Nuclear S-nitrosylation defines an optimal zone for inducing pluripotency

Abstract

Background:

We found that cell-autonomous innate immune signaling causes global changes in the expression of epigenetic modifiers to facilitate nuclear reprogramming to pluripotency. A role of S-nitrosylation by inducible NO synthase (iNOS), an important effector of innate immunity, has been previously described in transdifferentiation of fibroblasts to endothelial cells. Accordingly, we hypothesized that S-nitrosylation might also have a role in nuclear reprogramming to pluripotency.

Methods:

We used murine embryonic fibroblasts containing a doxycycline-inducible cassette encoding the Yamanaka factors (Oct4, Sox2, Klf4 and c-Myc), and genetic or pharmacological inhibition of iNOS together with Tandem Mass Tag (TMT) approach, chromatin immunoprecipitation-qPCR, site directed mutagenesis, and micrococcal nuclease assay to determine the role of S-nitrosylation during nuclear reprogramming to pluripotency.

Results:

We show that an optimal zone of innate immune activation, as defined by maximal yield of induced pluripotent stem cells (iPSCs), is determined by the degree of NFkB activation; nitric oxide (NO) generation; S-nitrosylation of nuclear proteins; and DNA accessibility as reflected by histone markings and increased mononucleosome generation in a micrococcal nuclease assay. Genetic or pharmacological inhibition of iNOS reduces DNA accessibility and suppresses iPSC generation. The effect of NO on DNA accessibility is mediated in part by S-nitrosylation of nuclear proteins, including MTA3, a subunit of Nucleosome Remodeling Deacetylase (NuRD) complex. S-nitrosylation of MTA3 is associated with decreased NuRD activity. Overexpression of mutant MTA3, in which the two cysteine residues are replaced by alanine residues, impairs the generation of iPSCs.

Conclusion:

This is the first report showing that DNA accessibility and iPSC yield depends on the extent of cell-autonomous innate immune activation and NO generation. This “Goldilocks zone” for inflammatory signaling and epigenetic plasticity may have broader implications for cell fate and phenotypic fluidity.

Introduction

The discovery that the forced expression in somatic cells of four transcriptional factors Oct4, Sox2, Klf4 and c-Myc (OSKM) was sufficient to induce nuclear reprogramming to pluripotency galvanized the scientific community and was recognized by the Nobel Prize for Physiology or Medicine in 2012^{1, 2}. The ability to generate induced pluripotent stem cells (iPSCs) from somatic cells holds great promise for regenerative medicine, and has facilitated studies of development and differentiation; promoted insights into pathobiology; and generated a novel platform for drug discovery and testing^{3, 4}.

Surprisingly, we observed that the retroviral vector used by Yamanaka was more than a vehicle for the genes encoding OSKM⁵. By activating cell-autonomous innate immune signaling, the viral vector caused global changes in the expression of epigenetic modifiers so as to favor activating histone marks. Stimulation of pattern recognition receptors (PRRs) such as TLR3 and RIG1 activated NFkB and IRF3, down regulated histone deacetylases (HDACs) and upregulated histone acetylases (HATs)^{5, 6}. Inhibition of innate immune signaling (with shRNA knockdown of elements of the PRR signaling cascade or with decoy oligonucleotides to p65) abrogated nuclear reprogramming to pluripotency. To date, every approach to nuclear reprogramming that we have studied (viral vectors; modified message RNA; doxycycline-inducible cassettes of the Yamanaka factors) also activates cell-autonomous innate immune signaling, which appears to be a requirement for induction of pluripotency.

In this report, we show for the first time the presence of a “Goldilocks zone” of cell-autonomous innate immune signaling and DNA accessibility where efficiency of nuclear reprogramming to pluripotency is maximum. This optimal zone for generating iPSCs is associated with a maxima of iNOS induction, NO generation and S-nitrosylation of nuclear proteins. One of these nuclear proteins is MTA3, a component of the NURD complex. The S-nitrosylation of MTA3 reduces histone deacetylase activity, enhances DNA accessibility and increases iPSC generation. Too little or excessive innate immune activation (suboptimal zones of activation) are associated with lesser iNOS activity, DNA accessibility and iPSC yield. To summarize, we provide new insight into the mechanisms by which cell-autonomous innate immune signaling promotes epigenetic plasticity and nuclear reprogramming.

Methods

The data, analytical methods, and study materials will be made available to other researchers from the corresponding author on reasonable request for purposes of reproducing the results or replicating the procedure on request.

Animal Husbandry

All animals were maintained in a specific pathogen-free facility at Houston Methodist Research Institute in Houston, TX. Animal use and care were approved by the Houston Methodist Animal Care Committee in accordance with institutional animal care and use guidelines.

Chemicals and Reagents.

All chemicals were purchased from Sigma-Aldrich (St Louis, MO) unless otherwise stated. Other reagents including FBS, Taqman RT-PCR primers, Taqman master mix, DAF-FM diacetate (Life Technologies, Carlsbad, CA); TLR3 agonist polyinosinic:polycytidilic acid (PIC) (Invivogen, San Diego, CA); p65i (an NFκB oligonucleotide decoy, Novus biologicals, Littleton, CO); antibody against inducible nitric oxide synthase (iNOS), HRP conjugated or normal goat anti mouse or rabbit antibodies (Santa Cruz Biotechnology, Santa Cruz, CA).

Cells and Cell culture.

BJ human neonatal foreskin fibroblast cells (BJ fibroblasts) (Stemgent), WT and iNOS^−/− murine embryonic fibroblasts (MEFs; Cell Biologics Inc, Chicago IL) were cultured and maintained in Dulbecco’s modified Eagle’s medium (DMEM; Invitrogen) with 10% fetal bovine serum and 1% penicillin/streptomycin (5% CO2, 37°C). For isolation of secondary dox-inducible MEFs, chimeric embryos were obtained from transgenic R26rtTA; Col1a12lox-4F2A mice expressing the loxP-flanked, dox-inducible polycistronic 4F2A cassette (Oct4, Sox2, Klf4, c-Myc; Jackson Laboratory). Secondary MEFs were isolated as previously described⁷, and expanded for two passages before freezing. Passage 3 cells were used in all the experiments unless indicated otherwise. Culture plates were coated with EmbryoMax® 0.2% Gelatin Solution (Millipore, US) for 30 min before use. All cells were cultured in ES medium under standard condition (5% CO₂, 37°C) unless stated otherwise.

Nuclear reprogramming.

MEFs expressing doxycycline inducible polycistronic cassette of Oct4, Sox2, KLF4 and c-Myc (Dox-MEFs; 7×10⁴cells/well) were induced to reprogram to pluripotency by addition of doxycycline (2 μg/ml daily for up to 7d), in the absence or presence of PIC (0–10000 ng/ml; daily up to 12 days). In some experiments, Dox-MEFs were transfected with plasmid constructs encoding wild type or mutant versions of MTA3 (see below) using lipofectamine and selected in presence of 10μg/ml of blasticidin for 7 days. The selected Dox-MEFs were then subjected to nuclear reprogramming by adding doxycycline in the presence or absence of PIC30ng/ml as mentioned above. BJ fibroblasts were subjected to nuclear reprogramming as described earlier⁵ in presence or absence of PIC (300 ng/ml). WT and iNOS^−/− MEFs were subjected to nuclear reprogramming using STEMCCA Lentivirus Reprogramming Kit (EMD Millipore) as per manufacturer’s protocol. Number of iPSC colonies was counted on day 21 using alkaline phosphatase staining.

