EMBRYONIC STEM CELLS/INDUCED PLURIPOTENT STEM CELLS a Global Proteomic and Methylome Analysis Molecular Genetics and Control of Pluripotency Laboratory, Faculty of Medicine, Department of in Human Induced Pluripotent Stem Cells Reveals Biomedicine, University of Barcelona, Barcelona, Spain; Overexpression of a Human TLR3 Affecting Proper b Genetics Unit, Department of Biomedicine, Faculty of Medicine, Innate Immune Response Signaling University of Barcelona, IDIBAPS and Hospital Clinic, Barcelona, Spain; c Cancer Epigenetics and Biology JORDI REQUENA,a ANA BELEN ALVAREZ-PALOMO,a MONTSERRAT CODINA-PASCUAL,b Program (PEBC), Bellvitge Biomedical Research Institute RAUL DELGADO-MORALES,c,d SEBASTIAN MORAN,c MANEL ESTELLER,c,e,f MARTÍ SAL,a (IDIBELL), L’Hospitalet de Llobregat, MANEL JUAN,g ANNA BORONAT BARADO,g ANTONELLA CONSIGLIO,h,i,j ORLEIGH ADDELECCIA BOGLE,b Barcelona, Catalonia, Spain; d Department of Psychiatry & ERNST WOLVETANG,k DMITRY OVCHINNIKOV,k INAKI ALVAREZ,l DOLORES JARAQUEMADA,l Neuropsychology, School for Mental JOVITA MEZQUITA-PLA,a,e RAFAEL OLIVA ,b,e MICHAEL J. EDEL a,m Health and Neuroscience (MHeNs), Maastricht University, Maastricht, The Netherlands; eDepartment of Key Words. Cytokine • Human leukocyte antigen • Immune response • Induced pluripotent stem cells • Biomedicine, School of Medicine, Inflammation • MHC-I • Neural stem cells • Toll like receptor University of Barcelona, Barcelona, Catalonia, Spain; fInstitució Catalana de Recerca i Estudis Avançats (ICREA), Barcelona, Catalonia, Spain; g ABSTRACT Service of Immunology, Hospital Clinic, Hospital Sant Joan de Déu, When considering the clinical applications of autologous cell replacement therapy of human Barcelona, Spain; hDepartment of induced pluripotent stem cells (iPSC)-derived cells, there is a clear need to better understand what Pathology and Experimental Therapeutics, Bellvitge University the immune response will be before we embark on extensive clinical trials to treat or model human Hospital-IDIBELL, Hospitalet de disease. We performed a detailed assessment comparing human fibroblast cell lines (termed F1) Llobregat 08908, Spain; iInstitute of reprogrammed into human iPSC and subsequently differentiated back to fibroblast cells (termed F2) Biomedicine of the University of or other human iPSC-derived cells including neural stem cells (NSC) made from either retroviral, Barcelona (IBUB), Barcelona 08028, Spain; jDepartment of Molecular episomal, or synthetic mRNA cell reprogramming methods. Global proteomic analysis reveals the and Translational Medicine, main differences in signal transduction and immune cell protein expression between F1 and F2 cells, University of Brescia, Brescia 25121, implicating wild type (WT) toll like receptor protein 3 (TLR3). Furthermore, global methylome analy- Italy; kStem Cell Engineering Group, sis identified an isoform of the human TLR3 gene that is not epigenetically reset correctly upon Australian Institute for Bioengineering and Nanotechnology, differentiation to F2 cells resulting in a hypomethylated transcription start site in the TLR3 isoform University of Queensland, promoter and overexpression in most human iPSC-derived cells not seen in normal human tissue. Queensland, Brisbane, Australia; The human TLR3 isoform in human iPSC-NSC functions to suppress NF-KB p65 signaling pathway in l Immunology Unit, Department of Cell Biology, Physiology and response to virus (Poly IC), suggesting suppressed immunity of iPSC-derived cells to viral infection. Immunology and Institute of The sustained WT TLR3 and TLR3 isoform overexpression is central to understanding the altered Biotechnology and Biomedicine, immunogenicity of human iPSC-derived cells calling for screening of human iPSC-derived cells for Autonomous University of Barcelona, Bellaterra, Spain; TLR3 expression levels before applications. STEM CELLS 2019;37:476–488 m Department of Physiology, Anatomy and Genetics, Oxford University, Oxford, United Kingdom SIGNIFICANCE STATEMENT Correspondence: Michael J. Edel, The cell reprogramming process alters the immune response of most induced pluripotent stem Ph.D., University of Barcelona, Faculty of Medicine, 143 Casanova, 08036 cells (iPSC)-derived cells highlighting human toll like receptor protein 3 (TLR3) isoform overex- Barcelona, Spain. Telephone: 0034 pression that functions to suppress innate immune signal transduction and provides an insight 934024518; for future clinical applications of autologous human iPSC-derived cell transplantation. The e-mail: edel.michael@gmail.com authors propose that screening human iPSC clones and their derived cells for alterations in the Received September 20, 2018; genome of patient cells that include expression levels of human TLR3 and its shorter TLR3 iso- accepted for publication December 21, 2018; first published online in form identified in this work is essential before embarking on “first in man” clinical trials or STEM CELLS EXPRESS January 3, 2019. human disease modeling using human iPSC. http://dx.doi.org/ 10.1002/stem.2966 This is an open access article under INTRODUCTION major challenges that must be overcome before the terms of the Creative Commons reprogrammed stem cells can be used for clinical Attribution-NonCommercial- The discovery of methods to reprogram somatic NoDerivs License, which permits use applications are genetic instability, proper differ- and distribution in any medium, human cells into tissue specific adult stem cells provided the original work is directly or by full reprogramming, to produce entiation to relevant cell types to regenerate properly cited, the use is non- damaged tissues and the immune response of commercial and no modifications or induced pluripotent stem cells (iPSC) has pre- adaptations are made. sented the possibility of an easy and abundant transplanted cells. The interaction of the human source of cells for autologous stem cell therapy immune system with autologous reprogramed to treat many human diseases [1–3]. Three human cells remains an unexplored question. STEM CELLS 2019;37:476–488 www.StemCells.com ©2019 The Authors. STEM CELLS published by Wiley Periodicals, Inc. on behalf of AlphaMed Press 2019 Requena, Alvarez-Palomo, Codina-Pascual et al. 477 It is well accepted that Toll receptors are a major part of the tissues or cells. We confirm in three retroviral made and three innate immune system essential for life [4]. Likewise, regulation of episomal made human iPSC from healthy control patient cells inflammation cytokines by toll-like receptor 3 (TLR3 - virus) and that the TLR3 isoform is over expressed in five from six matched toll like receptor 4 (TLR4 - bacteria) is well established [5, 6]. pairs of cells (approx. 85% of human iPSC overexpress the Recent publications in animal studies (mice and non-human pri- human TLR3 isoform). We show that the human TLR3 isoform mates) indicate that there may be some immune response to functions to suppress NF-KB p65 signaling pathway in response autologous reprogramed cells [7–14]. A recent review identi- to virus (Poly IC), suggesting suppressed immunity of human fied that many aspects of the adaptive and innate immune sys- iPSC-derived cells to viral infection. Taken together the data tem mechanisms are yet to be explored in detail for human demonstrate that the cell reprogramming process alters spe- cells [3, 15]. Data from human transplanted autologous repro- cific aspects of the immune response of in vitro generated gramed cells in vivo is limited and only one current clinical trial human iPSC-derived cells, suggesting a reduction of innate immu- for iPSC-RPE cells exists that found edema in the patient retina nogenicity of human iPSC-derived cells, calling for screening of most likely from the surgical intervention for the transplanted human iPSC clones for human TLR3 expression levels before use iPSC-RPE cell sheet [16–19]. As the field grows and moves in disease modeling or clinical applications. closer to clinical application the need to understand the extent of the immune response to transplanted human iPSC-derived cells becomes paramount. MATERIALS AND METHODS Of the studies conducted to date in mouse models on the immune response to autologous iPSC-derived cells, the group iPSC Cell Culture from Zhao et al. (2011) studied the immunogenicity of both ret- Established human iPSC clones were obtained from collaborators roviral vector reprogrammed mouse iPSC and episomal vector for this project and previously published. Six human skin fibroblast reprogrammed integration free mouse