gene set enrichment analysis of human VNPs and SIV-infected sooty mangabeys. To examine whether the phenotype maintained by VNPs and natural host species was due to a shared, underlying molecular mechanism, we used gene set enrichment analysis (GSEA) of the human transcriptome data sets with gene sets derived from the analysis of sooty mangabeys and rhesus macaques (9). GSEA tests the relative position of a collection of genes (“query gene set”) within an independent, ranked data set (“reference gene set”). Because GSEA relies on an additive signal of multiple genes within a data set, it is less dependent on arbitrary cutoffs, such as fold change of specific P values, making its ability to detect an underlying process within transcriptome data potentially more sensitive than a “single-gene” approach using traditional statistics. The use of rank data rather than absolute intensity measurements in GSEA also affords greater flexibility to make comparisons between diverse gene-expression data (i.e., between tissues, species, or array platforms) (15).

As presented in Table 1 and in Supplemental Figure 5, the query set of ISGs identified as differentially expressed in the rhesus macaque was associated with enrichment in human RPs, although the P values were only consistent with a statistical trend. The CD8+ T cell expression data was particularly enriched for the ISGs; the data set comprised 15,879 nonredundant genes, and the lowest-ranked ISG was at position 10,562, well above the phenotype threshold at position 7,556, below which genes demonstrated higher expression in VNPs, and 12 out of 20 queried ISGs were higher than position 14,500 (Supplemental Figure 5). Genes found to be correlated with immune activation in rhesus macaques were also enriched in the RP phenotype in humans in both CD4 and CD8+ T cell data (Table 1 and Supplemental Figure 5). The enrichment of immune activation genes in RPs would indicate that VNPs have reduced cellular activation/proliferation relative to RPs. Taken together, these data suggest that VNPs, at least at the transcriptional level, are able to reduce the chronic immune activation seen in pathogenic HIV/SIV infection and that this attenuation largely overlaps with comparisons between sooty mangabeys and rhesus macaques. Because the human VNP and RP samples were obtained from the postacute phase of infection, we reasoned that genes found to be differentially expressed between sooty mangabeys and rhesus macaques during chronic infection may be enriched in the VNP phenotype. When we performed GSEA using genes found to be significantly higher in sooty mangabeys than rhesus macaques during chronic infection against the human data sets, we found that there was no significant enrichment in either phenotype in CD8+ T cells, but that there was significant enrichment in the VNP phenotype of CD4+ T cells (Table 1). The enrichment was largely driven by a single gene, SV2A, that ranked extremely high in the VNP phenotype. Taken together, these results suggest that sooty mangabeys and VNPs share some similarities in expression during chronic SIV/HIV infection; however, these similarities were not statistically significant.

Table 1 Analysis of gene sets of the primate model

Detailed analysis of genes identified in nonpathogenic primate models of natural infection. We extended the above analysis to examine in detail a list of genes reported by Bosinger et al. (9). We used a heuristic approach to inform this list (see Methods) by assessing (a) the consistency and direction of the association (downregulation or upregulation) between the primate model and the human expression profile, (b) the general correlation between CD4+ T cell and CD8+ T cell observations, and (c) the statistical support for the different associations in this subanalysis. Six genes fulfilled the criteria; genes CASP1, CD38, LAG3, and TNFSF13B presented lower expression levels in VNPs and in the nonpathogenic animal model, and SOCS1 and EEF1D presented greater expression levels in VNPs and in the nonpathogenic animal model of infection. The short list of genes was constituted into a signature to be evaluated in an independent set of data. For this, we used the large data set of CD4+ T cell expression (16) to assess the association of the signature genes with viral load and with progression of immunosuppression (as defined by time to fewer than 350 CD4+ T cells/µl). In unadjusted regression, the following genes showed statistically significant association with time to progression to fewer than 350 CD4+ T cells/µl: CASP1, LAG3, CD38, TNFSF13B, and EEF1D (Figure 4). A multigene model explained 19.5% of the variance in disease progression (P = 0.0003). Inclusion of viral load in the model improved the proportion of variance explained to 26% (P = 4.8 × 10–7). However, there was significant colinerarity with viral load and, after its inclusion in the model, only EEF1D remained as an independent variable (P = 0.013).

Figure 4 Analysis of the candidate VNP signature in an independent CD4+ T cell expression data set. The signature associated with the VNP profile upon transcriptome analysis in humans and nonhumans was tested in an independent validation set of 153 individuals, contributing CD4+ T cell expression data across all levels of viral set point after seroconversion. Correlations with individual gene expression levels and viral set point after seroconversion are shown in the 6 top panels. Correlations with disease progression, as indicated by time to CD4+ T cell count depletion to fewer than 350 cells/µl, are shown in the 6 bottom panels. Multiple probes for 1 gene are shown in different colors: orange/red is used for genes differentially upregulated in RPs, and blue/light blue is used for genes differentially upregulated in the VNPs. Where there are 2 P values, the first value represents the red/blue lines, and the second value represents the orange/light blue lines. Each dot represents an individual. The regression lines from the linear models are shown.

Association of soluble CD14 levels with clinical groups. To further assess whether the differences between RPs and VNPs reflected differences in mechanisms of pathogenesis, we assessed plasma levels of soluble CD14 (sCD14), which is produced by monocytes on becoming activated by LPS. Thus, plasma sCD14 levels reflect the host response to translocated bacterial products and are a significant independent predictor of mortality in HIV infection (17, 18). We analyzed samples from 24 RPs and 4 VNPs collected within 3 years after seroconversion. To contextualize these data, we measured plasma sCD14 levels in healthy volunteers and from chronic progressors. sCD14 levels were significantly higher in the plasma samples from RPs than in samples from chronic progressors, healthy donors, and for 3 out of 4 VNP samples analyzed (median 6,235 ng/ml [IQR, 5,069–8,808 ng/ml], median 6,065 ng/ml [IQR, 4,973–7,043 ng/ml], median 4,516 ng/ml [IQR, 3,972–5,304 ng/ml], and median 4,852 ng/ml [IQR, 4,069–8,612 ng/ml]); the differences between RPs versus healthy controls and chronic progressors versus healthy controls were significant (P < 0.0001) (Figure 5A). Additional plasma samples of the fourth VNP were consistently elevated. There was a trend toward increasing levels of sCD14 for individuals sampled at the time of advanced immunosuppression, with CD4+ T cell counts of below 350 cells/µl (Figure 5B).

Figure 5 Analysis of sCD14 plasma levels. (A) sCD14 levels were measured during the 3-year period after seroconversion and/or transcriptome analysis in RPs (n = 24) and VNPs (n = 4). Chronic progressors (CP; n = 39) and healthy donors (HD; n = 38) contributed reference data. The gray line represents the median values. (B) RP sCD14 levels were higher at lower CD4+ T cell counts. The gray line represents the LOWESS curve fitted to the sample population. Each dot represents an individual

The current study defines 2 presentations of HIV infection that share a similar level of high viral replication but differ in the degree of immunological damage and in the pattern of clinical evolution, i.e., RPs and VNPs. The proportion of individuals with rapidly progressive disease was estimated in the SHCS (19). In this nationwide and representative cohort, 7.9% of HIV-infected individuals with a known seroconversion date fulfilled the criteria of RPs. Severity of the disease, rapid initiation of treatment, and the need for precise knowledge of the seroconversion window hampered recruitment of RPs into clinical cohorts and research protocols in the past. VNPs constitute a group of individuals that sustain prolonged periods of high viral load, in the range of 100,000 copies/ml, while maintaining stable CD4+ T cell counts. VNPs represent a very uncommon pattern of disease progression; prevalence estimates in the SHCS indicate that only 0.1% of HIV-infected individuals would fulfill the strict definition of VNPs used in the current work. However, the selected individuals likely represent the extreme of the distribution of VNPs, and relaxed criteria compared with those used in the present study will lead to different estimates of frequency.

The various genomic analyses in this study associate rapid progression with an enrichment for HLA alleles linked to adverse prognosis and a depletion of protective alleles. This pattern validates the phenotypic set of criteria elaborated to define rapid progression. In contrast, we found no association of the RP cohort with KIR alleles or KIR/HLA combinations previously related to disease progression or viremia (20). The genome-wide association study was conducted to exclude a major impact of common variants and to assess the candidates from a previous study of similar power (4) that could not be validated here. The transcriptome profile did confirm the deregulation of the ISGs in CD8+ T cells in RPs, as previously documented for CD4+ T cells (16, 21, 22) and in lymphatic tissue (23). It also identified a characteristic pattern of upregulation in CD8+ T cells of RPs for genes involved in cell proliferation and cell division as well as in the immunoproteasome. RPs shared a number of features with the chronic SIV infection of rhesus macaques, in particular the prominent expression of a ISG and of immune activation markers. The absence of persistent immune activation during chronic SIV infection is a key characteristic of natural host species, such as the sooty mangabeys (24), and the presence of proliferation/activation markers on CD4+ and CD8+ T cells is an accurate predictor of disease in HIV-infected individuals (25). The immune activation gene set assessed in the present study was originally identified as being correlated with CD8+ T cells expressing the activation marker Ki67 in SIV-infected rhesus macaques but was not expressed in SIV-infected sooty mangabeys (9).

More remarkable were the observations in VNPs. While the study did not have the power to allow a discovery that was not a priori, it permitted the assessment of a number of characteristics that have been previously described in SIV-infected sooty mangabeys. Individuals with the VNP profile display a limited deregulation of the ISG when compared with RPs, particularly in CD8+ T cells. It should be stressed that these differences were present despite greater levels of viremia among VNPs than in RPs. In addition, to assess whether VNPs demonstrated lower immune activation and/or chronic interferon responses relative to RPs, we ranked the CD4+ and CD8+ expression data sets according to the significance value determined by the Bayesian analysis and used GSEA to test the relative position of ISGs/immune activation genes and genes differentially expressed in SIV-infected sooty mangabeys and rhesus macaques. This analysis supported the notion that the human profile of VNPs shares common features, at the transcriptome level, with the nonpathogenic model of SIV infection in the natural host. Reduced ISG expression is a consistent feature of natural host infection and not due to temporal fluctuation (10). Although the observation of reduced ISGs in VNPs in the current study is cross-sectional, it was consistent in showing ISG reduction relative to RPs. How differences in transcription levels of the ISGs translates into protein and the mechanisms of regulation should be the focus of future research (26).