NF-kB Luciferase Assay.

Dox-MEFs were subjected to nuclear reprogramming in absence or presence of different concentrations of PIC and transfected with pNF-kB-Luc plasmid by using Lipofectamine 2000, and after 24 hr cells were studied by using the Bright-Glo Luciferase Assay System and a luminometer.

Alkaline phosphatase (AP) staining.

At day 21 of reprogramming, media was removed and 4% Paraformaldehyde in 1 × PBS was added. After 1 minute, cells were washed with TBST (20mM Tris-HCl, pH 7.4, 0.15M NaCl, 0.05% Tween-20) twice. Alkaline phosphatase staining solution (prepared as recommended by the manufacturer) was incubated with the cells for 15 minutes at room temperature prior to imaging.

MTT assay.

Dox-MEFs were subjected to nuclear reprogramming (see protocol above) in absence or presence of PIC at different concentrations up to 7 days followed by measurement of cell viability using MTT cell growth kit (EMD Millipore) as per manufacturer’s instruction.

RNA extraction and Real-Time PCR.

Cellular mRNA was extracted using RNeasy Mini Kit (Qiagen) and reverse transcribed into cDNA and then specific Taqman primers for targeted genes were used for real-time PCR using QuantStudio real-time PCR system from Life Technologies. Gene expression data using the ΔΔCt method was normalized to the expression of the control gene and plotted as relative changes.

Western blot and immunoprecipitation.

Western blot analysis was performed as previously described⁵. Immunoprecipitation was performed using the Pierce Classic Magnetic IP/Co-IP Kit (Thermo Scientific, Waltham, MA) following manufacturer’s manual.

Chromatin immunoprecipitation (ChIP)–qPCR.

ChIP was performed using a previously established method⁵. Results were normalized using signal over background method. Primers of the Oct4 promoter were purchased from Cell Signaling Technology (CST).

Determination of NO generation using DAF-FM diacetate.

To assess intracellular NO, relative fluorescence intensity of NO generation in Dox-MEFs undergoing nuclear reprogramming were measured using 4-Amino-5-methylamino-2′, 7′- difluorofluorescein Diacetate (DAF-FM DA). After the Dox-MEFs had been exposed to different experimental conditions, they were incubated with DAF-FM DA (10 μM) at 37 °C for 30 min. Fluorescence was quantified by measuring fluorescence intensity at 495 nm excitation and 515 nm emission by a fluorescence plate reader (Tecan M1000 PRO).

Griess assay.

Total concentration of nitrate and nitrite was measured using nitrate/nitrite fluorometric assay kit (Cayman Chemical, MI). 20 μl of media was used to measure the nitrate generated under different conditions of nuclear reprogramming following manufacturer’s protocol.

NOS inhibitor assay.

Nuclear reprogramming of Dox-MEFs were performed under different experimental conditions in presence or absence of general NOS inhibitor like L-NAME (10μM) or iNOS specific inhibitor like 1400W (10 μM) and BYK-hydrochloride (10μM) at IC_90. Inhibitors were present up to 3-days and re-added when the reprogramming medium was replenished. Number of alkaline positive colonies were counted on day 21.

Protein extraction, S-nitrosylation detection and Western blotting.

Nuclear and cytoplasmic proteins were extracted on day 4 of nuclear reprogramming from Dox-MEFs under different experimental conditions using NE-PER Nuclear and Cytoplasmic Extraction Reagents from Thermo Scientific kit. The total protein concentration was quantified using BCA assay (Pierce). For the detection of S-nitrosylated proteins, S-nitrosylation Western blot Kit (Pierce) was used. In brief, total protein (100μg in 100 μl) was taken for each sample and unmodified cysteines were first blocked using methyl methanethiosulfonate (20mM; MMTS), a sulfhydryl reactive compound. Excess MMTS was removed by precipitating proteins with six volumes of pre-chilled acetone and freezing at −20°C for 2 hours. Samples were centrifuged, air dried and re-suspended in HENS buffer (100 μl). Next, selective reduction of S-nitrosylated cysteines were performed with ascorbate (40mM) and labeled with iodo-TMTzero reagents (0.8 mM) which irreversibly bind to the cysteine thiol that was S-nitrosylated. TMT reagent-modified proteins were resolved on 4–20% gradient SDS-polyacrylamide gels. The gel was transferred into nitrocellulose membrane using an iBlot System (Invitrogen) followed by detection of S-nitrosylated proteins using an anti-TMT antibody.

Analysis of S-nitrosylation by mass spectrometry (MS).

Nuclear proteins isolated on day 4 from Dox-MEFs undergoing nuclear reprogramming in the presence of PIC (30ng/ml) were labelled with iodo-TMTzero (as described in previous section) and immunoprecipitated using anti-TMT antibody, resolved on 4–20% gradient SDS/PAGE gel and stained with Coomassie Blue. After destaining, one band below 70 kD was excised and sent to the mass spectrometry-proteomics core facility (Baylor College of Medicine). The gel bands were subjected to in-gel double digestion with trypsin and GluC (NEB) followed by extraction of peptides from the gel, quantification and analysis of the sample (1 μg) using Q Exactive Plus mass spectrometers (Thermo Fisher Scientific, Rockford, IL, USA) coupled with an Easy-nLC 1000 nanoflow LC system (Thermo Fisher Scientific). The RAW files were fed into Peaks software version 7.5 with Post Translational Modification (PTM) analysis with TMT-zero set as the target search. The data were also searched against a decoy database so that protein identifications were accepted at a false discovery rate of 1%. The organism was set as mouse and the database was NCBI non-redundant database.

HDAC1/2 activity assay.

Nuclear proteins were isolated from Dox-MEFs undergoing nuclear reprogramming in absence or presence of poly I:C on day 4 and HDAC1/2 activity was measured for each sample following manufacturer’s (Epigentek, P-4003) protocol. Nuclear proteins (5 μg) from each sample were used for the assay.

Histone H3 modification multiplex assay.

Histone proteins were isolated from Dox-MEFs undergoing nuclear reprogramming at minimal (Dox+Bay11), optimal (Dox+PIC30) and maximal (Dox+PIC1000) innate immune activation on day 4 and histone H3 assessed for 21 different epigenetic marks (H3K4me1, H3K4me2, H3K4me3, H3K9me1, H3K9me2, H3K9me3, H3K27me1, H3K27me2, H3K27me3, H3K36me1, H3K36me2, H3K36me3, H3K79me1, H3K79me2, H3K79me3, H3K9ac, H3K14ac, H3K18ac, H3K56ac, H3ser10P, H3ser28P) for each sample following manufacturer’s (Epigentek, P-3100) protocol. Nuclear proteins (5ng) from each sample were used for the assay. Bay11 (10 μM) inhibits NFkB activity.

HAT activity assay.