iPSC. They transplanted cell lines (termed F1) that have previously been reprogrammed into the mouse cells under the skin to evaluate the immunogenicity iPSC, three by retroviral methods [2, 20, 21] and three by non-viral of mouse iPSC-derived cells in syngeneic mice and found episomal methods [22] were used. Synthetic mRNA reprogramming issues with immogenicity [8]. Studies that transplanted mouse to iPSC was performed using Stemgent’s microRNA-enhanced iPSC-derived cells into the neural system and kidney found no mRNA reprogramming kit (STEMGENT, #00-0071, UK) protocol apparent immunogenicity [7, 9]. Further work by Todorova et al. using Thermo Fischer defined cell culture media products. Differen- identified one possible mechanism underlying these conflicting tiation to various cell lines and the mRNA transfection method is conclusions [14]. They show that the lack of immunogenicity of described in full in a separate manuscript. All hiPSC were differenti- mouse iPSC-derived cells in syngeneic host when transplanted ated between passages 10–15. Feeder-free “clinical grade” synthetic into the kidney is due to the absence of functional antigen mRNA iPSC were cultured in defined Flex E8 culture medium (Life presenting cells in mouse kidney, highlighting that the site of Technologies, A2858501, Spain) on Vitronectin (Life Technologies, iPSC-derived cell transplantation in vivo is important in experi- A14700, Spain) coated plates. Colonies were passaged by gentle dis- mental design [14]. A more recent study was performed in sociation with 0.5 mM EDTA when they were 75%–80% confluence. non-human primates testing iPSC-derived neurons found no Cell culture conditions for hiPSC can be found for Episomal and Ret- major immune response, although a small immune response roviral made hiPSC in the papers from Refs. 20–22. was observed [10]. In the study by Zhao et al. (2015), iPSC-derived smooth muscle cells and retinal pigmented epithelial cells were Proteomic Analysis for Global and MHC-1 Peptide transplanted into humanized mice reconstituted with autolo- Analysis gous human immune system, demonstrating differential immu- nogenicity in vivo [12]. Interestingly, IL10 secretion from mouse Protein Solubilization and Quantification. Protein solubiliza- iPSC-derived cells has been found to suppress the immune tion was performed independently for each of the eight pel- response, and IL10 is known to suppress the activity of activated leted cell samples, five with patients fibroblasts (F1) and three T cells, suggesting a mechanism for iPSC-derived cells to by-pass with iPSC-fibroblasts (F2), with 1% (vol/vol) SDS lysis buffer. immune system to survive and engraft [11]. To eliminate viscosity, the lysates were heated to 60 C for There is a clear need to better understand what the 15 minutes and sonicated at 45% amplitude for 4 seconds. immune response to human autologous reprogramed cells will Protein concentration of each sample was estimated using be before we embark on extensive clinical trials to treat BCA Reagent Compatible kit (Thermo Scientific, Rockford, IL) human disease. We have performed a detailed global proteo- following the manufacturer’s instructions. mic and methylome analysis of the possible immune response of matched human patient fibroblast cells (termed F1) repro- Peptides Preparation and Labeling with Isobaric Tags (Tandem gramed to iPSC and compared to iPSC-derived cells (termed F2) Mass Tag 10-Plex). For global analysis, 50 g of proteins from for retroviral, episomal, and synthetic mRNA transfection methods each sample were digested and labeled with tandem mass tag to reprogram human fibroblasts to human iPSC. The data shows (TMT) 10-plex isotopic label reagent set (Thermo Scientific, that in most aspects the immune biology is similar for either viral Rockford, IL) following the manufacturer’s instructions. Nine of or non-viral made iPSC-derived cells; however, a significantly dif- the 10 TMT label reagents (tags: 126; 127N; 127C; 128N; 128C; ferent higher expression of wild type (WT) human TLR3 in iPSC 129N; 129C; 130N; 131) were used. For normalization purposes, derived cells. Global methylome analysis in synthetic mRNA made an internal control consisting on a mixture of 6.25 g of each of human iPSC revealed a shorter isoform of human TLR3 is over the eight samples was included and labeled with TMT126. The expressed in human reprogrammed cells (the TLR3 isoform is TMT labeled samples including control were then combined at not expressed in mice) not seen in normal human physiological a 1:1:1:1:1:1:1:1:1 ratio. Mixture of samples was cleaned by www.StemCells.com ©2019 The Authors. STEM CELLS published by Wiley Periodicals, Inc. on behalf of AlphaMed Press 2019 478 iPSC Overexpress Human TLR3 Affecting Immunity Pierce C18 Spin columns (Thermo Scientific, Rockford, IL) and only unique peptides were considered, and protein ratios were stored at −80 C. normalized to protein median. Normalized values for each pro- tein were exported to Microsoft Excel (v14.0.0), and only those Liquid Chromatography–Tandem Mass Spectrometry proteins with quantification values for all of the samples For global analysis, labeled peptides were separated by means (i.e., eight samples and the internal control) were considered of nanoliquid chromatography using a nanoLC ULTRA AS2 for further analysis. (Eksigent, Australia) with a flow rate of 400 nl/min, a C18 Pep- map 100 trap column (5 μm, 100 Å, 300 μm i.d. × 5 mm in Illumina Infinium Methylation EPIC BeadChip length), and an Acclaim PepMap C18 analytical column (3 μm, 100 Å, 75 μm i.d. × 15 cm in length). All DNA samples were assessed for integrity, quantity, and The following linear gradient, using solvent B (97% ACN, purity by electrophoresis in a 1.3% agarose gel, picogreen 0.1%formic acid) and solvent A (3% ACN, 0.1% formic acid), was quantification, and nanodrop measurements. Bisulfite conver- used: 5%–30% buffer B (140 minutes). MS/MS analysis was per- sion of 500 ng of genomic DNA was performed using EZ DNA formed using an LTQ Orbitrap Velos (ThermoFisher Scientific, methylation kit (Zymo Research. Irvine) following manufac- Rockford, IL) with a nanoelectrospray ion source with precursor turer’s instructions. Bisulfite (200 ng) converted DNA were ion selection in the Orbitrap at 30,000 of resolution, selecting used for hybridization on the Illumina Infinium Methylation- the 20 most intense precursor ions in positive ion mode. MS/MS EPIC BeadChip (Illumina, Inc. San Diego). Briefly, samples were data acquisition was completed using Xcalibur 2.1 (ThermoFisher whole genome amplified followed by an enzymatic end-point Scientific, Rockford, IL). For identification of TMT labeled pep- fragmentation, precipitation, and resuspension. The resus- tides, higher energy collisional dissociation (HCD) with 40% fixed pended samples were hybridized onto the beadchip for 16 hours collision energy was the fragmentation method used. at 48 C and washed. A single nucleotide extension with labeled dideoxy-nucleotides was performed and repeated rounds of Protein Identification and Quantification staining were applied with a combination of fluorescently labeled Data was processed using Proteome Discoverer 1.4.1.14 antibodies differentiating between biotin and DNP. Fluorescent (Thermo Fisher Scientific, Rockford, IL). For database searching, signal from the microarray was measured with a HiScan scanner raw mass spectrometry files were submitted to the in-house (Illumina, Inc. San Diego) using iScan Control Software (V 3.3.29). Homo sapiens UniProtKB/Swiss-Prot database (released 2015-02; A three step-based normalization procedure was performed 20,165 protein entries) using SEQUEST (Thermo FisherScientific, using the lumi [23] package available for bioconductor [24], Rockford, IL). The Percolator search node, a machine-learning under the R statistical environment, consisting in color bias supplement to the SEQUEST search algorithm that increases the adjustment, background level adjustment and quantile normali- sensitivity and specificity of peptide identifications, was also zation across arrays, as specified in [23]. Methylation level used. Searches were performed using the following settings: (β-value) for each of the 866,836 CpG sites were calculated as 2 maximum miss cleavage for trypsin, TMT-labeled in N-Termi- the ratio of methylated signal divided by the sum of methylated nal, lysine (+229.163 Da) and methionine oxidation (+15.995 Da) and unmethylated signals plus 100. All beta values with an asso- as dynamic