We investigated in detail a set of genes identified through a comparative analysis of human and nonhuman primate transcriptome data; CASP1, CD38, LAG3, and TNFSF13B were upregulated in rhesus macaque and in human RPs; SOCS1 and EEF1D were upregulated in sooty mangabeys and in human VNPs. The shared expression pattern between VNPs and sooty mangabeys supports their role in lentiviral pathogenesis. Caspase-1 precursor (CASP1) is a well-known intermediate of the inflammatory processes and apoptosis. The lymphocyte differentiation antigen CD38 is associated with immune exhaustion during immune activation and with adverse prognosis (27–29). LAG3 negatively regulates the expansion of activated T cells, and T cell homeostasis and is required for maximal regulatory T cell function (30) and has been demonstrated to associate with immune dysfunction/exhaustion of CD8+ T cells in LCMV infection (31). Tumor necrosis factor ligand superfamily member 13B (TNFSF13B) is a receptor involved in the stimulation of B and T cell function and the regulation of humoral immunity. Suppressor of cytokine signaling (SOCS1) is involved in a negative feedback loop in the regulation of cytokine signal transduction signaled through the JAK/STAT5 pathway. Although SOCS1 was downregulated in RPs compared with ECs and VNPs, its expression levels did not exhibit a significant association with viral set point or disease progression in the validation data set of CD4+ T cell transcription data (16).

We completed the study by the analysis of a biomarker of compromised intestinal mucosal barrier, the monocyte-expressed LPS receptor sCD14 (18). Our data show higher plasma levels among RPs, in particular during advanced immunosuppression, than for other clinical progression groups. Although only 4 VNPs could be tested, 3 presented low sCD14 plasma levels, a pattern fitting other observations of lesser immunopathogenesis in these individuals. The transcriptome and biomarker data thus complement the work of Choudhary et al. (7) on VNPs that presented less extreme viral loads. They identified a lower percentage of activated HLA-DR+CD38+CD4+ and CD8+ T cells and lower levels of proliferating Ki67-expressing CD4+ and CD8+ T cells in VNPs compared with those of progressors. In contrast, viral isolates from VNPs and progressors replicated to similar levels and shared the capacity to deplete CD4+ thymocytes or CD4+ T cells in secondary lymphoid tissue and were equally cytopathic.

Future studies should extend analyses to plasmacytoid dendritic cells, as they are key activators of the immune system in HIV and SIV infection. Assessment of this cell population is limited by the low percentage of these cells in fresh blood, in particular, in the infected individual (32). The study has limited power due to the rarity of the study phenotypes and inherent limitations in recruitment. However, this work highlights the importance of 2 poorly understood clinical patterns of disease progression that have been minimally studied in the past and provides working definitions that should help identifying additional individuals to allow greater power in future genomic and functional studies. In addition, this report of a strong phenotypic similarity between nonpathogenic SIV infection of sooty mangabeys and a subset of HIV-infected individuals emphasizes the importance of studying natural SIV infection as a model to better understand HIV/AIDS pathogenesis.

Ethics statement. All participating centers provided local institutional review board approval for genetic analysis, and each participant provided informed consent for genetic testing. The Institutional Review Boards are Comission d’Ethique de la Recherche Clinique, Faculté de Médecine, Université de Lausanne, Lausanne, Switzerland, and Comitè Etic d’Investigació Clínica, Hospital Germans Trias i Pujol, Badalona, Spain.

Patients and definition of clinical profiles. Study participants were followed in the SHCS ( www.shcs.ch) or at the HIVACAT. The selection criteria for RPs included a HIV seroconversion window of less than 1 year with documented negative and positive serology and either of the following possibilities: (a) more than 2 CD4+ T cell counts below 350 cells/µl within 3 years of seroconversion and no subsequent rise of CD4+ T cells above 350/µl in the absence of combination antiretroviral therapy or (b) beginning antiretroviral therapy within 3 years of seroconversion and a CD4+ T cell count within 1 month of starting antiretroviral therapy of less than 350/µl. CD4+ T cell values in the first 6 months after seroconversion were excluded to avoid the CD4+ T cell nadir during acute HIV infection.

The selection criteria of VNPs included more than 3 years of follow-up, median HIV viremia from more than 3 measurements of more than 100,000 cp/ml, HIV viremia consistently above 10,000 cp/ml, a CD4+ T cell count above 350/µl, and no HIV treatment during follow-up.

Study candidates were identified by a standardized database search. Subsequently, the individual CD4+ T cell profiles of all candidates were visually inspected before final inclusion. In addition to individuals fulfilling the definition of RP or of VNP, the study included 9 ECs as reference group.

Immunogenetic and genome-wide association analyses. High-resolution geno­typing of HLA-A, HLA-B, HLA-Cw, and DRB1 alleles was performed by sequence-based typing methods. KIR gene typing was performed by a sequence-specific oligonucleotide probe using the Luminex microbead technology. For genome-wide association analysis, participants were genotyped using Illumina BeadChips Human660W-Quad. For quality control purposes, SNPs were removed based on their absence (locus absence >5%), minor allele frequency (>1%), and Hardy-Weinberg Equilibrium deviation (P < 1 × 10–6). Participants were filtered based on call rate, gender check (heterozygosity testing), cryptic relatedness, and population structure (33).

Cell isolation, RNA extraction, and transcriptome profiling. For transcriptome analysis, we included all RPs (n = 27) for whom viable cells were available from the time of seroconversion (>6 months to 3 years from acute infection) and before initiation of antiretroviral treatment. Samples from all VNPs were included. The CD4 and viral load values at the time of transcriptome analysis are presented in Supplemental Table 1. CD4+ and CD8+ T cells were positively selected from frozen PBMCs (median time of cryopreservation was 1,485 [IQR, 821–2,558] days) using magnetically labeled CD4+ or CD8+ microbeads and subsequent column purification according to the manufacturer’s protocol (Miltenyi Biotec). The median CD4+ T cell purity, verified by flow cytometry, was 96.8% (range, 93.9%–98.9%), whereas the median CD8+ T cell purity was 88.8% (range, 84.8%–92.1%). CD4+ and CD8+ T cell viability was assessed by the trypan blue dye exclusion method using the Vi-CELL (Beckman Coulter). Total RNA was extracted from purified CD4+ and CD8+ T cells using the mirVana miRNA Isolation Kit (Ambion) according to the manufacturer’s protocol for total RNA extraction. The amount of RNA was estimated by spectrophotometry using the Nanodrop 1000 (Thermo Fisher). RNA quality was determined by the Agilent RNA 6000 Pico Kit on an Agilent 2100 Bioanalyzer. Samples were collected between 1993 and 2008 and investigated in 2009. The median of CD4+ T cell viability for samples that were successfully analyzed was 79% (IQR, 64%–87%). The median of CD8+ T cell viability for samples that were successfully analyzed was 82% (IQR, 74%–87%). Viability was minimally dependent on time of cryopreservation and more dependent on collection center. These covariates were assessed in the analyses. Target preparation was performed starting from 200 ng total RNA using the Illumina TotalPrep-96 RNA Amplification kit (Ambion). cDNA and cRNA were purified using the MagMAX Express Magnetic Particle Processor (Applied Biosystems). cRNA quality was assessed by capillary electrophoresis on the Agilent 2100 Bioanalyzer. Hybridization on HumanHT-12 v3 Expression BeadChips (Illumina) was carried out according to the manufacturer’s instructions.

Transcriptome data analysis. Bead summary data were the output from Illumina’s BeadStudio software without background correction, as this has previously been shown to have detrimental effects (34). Genes declared as nonexpressed (P > 0.01) were excluded from analysis. Data preprocessing, including quantile normalization and log2 transformation was completed in the Partek Genomics Suite package (Partek Inc.). Outliers were identified based on principal component analysis using 3 standard deviations as the cutoff for inclusion. For the differential expression analysis, we applied an empirical Bayes analysis approach, as implemented in the “limma” package of the R programming language, to model the variation profiles of all genes and used that information as prior knowledge to better estimate the variance of each gene expression (14).

The selected analysis included the following genes: APOBEC3H, BST2, EIF2AK2, IFI27, IFI35, IFI44, IFIH1, IFITM1, IFITM3, IRF1, IRF9, ISG15, JAK1, JAK2, MX1, MX2, OAS3, STAT2, TAP, TRIM22, TYK2, ZBP1, APOBEC3F, APOBEC3G, IFI6, IFIT1, IFIT3, OAS1, OAS2, OASL, PSMB8, PTPN2, RNASEL, STAT1, and TRIM5 as previously described (16).

Signature analysis and validation. Because of the rarity of individuals with a VNP profile, we used a heuristic approach to assessing possible genetic markers associated with the clinical profile. This approach included the analysis of a preliminary signature, including genes identified as possibly associated with the VNP profile upon transcriptome analysis because of concordant signals in both CD4+ and CD8+ T cells as well as genes identified as potentially relevant in studies of SIV infection in the natural host: sooty mangabey and African green monkey. The signature was tested in an independent validation set of 153 individuals from a previous transcriptome analysis (16).

Pathway and network analyses. STRING ( http://string.embl.de/) was used to identify known and predicted interactions (derived from 4 sources: genomic context, high-throughput experiments, coexpression, and previous knowledge). IPA ( http://www.ingenuity.com/) and KEGG ( http://www.genome.jp/kegg/pathway.html) were used for the analysis of pathway enrichment.