Nuclear proteins were isolated from Dox-MEFs undergoing nuclear reprogramming at minimal (Dox+Bay11), optimal (Dox+PIC30) and maximal (Dox+PIC1000) innate immune activation on day 4 and HAT activity was measured for each sample following manufacturer’s (Epigentek, P-4001) protocol. Nuclear proteins (5ug) from each sample were used for the assay.

Micrococcal nuclease sensitivity assay.

Dox-MEFs undergoing nuclear reprogramming at minimal (Dox+Bay11), optimal (Dox+PIC30) and maximal (Dox+PIC1000) on day 4 were washed with cold PBS and cross-linked with 1% formaldehyde for 10 minutes at room temperature followed by quenching the reaction with 1X (final) glycine. The cells were then resuspended in buffer A (Cell Signaling Technology; #9003), incubated on ice for 10 minutes followed by pelleting nuclei and resuspending them in buffer B (Cell Signaling Technology; #9003). The resuspended pellets were aliquoted in equal volumes and digested at 37 °C for 5, 10 and 20 min with MNase. One aliquot was kept as uncut control. The reaction was stopped by adding EDTA (0.5M; 10μl). DNA was purified after RNase and proteinase K treatment using spin columns (Cell Signaling Technology; #9003). The DNA was resuspended in DNA elution buffer. Same samples were also analyzed by 2100 Bioanalyzer using DNA1000 assay kits to determine the percentages of mono-, di-, and tri-nucleosome fractions in the samples. The percentage of total area under the curve present for each fraction (mono-, di-, and tri-nucleosomes) at a given time point of MNase digestion for each sample were assessed by the 2100 expert software and mono-nucleosome to di-nucleosome (M/D) and mono-nucleosome to tri-nucleosome (M/T) ratios were determined by comparing the percentage of total areas of each fraction at different time points of MNase digestion for each sample.

Construction of MTA3 overexpression plasmid vector and site directed mutagenesis.

The mouse MTA3 overexpression plasmid vector MR208244 (Origene) was used to develop pCMV6-A-BSD-MTA3-wt, pCMV6-A-BSD-MTA3-C402A, pCMV6-A-BSD-MTA3-C405A, and pCMV6-A-BSD-MTA3-C402A-C405A (BSD-blasticydine) plasmid vectors overexpressing the mutant version of MTA3, respectively carrying mutation of cysteine to alanine residue at 402, 405 and both at 402 and 405 positions. Site directed mutagenesis was carried out using QuikChange II Site-Directed Mutagenesis Kit from Agilent following manufacturer’s protocol. Briefly, PCR reaction was performed using primer pair: C402AMutation-Forward (5’-CATGCAGTGTAGACTCGCCGCGACCTGTTGGCTG-3’) and C402A Mutation-Reverse (5’-CAGCCAACAGGTCGCGGCGAGTCTACACTGCATG-3’); C405A Mutation-Forward (5’-GTGTAGACTCTGCGCGACCGCTTGGCTGTATTGGAAAAAG-3’) and C405A Mutation-Reverse (5’-CTTTTTCCAATACAGCCAAGCGGTCGCGCAGAGTCTACAC-3’); C402AC405A Mutation-Forward (5’-CTAACATGCAGTGTAGACTCGCCGCGACCGCTTGGCTGTATTGGAAAAAGT-3’) and C402AC405A Mutation-Reverse (5’-ACTTTTTCCAATACAGCCAAGCGGTCGCGGCGAGTCTACACTGCATGTTAG-3’). The PCR products were treated with DpnI and transformed into E. coli Top10 competent cells by heat shock. Single clones were sequenced to examine the C402A, C405A and C402A-C405A mutations.

Data analysis

All experiments were repeated at least three times. Results are expressed as the mean±SEM. Statistical comparisons between 2 groups were performed via Student t test. One-way ANOVA was used to compare the means of multiple groups. P<0.05 was considered statistically significant (*, p< 0.05; **, p<0.01). All analyses were performed with GraphPad Prism 6.0 (La Jolla, CA).

Results

An optimal zone of innate immune signaling for generating iPSCs

To determine if boosting the innate immune signal could further increase iPSC yield, BJ fibroblasts were transduced with retroviral vectors expressing OSKM factors alone (as control) or in combination with the TLR3 agonist PIC (300 ng/ml). We observed that the addition of PIC to the retroviral vectors markedly reduced the yield of iPSC colonies (Figure 1A). This result shows that excessive activation of innate immune signaling may impair nuclear reprogramming to pluripotency and compelled us to determine the range of innate immune activation that is optimal for iPSC generation.

Figure 1. Overactivation of innate immunity reduces nuclear reprogramming to iPSCs: Presence of Goldilocks zone.

(A) BJ fibroblasts were treated with retroviral vectors carrying OSKM, in the presence of PIC (300 ng/ml) or vehicle. The number of alkaline phosphatase positive (AP⁺) iPSC colonies were counted on day 30. (B) Secondary Dox MEFs (passage #3) were plated and treated with dox (2 μg/ml) in absence or presence of PIC (0–10000 ng/ml) or NFkB oligonucleotide decoy p65i (10 or 50 μM). The generation of iPSC colonies (detected by SSEA-1 cell surface markers) was scored at day 21. In parallel, innate immune activation was monitored by luminescence using a luciferase construct driven by an NFkB promoter. SSEA-1, stage-specific embryonic antigen-1. (C) Alkaline phosphatase (AP) staining of iPSCs on day 21 after treatment with doxycycline in the presence of vehicle (Dox); PIC 30ng/ml (DP30); or PIC 1000ng/ml (DP 1000). (D) MTT assay showing cell viability is not compromised at higher concentration of PIC. Data are shown as the means ± SEM and are representative of 3 independent experiments (**, p<0.01; ns=non-significant).

Accordingly, we developed a model of nuclear reprogramming that is more amenable to define the boundaries of this optimal zone. We used murine embryonic fibroblasts containing a doxycycline-inducible cassette encoding the Yamanaka factors (Oct4, Sox2, Klf4 and c-Myc) that can be reproducibly reprogrammed to iPSCs simply by adding doxycycline (Dox) to the medium. Furthermore, we transfected the Dox-inducible MEFs with an NFκB promoter luciferase construct to assess the level of innate immune activation. We then added Dox to induce reprogramming, together with agents to increase or reduce innate immune activation in a dose-dependent manner. As shown in Figure 1B, the TLR3 agonist polyinosinic-cytidilic acid (PIC) increased NFκB activation in a dose-dependent manner (from 3 to 10,000ng/ml) as manifested by increasing luciferase activity. By contrast, in this same range of PIC stimulation, we observed a biphasic effect on iPSC colony generation. In a lower range (from 3–300 ng/ml), PIC increased colony generation. However, higher doses of PIC (1000 and 10000 ng/ml) were associated with less colony formation (Figure 1B, C). These data provided evidence for an optimal zone (Goldilocks zone) of innate immune activation where nuclear reprogramming to pluripotency is maximum. The reduced yield of iPSCs at higher doses of PIC was not due to cell death, as cells remained viable at high concentration of PIC (by MTT assay; Figure 1D). The concentration of poly I:C which generated maximal iPSCs was 30ng/ml. Interestingly, Dox-MEFs manifested a modest activation of NFκB even in absence of PIC suggesting a basal level of innate immune activation in these cells that facilitates reprogramming. Treatment of Dox-MEFs with increasing concentration of an NFκB oligonucleotide decoy (p65i) reduced NFκB activity below basal level, and reduced iPSC yield in a dose dependent manner (Figure 1B). Together, these results indicate the existence of an optimal zone of innate immune activation where the efficiency of iPSC generation is maximum.