modifications, cysteine carbamidomethylation ciated p value greater or equal to 0.01 were removed from the (+57.021 Da) as static modification, 20 ppm precursor mass toler- analysis. ance, 0.05mmu fragment mass tolerance, 5mmu peak integration tolerance, and most confident centroid peak integration method. Bisulfite Pyrosequencing The criteria used to accept protein identification included a mini- To validate the methylome array, we performed bisulfite pyrose- mum of two peptides matched per protein, with a false discovery quencing following published protocols [25]. Briefly, specific set of rate of 1%. To avoid the possible ambiguity associated with dif- primers for PCR amplification and sequencing were designed using ferent isoforms of the same protein during the quantification, all the PyroMark assay design program, version 2.0 (Qiagen, Nether- of the proteins were treated as ungrouped. lands); amplification primers hybridize with CpG-free sites to Quantitative analysis of the TMT experiments was per- ensure a methylation-independent reaction and one primer formed simultaneously to protein identification using Prote- (opposite the sequencing primer) is biotinylated to convert ome Discoverer software. Peptide spectral matches (PSMs) the PCR product to single-stranded DNA templates. We used were filtered by PD1.4 software with parameters: Min. Precur- 1 μl of bisulfite-treated DNA for each PCR. The EZ DNA methyl- sor Mass: 300 Da, Max. Precursor Mass: 5,000 Da. Reporter ation kit (Zymo Research, Irvine) was used for bisulfite conver- ion quantification of HCDMS2 spectra was enabled, and TMT- sion of all DNA samples used [25]. To prepare single-stranded 10 plex was set as quantification method. To avoid overlap PCR products, we used the Vacuum Prep Tool (Qiagen, Neth- between the isobaric reporter ions a resolution greater than erlands), following the manufacturer’s instructions. Pyrose- 60,000 was used from m/z 126.1 to m/z 131.1. The quantifica- quencing reactions and methylation quantification were tion values were extracted from the ration between the performed in a PyroMark Q24 System version 2.0.6 (Qiagen, reporter ions intensities (in; where n represents the reporter Netherlands). ions ranging from m/z 127.1 to m/z 131.1) and the reporter ion intensity from m/z 126.1, and were corrected according to the isotopic purities of the reporter ions provided by the manufac- Sequencing turer. Quantification values were rejected whenever any quanti- Human TLR3 isoform was sequenced using 50 ng of DNA follow- fication channel was missing or whenever variability of the ing a BigDye protocol published by the University of Pompeu reporter intensity was higher than 50% among the different Fabra Genomics Core Facility in Barcelona. Primers used listed PSMs of the same protein. For protein quantification purposes, in Supporting Information Figures. ©2019 The Authors. STEM CELLS published by STEM CELLS Wiley Periodicals, Inc. on behalf of AlphaMed Press 2019 Requena, Alvarez-Palomo, Codina-Pascual et al. 479 Lipopolysaccharide and Poly (I:C) Stimulation Spain). The detection was performed with the Western Breeze F1 and F2 cells were seeded in 24-well plate and once 80% con- Immunodetection Kit (Invitrogen, Spain). fluent they were stimulated with 100 ng/ml lipopolysaccharides (LPS) and 1 μg/ml Polyinosinic:polycytidylic acid, (Poly(IC)), to Flow Cytometry stimulate an innate immune response. After 24 hours superna- tants were collected and analyzed with an ilumina cytokine Cells were stained with antibodies listed in figures. All analyses kit. Cells were also pelleted and protein lysates were extracted were performed with a BD FACSCanto-II flow cytometer or and separated into cytoplasmic and nuclear fraction. MoFlo cell sorter (Dako Cytomation, USA) running Summit soft- ware. See Supporting Information Figure S9 for list of anti- Cytoplasmic and Nuclear Protein Fraction Extraction bodies used. Cells were resuspended on 80 μl of Buffer A (HEPES 10 mM pH 7.9, KCl 10 mM, MgCl2 1.5 mM, DTT 1 mM) with protease Real Time PCR inhibitors (PMSF 0.5 mM, aprotinin 5 μg/ml, leupeptin 2 μg/ml, Total mRNA was isolated using TRIZOL and 1 ug was used to sodium pyrophosphate decahydrate 2 mM, sodium beta- synthesize cDNA using the Invitrogen Cloned AMV First-Strand glycerophosphate pentahydrate 2 mM, sodium orthovanadate cDNA synthesis kit. One to two ul of the reaction was used to 1 mM). It was kept on ice for 15 minutes, then NP40 0.6% was quantify gene expression by qPCR for genes listed in the table added and centrifuged 30 seconds 16,000g at 4 C. The superna- below. tants (cytoplasmic fraction) were transferred to a new tube. Cell pellets (nuclear fraction) was resuspended in 50 μl of Buffer B (HEPES 20 mM pH 7.9, NaCl 0.46 M, EDTA 1 mM, EGTA 1 mM, Human TLR3 CCTGGTTTGTTAA TGAGGTGGA DTT 1 mM) with PMSF. TTGGATTAACGA GTGTTGCAAAGG Human TLR4 CCAGTGAGGATGATG GCCATGGCTGGATCA CCAGAAT GAGT Illuminex Cytokine Detection Array Human TLR3 AAGACACAACCA GCTTCTCTGACCTTCCA Cytokine levels were assessed by Luminex (Millipore, Billerica, short lsoform GGAACTGCC GTCC MA) following the manufacturer’s instructions. Briefly, superna- Human human TCCTTGGAGCTGTGA AAGGGCAGGAACAA leukocyte TCGCT CTCTTG tants were incubated for 2 hours with corresponding anticyto- antigen (HLA) A kine magnetic beads, and then washed with 1× washing buffer Human HLA B TCCTAG CAGTTGTGG TCAAGCTGTGAGAGA and stained with detection antibodies (provided) for 1 hour. TCATC CACAT Strepatividin-PE was then added for 30 more minutes. During all Human HLA C TCCTGGTTGTCCTAG CAGGCTTTACAAGTGA CTGTC TGAG incubation steps, the plate was agitated at 650 rpm. After wash- Human B2M TGACTTTGTCACAG AATCCAAATGCG G CA ing, plate was agitated for 15 minutes at 650 rpm and read in CCCAAGATA TCTTC the xMAP Luminex reader (Waltham, MA). Human GAPDH GCACCGTCAAGGC AGGGATCTCGCTCC TGAGAAC TGGAA Lentiviral Production Human WT TLR3 Lentiviral Vector was purchased (pLenti-GIII- TLR3 isoform RT-PCR primers were tested in both human UbC) (Abcam, LV335740, Cambridge, UK) and was used to over- foreskin fibroblasts (HFF) cells and NSC with R2 of 0.78465 and express human TLR3 full length in iPSC-derived neural stem 0.92829, respectively (Supporting Information Fig. S10). All cells (NSC). Viral particles were produced with a 293T packag- other primer sets are previously published. ing cell line. NSC were infected once with viral supernatants. At day 3, infected NSC were selected with puromycin for 7 days and placed in experiments. Overexpression of TLR3 was checked Statistical Analyses by RT-PCR, FACS, and Western blot. Unpaired nonparametric t tests were used to evaluate statistical significance at p ≤ .05. All quantitative and qualitative analyses Western Blot are representative of at least n = 3 biological repeats for all Protein extracts of cells collected by centrifugation, washed twice experiments to give significant statistical power for reliability and in PBS, lysed in 1× lysis buffer (50 nM Tris–HCl, 70 mM validity of data. Only significantly different data in the figures 2-mercaptoethanol, and 1% SDS) and the concentration of total are given a p value. Differential Protein Annotation and Bioinfor- protein was measured by the Bradford assay. Lysates were then matics Analyses: Proteins identified in different amounts in each boiled for 5 minutes, and subjected to 12% polyacrylamide SDS of the two experiments were analyzed using the bioinformatics gels or 4%–12% SDS resolving gels (Invitogen). After electropho- tool DAVID v6.7 (Database for Annotation, Visualization and Inte- resis, proteins were transferred to a nitrocellulose membrane grated Discovery; http://david.abcc.ncifcrf.gov/) to identify over- using a submerged transfer apparatus (BioRad), filled with represented biological pathways [26, 27]. Similarly, differential 25 mM Tris Base, 200 mM glycine, and 20% methanol. After proteins were analyzed for protein–protein interaction networks blocking with 5% non-fat dried milk in TBS-T (50 mM Tris–HCl using the online tool STRING v10 [28]. To determine differentially [pH 8.0], 150 mM NaCl, and 0.1% Tween 20) the membrane expressed quantified proteins between the groups (F1 vs. F2) for was incubated with the primary antibodies diluted in TBS-T and small peptide analysis, statistical analysis was carried out using washed extensively. For the list of primary antibodies used see SPSS for Windows (version 21.0, Chicago, IL). All variables were Supporting Information Figure S10. The membrane was washed checked for Levene’s test to check