GSEA and gene set selection. The GSEA algorithm uses a Kolgorimov-Smirnov statistic to determine the significance of distribution of a set of genes within a larger, ranked data set (35). To evaluate the enrichment of SIV-inducible genes in the rhesus macaques and sooty mangabeys and in our human data set, we performed GSEA as follows: transcriptome data from VNPs and RPs were ranked according to their calculated Bayesian statistic; genes in which the mean was greater in VNPs were classified as positive, and genes with a greater mean in RPs were classified as negative. The data were ranked by the inverse Bayesian P value, resulting in a data set in which the most significant genes, overexpressed in VNPs, were listed at the top, and the most significant genes, overexpressed in RPs, were listing at the bottom. We next defined discrete query gene sets (Supplemental Table 6) from a large microarray data set, detailing longitudinal SIV infection in rhesus macaques, which develop disease, and sooty mangabeys, a nonpathogenic, natural host species, described previously (9). The ISG set comprised genes known to be regulated by type I interferon that were found to be differentially expressed in SIVmac239-infected rhesus macaques after 180 days of infection. The immune activation gene set was defined by multiple criteria: significant correlation of expression with lymphocyte activation assessed by circulating levels of Ki67+CD8+ T cells in SIVmac239-infected macaques (FDR = 0.0106), significant induction of expression assessed by ANOVA (FDR = 0.0075), a minimum of 2-fold upregulation in macaques at 1 or more time points, and expression in sooty mangabeys not exceeding 1.5× at any interval. To determine whether gene expression maintained chronically in VNPs shared similarity with that of sooty mangabeys, we defined the sooty mangabey chronic query gene set as follows: robust multiarray average log10 intensity values from baseline samples were subtracted from chronic time points for individual animals of both species, and 2-sample t test was performed on the subsequent fold-change data; genes with a higher average fold change in sooty mangabeys relative to that in rhesus macaques were ranked according to P value, with the top 50 most significantly overexpressed genes selected for gene set inclusion. GSEA was performed using the desktop module available from the Broad Institute ( www.broadinstitute.org/gsea/). GSEA was performed on the pre-ranked human data sets using 1,000 permutations, median collapse of duplicates, and random seeding.

Analysis of sCD14 levels. sCD14 levels were quantified in plasma samples using a commercially available ELISA assay (Diaclone). Plasma samples were diluted (1:50 or 1:100) and tested in duplicate. Plasma aliquots were collected either in EDTA (n = 55) or BD Vacutainer CPT Cell Preparation Tube with Sodium Citrate (CPT) tubes (n = 12). The CPT tubes contained a nonnegligible amount of molar sodium citrate solution (1 ml for the tubes, 8 ml draw capacity) and polysaccharide/sodium diatrizoate solution (FICOLL Hypaque solution; 2 ml for the tubes, 8 ml draw capacity), therefore samples collected with these tubes were considered to be diluted 1.44 times, and values were corrected accordingly.

Statistics. Comparisons of clinical and demographic characteristics used Fisher’s exact tests for dichotomous variables and the Wilcoxon rank-sum test for continuous variables (STATA SE, release 11; StataCorp LP). In genome-wide association studies, association between genotype and phenotype (rapid progression) was tested using logistic regression, including top population principal components as covariates to correct for stratification. Genome-wide significance was assessed, using a cutoff of P < 5 × 10–8 to correct for multiple tests. In transcriptome analysis, we used a FDR method (36) to control for multiple testing. Probes selected for further analysis had an FDR-adjusted P value of less than 0.05. Statistical analyses dedicated to GSEA are detailed in the relevant section (see GSEA and gene set selection). Multiple regression analyses and graphical representations were performed by using the statistics package R ( www.r-project.org).

Microarray data accession number. Microarray results have been deposited in the Gene Expression Omnibus database; the accession number is GSE28128.

View Supplemental data

View Supplemental Excel file 1

View Supplemental Excel file 2

View Supplemental Excel file 3

This work has been financed in the framework of the SHCS and supported by the Swiss National Science Foundation (SNF) (grant no. 33CSC0-108787, project no. 587), by SNF grant (310000-110012, to A. Telenti), by the HIVACAT, and, in part, by the Ragon Institute and the Spanish AIDS network (RD06/0006). S.E. Bosinger is a recipient of a Canadian Institutes of Health Research HIV/AIDS Research Initiative Fellowship (HFE-85139). We thank S. Colombo, M. Rickenbach, I. Fernández, and J. Puig for study coordination; Y. Vallet for software support; John Werry for the IA gene set; and E. Grau, R. Ayen, and T. González for technical assistance. The members of the SHCS are M. Battegay, E. Bernasconi, J. Böni, H.C. Bucher, Ph. Bürgisser, A. Calmy, S. Cattacin, M. Cavassini, R. Dubs, M. Egger, L. Elzi, P. Erb, M. Fischer, M. Flepp, A. Fontana, P. Francioli (President of the SHCS, Centre Hospitalier Universitaire Vaudois, CH-1011 Lausanne), H. Furrer (Chairman of the Clinical and Laboratory Committee), C. Fux, M. Gorgievski, H. Günthard (Chairman of the Scientific Board), H. Hirsch, B. Hirschel, I. Hösli, Ch. Kahlert, L. Kaiser, U. Karrer, C. Kind, Th. Klimkait, B. Ledergerber, G. Martinetti, B. Martinez, N. Müller, D. Nadal, M. Opravil, F. Paccaud, G. Pantaleo, A. Rauch, S. Regenass, M. Rickenbach (Head of Data Center), C. Rudin (Chairman of the Mother and Child Substudy), P. Schmid, D. Schultze, J. Schüpbach, R. Speck, P. Taffé, P. Tarr, A. Telenti, A. Trkola, P. Vernazza, R. Weber, and S. Yerly. The members of the HIVACAT involved in this study are B. Clotet, J. Dalmau, I. Erkizia, J.M. Gatell, C. Ligero, M. López-Diéguez, C. Manzardo, J. Martinez-Picado, and J.M. Miro.

Conflict of interest: The authors have declared that no conflict of interest exists.

Citation for this article: J Clin Invest doi:10.1172/JCI45235.

Deeks SG, Walker BD. Human immunodeficiency virus controllers: mechanisms of durable virus control in the absence of antiretroviral therapy. Immunity. 2007;27(3):406–416. Migueles SA, Connors M. Long-term nonprogressive disease among untreated HIV-infected individuals: clinical implications f understanding immune control of HIV. JAMA. 2010;304(2):194–201. International HIV Controllers Study, et al. The major genetic determinants of HIV-1 control affect HLA class I peptide presentation. Science. 2010;330(6010):1551–1557. Le Clerc S, et al. Genomewide association study of a rapid progression cohort identifies new susceptibility alleles for AIDS (ANRS Genomewide Association Study 03). J Infect Dis. 2009;200(8):1194–1201. Casado C, et al. Host and viral genetic correlates of clinical definitions of HIV-1 disease progression. PLoS ONE. 2010;5(6):e11079. Dalmau J, et al. Contribution of immunological and virological factors to extremely severe primary HIV type 1 infection. Clin Infect Dis. 2009;48(2):229–238. Choudhary SK, et al. Low immune activation despite high levels of pathogenic human immunodeficiency virus type 1 results in long-term asymptomatic disease. J Virol. 2007;81(16):8838–8842. Lederer S, et al. Transcriptional profiling in pathogenic and non-pathogenic SIV infections reveals significant distinctions in kinetics and tissue compartmentalization. PLoS Pathog. 2009;5(2):e1000296. Bosinger SE, et al. Global genomic analysis reveals rapid control of a robust innate response in SIV-infected sooty mangabeys. J Clin Invest. 2009;119(12):3556–3572. Jacquelin B, et al. Nonpathogenic SIV infection of African green monkeys induces a strong but rapidly controlled type I IFN response. J Clin Invest. 2009;119(12):3544–3555. Durudas A, Milush JM, Chen HL, Engram JC, Silvestri G, Sodora DL. Elevated levels of innate immune modulators in lymph nodes and blood are associated with more-rapid disease progression in simian immunodeficiency virus-infected monkeys. J Virol. 2009;83(23):12229–12240. Taaffe J, et al. A five-year longitudinal analysis of sooty mangabeys naturally infected with simian immunodeficiency virus reveals a slow but progressive decline in CD4+ T-cell count whose magnitude is not predicted by viral load or immune activation. J Virol. 2010;84(11):5476–5484. Moir S, Chun TW, Fauci AS. Pathogenic mechanisms of HIV disease. Annu Rev Pathol. 2011;6:223–248. Smyth GK. Linear models and empirical bayes methods for assessing differential expression in microarray experiments. Stat Appl Genet Mol Biol. 2004;3:Article3. Haining WN, Wherry EJ. Integrating genomic signatures for immunologic discovery. Immunity. 2010;32(2):152–161. Rotger M, et al. Genome-wide mRNA expression correlates of viral control in CD4+ T-cells from HIV-1-infected individuals. PLoS Pathog. 2010;6(2):e1000781. Brenchley JM, et al. Microbial translocation is a cause of systemic immune activation in chronic HIV infection. Nat Med. 2006;12(12):1365–1371. Sandler NG, et al. Plasma levels of soluble CD14 independently predict mortality in HIV infection. J Infect Dis. 2011;203(6):780–790. Swiss HIV Cohort Study, et al. Cohort profile: the Swiss HIV Cohort study. Int J Epidemiol. 2010;39(5):1179–1189. Carrington M, Martin MP, van Bergen J. KIR-HLA intercourse in HIV disease. Trends Microbiol. 2008;16(12):620–627. Hyrcza MD, et al. Distinct transcriptional profiles in ex vivo CD4+ and CD8+ T cells are established early in human immunodeficiency virus type 1 infection and are characterized by a chronic interferon response as well as extensive transcriptional changes in CD8+ T cells. J Virol. 2007;81(7):3477–3486. Sedaghat AR, et al. Chronic CD4+ T-cell activation and depletion in human immunodeficiency virus type 1 infection: type I interferon-mediated disruption of T-cell dynamics. J Virol. 2008;82(4):1870–1883. Smith AJ, Li Q, Wietgrefe SW, Schacker TW, Reilly CS, Haase AT. Host genes associated with HIV-1 replication in lymphatic tissue. J Immunol. 2010;185(9):5417–5424. Silvestri G, et al. Nonpathogenic SIV infection of sooty mangabeys is characterized by limited bystander immunopathology despite chronic high-level viremia. Immunity. 2003;18(3):441–452. Giorgi JV, et al. Shorter survival in advanced human immunodeficiency virus type 1 infection is more closely associated with T lymphocyte activation than with plasma virus burden or virus chemokine coreceptor usage. J Infect Dis. 1999;179(4):859–870. Witwer KW, Sisk JM, Gama L, Clements JE. Micro­RNA regulation of IFN-beta protein expression: rapid and sensitive modulation of the innate immune response. J Immunol. 2010;184(5):2369–2376. Biancotto A, et al. Abnormal activation and cytokine spectra in lymph nodes of people chronically infected with HIV-1. Blood. 2007;109(10):4272–4279. Onlamoon N, Tabprasit S, Suwanagool S, Louisirirotchanakul S, Ansari AA, Pattanapanyasat K. Studies on the potential use of CD38 expression as a marker for the efficacy of anti-retroviral therapy in HIV-1-infected patients in Thailand. Virology. 2005;341(2):238–247. Li Q, et al. Microarray analysis of lymphatic tissue reveals stage-specific, gene expression signatures in HIV-1 infection. J Immunol. 2009;183(3):1975–1982. Workman CJ, Dugger KJ, Vignali DA. Cutting edge: molecular analysis of the negative regulatory function of lymphocyte activation gene-3. J Immunol. 2002;169(10):5392–5395. Blackburn SD, et al. Coregulation of CD8+ T cell exhaustion by multiple inhibitory receptors during chronic viral infection. Nat Immunol. 2009;10(1):29–37. Brown KN, Wijewardana V, Liu X, Barratt-Boyes SM. Rapid influx and death of plasmacytoid dendritic cells in lymph nodes mediate depletion in acute simian immunodeficiency virus infection. PLoS Pathog. 2009;5(5):e1000413. Price AL, Patterson NJ, Plenge RM, Weinblatt ME, Shadick NA, Reich D. Principal components analysis corrects for stratification in genome-wide association studies. Nat Genet. 2006;38(8):904–909. Dunning MJ, Barbosa-Morais NL, Lynch AG, Tavaré S, Ritchie ME. Statistical issues in the analysis of Illumina data. BMC Bioinformatics. 2008;9:85. Subramanian A, et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A. 2005;102(43):15545–15550. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Statist Soc B. 1995;57(1):289–289.