Generation of NO defines the Goldilocks zone during nuclear reprogramming

The role of iNOS as an effector of innate immunity has been extensively studied^{8, 9}. Furthermore, we have shown that innate immune signaling is required for transdifferentiation of fibroblasts to endothelial cells¹⁰ and that iNOS plays a role in this phenomenon¹¹. Specifically, iNOS translocates to the nucleus to S-nitrosylate RING1A, a component of the polycomb repressive complex 1 (PRC1). Nitrosylation of RING1A reduces its binding to chromatin and decreases suppressive histone markings of H2K119ub and H3K27me3¹¹.

However, with respect to nuclear reprogramming to pluripotency, a role for S-nitrosylation of epigenetic modifiers has not been shown. To determine if there was a role for NO in reprogramming to pluripotency, we began by measuring NO generation using DAF-FM DA reagent. DAF-FM DA is essentially non-fluorescent until it reacts with NO to form a fluorescent benzotriazole¹². Reprogramming was induced by the addition of doxycycline to the medium, in the presence or absence of increasing doses of PIC. Nuclear reprogramming was associated with an increase in NO generation during the initial phase. The generation of NO was greatest at the concentration of PIC (Dox+PIC30 ng/ml; DP30) that also generated the most iPSC colonies. With less activation of innate immunity (Dox alone) or with excessive activation of innate immunity (Dox+PIC 1000 ng/ml; DP1000), NO generation was reduced (Figure 2A, S1A). This result was further confirmed by Griess assay (Figure S1B) which showed that level of nitrogen oxides were highest at DP30.

Figure 2. Role of NO in defining the Goldilocks zone during nuclear reprogramming.

(A) Dox-MEFs were subjected to nuclear reprogramming with doxycycline in the presence of vehicle (dox) or PIC 30ng/ml (DP30) or PIC 1000ng/ml (DP1000). The generation of NO was measured on day 3 using DAF-FM DA. Fluorescence intensity was measured at 515 nm by using a fluorescent plate reader. Data are shown as the means ± SEM and are representative of 3 independent experiments. (B) These studies were repeated in the presence of vehicle or L-NAME (10μM) or (C) in the presence of iNOS specific inhibitors 1400W (10μM) and BYK-hydrochloride (10μM), and the generation of iPSC colonies assessed at 21days using alkaline phosphatase (AP) staining. (D) Nuclear reprogramming of wt-MEFs and iNOS deficient MEFs using lentiviral overexpression of the OSKM factors. (E) Relative expression of iNOS during nuclear reprogramming with doxycycline on day 4 in presence of vehicle (Dox), PIC 30ng/ml (DP30), or PIC 1000ng/ml (DP1000). (F) iNOS protein expression in nuclear and cytoplasmic fractions obtained from Dox-MEFs undergoing nuclear reprogramming with doxycycline on day 4 in absence or presence of PIC (30 or 1000 ng/ml). Histograms display data as the means ± SEM and are representative of 3 independent experiments (*, p< 0.05; **, p<0.01).

To determine if generation of NO was required for nuclear reprogramming, we performed nuclear reprogramming with Dox-MEFs in the presence or absence of the NOS inhibitor L-NAME (Figure 2B). The generation of iPSC colonies induced by DP30 were reduced by L-NAME, an effect that was confirmed by using two different iNOS specific inhibitors 1400W or BYK-hydrochloride (Figure 2C). Each of the inhibitors could reduce nuclear reprogramming efficiency.

To confirm the role of iNOS in nuclear reprogramming to pluripotency in a different model system, we assessed the efficiency of nuclear reprogramming of wild type and iNOS ^−/− MEFs using lentiviral overexpression of the Yamanaka factors. We observed that nuclear reprogramming was nearly abrogated in the iNOS^−/− MEFs compared to wild type MEFs (Figure 2D).

Subsequently, we measured iNOS expression from Dox-MEFs undergoing nuclear reprogramming at optimal and sub-optimal zones of innate immune activation (on day 4 of reprogramming; Figure 2E). These data suggested that total iNOS expression was slightly greater at DP30 compared to the sub-optimal zones of nuclear reprogramming. Moreover, when we examined the compartmentalization of iNOS, we found greater differences. Specifically, by comparison to the DP1000 condition, there was more nuclear iNOS expression with the DP30 condition (Figure 2F). This is significant as we have previously shown that in transdifferentiation of one somatic cell to another lineage, the nuclear translocation of iNOS is associated with binding to and S-nitrosylation of Ring 1A (a component of PRC1), together with a reduction in the suppressive histone markings associated with PRC1¹¹.

S-nitrosylation of nuclear proteins participates in nuclear reprogramming

Optimal zone for S-nitrosylation of nuclear proteins.

It has been well established that NO can modulate the activity of cytoplasmic signaling proteins by S-nitrosylation¹³. Several nuclear proteins are known to be S-nitrosylated¹⁴ (Supplemental Table 1), but the role of S-nitrosylation in nuclear reprogramming to pluripotency has not been studied. Accordingly, we assessed the extent of nuclear protein S-nitrosylation using Tandem Mass Tag (TMT) approach, a modification of the biotin switch assay. We observed that S-nitrosylation of nuclear proteins is greatest at the same concentration of PIC that yields the most iPSC colonies (Figure 3A). Addition of L-NAME during optimal nuclear reprogramming reduced the extent and intensity of the S-nitrosylated nuclear proteins.

Figure 3. S-nitrosylation of MTA3 defines the Goldilocks zone.

(A) Nuclear proteins were isolated from Dox-MEFs undergoing nuclear reprogramming on day 4. S-nitrosylated proteins in different treatments were detected using iodo-TMT reagent as described in materials and methods. (B) Representative MS/MS fragmentation spectrum for the Cys402 and Cys405-containing peptide of MTA3. The spectrum confirms the identity of the peptides LCATCWLYWK and the labeled C as S-nitrosylated cysteine. (C) Co-immunoprecipitation of iNOS and MTA3 from nuclear proteins during day 4 of nuclear reprogramming (D) Extent of S-nitrosylation of MTA3 (detected with TMT antibody after labeling with TMT) in nuclear fractions of Dox-MEFs undergoing doxycycline induced nuclear reprogramming on day 4 in absence or presence of PIC. (E) Protein expression level of MTA3 in nuclear and cytoplasmic fractions of Dox-MEFs undergoing doxycycline induced nuclear reprogramming on day 4 in absence or presence of PIC. (F) Protein expression level of other subunits of NuRD complex in nuclear fractions of Dox-MEFs undergoing doxycycline induced nuclear reprogramming on day 4 in absence or presence of PIC. (*, p< 0.05; **, p<0.01).