variance homogeneity. Com- 3 times with TBS-T and then incubated with the appropriate parison between groups (i.e., F1 vs. F2) was performed using horseradish peroxidase linked secondary antibody (Amersham, Student’s t test. p ≤ .05 was considered to be significant. www.StemCells.com ©2019 The Authors. STEM CELLS published by Wiley Periodicals, Inc. on behalf of AlphaMed Press 2019 480 iPSC Overexpress Human TLR3 Affecting Immunity Figure 1. Global proteomic analysis of patient (F1) and human iPSC derived fibroblasts (F2). (A): Panels left to right: Photos representative of the healthy control human fibroblasts termed F1 (human skin fibroblasts), iPSC clone (from previously published studies), differentiation to embryoid bodies and iPSC-derived fibroblasts, termed F2, at two different passages. A total of six sets of normal healthy control patient cells from previously published studies were used: three RV and three EPI (×100). Five F1 (n = 5) and three F2 (n = 3) cell populations was analyzed for proteomics between passage 3 to 9. (B): Table showing proteins characteristic of human MSC, hematopoietic cells and fibro- blasts that have either been identified or not in the F1 and F2 cells. The p value indicates that there is no difference in the amount of these proteins between F1 and F2 indicating that the F1 and F2 cell types are similar. (C): Table showing the differential proteins between F1 and F2 involved in the immune response according to Reactome and Panther. (D): Table showing the differential proteins between F1 and F2 involved in signal transduction according to Reactome. (E): Schematic diagram of the main pathways overrepresented (in yellow) of all pro- teins detected (Reactome). (F): Interactions among the differential proteins involved in signal transduction and the TLR3, TLR4, IL6, and CCL2 proteins (highest confidence = minimum interaction score of 0.900). (G): Interactions among the differential proteins involved in immune system and the TLR3, TLR4, IL6, and CCL2 proteins (highest confidence = minimum interaction score of 0.900). (H): Validation of proteomic screen in three (n = 3) F1 and three F2 cell populations by flow cytometry for fibroblast markers CD90, CD44 and negative control marker CD34 (MSC marker) in F1 and F2 cells, demonstrating that the two populations of fibroblasts are similar. (I): Validation of proteomic screen in three (n = 3) F1 and three F2 cell populations by Western blot for signal transduction proteins AKT Ser 478, total AKT and loading control Actin. Two from three matched pairs of cells demonstrated increased AKT signaling in F2 cells (EPI-2 the exception, see Fig. 4B). Abbrevia- tions: EPI, episomal made iPSC; iPSC, induced pluripotent stem cells; MSC, mesenchymal stem cells; RV, retroviral made iPSC. RESULTS Overall, we detected 1,508 quantifiable proteins of which 86 pro- teins differed significantly in the expression levels between F1 Global Proteomic Analysis of Patient (F1) and and F2 cells (Supporting Information Figs. S1, S2). Categorizing iPSC-Derived Fibroblasts (F2) proteins based on function, we first demonstrate no significant Six human skin fibroblast cell lines (termed F1) that have previ- difference in fibroblast markers CD90, Fibronectin and CD44 ously been reprogrammed into iPSC, three by retroviral methods between F1 and F2 cells, indicating that they are similar [2, 20, 21] and three by non-viral episomal methods [22], then fibroblast cell populations for both viral and non-viral made differentiated back to fibroblasts (termed F2) were analyzed for iPSC-derived cells (Fig. 1B). CD90, CD44, and Fibronectin are their potential immune response (Fig. 1A). established markers of fibroblasts as defined by Pilling et al. First, we performed a global proteomic analysis comparing [29]. The main differences in protein expression in F2 cells are F1 to F2 cells and mapped the protein expression profile. signal transduction, cell cycle, and immune system proteins ©2019 The Authors. STEM CELLS published by STEM CELLS Wiley Periodicals, Inc. on behalf of AlphaMed Press 2019 Requena, Alvarez-Palomo, Codina-Pascual et al. 481 Figure 2. Characterization of adaptive and innate immune system in human iPSC. (A): Genotype of the HLA of the six control healthy patient human fibroblasts (n = 6) previously reprogramed into hiPSC in other published studies. (B): Graph of the RT-PCR HLA gene expres- sion levels (n = 6) in healthy control patient fibroblasts (F1), subsequent iPSC clones and iPSC-derived fibroblasts (F2) demonstrating no dif- ference in HLA gene expression between F1 and F2 cells. (C): Flow cytometry histograms of the six patient iPSC derived fibroblasts (F2) (n = 6) for MHC-I protein expression in three RV and three EPI. (D): Graph representing expression levels of MHC-I (n = 3) by flow cyto- metry for F1 cells and F2 cells stimulated with 10 ng/ml INF-gamma overnight. (E): Table of RT-PCR analysis of TLR3 gene expression levels in F1 and F2 cells of pairs of cells (n = 10) demonstrating that the wild type (WT) human TLR3 gene is overexpressed between twofold and ninefold more cells in F2 iPSC-derived cells in eight from 10 matched healthy cells (80%). (F): Representative flow cytometry histograms of F1 and F2 cells stained for TLR3 receptor expression confirming that WT human TLR3 is overexpressed in twofold to fivefold more cells in F2 cells for one RV-made pair, one EPI-made pair (for RV and EPI, F1 and F2 cells are fibroblasts) and one synthetic mRNA made iPSC-made pair (F1 are HFF cells and F2 are neural stem cells) (n = 3). Abbreviations: EPI, episomal made iPSC; HLA, human leukocyte antigen; iPSC, induced pluripotent stem cells; MHC, major histocompatibility complex; RV, retroviral made iPSC; TLR3, toll like receptor protein 3. (Fig. 1C–1G; Supporting Information Fig. S1A). Interestingly, the Characterization of Adaptive and Innate Immune proteomic screen revealed many proteins interacting with the System in Human iPSC TLR3 and TLR4 pathway (Fig. 1F, 1G). Analysis of HLA type by deep sequencing methods was per- To verify the global proteomic analysis, we performed flow formed to confirm the HLA type for all cell types used (Fig. 2A). cytometry analysis of the F1 and F2 populations and confirmed MHCI mRNA gene expression (HLA-A, B, C, and B2M) for F1, iPSC that the two fibroblast cell populations are similar for two and F2 cells by RT-PCR and demonstrates no significant change fibroblast cell markers CD90, CD44 and negative for stem cell in expression between F2 cells and F1 cells (Fig. 2B), supported marker CD34 (Fig. 1H). This was the same for retrovirally made by similar results in a previously published report [30]. MHC-1 iPSC-derived cells or non-virally made iPSC-derived cells and so protein expression was assessed by flow cytometry in all F2 cell they are both grouped together (Supporting Information Fig. S3). lines and no significant difference in expression levels of MHCI Given that signal transduction proteins were the most represented was observed. Expression of MHC-1 was not significantly different differentially expressed proteins between F1 and F2 cells, we between retrovirally made iPSC-derived cells (RV1-3) and non- also verified that signal transduction pathways for AKT is over virally made iPSC-derived cells (Epi1-3) (Fig. 2C). expressed in F2 cells compared to F1 cells by Western blot To mimic the possible in vivo functional response, we analysis (Fig. 1I). We next assessed the immune response given assessed MHCI expression response to inflammation cytokine the differences identified in the proteomic screen. INF-gamma (10 ng/ml over 24 hours) in F1 and F2 cells from www.StemCells.com ©2019 The Authors. STEM CELLS published by Wiley Periodicals, Inc. on behalf of AlphaMed Press 2019 482 iPSC Overexpress Human TLR3 Affecting Immunity both viral and non-viral made iPSC by flow cytometry. Both F1 isoform of full length TLR3 gene, and not other TLR gene family and F2 cells responded with an upregulation of MHCI expres- members (Supporting Information excel methylome data), is sion in response to INF-gamma and there was no significance hypomethylated in human iPSC and derived NSC (F2), not seen difference indicating that the regulation of MHCI expression is in the starting parent HFF cells (F1) (Fig. 3B). We confirmed the not affected in iPSC derived cells (Fig. 2D, Supporting Informa- CpG hypomethylation of the TLR3 isoform by bisulfide pyro- tion Fig. S4). This suggests further that MHCI expression is sequencing (Fig. 3C, Supporting Information Fig. S7). Bioinfor- functionally