View the original article here

read more

The histone trimethyllysine demethylase JMJD2A promotes cardiac hypertrophy in response to hypertrophic stimuli in mice


J Clin Invest. doi:10.1172/JCI46277.
Copyright © 2011, The American Society for Clinical Investigation. Qing-Jun Zhang1, Hou-Zao Chen2, Lin Wang1, De-Pei Liu2, Joseph A. Hill1 and Zhi-Ping Liu1

1Departments of Internal Medicine and Molecular Biology, University of Texas Southwestern Medical Center, Dallas, Texas, USA.
2National Laboratory of Medical Molecular Biology, Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China.

Address correspondence to: Zhi-Ping Liu, Departments of Internal Medicine and Molecular Biology, University of Texas Southwestern Medical Center, 5323 Harry Hines Blvd., Dallas, Texas 75390-9148, USA. Phone: 214.648.1485; Fax: 214.648.1450; E-mail: Zhi-Ping.Liu@utsouthwestern.edu.

Authorship note: Qing-Jun Zhang and Hou-Zao Chen contributed equally to this work.

Published May 9, 2011
Received for publication January 3, 2011, and accepted in revised form March 23, 2011.

Cardiac hypertrophy and failure are accompanied by a reprogramming of gene expression that involves transcription factors and chromatin remodeling enzymes. Little is known about the roles of histone methylation and demethylation in this process. To understand the role of JMJD2A, a histone trimethyl demethylase, in cardiac hypertrophy, we generated mouse lines with heart-specific Jmjd2a deletion (hKO) and overexpression (Jmjd2a-Tg). Jmjd2a hKO and Jmjd2a-Tg mice had no overt baseline phenotype, but did demonstrate altered responses to cardiac stresses. While inactivation of Jmjd2a resulted in an attenuated hypertrophic response to transverse aortic constriction–induced (TAC-induced) pressure overload, Jmjd2a-Tg mice displayed exacerbated cardiac hypertrophy. We identified four-and-a-half LIM domains 1 (FHL1), a key component of the mechanotransducer machinery in the heart, as a direct target of JMJD2A. JMJD2A bound to the FHL1 promoter in response to TAC, upregulated FHL1 expression, and downregulated H3K9 trimethylation. Upregulation of FHL1 by JMJD2A was mediated through SRF and myocardin and required its demethylase activity. The expression of JMJD2A was upregulated in human hypertrophic cardiomyopathy patients. Our studies reveal that JMJD2A promotes cardiac hypertrophy under pathological conditions and suggest what we believe to be a novel mechanism for JMJD2A in reprogramming of gene expression involved in cardiac hypertrophy.

Heart failure is a major public health problem and a leading cause of mortality in Western countries. It is the final common pathway for a wide spectrum of cardiovascular diseases including hypertension, coronary artery disease, myocardial ischemia, valve abnormalities, and inherited or acquired cardiomyopathies. There is no single unifying mechanism that explains the development and progression of heart failure. Nevertheless, heart failure is frequently preceded by left ventricular hypertrophy (LVH), and LVH is a major predictor for progressive heart disease and an adverse prognosis (1).

Myocardial hypertrophy is an adaptive response of cardiac muscle to altered conditions caused by a large number of physiological and pathological stimuli (2). Cardiac hypertrophy, even though it is physiological or compensatory in the beginning, can become pathological or maladaptive, leading to heart failure if left untreated. Pathological hypertrophy and heart failure are accompanied by a reprogramming of cardiac gene expression and activation of “fetal” genes that correlate with loss of cardiac functions (3–5). Therefore, elucidation of the mechanism or mechanisms involved in the reprogramming of cardiac gene expression in hypertrophic and failing hearts could provide us information for designing drugs to manipulate and normalize cardiac gene expression in a “transcriptional therapy” for cardiac hypertrophy and failure (6).

The transcriptional regulation of gene expression involves not only transcription factors but also posttranslational modifications of the histone tails. Histone modifications can alter chromatin conformations that allow the accessibility of transcription factors and the recruitment of the transcriptional complex on the promoter/enhancer and transcriptional regions of the genes. There are at least 7 distinct types of modifications found on histones, including methylation and demethylation (7–9). Different histone methylation patterns can provide specialized binding surfaces that recruit protein complexes containing chromatin remodeling and transcriptional activation/repression activity. Multiple histone methyltransferase and demethylases have been identified. Studies have suggested their roles in multiple aspects of development across various species and in diseases such as cancer and neurological disorders (10–18). However, the role of histone demethylases in the reprogramming of cardiac gene expression in hypertrophic and failing hearts remains elusive. A genome-wide histone methylation profile for heart failure showed a differential marking of trimethylation of H3K4 and H3K9 (H3K4me3 and H3K9me3) in cardiomyocytes during development of heart failure in both animal models and human (19), suggesting that the enzymes responsible for methylation and demethylation of H3K4me3 and H3K9me3 may play a role in cardiac hypertrophy and heart failure.

JMJD2A/KDM4A is a member of the JmjC domain–containing family JMJD2 of histone demethylases, including JMJD2B, JMJD2C, JMJD2D, and JMJD1B (12, 14, 18). JMJD2 proteins are lysine trimethyl–specific histone demethylases that catalyze the demethylation of trimethylated H3K9 (H3K9me3) and H3K36 (H3K36me3). Genome-wide studies show that H3K9me3 is enriched in heterochromatin (15), which predicts that the H3K9me3 demethylase activities of JMJD2 proteins may act as transcriptional activators. Consistent with this model, JMJD2A and JMJD2D were reported to be coactivators for the androgen receptor (20). On the other hand, JMJD2A has also been reported to repress N-CoR–regulated transcription in cell-culture studies (13, 21, 22). These results suggest that the function of JMJD2A may be context dependent and emphasize the need for direct analysis of JMJD2A function in specific physiological settings. Of relevance to the study presented here, it is notable that the roles of histone methylation/demethylation and the corresponding enzymes in the reprogramming of cardiac gene expression during cardiac hypertrophy and failure have not been elucidated.

To understand the role of JMJD2A in the reprogramming of gene expression in hypertrophic hearts, we generated conditional heart-specific Jmjd2a KO (Jmjd2a hKO) mice and mouse lines that overexpress JMJD2A in the postnatal heart (Jmjd2a-Tg). Our studies demonstrate a role for JMJD2A in promoting cardiac hypertrophy under pathologic conditions. We identified four-and-a-half LIM domains 1 (FHL1) as a target of JMJD2A. We have found that JMJD2A binds to the FHL1 promoter in response to hypertrophic stimuli and upregulates the transcription of FHL1. We have further observed that the transcriptional activity of JMJD2A is mediated through SRF/myocardin. JMJD2A promotes recruitment of SRF/myocardin to the SRF-targeted genes. JMJD2A is upregulated in human hypertrophic cardiomyopathy (HCM) patients. Taken together, our studies show that JMJD2A promotes cardiac hypertrophy and suggest what we believe is a novel mechanistic interplay among SRF, myocardin, and JMJD2A in the reprogramming of gene expression during hypertrophic remodeling.

Generation and characterization of Jmjd2a hKO. JMJD2A is expressed ubiquitously and higher in the heart, the skeletal muscle, and the liver in mice (Supplemental Figure 1; supplemental material available online with this article; doi: 10.1172/JCI46277DS1). To study the biological function of JMJD2A in the heart, we generated mice with conditional alleles for Jmjd2a and mice with deleted Jmjd2a in the heart using a-MHC–Cre (ref. 23 and Figure 1, A–D). The deletion of exon 3 in the Jmjd2afl/fl×a-MHC–Cre mouse (Jmjd2a hKO) resulted in a significant reduction of JMJD2A protein in the heart compared with control mice (Jmjd2afl/fl) (Figure 1D). The residual JMJD2A in hKO hearts may come from the contribution of noncardiomyocytes in the heart and/or incomplete deletion of Jmjd2a in the cardiomyocyte by the a-MHC–Cre transgene. Jmjd2a hKO mice are viable and no overt baseline cardiac phenotypes were observed for mice up to 6 months of age. The cardiac function of adult Jmjd2a hKO mice was indistinguishable from that of control mice as assayed by echocardiograph (data not shown).