MTA3 is S-nitrosylated during nuclear reprogramming

We observed several S-nitrosylated bands in the western analysis of DP30 treated Dox-MEF nuclear extract (Figure 3A). We focused initially on one discrete band (below 70kD region, indicated by arrow, Figure 3A) as the intensity of this band was significantly reduced in presence of L-NAME. We performed mass spectrometry and found that one of the proteins from this band was S-nitrosylated at cysteine 402 and cysteine 405 residues (Figure 3B). The protein was identified as Metastasis Associated 1 Family Member 3 (MTA3). This protein is a subunit of the Nucleosome Remodeling Deacetylase (NuRD) complex. To determine if iNOS interacts with MTA3 during S-nitrosylation, we subjected nuclear lysate from reprogramming cells to co-immunoprecipitation assay and found that iNOS and MTA3 interact. Their interaction is maximal with the DP30 condition (Figure 3C).

Optimal zone for S-nitrosylation of MTA3

Next, we assessed the extent of S-nitrosylation of MTA3 throughout the concentration range of innate immune stimulation. We pulled down MTA3 from nuclear fractions during nuclear reprogramming on day 4 from cells treated with different concentrations of PIC. As shown in Western analysis, MTA3 is heavily S-nitrosylated in those cells treated with a concentration of PIC 30ng/ml, by comparison to cells treated with lower or higher doses (Figure 3D). Furthermore, the level of MTA3 protein expression in the nuclear fraction during optimal nuclear reprogramming (i.e. in the presence of PIC 30ng/ml) is significantly less by comparison to its level in cells treated with lower or higher concentrations of PIC (Figure 3E). In contrast, the cytoplasmic expression level of MTA3 is greatest during optimal nuclear reprogramming. However, we found little difference in the RNA (Figure S2A, S2B) and protein expression levels of other subunits (HDAC1, HDAC2, RBAP46, MBD3, CHD4) of the NuRD complex in the nuclear fraction throughout the concentration range of PIC (Figure 3F). These results suggest that differences in the extent of S-nitrosylation and nuclear compartmentalization of MTA3 may be critical for optimal nuclear reprogramming.

MTA3 S-nitrosylation reduces NuRD deacetylase activity

The NuRD complex is a barrier to nuclear reprogramming to pluripotency; knockdown of elements of the NURD complex increases, whereas overexpression of the NURD complex reduces, the yield of iPSCs^{15, 16}. The NURD complex consists of several subunits including MTA1 or MTA2 or MTA3; the exact composition of the NURD complex varies in different cell types¹⁷. The NuRD complex exerts its histone deacetylase activity by its two subunits HDAC1 and HDAC2^{18, 19}. As MTA3 is known to interact with both HDAC1 and HDAC2^{20, 21}, we were interested to see if S-nitrosylation of MTA3 has any effect on histone deacetylase activity. Accordingly, we assessed the intensity of HDAC activity throughout the concentration range of innate immune stimulation. We observed that HDAC1 and HDAC2 activities were least at the concentration of PIC (30ng/ml) that yields the most iPSC colonies (Figure 4A, B). A reduction in the activities of HDAC1 and HDAC2 should increase H3K27ac marks²². Indeed, ChIP-qPCR studies revealed that H3K27ac marks were increased in the promoter region of the major pluripotency gene Oct4 when Dox-MEFs were treated with PIC 30ng/ml during reprogramming (Figure 4C). To further confirm that optimal innate immune signaling is associated with the greatest suppression of HDAC activity, we investigated H3K27ac and H3K27me3 marks on the promoter region of additional pluripotency genes. By ChIP-qPCR we found that the activation mark H3K27ac increases whereas the repressive mark H3K27me3 decreases in the promoter region of Sox2 and Nanog, to the greatest degree in the optimal range of innate immune activation (Figure 4D–G). These observations were consistent with our finding by ChIP-qPCR assay that HDAC1 binding to the Oct4 promoter was significantly decreased in the optimal zone of innate immune stimulation (Figure 4H). In a co-immunoprecipitation assay, we also found that interaction of MTA3 with HDAC2 is least within the optimal zone of innate immune stimulation (Figure 4I). We further determined if S-nitrosylation of MTA3 has any effect on its occupancy at Oct4 promoter during nuclear reprogramming. The result (Figure 4J) showed decreased occupancy in the optimal zone compared to the suboptimal zones of innate immune stimulation. Taken together, these data suggest that S-nitrosylation of MTA3 during optimal nuclear reprogramming disrupts the interaction with HDAC2 to impair the deacetylase activity of the NuRD complex. The reduction in deacetylase activity would be expected to result in an increase in the activating epigenetic mark H3K27ac at the promoter of the pluripotency genes Oct4, Sox2 and Nanog.

Figure 4. S-nitrosylation of MTA3 reduces NuRD activity.

HDAC1 (A) and HDAC2 (B) activity of Dox-MEFs undergoing nuclear reprogramming in the presence or absence of PIC. ChIP-qPCR in presence or absence of PIC for analyzing enrichment of H3K27ac marks on Oct4 (C), Sox2 (D) and Nanog (E) promoter; H3K27me3 marks on Sox2 (F) and Nanog (G) promoter; and HDAC1 on Oct4 (H) promoter on day 4 of nuclear reprogramming (I) Interaction of HDAC2 with MTA3 in the presence or absence of PIC. (J) ChIP-qPCR in presence or absence of PIC for analyzing occupancy of MTA3 on Oct4 promoter on day 4 of nuclear reprogramming. (*, p< 0.05; **, p<0.01).

Both C402 and C405 S-nitrosylation of MTA3 is important for optimal nuclear reprogramming

To determine whether the S-nitrosylation of the cysteine residues (C402 or C405) of MTA3 are mediating the effects on nuclear reprogramming, we performed site directed mutagenesis. We first constructed plasmid vectors containing wild type MTA3 (pCMV6-A-BSD-MTA3-wt) or mutant MTA3 [C402A, C405A, or the double mutant C402A-C405A]. The pCMV6-A-BSD-MTA3-C402A construct encoded a mutant MTA3 with the TGC codon encoding the 402 position cysteine mutated into a GCC codon encoding alanine; the pCMV6-A-BSD-MTA3-C405A construct encoded a mutant MTA3 with the TGT codon encoding the 405 position cysteine mutated into a GCT codon encoding alanine; and the pCMV6-A-BSD-MTA3-C402AC405A construct encoded the double mutant (dm) MTA3 with both codons mutated. The mutations and expression level of the constructs were confirmed by sequencing and qPCR (Figure 5A, S3). Then these plasmid vectors, or the empty vector pCMV6-A-BSD, were separately transfected into Dox-MEFs, which were then subjected to nuclear reprogramming in the optimal range of innate immune stimulation (i.e. in the presence of PIC 30 ng/ml). As shown in Figure 5B, the overexpression of C402AC405A double mutation of MTA3 significantly reduced nuclear reprogramming. Furthermore, the fold enrichment of HDAC1 and HDAC2, determined by ChIP-qPCR assay, significantly increased at the Oct4 promoter on day 4 in the Dox-MEF cells overexpressing the double mutants (MTA3-dm) during nuclear reprogramming (Figure 5C, D). In order to determine if S-nitrosylation of MTA3 is important for epigenetic modifications, we assessed the amount of H3K27ac, H3K9me3, H3K27me3 and H3K79me3 in MTA3-wild type and MTA3-dm cell lines on day 4 of optimal nuclear reprogramming. We found an increase in the repressive H3K9me3, H3K27me3 and H3K79me3 marks and a decrease in the H3K27ac activation mark for Dox MEFs overexpressing MTA3-dm by comparison to Dox- MEFs overexpressing the MTA3-wild type (Figure 5E). Altogether, these data suggest that S-nitrosylation of MTA3 at both the 402 and the 405 cysteine is critical for epigenetic changes necessary for optimal nuclear reprogramming.