normal in iPSC-derived cells. matics analysis of the human TLR3 WT and TLR3 Isoform To further characterize the F1 and F2 cells, we analyzed demonstrate that the TSS site is associated with Acetylation and the expression of small peptides presented by MHC-1 using DNAse hypersensitivity sites associated with gene expression and mass spectrometry for RV1 clone that revealed differences in that the TLR3 isoform is uniquely expressed in humans and not the number and type of small peptide expressions between F1 in mice or rat (Fig. 3D, 3E). and F2 cells that could implicate functional differences in the adaptive immune response of transplanted iPSC-derived cells Human TLR3 Isoform Is Overexpressed in Most Human (Supporting Information Fig. S5). iPSC-Derived Cells Not Seen in Normal Human Tissue We next analyzed the innate immune response by asses- or Cells sing the expression of 25 cytokines routinely assed by clinical We next assessed gene expression levels of human TLR3 iso- immunologists using the ELISA method in two matched cell form in human F1 and F2 cells from the healthy patient con- lines (F1 vs. F2 cells) for viral (RV) or non-viral (EPI) made iPSC trol cells (Episomal and Retroviral) described in Figures 1 and 2 derived cells, with or without viral ligand stimulation (Poly lC) and the synthetic mRNA made human iPSC cells described in or bacterial ligand stimulation (LPS) identified the main cyto- Figure 3. We next performed PCR of genomic DNA for sequenc- kines involved (Supporting Information Fig. S6A). Further anal- ing in F2 cells from synthetic mRNA made human iPSC-derived: ysis in another focused cytokine array (six cytokines) assessing NSC and cardiomyocytes (CM) to demonstrate the unique 50 four matched iPSC lines (F1 vs. F2 cells), we found no signifi- UTR of the human TLR3 isoform in F2 cells. Sequencing data of cant difference in expression of cytokines between viral and the PCR product for the 50 UTR confirms the human TLR3 iso- non-viral made F2 cells when compared to F1 cells. In general, form (in Red) in human F2 cells with whole sequence found at we found high variation between patient cell lines in cytokine UC011CKZ.2 or XM_017008577.1 transcript variant X1 (Fig. 4A). levels that suggested altered expression in F2 cells (Supporting We next tested protein expression of the human WT TLR3 and Information Fig. S6). shorter TLR3 isoform by Western blot in F1 and F2 cells (from We found proteins associated with the innate immune sys- synthetic mRNA made human iPSC-derived) demonstrating tem with higher WT human TLR3 and WT TLR4 expression expression of the WT TLR3 at 107 KDa and the putative found in the proteomic screen in F2 cells (Fig. 1) suggesting a shorter TLR3 isoform at approximately 80 KDa missing the dysfunctional innate cell signaling pathway, prompted us to LRR1-8 region important in receptor stabilization (Fig. 4A, Sup- investigate further human TLR3 expression levels in the healthy porting Information Fig. S8). patient control matched F1 and F2 cells (Fig. 1, Supporting Infor- We next performed a comprehensive analysis of human mation Figs. S1, S2). We observed that four from five patient TLR3 gene expression levels in normal endogenous cell lines or (80%) control F1 matched F2 cells have increased TLR3 gene normal human neural tissue (frontal cortex or Hippocampus) expression in F2 cells by RT-PCR analysis (Fig. 2E, Supporting compared to various human iPSC derived cell types (from syn- Information Fig. S7B). We next confirmed by flow cytometry thetic mRNA method) to further confirm overexpression in analysis TLR3 receptor protein expression is independent of human iPSC derived cells (F2) and that differentiation or method used to reprogram cells and overexpressed in F2 cells reprograming method does not affect TLR3 expression levels (Fig. 2F). Taken together the data demonstrate that the WT (Fig. 4B). RT-PCR gene expression analysis revealed that the TLR3 receptor expression is higher in healthy patient control shorter TLR3 isoform is lower expressed in F1 cells than WT iPSC-derived cells, suggesting that reprogramed cells have an TLR3 and is significantly higher upregulated in all reprogrammed altered TLR3 pathway. differentiated F2 cell types (CM, p = .0093; Definitive Endoderm – DE, p = .033; NSC, p = .0044 and Motor Neurons – MN, p = .013) Human TLR3 Isoform Transcription Start Site Is (Fig. 4C). Interestingly, in addition to HFF other normal endogenous Hypomethylated in Human iPSC-Derived Cells and cells such as bone marrow mesenchymal stem cells (BM MSC), Exclusively Expressed in Humans umbilical stem cells, or normal human brain tissue (hippocampus To investigate further the increase in TLR3 expression in a and frontal cortex) have low levels of expression of WT TLR3 and wide range of iPSC-derived cell types, we reprogrammed using the shorter isoform TLR3, suggesting only reprogramed iPSC- non-viral synthetic mRNA transfection methods HFF (F1) and derived cells have elevated levels of the TLR3 isoform (Fig. 4C). characterized six iPSC clones differentiated to various cell types Therefore, we demonstrate that the overexpression of the (F2) (iPSC characterization data in a separate submitted paper) shorter TLR3 isoform in iPSC-derived cells is independent of (Figs. 3–5). We then performed a methylome array to further the methodology used because high levels of the isoform have characterize and better define the differences between F1 and been found in most human iPSC generated by three different F2 cells (Fig. 3A). Using the following bioinformatic criteria: methods (retrovirus, episomal vectors, or mRNA transfections). (a) more than 20% difference in CpG methylation levels, (b) more Moreover, higher expression of the TLR3 isoform is indepen- than one CpG island affected, and (c) at least two hiPSC clones dent of the starting parent cell types because all somatic cell affected, we analyzed the methylation levels for Toll-like recep- types we tested, including HFF, cord blood cells, patient skin tors (Fig. 3B). Interestingly, the methylome array detected that a fibroblasts, mesenchymal stem cells had low levels of expres- transcription start site (TSS) site in the CpGs of a shorter genetic sion of the human TLR3 isoform. Finally, the differentiation ©2019 The Authors. STEM CELLS published by STEM CELLS Wiley Periodicals, Inc. on behalf of AlphaMed Press 2019 Requena, Alvarez-Palomo, Codina-Pascual et al. 483 Figure 3. Human TLR3 isoform transcription start site is hypomethylated in human iPSC-derived cells and exclusively expressed in humans. In addition to human iPSC used in Figures 1 and 2, a HFF cell line (Life Technologies) was reprogrammed by transfection of syn- thetic mRNA of the reprogramming factors to make six human iPSC lines and differentiated to NSC for further in depth analysis (charac- terization of mRNA made hiPSC in a separate publication). (A): Methylome array raw data for all toll genes for F1 (HFF), six human iPSC, and six F2 (NSC) cell lines. (B): Methylation profile of HFF (F1), iPSC, and NSC (F2) demonstrating four CpGs that containing a TSS in TLR3 is hypo-methylated in iPSC and NSC (F2 cells) but not in stating HFF (F1 cells). Methylation regions were considered significant using the following bioinformatic criteria: (a) more than 20% difference in CpG methylation levels between F1 and F2, (b) more than one CpG island affected, and (c) at least two hiPSC or NSC clones affected, (C) validation of hypomethylted TSS by pyrosequencing for two CpGs in human iPSC confirming the CpG TSS region of an isoform TLR3 is hypomethylated. (D): Sketch of bioinformatics of the TLR3 WT and TLR3 isoform demonstrating the TSS site (red) associated with acetylation (yellow) and DNAse hypersensitivity sites (dark blue) that indicate gene expression and homology to other species (green). TLR3 isoform is uniquely expressed in humans and rhesus monkey. Lower panel: Detailed sketch of the exons of the shorter TLR3 isoform discovered from the methylome array, compared to WT TLR3 for comparison. Brown and blue arrows show gene regions showing RT PCR primer sequences. Abbreviations: HFF, human fibroblast; iPSC, induced plurip- otent stem cells; NSC, neural stem cells; TLR3, toll like receptor protein 3; TSS, transcription start site; WT, wild type. method did not affect levels of the TLR3 isoform expression overnight) in iPSC-derived cells (Fig. 4D). This suggests that because we generated many different cell types from iPSC transplanted human iPSC-NSC would act as a cellular pump (F2 cells