Figure 1 Generation and assessment of a cardiac-specific Jmjd2a KO mouse model. (A) Strategy used to delete Jmjd2a in mouse cardiomyocytes and generate the Jmjd2a hKO model. Schematic protein structure of JMJD2A (14) and maps of the WT Jmjd2a locus, the targeting vector, the floxed allele, and the excised allele are shown. Exons are shown in bars. Exon 3 is flanked by loxP sites. (B) Southern blot of tail DNA from a WT (+/+) and a heterozygous (fl/+) mouse using 5' probe shown in A after digestion with BclI. (C) PCR genotyping on genomic DNA isolated from a WT (+/+), a heterozygous (fl/+), and a homozygous floxed (fl/fl) mouse using primers p1/p2 indicated in A. (D) Western blot demonstrating significantly reduced JMJD2A protein in the heart homogenate of Jmjd2a hKO mice at 4 weeks. (E) H&E stains of representative paraffin sections of control and hKO hearts 3 weeks after sham and TAC operation. (F) HW/BW ratios of control and hKO mice 3 weeks after sham and TAC operation (n = 10–14 per group). The HW/BW ratio after TAC operation was increased 79% compared with sham operation in control mice, whereas it was only increased 42% in hKO mice. (G) Myocyte cross-sectional areas from control and hKO mice (>200 myocytes per heart in randomly selected filed) were assessed (n = 6–8 per group). (H) mRNA transcripts of ANP, BNP, and myh7 in control and hKO mouse hearts after 3 weeks sham and TAC operation. RNAs were normalized to internal GAPDH expression and presented as the fold change relative to control sham samples (n = 4–6 per group). Values are mean ± SEM. *P < 0.05.

Jmjd2a hKO mice have attenuated hypertrophic responses. To determine whether Jmjd2a hKO mice have altered cardiac responses under pathological conditions, we subjected hKO and control (Jmjd2afl/fl) littermates to transverse aortic constriction (TAC), which causes pathological cardiac hypertrophy due to increased afterload (24). Three weeks after TAC, Jmjd2a hKO hearts were significantly smaller than those of control mice (Figure 1E). This is reflected in a significantly smaller increase in heart weight/body weight (HW/BW) ratio (Figure 1F) and cross-sectional cardiomyocyte areas (Figure 1G) in hKO mice compared with those of controls. hKO mice also showed better preserved cardiac function after TAC compared with TAC-operated control mice, as indicated by echocardiograph (Table 1). A blunted hypertrophic response in hKO mice was further indicated by significantly decreased expression of the “fetal gene” markers ANP, BNP, and myh7 in response to TAC-induced pressure overload compared with that of control mice (Figure 1H). To address the possible effect of Cre-transgene on the cardiac phenotype, we also subjected Jmjd2a+/+×a-MHC–Cre mice to sham and TAC operation and compared the phenotype with those of Jmjd2afl/fl littermates. No significant differences of hypertrophic responses were observed between the 2 genotypes (Supplemental Figure 2), suggesting that the hypertrophic response we observed in hKO mice is not due to the effect of the a-MHC–Cre transgene.

Table 1 Echocardiographic analysis of control and Jmjd2a hKO mice after sham and TAC operations

Overexpression of Jmjd2a in postnatal hearts exacerbates the hypertrophic response to TAC-induced hypertrophy. To determine whether gain of function of JMJD2A promotes cardiac hypertrophy, we generated transgenic mice in FVB background with JMJD2A expressed specifically in the postnatal heart (Figure 2A). Four founder mice were obtained. Two transgenic lines (Tg-A and Tg-B) with modest JMJD2A expression shown by Western blot analysis (Figure 2A) were established. The results presented here are those obtained with the transgenic line B. In the absence of stress, Jmjd2a-Tg mice displayed normal cardiac morphologies and functions similar to those of WT littermates (data not shown). Three weeks after TAC, Jmjd2a-Tg mice manifested an exacerbated hypertrophic response (Figure 2B) compared with that of WT littermates with significantly increased HW/BW ratio (Figure 2C). Analysis of cross-sectional cardiomyocyte areas also revealed a further increase in cardiomyocyte size after TAC in Tg mice compared with that of WT littermates (Figure 2D) and more fibrosis (data not shown). TAC Jmjd2a-Tg mice also showed significant loss of cardiac function compared with TAC WT mice (Table 2). An increase in hypertrophic response in TAC Jmjd2a-Tg mice is further indicated by the significantly larger increases of the expression of cardiac fetal gene markers such as ANP, BNP, and myh7 compared with those of WT littermates (Figure 2E). Similar exacerbated hypertrophic responses to TAC were also observed with the transgenic line A (Supplemental Figure 3), suggesting that the observed phenotypes are specifically due to JMJD2A. We also noticed a difference in the TAC-induced hypertrophic response between the WT mice in Figure 2C and control/WT mice in Figure 1F. This may be due to the difference of the genetic background of the KO (129/C57BL6) and transgenic (FVB) mice.

Figure 2 JMJD2A promotes cardiac hypertrophy in response to TAC-induced pressure overload. (A) Schematic diagram of cDNA encoding a flag-tagged JMJD2A in an expression vector containing the a-MHC promoter (left panel). 4 founder mice were obtained. 2 transgenic lines (Tg-A and Tg-B) with modest JMJD2A expression shown by Western blot analysis with anti-JMJD2A antibody (right panel) were established. The exogenous JMJD2A protein level is about 2-fold higher in the Tg-A line and 8-fold higher in the Tg-B line compared with that of endogenous JMJD2A. (B) H&E stains of paraffin section of WT and Jmjd2a-Tg (Tg) mouse hearts 3 weeks after sham and TAC operations. (C) HW/BW ratio (n = 6–7 per group) and (D) relative areas of cardiomyocytes (n = 3–4 per group) in WT and Jmjd2a-Tg sham- and TAC-operated animals. (E) Transcript levels of fetal gene markers, including ANP, BNP, and myh7, were quantified by real-time qRT-PCR, normalized against levels of internal GAPDH, and expressed as the fold change relative to that of sham-operated WT animals (n = 3 per group). The HW/BW ratio after TAC operation was increased 50% compared with sham operation in WT mice, whereas it was increased 100% in Tg mice. Values are mean ± SEM. *P < 0.05.

Table 2 Echocardiographic analysis of WT and Jmjd2a-Tg mice after sham and TAC operations

To determine whether JMJD2A is pathophysiologically relevant to human cardiac hypertrophy, we examined the protein levels of JMJD2A in the hearts of patients with HCM and non-HCM individuals by Western blot analysis. As shown in Figure 3, JMJD2A protein was significantly upregulated in HCM patients, and the upregulation was paralleled with that of BNP. An additional band was detected in human JMJD2A Western blot analysis. Whether this band is an alternative splice variant or degradation of JMJD2A remains to be determined. We also performed quantitative RT-PCR (qRT-PCR) analysis on the transcript of JMJD2A and BNP and observed similar upregulation of JMJD2A and BNP in HCM patients compared with controls. These data, together with those of loss-of- and gain-of-function studies of JMJD2A in mice, suggest that JMJD2A is a prohypertrophic factor under pathological conditions.

Figure 3 JMJD2A is upregulated in human hypertrophic hearts. (A) Heart tissue samples of human HCM patients and unmatched non-HCM controls (con) were subjected to Western blot analysis using antibodies against JMJD2A, BNP, and GAPDH. Consistent with the HCM phenotype, BNP was upregulated in HCM samples. The lanes were run on the same gel but were noncontiguous. (B) More importantly, JMJD2A was significantly upregulated in HCM samples (n = 7) compared with samples from the non-HCM patients (n = 4). GAPDH was used as the loading control. (C) Relative mRNA levels of JMJD2A and BNP in control (n = 3) and HCM patients (n = 7) measured by qRT-PCR. Values are mean ± SEM. *P < 0.05.

Jmjd2a upregulates the expression of FHL1 and binds to the FHL1 promoter in vivo in response to TAC. To understand the molecular mechanism or mechanisms that underlie the prohypertrophic function of JMJD2A, we performed gene profiling experiments with cDNAs from the hearts of WT, Jmjd2a hKO, and Jmjd2a-Tg mice. Of the genes whose expressions were altered in the hKO and transgenic mouse hearts in comparison with those of control/WT mice, we noticed that FHL1 was significantly changed (Figure 4). FHL1 was upregulated in response to TAC, which was significantly reduced in Jmjd2a hKO mice compared with controls (Figure 4, A and C) and exacerbated in transgenic mouse hearts (Figure 4, B and C). No change of expression for FHL2, a close member of the FHL protein family, was observed in either hKO or Tg mouse hearts compared with either control or WT mice, respectively (Figure 4C). We also noticed that, although both mRNA and protein levels of FHL1 were consistently altered in Jmjd2a hKO and/or in Jmjd2a-Tg mouse hearts when compared with their respective controls, the fold change in protein level was less than that of mRNA, suggesting an alternative posttranslational mechanism for FHL1 regulation independent of JMJD2A.

Figure 4 JMJD2A upregulates the expression of FHL1. qRT-PCR analysis of transcript levels of FHL1 from (A) control (Jmjd2afl/fl) and Jmjd2a hKO mouse hearts (n = 4–6) and (B) WT and Jmjd2a-Tg mouse hearts (n = 3–4) 21 days after sham and TAC surgery. Values are mean ± SEM. *P < 0.05. (C) Western blot of heart lysates of sham- and TAC-operated control and Jmjd2a hKO mice (left panel) and sham- and TAC-operated WT and Jmjd2a-Tg mice (right panel) using antibodies against FHL1, FHL2, phospho-ERK1/2, ERK1/2, phospho-AKT, and AKT. Representative of 4–6 experiments was shown. The relative fold changes for FHL1 and p-ERK were quantified by ImageJ. GAPDH was used as the loading control.

Sheikh et al. have shown that FHL1 is a biomechanical stress sensor that mediates the MAPK-activated signaling pathway. FHL1 is required for pressure overload–induced cardiac hypertrophy, as deletion of FHL1 blunts the TAC-induced hypertrophic response (25). Upregulation of FHL1 leads to activation of the MAPK pathway (25). Consistent with this notion, we observed upregulation of p-ERK1/2 at baseline (Figure 4C) and a further increase in transgenic mouse hearts upon TAC (Figure 4C). Upregulation of p-ERK1/2 in response to TAC was attenuated in TAC-operated Jmjd2a hKO mice (Figure 4C). We also observed upregulation of fetal gene expression in FHL1-transduced neonatal cardiomyocytes (Supplemental Figure 4), suggesting that FHL1 can promote cardiac hypertrophy under cardiac stresses. In light of the functional importance of FHL1 in promoting cardiac hypertrophy, we next focused on investigating whether and how FHL1 is regulated by JMJD2A.