Figure 5. S-nitrosylation of C402 and C405 residues of MTA3 is important for optimal nuclear reprogramming.

(A) Sequence confirmation of site directed mutagenesis of cysteine 402 and C405 for plasmid vectors pCMV6-A-BSD-MTA3-C402A, pCMV6-A-BSD-MTA3-C405A, and pCMV6-A-BSD-MTA3-C402A-C405A plasmid vectors. (B) Dox-MEFs were separately transfected with pCMV6-A-BSD-MTA3-wt, pCMV6-A-BSD-MTA3-C402A, pCMV6-A-BSD-MTA3-C405A, and pCMV6-A-BSD-MTA3-C402A-C405A (MTA3-dm) plasmid vectors and then subject to nuclear reprogramming in the presence of PIC 30ng/ml. At day 21, number of iPSCs were calculated using AP staining. HDAC1 (C) and HDAC2 (D) binding to Oct4 promoter analyzed by ChIP-qPCR in Dox-MEFs overexpressing pCMV6-A-BSD-MTA3-C402A-C405A (MTA3-dm) or MTA3-wt plasmid vector and undergoing nuclear reprogramming in the presence of PIC. (E) H3K27ac, H3K9me3, H3K27me3 and H3K79me3 marks in MTA3-wild type and MTA3-dm cell lines during day 4 of nuclear reprogramming. (**, p<0.01)

DNA accessibility depends on the extent of innate immune activation

While histone deacetylation favors a closed chromatin state, histone acetylation promotes DNA accessibility²³. Accordingly, we determined HAT activity in nuclear lysate isolated on day4 from Dox-MEFs undergoing nuclear reprogramming at low (Dox +Bay 11), optimal (Dox +PIC30) and excessive (Dox+PIC1000) innate immune activation (Figure 6A). We observed that HAT activity was greatest at PIC30 ng/ml. The greater HAT activity at PIC30ng/ml was associated with the greatest decrease in the repressive marks of H3K9me3, H3K27me3 and H3K79me3 on day 4 of nuclear reprogramming (Figure 6B). No significant changes were observed in other epigenetic marks (Figure S4).

Figure 6. DNA accessibility depends on extent of innate immune activation: Goldilocks zone of DNA accessibility.

(A) HAT activity and (B) determination of epigenetic marks of Dox MEFs undergoing nuclear reprogramming at low (Dox+Bay 11), optimal (Dox+PIC30; DP30) and high (Dox+PIC 1000 ng/ml; DP1000) innate immune activation on day 4. (C) Mono(M)-, di (D)- and tri(T) nucleosome fractions at 20 minutes of MNase digestion of chromatin obtained from Dox MEFs undergoing nuclear reprogramming at low, optimal and high innate immune activation on day 4. “L” and “U represents the lower and upper marker” respectively. (D) Mono (M)- to tri(T)nucleosome ratio at 20 minutes of MNase digested chromatins from Dox MEFs undergoing nuclear reprogramming at low, optimal and high innate immune activation on day 4. (E) Mono(M)-; di (D)- and tri(T)nucleosome fractions at 20 minutes of MNase digestion of chromatin obtained from Dox MEFs undergoing optimal nuclear reprogramming (day 4) in absence or presence of L-NAME. “L” and “U represents the lower and upper marker” respectively. (F) Mono(M)- to tri(T)nucleosome ratio at 20 minutes of MNase digested chromatin from Dox MEFs undergoing optimal nuclear reprogramming (day 4) in absence or presence of L-NAME. (G) Mono(M)-; di (D)- and tri(T)nucleosome fractions at 20 minutes of MNase digestion of chromatin obtained from Dox MEFs undergoing optimal nuclear reprogramming (day 4) overexpressing plasmid vector pCMV6-A-BSD-MTA3-C402A-C405A (MTA3-dm) or MTA3-wt. “L” and “U represents the lower and upper marker” respectively. (H) M/T ratio at 20 minutes of MNase digested chromatins from Dox MEFs undergoing optimal nuclear reprogramming (day 4) overexpressing plasmid vector pCMV6-A-BSD-MTA3-C402A-C405A (MTA3-dm) or MTA3-wt. (*, p< 0.05; **, p<0.01).

Recent studies showed that DNA accessibility is important for cellular reprogramming to iPSCs^{24, 25}. To gain more insight regarding DNA accessibility throughout the range of innate immune activation, we performed micrococcal nuclease (MNase) digestion assay of chromatin isolated on day 4 from the reprogramming Dox-MEFs. With greater DNA accessibility, one should observe a greater increase in the mononucleosome peak with MNase digestion. We observed that the mononucleosome fraction was greatest in reprogramming MEFs that had been exposed to PIC 30ng/ml (Figure 6C, S5B, S5C). To normalize the data we calculated the mono(M) to di(D), as well as the mono to tri(T)-nucleosome ratios (M/D, M/T ratios) as a reflection of the extent of DNA accessibility²⁶. We found that the M/T (Figure 6D, S5D) and M/D (Figure S5E) ratios are highest in the optimal zone of innate immune stimulation.

To determine if NO generation could influence DNA accessibility, we repeated these studies with the NOS inhibitor L-NAME. As expected, treatment with L-NAME decreased DNA accessibility, as reflected by the reduced area of the mono-nucleosome fractions (Figure 6E, S5G), as well as the reduced M/T (Figure 6F, S5H) and M/D (Figure S5I) ratios. To determine if the effect of NO generation on DNA accessibility was mediated in part by S-nitrosylation of MTA3, we repeated these studies in Dox-MEFs overexpressing MTA3-dm (mutated at both S-nitrosylation sites) and MTA3-wt. We observed that DNA accessibility during reprogramming of the MTA3-dm Dox-MEFs was reduced by comparison to MTA3 wild-type (MTA3-wt) Dox-MEFs (Figure 6G, S5K) as quantified by the M/T (Figure 6H, S5L) and M/D (Figure S5M) ratios. Taken together, these results indicate that the DNA accessibility of reprogramming Dox-MEFs is maximal at the same intensity of innate immune stimulation at which iPSC yield, NO generation, and S-nitrosylation are maximal.

Discussion

Seminal findings.