included fibroblasts, NSC, MN, CM, and DE cells) and producing physiologically higher levels of IL6 cytokine than is most had significantly higher levels of expression of the TLR3 normal in circulating blood and tissues and the cells would not isoform (Fig. 4B, 4C). respond to viral infection (Fig. 4D). Given that it has been previously published that a short isoform of mouse TLR3, but different sequence and size to the human TLR3 isoform described in this study can act as a sup- Human TLR3 Isoform Functions to Suppress NF-KB p65 pressor of the TLR3 innate immune pathway by competing Signaling Pathway in Response to Virus (Poly IC) with the full length TLR3 protein in mouse astrocytes [31–33], Next, we performed a gain of function test in human iPSC-NSC we next assessed the functional immune response of inflam- (F2 cells) to demonstrate the possible functional effect of the mation cytokine IL6 in human iPSC-NSC to Poly (I:C) and LPS. human TLR3 isoform. We overexpressed WT TLR3 by lentiviral Assessment of IL6 expression levels by ELISA demonstrated methods in five synthetic mRNA made iPSC-NSC clones to out- that iPSC-derived cells (F2 cells) have significantly higher levels compete the TLR3 isoform and all NSC clones demonstrated high of expression of IL6 compared to F1 cells and a suppressed levels of WT human TLR3 expression after infection (Fig. 5A). innate immune response to viral infection (Poly IC 1ug/ml Analysis of the expression of nuclear NF-KB p65 protein www.StemCells.com ©2019 The Authors. STEM CELLS published by Wiley Periodicals, Inc. on behalf of AlphaMed Press 2019 484 iPSC Overexpress Human TLR3 Affecting Immunity Figure 4. Human TLR3 isoform is overexpressed in most human iPSC derived cells not seen in normal human tissue or cells. (A): Expres- sion of the shorter human isoform TLR3 with unique 50 UTR in red. Below the PCR of genomic DNA from human iPSC-derived NSC and CM of the unique 50 UTR of the TLR3 isoform. Sequencing data of the PCR product for the 50 UTR of the human TLR3 isoform (in red) with whole sequence found at or XM_017008577.1 transcript variant X1. Underlined regions are RT-PCR primer sequences used to detect human isoform TLR3. Lower left panel: Western blot of TLR3 expression demonstrating expression of the WT TLR3 at 107 KDa and the putative shorter TLR3 isoform at approximately 80 KDa in F1 cells (HFF) and F2 cells iPSC-derived NSC with schematic drawing of WT and isoform human TLR3 missing the LRR1-8 region important in binding viral dsRNA and receptor stability. (B): Table of individual F1 and F2 cell lines for RT-PCR analysis of gene expression demonstrating that the human TLR3 isoform gene is overexpressed between 3 and 81 fold more in F2 in five from six matched healthy cells (83%). In red, it is highlighted the exception healthy control patient sample (Epi-2) that also did not have elevated AKT signal transduction protein levels in F2 cells seen in Figure 1. (C): Graph comparing WT TLR3 (FL: full length) and isoform TLR3 gene expression in five normal healthy cells and tissues compared to synthetic mRNA made iPSC-derived cells (dotted line divides normal from iPSC cells analyzed). Normal cells/tissues include: starting fibroblasts (HFF), normal endogenous BM MSC, cord blood USSC, and normal tissues: normal hippocampus and frontal cortex human brain tissue demonstrating, demonstrating low levels of gene expression of the shorter TLR3 isoform in normal tissues. Dotted line: Compared to the synthetic mRNA made repro- gramed human iPSC differentiated into three different iPSC-derived cells: CM, NSC, ENDO, and MN demonstrating a significant increase in gene expression of the shorter TLR3 isoform in all types of iPSC-derived cell types tested. (D): Graph of an ELISA assay of the functional response of IL6 in human foreskin fibroblasts (F1, n = 2) and their mRNA made iPSC-derived neural stem cells (F2, n = 5) to LPS (100 ng/ml over 24 hours) and Poly (I:C) stimulation (1 ug/ml) in triplicate, demonstrating a significantly suppressed IL6 immune response to Poly IC (1 ug/ml over 24 hours) in iPSC-derived cells (F2 cells). Note: Synthetic mRNA made iPSC and derived NSC, CM, and DE have been charac- terized and described in another paper currently under consideration for publication. Abbreviations: BM, bone marrow; CB, cord blood; CM, cardiomyocytes; EPI, episomal made iPSC; HFF, human fibroblast; iPSC, induced pluripotent stem cells; LPS, lipopolysaccharide; MN, motor neurons; MSC, mesenchymal stem cells; NSC, neural stem cells; RV, retroviral made iPSC; TLR3, toll like receptor protein 3; TSS, transcription start site; USSC, umbilical stem cells; WT, wild type. phosphorylation by Western blot, an established marker of a return to sensitivity to Poly IC stimulation supports the model functional human TLR3 pathway, demonstrated that overex- that the shorter TLR3 isoform is suppressing normal TLR3 pression of human WT TLR3 in NSC could return sensitivity to function in NSC to viral insult (F2 cells) (Fig. 5B). Taken Poly (IC) treatment of NSC (Fig. 5B). Note that in control NSC together, we propose the model that in human iPSC-derived cells and in empty lentiviral control NSC (UB promoter only) cells that overexpress the shorter isoform of TLR3 competes the NSC do not respond to Poly IC viral stimulation demon- with full length TLR3 to suppress the innate inflammation sig- strated by no change in NF-KB p65 phosphorylation levels nal transduction response in F2 cells to Poly IC (viral infection) (Fig. 5B). Overexpression of WT TLR3 in NSC and resulting (Fig. 5C). ©2019 The Authors. STEM CELLS published by STEM CELLS Wiley Periodicals, Inc. on behalf of AlphaMed Press 2019 Requena, Alvarez-Palomo, Codina-Pascual et al. 485 Figure 5. Human TLR3 isoform functions to suppress NF-KB p65 signaling pathway in response to virus (Poly IC). (A): Graph of RT-PCR for WT human TLR3 gene lentiviral overexpression in F2 cells (NSC that have high levels of expression of TLR3 isoform, (Black bars)). WT human TLR3 lentivirus was used to overexpress WT TLR3 in five (n = 5) F2 (NSC) clones to functionally outcompete with the TLR3 isoform observed to be overexpressed in F2 cells (B). Western blot analysis of phosphorylated p65 (p-p65) in two NSC (F2 cells) clones. Gain of function with WT TLR3 in F2 cells demonstrates the immune suppressive function of the short TLR3 isoform on TLR3 signal transduction in response to viral infection (Poly IC). After lentiviral overexpression of full length TLR3, NSC recover responsiveness to stimulation by the viral analog Poly (IC), as seen by the increase of phospho-p65 in the nuclear fraction. Control NSC were infected with a lentiviral empty vector (ubiquitin promoter only (LV-Ct) that contained the same ubiquitin promoter as the wild type TLR3-LV vector). Loading controls are tubulin nuclear protein fraction. Representative pictures of five NSC tested. (C): Schematic proposed model of full length TLR3 and the role of the shorter TLR3 isoform predominately expressed in reprogrammed cells that acts to suppress the immune response. All experiments performed with n = 6 at least twice, unless stated. Abbrevia- tions: iPSC, induced pluripotent stem cells; LV, lenti virus; NSC, neural stem cells; TLR3, toll like receptor protein 3; WT, wild type. DISCUSSION a. Unstimulated state: increased IL6 associated with increased WT TLR3 expression in F2 cells (iPSC-derived cells) (Fig. 4C, 4D). There is a clear need to better understand the immune response b. Stimulation with Poly IC: suppressed response to Poly IC of autologous reprogramed human cells generated in vitro (Virus) in F2 cells due to overexpression of an isoform TLR3 that before we embark on extensive clinical trials using human iPSC can be rescued by overexpression of WT TLR3 (Fig. 5A–5C). to study and treat human disease. With that in mind we posit the question: do patient cells that have been reprogramed into The higher levels of expression of the human TLR3 isoform iPSC-derived cells cause an immune response? is central to understanding the altered immunogenicity of This study is based on a global proteomic and methylome human iPSC-derived cells. To mediate innate immune response screening we performed using human skin cells (termed F1 to virus infection and consequently the virus immunogenicity, cells) compared to their iPSC-derived cells (F2 cells) made from double-stranded RNA (dsRNA), a frequent by-product of virus three different reprogramming methods (retroviral, episomal, infection, should be recognized by TLR3. Simultaneously, the