We first determined whether JMJD2A regulates FHL1 transcription by binding to the FHL1 promoter in vivo and in response to pressure overload using ChIP assays. As shown in Figure 5A, JMJD2A binds the FHL1 promoter in response to TAC in either WT or control (Jmjd2afl/fl) mice (Figure 5A). Binding of JMJD2A to the FHL1 promoter is specific, as a JMJD2A ChIP assay on the GAPDH promoter (a negative control) showed very little binding of JMJD2A to the GAPDH promoter (data not shown). Furthermore, ChIP assays with anti-JMJD2A antibodies in Jmjd2a hKO hearts showed very little binding (Figure 5A), which was similar to the results obtained using IgG control (data not shown). Significant amounts of JMJD2A were also found to bind the FHL1 promoter in Jmjd2a-Tg mouse hearts at baseline and were further augmented in TAC-operated Jmjd2a-Tg mouse hearts compared with those of WT counterparts (Figure 5A). ChIP assays with anti-H3K9me3 antibody indicated that binding of JMJD2A to the FHL1 promoter was associated with decreased levels of H3K9me3 (Figure 5B). The H3K9me3 level was also significantly downregulated in Jmjd2a-Tg mice and further decreased upon TAC (Figure 5B). Consistent with the role of JMJD2A in regulating the level of H3K9me3, no significant reduction of H3K9me3 was observed in association with TAC operation in Jmjd2a hKO mice (Figure 5B).

Figure 5 JMJD2A binds to the FHL1 promoter in vivo. (A) JMJD2A-chromatin complexes were immunoprecipitated from lysates of sham- and TAC-operated WT, control (Jmjd2afl/fl), Jmjd2a-Tg, and Jmjd2a hKO mouse hearts, with anti-JMJD2A antibody and quantified by qPCR with primers in the FHL1 promoter. The amounts of immunoprecipitated complex were normalized against DNA purified from sonicated cell lysates (input) and expressed as relative to that of sham-operated WT/control. (B) ChIP was performed on aliquots of lysates from A using an anti-H3K9me3 antibody. The amounts of chromatin in the FHL1 promoter associated with H3K9me3 were measured by qPCR, normalized against input, and expressed relative to that of sham-operated WT/controls. JMJD2A binds the FHL1 promoter in response to TAC, and binding of JMJD2A is associated with decreased levels of H3K9me3. Values are expressed as mean ± SEM of 3 independent experiments. *P < 0.05.

Jmjd2a upregulates the transcription of FHL1 through SRF and myocardin. To understand how JMJD2A is recruited to the FHL1 promoter, we examined the 5'-upstream genomic sequences of the FHL1 promoter and identified a conserved serum responsive factor–binding (SRF-binding) site (Supplemental Figure 5), known as the CArG box (26). A gel mobility shift assay indicated that this CArG box is a functional SRF-binding site, as it formed a specific nucleotide-protein complex with SRF (Figure 6A). We next cloned the 1.9-kb promoter sequence containing this CArG box in front of a luciferase reporter plasmid and tested the ability of JMJD2A to transactivate the reporter. Myocardin (myocd) is a SRF cofactor and activates transcription of SRF-dependent genes in cardiac and smooth muscle cells (26–30). Myocardin activated the transcription of FHL1, which was further upregulated by JMJD2A (Figure 6B). The ability of JMJD2A to increase the transcription of FHL1 requires its demethylase activity, as a demethylase-inactive mutant of JMJD2A, 2A (H188A), failed to augment myocardin-activated FHL1 transcription (Figure 6B). Activation of FHL1 transcription by myocardin and JMJD2A was SRF-dependent, as mutation of the CArG box abolished the transactivation of FHL1 by myocardin or myocardin plus JMJD2A (Figure 6C). Furthermore, the ability of myocardin/JMJD2A to activate the transcription of FHL1 was abolished in SRF-null cells (Figure 6D). These results indicate that JMJD2A is a cofactor of SRF/myocardin and its catalytic activity is required for this coactivation. Since the cardiac “fetal” genes ANP and sm22 are known SRF/myocardin-targeted genes, we tested to determine whether JMJD2A can augment the transcription of these genes. As expected, JMJD2A upregulated myocardin-activated transcription of ANP and sm22 in a demethylase-dependent manner as well (Figure 6, E and F, respectively).

Figure 6 JMJD2A upregulates FHL1 transcription synergistically with SRF/myocardin. (A) A 32P-labeled oligonucleotide probe containing the CArG box from the FHL1 promoter was incubated with the in vitro–translated (ivt) SRF in the presence or absence of anti-SRF antibody or a 100-fold molar excess of unlabeled, WT, or mutant (mut) oligonucleotides. SRF forms a complex with the WT oligonucleotide probe (SRF-CArG) that can be super-shifted (ss) by an anti-SRF antibody. ns, nonspecific nucleoprotein complex. (B) A 1.9-kb FHL1 promoter–driven luciferase construct was transfected along with the plasmids indicated into QBI293A cells. FHL1-luciferase activities were measured 24 hours after transfection and normalized against cotransfected ß-galactosidase. (C) Responsiveness of the deletion and site-specific mutants of the FHL1-luc reporter to myocardin and JMJD2A in QBI293A cells. Deletion or site-specific mutation of the SRF-binding site CArG box impairs the responsiveness of the promoter to myocardin and JMJD2A. (D) Relative activity of the FHL1-luc reporter in SRF-null and WT ES cells in the presence and absence of cotransfected expression plasmids as indicated. Representative of 3 independent experiments is shown. (E) Relative luciferase activity of the ANP-luc reporter and (F) the sm22-luc reporter in QBI293A cells transfected with the plasmids indicated. Myocardin-activated ANP- and sm22-luc reporters that can be further upregulated by WT but not mutant JMJD2A. Data shown are mean ± SEM of 3 independent experiments. *P < 0.05.

Hebbar and Archer have shown that all 4 core histones (H2A, H2B, H3, and H4) and the linker histone H1 are associated with transiently transfected DNA despite altered histone H1 stoichiometry and an absence of nucleosome positioning on transfected DNA (31). As demethylated H3K9me3 could provide binding sites to recruit transcription factors, the inability of the demethylase-inactive mutant JMJD2A (H188A) to augment myocardin-activated FHL1-luc reporter suggests that there may be less myocardin binding on the FHL1-luc promoter. We performed ChIP assays to test this hypothesis. As shown in Figure 7A, although the expression levels of myocardin in transfected cells were similar (data not shown), there was more myocardin bound on the FHL1 promoter in cells transfected with both JMJD2A and myocardin than in cells transfected with myocardin alone or myocardin with the catalytically inactive mutant JMJD2A (Figure 7A). An increased amount of SRF was also found to be associated with the endogenous FHL1 promoter chromatin in vivo in either Jmjd2a-Tg mouse heats or hypertrophic mouse hearts after TAC surgery (Figure 7B).

Figure 7 JMJD2A enhances the binding of SRF/myocardin to the FHL1 promoter. (A) QBI293A cells were transfected with the plasmids indicated. ChIP assays were performed 24 hours later using anti-myc and anti-JMJD2A antibodies as indicated. (B) ChIP assays were performed with lysates of sham- and TAC-operated WT mouse hearts and with WT and Jmjd2a-Tg mouse hearts using anti-SRF antibody. Chromatins associated with SRF on the FHL1 promoter were quantified using qPCR. Data are expressed as mean ± SEM of 3 independent experiments. *P < 0.05.

To test whether the demethylase activity of JMJD2A is required for activation of FHL1 transcription in vivo, we performed gene knockdown (KD) experiments in rat neonatal cardio­myocytes using JMJD2A-specific siRNA duplexes. The mRNA level of JMJD2A was significantly KD by JMJD2A-specific siRNAs compared with control nonspecific siRNA (80% reduction; Figure 8A). Consistent with what we observed in Jmjd2a hKO mice, downregulation of JMJD2A in cardiomyocytes resulted in a decrease in the FHL1 transcription at baseline (Figure 8B). Reexpression of JMJD2A in the JMJD2A KD cardiomyocytes using adenoviral transduction restored FHL1 expression (Figure 8B), whereas the mutant JMJD2A (H188A) did not (Figure 8B). We also determined whether JMJD2A is involved in phenylephrine-stimulated (PE-stimulated) myocyte hypertrophy. PE had no effects on the transcript level of JMJD2A (Figure 8A) and significantly upregulated FHL1 transcription (Figure 8C). Upregulation of FHL1 by PE was significantly attenuated in JMJD2A KD cells (Figure 8C). Upregulation of fetal genes by PE, including ANF, BNP, and myh7 (Figure 8C), and of the cardiomyocyte size (Figure 8D) in JMJD2A KD cells were significantly attenuated, suggesting that JMJD2A mediates the hypertrophic effect of PE on cardiomyocytes.

Figure 8 Downregulation of JMJD2A attenuates the expression of FHL1 and PE-activated fetal gene program in cardiomyocytes. Neonatal rat ventricular myocytes were transfected with nonspecific control siRNA (cRNAi) or JMJD2A-specific siRNA (2A-RNAi). Forty-eight hours after transfection, cells were treated with PBS (vehicle) or with PE (10 µM) (A, C, and D) or with various adenoviruses as indicated (B). Relative transcript levels of JMJD2A (A), FHL1 (B, lanes 1–5, and C, lanes 1–4), and fetal genes (C, lanes 5–16) were determined 24 hours after the treatment using qRT-PCR, normalized against internal GAPDH, and expressed relative to those of the vehicle-treated cRNAi transfected cells. (D) Cardiomyocytes were photographed, and areas were quantified using ImageJ and expressed relative to those of the vehicle-treated cRNAi transfected cells. Data are expressed as mean ± SEM of 3 independent experiments. *P < 0.05.