In this study we show that efficient reprogramming to pluripotency requires an optimal activation of innate immune signaling. We observe a biphasic response in iPSC generation based on the degree of concomitant innate immune activation during reprogramming (Figure 1B). The biphasic response in iPSC yield is closely paralleled by biphasic changes in iNOS expression, NO generation, S-nitrosylation of nuclear proteins, HDAC and HAT activities, repressive and activating histone markings, as well as DNA accessibility (Figure 2A, S1A, S1B, 2E, 3A, 4A–G, 6A–D, S5B–E). The changes in HDAC activity and DNA accessibility are mediated in part by S-nitrosylation of MTA3 of the NuRD complex (Figure 3B). This unique modification of an epigenetic element reduces the activity of the NuRD complex (Figure 4A, 4B) and its association with the chromatin (Figure 4I, 4J). In sum, our work reveals that optimal innate immune signaling is required for efficient iPSC generation, and is mediated to a significant degree by iNOS activity and S-nitrosylation of nuclear proteins to increase DNA accessibility. At suboptimal (little or excessive) innate immune activation, NO generation as well as DNA accessibility is less, resulting in impaired nuclear reprogramming to pluripotency.

Role of innate immune activation in cell fate transitions.

Two recent studies showed that DNA accessibility is dynamic during iPSC reprogramming^{24, 25}. Nuclear reprogramming with retroviral vectors encoding pluripotency genes Sox2, Klf4 and Oct4 drives chromatin from a closed to an open state²⁴. However, these studies did not elucidate the driving force promoting the open chromatin configuration at the promoter region of pluripotency genes. We have shown that nuclear reprogramming (using a viral vector or mRNA) activates PRRs that mediate activation of NFkB and IRF3. These transcriptional effectors of inflammatory signaling induce global changes in the expression of epigenetic modifiers that promote histone markings and favor an open chromatin state^{5, 6}. Pharmacological antagonism or genetic knockdown of elements in the inflammatory signaling cascade reduce or abrogate the generation of iPSCs^{6, 27}.

We have found that a similar mechanism regulates the transdifferentiation of one somatic cell into another somatic cell lineage. Whereas most investigators who study therapeutic transdifferentiation use viral vectors encoding lineage-determining transcription factors^{28, 29}, we showed that pharmacological activation of PRRs, together with instructional factors in the medium, are sufficient to induce transdifferentiation. Specifically, human fibroblasts can be induced to transdifferentiate into endothelial cells, by activating innate immune signaling, in the presence of “endothelial instructional factors” (i.e. VEGF, FGF, BMP4 and 8BrcAMP) in the media^{10, 11}. This transdifferentiation does not occur in the absence of innate immune activation, and is attenuated with pharmacological or genetic knockdown of elements of the inflammatory signaling cascade. The pharmacological manipulation of innate immune signaling and epigenetic plasticity is a first step toward a small molecule approach to therapeutic transdifferentiation.

We also found that transdifferentiation of fibroblasts to endothelial cells involved induction of iNOS and its translocation to the nucleus¹¹. There it binds to, and S-nitrosylates RING1A of the polycomb complex 1 (PRC1), reducing the chromatin binding and action of PRC1, the chromatin, associated with a global reduction in H3K27me3. We now show that a similar mechanism is operative in nuclear reprogramming to pluripotency. During the induction of pluripotency, we observe an increase in NO generation (Figure 2A, S1A) and S-nitrosylation (Figure 3A) of nuclear proteins in association with induction of iNOS (Figure 2E). We observe that MTA3 of the NuRD complex is S-nitrosylated (Figure 3B), in association with reduced association of MTA3 with chromatin (Figure 4I, 4J) and a reduction in HDAC activity (a function of the NuRD complex) (Figure 4A, 4B). Pharmacological antagonism or genetic deficiency of iNOS markedly attenuates iPSC yield during nuclear reprogramming (Figure 2B–D). Interestingly, administration of sin1 (NO donor) exogenously does not rescue iPSC generation in iNOS deficient MEFs (data not shown). Similarly, exogenous NO donors did not rescue transdifferentiation of iNOS deficient fibroblasts to endothelial cells¹¹. These data suggest that the function of iNOS may require its translocation to the nucleus, where it binds to and S-nitrosylates epigenetic modifiers (Figure 2F, 3C). In this regard, mutation of the cysteine residues of MTA3 (to disrupt its S-nitrosylation) (Figure 5A) reduces DNA accessibility (Figure 6G, 6H, S5K–M) and iPSC generation (Figure 5B).

We observed multiple bands of S-nitrosylated nuclear proteins (Figure 3A), but focused on MTA3 in this study. It is likely that other nuclear proteins are S-nitrosylated, and some of these may be involved in epigenetic modification. Recently, a consensus binding sequence has been described for cytoplasmic proteins that are bound by iNOS, a conserved I/L-X-C-X₂-D/E motif³⁰, that is recognized by an S-nitrosylase complex consisting of iNOS, S100A8, and S100A9. We performed a bioinformatics analysis revealing that 20% of epigenetic modifiers harbor this conserved motif (data not shown). However, protein sequence analysis of MTA3 indicated that the I/L-X-C-X₂-D/E motif is absent in MTA3. Thus, our bioinformatics analysis may underestimate the epigenetic modifiers that may be S-nitrosylated.

A Goldilocks zone for innate immune stimulation and iPSC generation.

We show that within a physiological range of innate immune stimulation there is an optimal zone for inducing pluripotency (Figure 1B, 1C). To define the boundaries of this zone, we used murine embryonic fibroblasts carrying a doxycycline-inducible cassette of the Yamanaka factors (dox-MEF) and a luciferase reporter. We induced reprogramming by adding doxycycline to the medium, and monitored the level of innate immune activation using the luciferase reporter (Figure 1B). We found that the Dox-MEFs had a basal level of innate immune activation. When we abrogated this basal level of innate immune activation using a decoy oligonucleotide to p65, or the NFkB antagonist Bay 11, iPSC generation was nearly abolished. Alternatively, when we increased innate immune activation by adding the TLR3 agonist PIC, iPSC yield increased, with a maximal effect at a dose of PIC 30ng/ml. However, at higher doses of PIC (which increased NFkB activation) fewer iPSC colonies were observed. In the case of retroviral OSKM nuclear reprogramming of BJ fibroblasts exogenous PIC did not improve the nuclear reprogramming efficiency. This finding may be due to the fact that exposure to retroviruses provides for sufficient innate immune activation for reprogramming.

With respect to the determinants of the Goldilocks zone, we observed a biphasic response in NO generation (Figure 2A). A similar biphasic response was observed for iNOS induction (Figure 2E) as well as the intensity of S-nitrosylation of nuclear proteins (Figure 3A). These biphasic responses were mirrored by biphasic change in HAT activity (Figure 6A), and reciprocal changes in HDAC activity (Figure 4A, 4B). The activating (Figure 4C–E) or suppressive histone markings (Figure 4F, 4G, 6B) followed HAT or HDAC activity respectively. The data suggest that nuclear S-nitrosylation of epigenetic modifiers plays a major role in defining the optimal zone for nuclear reprogramming to pluripotency. This notion is supported by the studies of MTA3-dm Dox-MEFs. This cell line lacks the two sulfhydryl groups of cysteine that are targets of S-nitrosylation (Figure 5A). In these cells, reprogramming to pluripotency is nearly abolished, in association with a reduction in the epigenetic modifications that are required for DNA accessibility (Figures 5B–E). Thus it seems likely that iNOS-generated NO plays a major role to modulate epigenetic control of DNA accessibility during nuclear reprogramming.