and synthetic mRNA transfection). We reveal that WT TLR3 PI3K-Akt (Fig. 1I) and NF-KB (Fig. 5B) pathways play an essential and a isoform of TLR3 are overexpressed in the majority (83%) role in TLR3-mediated gene induction [35]. The loss of inflam- of F2 cells tested (five from six F1-F2 match lines tested) and mation response of cells to bacterial (LPS) or viral (Poly (IC) in all types of iPSC-derived cells (NSC, MN, DE, and CM) war- ranting screening of iPSC-derived cell clones and removal in infection has been shown before in specific mouse neural cell disease modeling studies or clinical applications (Figs. 2D, types that overexpress a different mutant mouse TLR3 isoform 3B, 3C). to the human TLR3 isoform described in this study (not expressed The data demonstrate that the cell reprogramming process in mice) [31–33]. Interestingly, a recent study found that differen- alters specific aspects of the innate immune response resulting tiation of human iPSC results in a loss of immunogenicity and in a possible suppression of immunogenicity, supporting similar leads to the induction of tolerance, despite expected antigen findings recently described for human iPSC-derived cells [11, 34]. expression differences between iPSC-derived versus original We discover two distinct states in human iPSC-derived cells somatic cells [11]. In further agreement with this study, it has related to overexpression WT TLR3 and Isoform TLR3 that could also been demonstrated that neural progenitor cells from human impinge on future clinical applications and disease modeling: iPSC generated less autogenous immune response, possibly by www.StemCells.com ©2019 The Authors. STEM CELLS published by Wiley Periodicals, Inc. on behalf of AlphaMed Press 2019 486 iPSC Overexpress Human TLR3 Affecting Immunity altered toll like receptor pathway signaling they did not assess remodeling and nuclear reprogramming. Furthermore, they con- [34]. This observation among others brings forward the interest- clude that activation of inflammatory pathways are required for ing possibility that human iPSC-derived cells may benefit from a efficient nuclear reprogramming in the induction of pluripotency suppressed immunogenic microenvironment favoring initial sur- [38]. Our finding suggests that the TLR3 pathway, important for vival and subsequent engraftment and tissue regeneration. achieving cell pluripotency, is not re-set to the correct expression Research in animal models is well documented and work levels in iPSC-differentiated cells (F2) and that this then results in by Araki et al., working mainly with skin and bone marrow altered cytokine secretion leading to possible immune suppression transplant in mice, demonstrates mouse iPSC-derived CM cause of transplanted iPSC-derived cells (Figs. 4, 5). We demonstrate that T cell infiltration when transplanted under the skin of syngeneic the mechanism of constitutive expression of the shorter TLR3 iso- mice [13]. Todorova et al., found that the location of iPSC trans- form is most likely at an epigenetic level, resulting in hypomethyla- plantation can affect the immune outcome because the kidney tion of a TSS producing a shorter TLR3 isoform that competes to is immune privileged, highlighting that differences between suppress normal full length TLR3 signal transduction function studies could be explained by experimental procedure and design (Figs. 3, 4). In conclusion, the data demonstrate that the cell reprogram- differences [14]. The group by Almeida et al., found that mouse ming process alters the immune response of most iPSC-derived iPSC-derived endothelial cells are intrinsically immunogenic in cells highlighting human TLR3 isoform overexpression that func- syngeneic mice, but are immune tolerated by overexpressing tions to suppress innate immune signal transduction and provides immune suppressive molecule IL-10; however, their study did an insight for future clinical applications of autologous human iPSC- not assess TLR3 expression levels [11]. derived cell transplantation. We propose that screening human The altered TLR3 innate immune response of human in vitro iPSC clones and their derived cells for alterations in the genome iPSC-derived cells raises a number of issues to be considered for of patient cells that include expression levels of human TLR3 future cell transplantation in humans. and its shorter TLR3 isoform identified in this work is essential i. This may allow for human iPSC-derived cells to be used not before embarking on “first in man” clinical trials or human only as an autologous cell therapy but also in a limited disease modeling using human iPSC. capacity as an allogenic cell therapy because of a sup- pressed immunogenic microenvironment favoring initial sur- vival and subsequent engraftment and tissue regeneration. ETHICAL STATEMENT ii. Evasion of the immune system is a key hallmark of cancer Use of human iPSC and tissue obtained by collaborators from pub- and suppression of the immune response of iPSC-derived lished work funded by approved projects 2012-StG-311736-PD- cells may impose a long-term cancer threat not previously HUMMODEL, BFU2016-80870-P), Red de Terapia Celular - TerCel considered. This consideration is considerably enhanced by RD16/0011/0024 and BFU2011–26596 [2, 20–22]. The generation recent work that demonstrates the AKT pathway induces a and/or use of human iPSC were approved by the Spanish compe- transcriptional program via C/EBP that promotes immune tent authorities (Commission on Guarantees concerning the Dona- suppression during inflammation and tumor growth [36]. tion and Use of Human Tissues and Cells of the Carlos III National Inhibition of the AKT driven signaling pathway prolongs NFκB Institute of Health). The human iPSC lines have been (or are in activation, which in turn inhibits C/EBPβ activation promoting the process of being) deposited at the Spanish National Stem Cell an immunostimulatory transcriptional program that restores Bank, according to the Spanish legislation. CD8+ T cell activation, cytotoxicity, and synergizes with check- point inhibitor therapy to promote tumor regression and extend survival in mouse models of cancer [36]. ACKNOWLEDGMENTS iii. Transplantation of iPSC-derived cells that constitutively express physiologically higher levels of inflammatory cytokine IL6 We acknowledge the support of the following people: Angel Raya (Fig. 4D) may result in long-term adverse effects in the body. Centre for Regenerative Medicine Barcelona (CMRB), John Christo- This may have important implications in future human iPSC doulou of the University of Sydney, Eduard Palou Rivera for HLA clinical trials such as the RIKEN AMD-RPE clinical trial [17, 18]. analysis, Josep Maria Estanyol of the Technological and Scientific iv. The need to screen patient iPSC and derived cells using Centre at the University of Barcelona, a member of the ProteoRed global gene analysis “omics” for alterations in the genome network (http://www.proteored.org/) for assistance in the proteo- of patient cells that include expression levels of TLR3 and mic analysis, Montserrat Pau for technical assistance, Carmen Bar- its shorter TLR3 isoform identified in this work. Indeed, rot for RT-PCR analysis. Isabel Crespo Torres, Responsable impaired TLR3 function to herpes simple virus (HSV-1) infec- Plataforma Citometria IDIBAPS. MJE is supported in part by the tion in the CNS has important implications in the pathogene- Program Ramon y Cajal (RYC-2010-06512), FBG project 307900, sis of HSV-1 encephalitis (HSE) in children [37]. MINECO project grant BFU2011–26596, BFU2014-54467-P, TV3 Marato project grant /FBG309768 and Talent Retention program, Interestingly, activation of innate immunity is required for University of Barcelona. The proteomic analysis was supported by efficient nuclear reprogramming to iPSC via the TLR3 receptor projects from the Ministerio de Economica y Competitividad pathway using either viral or non-viral methods [38]. The P113/00699 and EU-PF7-PEOPLE-2011-ITN289880 to RO. Dr. Manel authors find that by using gain and loss of function studies, Esteller laboratory is supported by, among other institutions, the the TLR3 pathway enables efficient induction of pluripotency EU Joint Programme – Neurodegenerative Disease Research by viral or non-viral synthetic mRNA transfection approaches. (JPND); Cellex Foundation; the Health and Science Departments of Stimulation of TLR3 causes rapid and global changes in the the Catalan Government (Generalitat de Catalunya); the E-Rare expression of epigenetic modifiers to enhance chromatin (ERA-Net for research programs on rare diseases) and EuroRETT ©2019 The Authors. STEM CELLS published by STEM CELLS Wiley Periodicals, Inc. on behalf of AlphaMed Press 2019 Requena, Alvarez-Palomo, Codina-Pascual et al. 487 (a European network on Rett syndrome, funded by the writing; R.D.