A key finding in the current study is that JMJD2A promotes cardiac hypertrophy under pathological conditions. This is supported by both gain-of- and loss-of-function studies presented here. While overexpression of JMJD2A in mice exacerbates the hypertrophic response to pressure overload, inactivation of Jmjd2a blunts it (Figures 2 and 1, respectively). Since JMJD2A is upregulated in human HCM patients (Figure 3), we speculate that this upregulation of JMJD2A may play an active role in promoting cardiac hypertrophy in humans under cardiac stress conditions. It is noted that the mechanism by which the JMJD2A activity is regulated in mice may be different from that in humans, since we did not observe upregulation of JMJD2A expression in TAC-induced hypertrophy in our mouse model (data not shown). Rather, we found enhanced binding of JMJD2A to the promoter of prohypertrophic genes such as FHL1 in response to TAC.

Our studies suggest that JMJD2A may promote cardiac hypertrophy through FHL1. This is supported by the following evidence: FHL1 is known to mediate pressure overload–induced cardiac hypertrophy (25), and overexpression of FHL1 in cardiomyocyte upregulates fetal gene expression (Supplemental Figure 4). JMJD2A activates the transcription of FHL1 both in vitro and in vivo and in response to hypertrophic stimuli (Figures 4–6). The phenotype of cardiac inactivation of Jmjd2a is consistent with that of FHL1 KO mice (Figure 1). The MAPK-signaling pathway regulated by FHL1 is also affected by JMJD2A (Figure 4). It remains to be determined whether FHL1 is the major downstream effector of JMJD2A. This could be determined by examining the phenotype of Jmjd2a-Tg and FHL1-null compound mice and testing to determine whether inactivation of FHL1 blunts the TAC-induced hypertrophy in Jmjd2a-Tg mice. It is likely that JMJD2A may have other transcriptional targets during hypertrophic remodeling. We have shown that JMJD2A can synergistically activate ANP and sm22 transcription with SRF/myocardin in vitro (Figure 6). Myocardin/SRF has been shown to be involved in cardiac development and hypertrophic remodeling (32–33). Whether these and/or other SRF/myocardin-targeted genes mediate the prohypertrophic function of JMJD2A in vivo remains to be determined in the future. A genome-wide ChIP followed by deep sequencing (ChIP-seq) with chromatins from Jmjd2a-Tg and Jmjd2a hKO hearts would provide means to identify further potential JMJD2A targets involved in cardiac hypertrophy and heart failure.

It is well established that histone modifications play important roles in gene transcription. Over the past decades, a great deal has been learned regarding histone acetylation and its role in cardiac remodeling (6). By comparison, little is known about the function of histone methylation even though it is the most abundant form of histone modifications. Unlike histone acetylation, which is usually associated with gene activation, histone methylation can lead to either activation or repression of gene transcription, depending on the lysine residues, the degree of methylation status (mono-, di-, or trimethylation), and chromatin location. It was shown previously that JMJD1A, a di/monomethyl demethylase, can activate the expression of contractile genes in smooth muscle cells (34). Our studies indicate that trimethyl demethylation of H3K9me3 at the FHL1 promoter by JMJD2A in cardiomyocytes promotes gene transcription. As JMJD2A upregulates other SRF-regulated genes such as ANP and sm22, it is tempting to speculate that demethylation of H3K9 on promoters of SRF-targeted genes activates gene transcription.

JMJD2A could activate the transcription of SRF-targeted genes by either providing binding sites that stabilize/increase the affinity for transcription factors or preventing binding of H3K9me3 to heterochromatin binding protein HP-1. Binding of HP-1 to H3K9me3 was known to mediate conversion of euchromatin to heterochromatin, resulting in silencing of gene transcription. Our studies favor the former possibility as KD of JMJD2A did not abolish the ability of SRF/myocardin to activate FHL1 transcription (Supplemental Figure 6), suggesting that FHL1 was not inactivated in the absence of JMJD2A. Furthermore, we observed an increased amount of SRF/myocardin binding to the FHL1 promoter that is associated with upregulation of JMJD2A and decreased levels of H3K9me3. Taken together, our data suggest a feed-forward mechanism for the synergy between SRF/myocardin and JMDJ2A; SRF/myocardin recruits JMJD2A to the FHL1 promoter. In turn, JMJD2A demethylates H3K9me3, generating a surface/binding area for further recruitment of SRF/myocardin.

JMJD2A is a global regulator of chromatin remodeling and gene expression, and yet deletion of JMJD2A affects the transcription of only a handful of genes. While this finding may seem paradoxical, it is not unexpected. Gene expression is regulated through the action of transcription factors and histone-modifying enzymes. Many different histone-modifying enzymes, including HDACs, HATs, HMTs, and HDMs, contribute to the dynamic regulation of chromatin structure and function, with concomitant impacts on gene transcription. Unlike transcription factors that often have on-off effects on gene transcription, the effects of histone-modifying enzymes on gene transcription are often modulatory. This modulatory effect can be context- and gene-dependent such that only those genes exceeding the threshold will yield a phenotype and be identified. Therefore, it is not surprising that deletion of JMJD2A in the heart resulted in changes of the transcription levels of only a few target genes. However, it is worth noting that even though the number of genes whose expressions are affected by deletion of Jmjd2a is small, the genome-wide H3K9me3 marks affected by Jmjd2a deficiency may still be large. It will be interesting to identify these marks using ChIP-seq and to further investigate the relationship between JMJD2A-regulated H3K9me3 marks and other chromatin marks, i.e., histone code, for such a relationship may ultimately determine the transcriptional state of the gene as either active, repressed, or poised for activation.

In summary, our studies indicate that JMJD2A promotes cardiac hypertrophy in response to hypertrophic stimuli. JMJD2A demethylates H3K9me3 and activates transcription of prohypertrophic genes synergistically with SRF/myocardin. It is noted that the effect of JMJD2A on the expression of SRF/myocardin-targeted genes is modulatory, as loss of function of Jmjd2a does not abolish the ability of SRF/myocardin to activate gene transcription. JMJD2A could be a potential drug target for transcriptional therapy against cardiac hypertrophy and heart failure. It will be worth determining in the future whether small molecules designed to target the demethylase activity of JMJD2A could reduce and normalize the expression of SRF/myocardin-targeted genes that are upregulated during cardiac remodeling without compromising the physiological cardiac growth and functions of these genes that are part of the adaptive response to altered conditions.

Gene targeting. A Jmjd2a-targeting vector was generated using the pGKNEO-F2L2-DTA vector that contains a neomycin resistance cassette flanked by frt and loxP sites and a diphtheria toxin gene cassette. The arms for homologous recombination were generated by high-fidelity PCR amplification of 129SvEv genomic DNA. The resulting vector was verified by restriction mapping and DNA sequencing. The targeting vector was linearized with PvuI and electroporated into 129SvEv-derived ES cells. ES cell clones were screened for homologous recombination by Southern blot analysis. Genomic DNA was digested with BclI, and successful loxP site incorporation was confirmed with a 5' probe and PCR strategy. Targeted ES cells were injected into the blastocysts of C57BL/6 females to generate chimeric mice. Chimeras were bred to C57BL/6 females to achieve germ-line transmission. The Jmjd2afl/fl mice were intercrossed to a-MHC–Cre transgenic mice to obtain mice with Jmjd2a-null allele specifically in the heart. All experiments on control and JMJD2A hKO mice were conducted on a 129SvEv/C57BL6 mixed background.

Generation of Jmjd2a-Tg mice. A fragment containing cDNA for human JMJD2A with an N-terminal Flag tag was subcloned into the SalI/HindIII sites between the mouse cardiac a-MHC promoter and the human growth hormone gene polyadenylation sequence (35). The plasmid was digested free of vector sequence with BamHI, purified, and microinjected into fertilized eggs of FVB mice to generate transgenic mice according to standard procedures. Founder and offspring mice were genotyped by PCR (Supplemental Table 1).

Surgical procedures and echocardiography. TAC and sham surgery were performed on control, Jmjd2a hKO, WT, and heterozygote transgenic mice at 8 to 10 weeks of age as described previously (24). Echocardiography was performed on mice before sacrifice. All animal experiments were approved by the Institutional Animal Care and Use Committee (IACUC) of the University of Texas Southwestern Medical Center.

Measurement of cardiomyocyte area. Cardiomyocyte size was assessed by measuring the mean cardiomyocyte cross-sectional area for each sham-operated and TAC-operated heart. More than 200 randomly selected cardiomyocytes cut in cross-section and with a central nucleus on H&E-stained heart sections were measured with ImageJ.

RNA and cDNA preparation, real-time qRT-PCR, and microarray analysis. Total RNA was isolated with TRIzol reagent (Invitrogen) following manufacturer’s protocol. First-strand cDNA was made using Superscript III Reverse Transcriptase (Invitrogen). SYBR-based qRT-PCR was used to examine relative levels of selected mRNAs. All data were normalized to an internal standard (GAPDH; ?CT method). Sequences for gene-specific primer pairs are listed in Supplemental Table 1. Microarray analyses were performed using RNAs of mouse hearts from WT sham, WT TAC, Jmjd2a-Tg sham, Jmjd2a-Tg TAC, control, and Jmjd2a hKO mice with Illumina bead array platform. All microarray data were deposited into the NCBI GEO repository (GSE27689).

ChIP. ChIP was performed as described with minor modifications (36). Protein-chromatin complexes were immunoprecipitated with antibodies against JMJD2A (Bethyl), trimethylated H3K9 (Abcam), and SRF (Santa Cruz Biotechnology Inc.). Immunoprecipitated chromatin fragments were quantified by SYBR-based qPCR, normalized using the percent input method (Invitrogen).

Protein and Western blot. Tissue protein was extracted with T-PER buffer (Pierce) according to the manufacture’s protocol. After SDS-PAGE separation, the protein was probed on nylon membrane (Bio-Rad). The following antibodies were used: JMJD2A (Bethyl), FHL1 (Abcam), FHL2 (Abcam), p-ERK1/2, ERK1/2, p-AKT, AKT (Cell Signaling), BNP (Santa Cruz Biotechnology Inc.), and GAPDH (Santa Cruz Biotechnology Inc.).