Physiological and clinical relevance of our findings.

Pattern recognition receptors (PRRs) are ubiquitous in mammalian cells and represent an early detection system for cellular challenges, being activated by damage associated molecular patterns (DAMPs) and pathogen associated molecular patterns (PAMPs). The well-characterized innate immune signaling cascade triggers the release of inflammatory cytokines, chemokines, and reactive oxygen species that initiate the first cellular response to damage or pathogens. We have discovered a new limb of this cascade that induces global changes in the expression and activity of epigenetic modifiers to increase DNA accessibility^{5, 11}. The increase in DNA accessibility provides the cell with the phenotypic fluidity that permits rapid adaptation to cellular challenges. We have termed this process “transflammation”. We show that DNA accessibility depends on the extent of innate immune activation (Figure 6C, 6D, S5B–E). The phenotypic fluidity in response to PAMPs and DAMPs may be an evolutionary mechanism for cell survival. However, in the presence of overwhelming PAMPs (as with a large multiplicity of infectious particles) and excessive activation of innate immune signaling, it may be advantageous for the cell to decrease DNA accessibility to avoid viral usurpation of cellular machinery.

We speculate that this Goldilocks zone is broadly relevant to scientific and clinical arenas as diverse as regenerative medicine and cancer biology. For example, clinicians have long recognized that some amount of inflammation is necessary for wound healing. Patients on steroids have an attenuated inflammatory response, and heal poorly after surgery^{31, 32}. By contrast, the patient with a diabetic foot ulcer has a ring of highly inflamed skin circumscribing the non-healing wound³³. Accordingly, methods to identify whether a patient or tissue is in an optimal zone of inflammatory signaling; and drugs or biologics that maintain the patient or tissue in an optimal zone; might improve surgical and medical therapies. Similarly, oncologists have long been aware that cancer often arises at sites of inflammation^{34, 35}, for example, the patient developing esophageal cancer in the setting of chronic esophagitis^{36, 37}. Based on the coupling of inflammatory signaling and DNA accessibility, one might imagine that there is increased DNA injury and mutation with increased exposure of the DNA template to reactive oxygen and nitrogen species. The combination of increased mutational burden and increased epigenetic plasticity would set the stage for malignancy. Early detection of chronic inflammatory signaling and DNA accessibility, and methods to reverse these processes, might guide cancer prevention.

Whereas some authors have provided data suggesting a positive role of innate immune activation in skin regeneration³⁸ or recovery from ischemia/reperfusion injury³⁹, others report adverse effects of innate immune signaling in regenerative processes^40–43. Our observation of an optimal zone for innate immune signaling for epigenetic plasticity may explain the contradictory reports in the literature regarding the role of innate immune signaling in tissue regeneration and repair.

Conclusions

In summary, our study identified an unrecognized optimal (Goldilocks) zone of innate immune activation for nuclear reprogramming to pluripotency. The boundaries of this Goldilocks zone seem to be determined in part by iNOS expression and activity that manifests a biphasic dose response curve to innate immune stimulation. This biphasic response is paralleled by a similar changes in S-nitrosylation of nuclear proteins; activity of epigenetic modifiers; and their respective histone markings. As paradigmatic of how iNOS activity may modify epigenetic control, we show that S-nitrosylation of MTA3 is associated with reduced chromatin binding and activity of HDAC2, favoring an open chromatin state (Figure 7). Our observations may have broad relevance for epigenetic control; for regenerative processes; and for the pathobiology of cancer and other diseases.

Figure 7. Mechanism of increased DNA accessibility within Goldilocks zone of innate immune activation.

Optimal activation of cell autonomous innate immune signaling leads to increased generation of iNOS which S-nitrosylates MTA3 resulting in decreased NuRD deacetylase activity and increased DNA accessibility. In this setting of increased DNA accessibility, the Yamanaka factors (Oct 4, Sox2, Klf4 and c-Myc) can provide transcriptional directionality to the induction of pluripotency. Poly IC: Polyinosinic-polycytidylic acid; TLR3: Toll-like receptor 3; TRAF6: Tumor necrosis factor receptor (TNFR)-associated factor 6; TRIF: TIR-domain-containing adapter-inducing interferon-β; TRAF3: Tumor necrosis factor receptor (TNFR)-associated factor 3; NFκB: Nuclear factor kappa-light-chain-enhancer of activated B cells; iNOS: Inducible nitric oxide synthase; SNO: S-nitrosothiol, formed by covalent addition of nitric oxide to cysteine thiol within a protein; NuRD: Nucleosome Remodeling Deacetylase; MTA3: Metastasis Associated 1 Family Member 3; Ac: acetyl group; HAT: Histone acetyl transferase; OSKM: Oct4, Sox2, Klf4, c-Myc; iPSCs: induced pluripotent stem cells.

Clinical perspective.

1). What is new?

• Our study identified an optimal zone (“Goldilocks zone”) of innate immune activation for nuclear reprogramming to pluripotency.

• This “Goldilocks zone” for nuclear reprogramming may have broad relevance for epigenetic control; for regenerative processes; and for the pathobiology of cancer and other diseases.

2). What are the clinical implications?

• This study may help to develop methods that identify whether a patient or tissue is in an optimal zone of inflammatory signaling; and drugs or biologics that maintain the patient or tissue in an optimal zone; might improve surgical and medical therapies.

• Early detection of chronic inflammatory signaling and DNA accessibility, and methods to reverse these processes, might guide cancer prevention.

• This study also explain the contradictory reports in the literature regarding the role of innate immune signaling in tissue regeneration and repair.

Non-standard Abbreviations and Acronyms

1400W

N-[[3-(aminomethyl)phenyl]methyl]-ethanimidamide dihydrochloride

BJ

Human foreskin fibroblasts

BYK HCl

2-[2-(4-methoxy-2-pyridinyl)ethyl]-3H-imidazo[4,5-b]pyridine dihydrochloride

ChIP

Chromatin immunoprecipitation

c-Myc

Cancer-related human homolog of avian myelocytomatosis viral oncogene

DAF-FM

4-Amino-5-methylamino-2’,7’-difluorofluorescein diacetate

DP30

Dox+PIC30

Dox

Doxycycline

DP1000

Dox+PIC1000

HAT

Histone acetyltransferase

HDAC

Histone deacetylase

iPSC

Induced pluripotent stem cell

iNOS

Inducible nitric oxide synthase

KLF4

Kruppel-like factor 4

L-NAME

Nω-Nitro-L-arginine methyl ester hydrochloride

MEFs

Mouse embryonic fibroblasts

MTA3

Metastasis Associated 1 Family Member 3

MTT

3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide

NFκB

Nuclear factor kappa-light-chain-enhancer of activated B cells

NO

Nitric oxide

NOS

Nitric Oxide Synthase

NuRD

Nucleosome Remodeling Deacetylase

Oct4

Octamer-binding transcription factor 4

PIC

Poly I:C or Polyinosinic-polycytidylic acid

RIG1

Retinoic acid-inducible gene1

ROS

Reactive oxygen species

Sox2

SRY (sex determining region Y)-box 2

TBP

TATA-binding protein

TLR3

Toll-like receptor 3