-M., S.M., M.E., and A.B.B.: collection and/or assem- European Commission under its 6th Framework Program since bly of data, data analysis and interpretation; M.J., and J.M.P.: col- 2006). Dr. Manel Esteller is an ICREA Research Professor. Anto- lection and/or assembly of data, data analysis and interpretation, nella Consiglio laboratory is supported by the European conception and design; E.W. and D.O.: collection and/or assembly Research Council-ERC (2012-StG-311736- PD-HUMMODEL), of data; D.J.: data analysis and interpretation, manuscript writing; the Spanish Ministry of Economy and Competitiveness- R.O.: collection and/or assembly of data, financial support, data MINECO (BFU2016-80870-P), Instituto de Salud Carlos III-ISCIII/ analysis and interpretation, manuscript writing; M.J.E.: conception FEDER (Red de Terapia Celular - TerCel RD16/0011/0024), and design, financial support, provision of study material collec- AGAUR (2017-SGR-899). Contact R. Delgado-Morales and M. tion and/or assembly of data, data analysis and interpretation, Esteller for comments on the methylome array. Contact R. manuscript writing, final approval of manuscript, coordination of Oliva for comments on the proteomic screen. national and international teams. AUTHOR CONTRIBUTIONS DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST J.R., A.B.A.-P., M.C.-P., I.A., M.S., A.C., and O.A.B.: collection and/ or assembly of data, data analysis and interpretation, manuscript The authors indicated no potential conflicts of interest. REFERENCES 12 Zhao T, Zhang ZN, Westenskow PD et al. down syndrome etiology. STEM CELLS 2013;31: Humanized mice reveal differential immuno- 467–478. 1 Takahashi K, Tanabe K, Ohnuki M et al. genicity of cells derived from autologous induced 23 Du P, Kibbe WA, Lin SM. lumi: A pipeline Induction of pluripotent stem cells from adult pluripotent stem cells. Cell Stem Cell 2015;17: for processing Illumina microarray. Bioinfor- human fibroblasts by defined factors. Cell 353–359. matics 2008;24:1547–1548. 2007;131:861–872. 13 Araki R, Uda M, Hoki Y et al. Negligible 24 Gentleman RC, Carey VJ, Bates DM et al. 2 Edel MJ, Menchon C, Menendez S et al. immunogenicity of terminally differentiated cells Bioconductor: Open software development for Rem2 GTPase maintains survival of human derived from induced pluripotent or embryonic computational biology and bioinformatics. embryonic stem cells as well as enhancing stem cells. Nature 2013;494:100–104. Genome Biol 2004;5:R80. reprogramming by regulating p53 and cyclin 14 Todorova D, Kim J, Hamzeinejad S et al. 25 Bibikova M, Le J, Barnes B et al. Genome- D1. Genes Dev 2010;24:561–573. Brief report: Immune microenvironment deter- wide DNA methylation profiling using Infinium 3 Palomo AB, Lucas M, Dilley RJ et al. The mines the immunogenicity of induced pluripo- (R) assay. Epigenomics 2009;1:177–200. power and the promise of cell reprogramming: tent stem cell derivatives. STEM CELLS 2016;34: 26 Huang da W, Sherman BT, Lempicki RA. Personalized autologous body organ and cell 510–515. Bioinformatics enrichment tools: Paths transplantation. J Clin Med 2014;3:373–387. 15 Hew M, O’Connor K, Edel MJ et al. The toward the comprehensive functional analysis 4 Hoffmann JA, Kafatos FC, Janeway CA et al. possible future roles for iPSC-derived therapy of large gene lists. Nucleic Acids Res 2009; Phylogenetic perspectives in innate immunity. for autoimmune diseases. J Clin Med 2015;4: 37:1–13. Science 1999;284:1313–1318. 1193–1206. 27 Huang da W, Sherman BT, Lempicki RA. 5 Romieu-Mourez R, Francois M, Boivin 16 Mandai M, Kurimoto Y, Takahashi M. Systematic and integrative analysis of large MN et al. Cytokine modulation of TLR expres- Autologous induced stem-cell-derived retinal gene lists using DAVID bioinformatics resources. sion and activation in mesenchymal stromal cells for macular degeneration. N Engl J Med Nat Protoc 2009;4:44–57. cells leads to a proinflammatory phenotype. 2017;377:792–793. 28 Szklarczyk D, Franceschini A, Wyder S J Immunol 2009;182:7963–7973. 17 Mandai M, Watanabe A, Kurimoto Y et al. et al. STRING v10: Protein-protein interaction Autologous induced stem-cell-derived retinal 6 Bai W, Liu H, Ji Q et al. TLR3 regulates networks, integrated over the tree of life. cells for macular degeneration. N Engl J Med mycobacterial RNA-induced IL-10 production Nucleic Acids Res 2015;43:D447–D452. 2017;376:1038–1046. through the PI3K/AKT signaling pathway. Cell 29 Pilling D, Fan T, Huang D et al. Identification 18 Kamao H, Mandai M, Okamoto S et al. Signal 2014;26:942–950. of markers that distinguish monocyte-derived Characterization of human induced pluripo- 7 Liu P, Chen S, Li X et al. Low immunoge- tent stem cell-derived retinal pigment epithe- fibrocytes from monocytes, macrophages, and nicity of neural progenitor cells differentiated lium cell sheets aiming for clinical application. fibroblasts. PLoS One 2009;4:e7475. from induced pluripotent stem cells derived Stem Cell Reports 2014;2:205–218. 30 Drukker M, Katz G, Urbach A et al. Char- from less immunogenic somatic cells. PLoS 19 Alvarez Palomo AB, McLenachan S, Chen acterization of the expression of MHC pro- One 2013;8:e69617. FK et al. Prospects for clinical use of repro- teins in human embryonic stem cells. Proc 8 Zhao T, Zhang ZN, Rong Z et al. Immuno- grammed cells for autologous treatment of Natl Acad Sci USA 2002;99:9864–9869. genicity of induced pluripotent stem cells. macular degeneration. Fibrogenesis Tissue Repair 31 Farina C, Krumbholz M, Giese T et al. Nature 2011;474:212–215. 2015;8:1–9. Preferential expression and function of Toll-like 9 Guha P, Morgan JW, Mostoslavsky G et al. 20 Sanchez-Danes A, Consiglio A, Richaud Y receptor 3 in human astrocytes. J Neuroimmu- Lack of immune response to differentiated cells et al. Efficient generation of A9 midbrain nol 2005;159:12–19. derived from syngeneic induced pluripotent dopaminergic neurons by lentiviral delivery of 32 Seo JW, Yang EJ, Kim SH et al. An inhibi- stem cells. Cell Stem Cell 2013;12:407–412. LMX1A in human embryonic stem cells and tory alternative splice isoform of Toll-like 10 Morizane A, Doi D, Kikuchi T et al. Direct induced pluripotent stem cells. Hum Gene receptor 3 is induced by type I interferons in comparison of autologous and allogeneic trans- Ther 2012;23:56–69. human astrocyte cell lines. BMB Rep 2015;48: plantation of iPSC-derived neural cells in the 21 Sanchez-Danes A, Richaud-Patin Y, Carballo- 696–701. brain of a nonhuman primate. Stem Cell Reports Carbajal I et al. Disease-specific phenotypes in 33 Yang E, Shin JS, Kim H et al. Cloning 2013;1:283–292. dopamine neurons from human iPS-based of TLR3 isoform. Yonsei Med J 2004;45:359–361. 11 de Almeida PE, Meyer EH, Kooreman NG models of genetic and sporadic Parkinson’s 34 Huang K, Liu PF, Li X et al. Neural et al. Transplanted terminally differentiated disease. EMBO Mol Med 2012;4:380–395. progenitor cells from human induced pluripo- induced pluripotent stem cells are accepted 22 Briggs JA, Sun J, Shepherd J et al. Integra- tent stem cells generated less autogenous by immune mechanisms similar to self-toler- tion-free induced pluripotent stem cells model immune response. Sci China Life Sci 2014;57: ance. Nat Commun 2014;5:3903. genetic and neural developmental features of 162–170. www.StemCells.com ©2019 The Authors. STEM CELLS published by Wiley Periodicals, Inc. on behalf of AlphaMed Press 2019 488 iPSC Overexpress Human TLR3 Affecting Immunity 35 Sarkar SN, Peters KL, Elco CP et al. 36 Kaneda MM, Messer KS, Ralainirina N et al. iPSC-derived TLR3-deficient CNS cells. Nature Novel roles of TLR3 tyrosine phosphoryla- PI3Kgamma is a molecular switch that controls 2012;491:769–773. tion and PI3 kinase in double-stranded RNA immune suppression. Nature 2016;539:437–442. 38 Lee J, Sayed N, Hunter A et al. Activation signaling. Nat Struct Mol Biol 2004;11: 37 Lafaille FG, Pessach IM, Zhang SY et al. of innate immunity is required for efficient 1060–1067. Impaired intrinsic immunity to HSV-1 in human nuclear reprogramming. Cell 2012;151:547–558. See www.StemCells.com for supporting information available online. ©2019 The Authors. STEM CELLS published by STEM CELLS Wiley Periodicals, Inc. on behalf of AlphaMed Press 2019