Gel shifts. Gel shifts were performed as described (37). Briefly, in vitro–translated SRF (Promega) was incubated with 32P-labeled oligonucleotide probe in binding buffer at room temperature for 20 minutes. DNA and protein complexes were separated on 8% nondenaturing polyacrylamide gels. Competition experiments were performed by addition of a 100-fold molar excess of non–radio-labeled WT or mutant oligonucleotide probes. Supershift assay was performed with addition of anti-SRF antibody for 15 minutes following incubation of SRF and radioactive probe in binding buffer.

Plasmids, cell culture, transfection, siRNA, and luciferase reporter assays. The expression vector of human JMJD2A was made by subcloning PCR-amplified inserts into Flag-tagged pcDNA3.1 (Invitrogen). FHL1-luciferase reporter was constructed by subcloning PCR-amplified inserts corresponding to the promoter sequence of FHL1 from mouse genomic DNA (–1105 to +832) into the pGL3-basic vector (Promega). Point mutation of the FHL1 promoter was performed using the QuikChange kit (Stratagene). Primers used in the cloning are listed in Supplemental Table 1. ANP- and sm22-luciferase reporter constructs, SRF and myocardin expression plasmids, and SRF-null ES cells were gifts from E. Olson (University of Texas Southwestern). QBI-293A cells (ATCC) were maintained in complete medium (DMEM supplemented with 10% fetal bovine serum, 2 mM glutamine, and penicillin/streptomycin). Rat neonatal ventricular cardiomyocytes were prepared as described (38). Cells were transfected with a combination of plasmids indicated for each experiment using Fugene 6 (Roche). siRNA duplex (smart pool) was purchased from Dharmacon and transfected using Lipofectamine 2000 (Invitrogen). Cell lysates were assayed for luciferase expression using a luciferase assay kit (Promega). Relative promoter activities were expressed as relative luminescence units normalized for cotransfected ß-galactosidase activities in the cell.

Adenoviruses expressing JMJ2A and FHL1 and neonatal cardiomyocyte culture. Adenoviruses harboring FHL1, WT, and mutant forms of JMJD2A were constructed and packaged with AdEasy XL Adenoviral system (Stratagene) according to the manufacturer’s protocol. The recombined constructs were used to package and amplify adenovirus in QBI-293A cells.

Patient samples and controls. Human hypertrophic heart tissue samples were obtained from patients undergoing surgical myectomy who were previously diagnosed with HCM. Control non-HCM adult human ventricular tissue samples were obtained from unmatched victims of motor vehicle accidents (2 cases) and patients previously diagnosed with atrial fibrillation (2 cases). Tissue procurement was based on the receipt of written patient-informed consent and approved by the institutional review boards (Peking Union Medical College). Western blot analysis of the tissue lysates was performed using standard protocols with antibody against JMJD2A (Bethyl), BNP (Santa Cruz Biotechnology Inc.), and GAPDH (Santa Cruz Biotechnology Inc.).

Statistics. All data are shown as mean ± SEM. When comparing multiple groups, 1-way ANOVA was used. Student’s t test (2-tailed) was used to compare the difference between 2 groups. P < 0.05 was considered statistically significant.

View Supplemental data

This study was supported by a Scientist Development grant from the American Heart Association, Texas Advanced Research Program 010019-0070-2007, and RO1 HL085749 from the National Heart, Lung, and Blood Institute (to Z.P. Liu). Work in J.A. Hill’s lab is supported by HL-075173, HL-080144, HL-090842, UO1-RFA-HL09-010, and T32 HL007360-31 from the NIH, and 0640084N and DeHaan Cardiac Myogenesis Research Center grants from the American Heart Association.

Conflict of interest: The authors have declared that no conflict of interest exists.

Citation for this article: J Clin Invest doi:10.1172/JCI46277.

Levy D, Garrison RJ, Savage DD, Kannel WB, Castelli WP. Prognostic implications of echocardiographically determined left ventricular mass in the Framingham Heart Study. N Engl J Med. 1990;322(22):1561–1566. Hill JA, Olson EN. Cardiac plasticity. N Engl J Med. 2008;358(13):1370–1380. Abraham WT, et al. Coordinate changes in Myosin heavy chain isoform gene expression are selectively associated with alterations in dilated cardiomyopathy phenotype. Mol Med. 2002;8(11):750–760. Blaxall BC, Tschannen-Moran BM, Milano CA, Koch WJ. Differential gene expression and genomic patient stratification following left ventricular assist device support.. Am J Coll Cardiol. 2003;41(7):1096–1106.1106 . Meyer K, Zhang L. Fetal programming of cardiac function and disease. Reprod Sci. 2007;14(3):209–216. McKinsey TA, Olson EN. Toward transcriptional therapies for the failing heart: chemical screens to modulate genes. J Clin Invest. 2005;115(3):538–546. Couture JF, Trievel RC. Histone-modifying enzymes: encrypting an enigmatic epigenetic code. Curr Opin Struct Biol. 2006;16(6):753–760. Bártová E, Krejcí J, Harnicarová A, Galiová G, Kozubek S. Histone modifications and nuclear architecture: a review. J Histochem Cytochem. 2008;56(8):711–721. Campos EI, Reinberg D. Histones: annotating chromatin. Annu Rev Genet. 2009;43:559–599. Rea S, et al. Regulation of chromatin structure by site-specific histone H3 methyltransferases. Nature. 2000;406(6796):593–599. Shi Y, et al. Histone demethylation mediated by the nuclear amine oxidase homolog LSD1. Cell. 2004;119(7):941–953. Trojer P, Reinberg D. Histone lysine demethylases and their impact on epigenetics. Cell. 2006;125(2):213–217. Klose RJ, et al. The transcriptional repressor JHDM3A demethylates trimethyl histone H3 lysine 9 and lysine36. Nature. 2006;442(7100):312–316. Whetstine JR, et al. Reversal of histone lysine trimethylation by the JMJD2 family of histone demethylases. Cell. 2006;125(3):467–481. Barski A, et al. High resolution profiling of histone methylations in the human genome. Cell. 2007;129(4):823–837. Lorbeck MT, et al. The histoe demethylase Dmel\Kdm4A controls genes required for life span and male–specific sex determination in Drosophila. Gene. 2010;450(1–2):8–17. Agger K, Christensen J, Cloos PA, Helin K. The emerging functions of histone demethylases. Curr Opin Genet Dev. 2008;18(2):159–168. Nottke A, Colaiácovo MP, Shi Y. Developmental roles of the histone lysine demethylases. Development. 2009;136(6):879–889. Kaneda R, et al. Genome-wide histone methylation profile for heart failure. Genes Cells. 2009;14(1):69–77. Shin S, Janknecht R. Activation of androgen receptor by histone demethylases JMJD2A and JMJD2D. Biochem Biophys Res Commun. 2007;359(3):742–746. Zhang D, Yoon HG, Wong J. JMJD2A is a novel N-CoR-interacting protein and is involved in repression of the human transcription factor achaete scute-like homologue 2 (ASCL2/Hash2). Mol Cell Biol. 2005;25(15):6404–6414. Gray SG, et al. Functional characterization of JMJD2A, a histone deacetylase- and retinoblastoma-binding protein. J Biol Chem. 2005;280(31):28507–28518. Oka T, et al. Cardiac-specific deletion of Gata4 reveals its requirement for hypertrophy, compensation, and myocyte viability. Circ Res. 2006;98(6):837–845. Berry JM, Naseem RH, Rothermel BA, Hill JA. Models of cardiac hypertrophy and transition to heart failure.. Drug Discovery Today: Disease Models. 2007;4(4):197–206.206 . Sheikh F, et al. An FHL1-containing complex within the cardiomyocyte sarcomere mediates hypertrophic biomechanical stress responses in mice. J Clin Invest. 2008;118(12):3870–3880. Miano JM. Serum response factor: toggling between disparate programs of gene expression. J Mol Cell Cardiol. 2003;35(6):577–593. McDonald OG, Owens GK. Programming smooth muscle plasticity with chromatin dynamics. Circ Res. 2007;100(10):1428–1441. Wang D, et al. Activation of cardiac gene expression by myocardin, a transcriptional cofactor for serum response factor. Cell. 2001;105(7):851–862. Wang DZ, Olson EN. Control of smooth muscle development by the myocardin family of transcriptional coactivators. Curr Opin Genet Dev. 2004;14(5):558–566. Miano JM, Long X, Fujiwara K. Serum response factor: master regulator of the actin cytoskeleton and contractile apparatus. Am J Physiol Cell Physiol. 2007;292(1):C70–C81. Hebbar PB, Archer TK. Altered histone H1 stoichiometry and an absence of nucleosome positioning on transfected DNA. J Biol Chem. 2008;283(8):4595–4601. Xing W, et al. Myocardin induces cardiomyocyte hypertrophy. Circ Res. 2006;98(8):1089–1097. Kontaraki JE, Parthenakis FI, Patrianakos AP, Karalis IK, Vardas PE. Myocardin gene regulatory variants as surrogate markers of cardiac hypertrophy – study in a genetically homogeneous population. Clin Genet. 2008;73(1):71–78. Lockman K, Taylor JM, Mack CP. The histone demethylase, Jmjd1a, interacts with the myocardin factors to regulate SMC differentiation marker gene expression. Circ Res. 2007;101(12):e115–e123. Subramaniam A, Jones WK, Gulick J, Wert S, Neumann J, Robbins J. Tissue-specific regulation of the alpha-myosin heavy chain gene promoter in transgenic mice. J Biol Chem. 1991;266(36):24613–24620. Nelson JD, Denisenko O, Bomsztyk K. Protocol for the fast chromatin immunoprecipitation (ChIP) method. Nat Protoc. 2006;1(1):179–185. Zhou W, et al. FoxO4 inhibits NF-kappaB and protects mice against colonic injury and inflammation. Gastroenterology. 2009;137(4):1403–1414. Liu ZP, Olson EN. Suppression of proliferation and cardiomyocyte hypertrophy by CHAMP, a cardiac-specific RNA helicase. Proc Natl Acad Sci U S A. 2002;99(4):2043–2048.

View the original article here

read more

Popular Posts

There was an error in this gadget
There was an error in this gadget