Plasmid-mediated gene transfer of Cas9 induces vector-related but not SpCas9-related immune responses in human retinal pigment epithelial cells

The clustered regularly interspaced short palindromic repeat (CRISPR)-associated protein 9 (Cas9) system represents a powerful gene-editing tool and could enable treatment of blinding diseases of the retina. As a peptide of bacterial origin, we investigated the immunogenic potential of Cas9 in models of retinal immunocompetent cells: human microglia (IMhu) and ARPE-19 cells. Transfection with Streptococcus pyogenes-Cas9 expression plasmids (SpCas9 plasmid) induced Cas9 protein expression in both cell lines. However, only ARPE-19 cells, not IMhu cells, responded with pro-inflammatory immune responses as evidenced by the upregulation of IL-8, IL-6, and the cellular activation markers HLA-ABC and CD54 (ICAM). These pro-inflammatory responses were also induced through transfection with equally sized non-coding control plasmids. Moreover, viability rates of ARPE-19 cells were reduced after transfection with both the SpCas9 plasmids and the control plasmids. Although these results demonstrate cell type-specific responses to the DNA plasmid vector, they show no evidence of an immunogenic effect due to the presence of Cas9 in models of human retinal pigment epithelial and microglia cells. These findings add another layer of confidence in the immunological safety of potential future Cas9-mediated retinal gene therapies.


Characterization of the immune-responsiveness of IMhu cells to stimulation with PRR ligands.
It has been shown that inherited retinal diseases caused by gene mutations can be corrected via plasmid-mediated Cas9 gene editing 7,10 . However, there is evidence that either Cas9 or DNA plasmids can induce immune responses 4,11,14 . Retinal microglia participate in inflammatory processes by secreting pro-inflammatory cytokines such as tumor necrosis factor-α (TNF-α), interleukin (IL)-1β, IL-6, and IL-18 18 . Moreover, murine microglia have been shown to release pro-inflammatory cytokines like TNF-α, IL-1β, IL-6 and the type I interferons (IFN) IFN-α and IFN-β in response to ligands of major PRRs 29 . This suggests that microglia could potentially also respond to the DNA plasmid vector and/or Cas9.
IMhu cells were used to analyze microglial immune responses to plasmid-mediated gene transfer of Cas9. This new microglia cell line is of validated human origin, as confirmed by sequencing and displays numerous similarities to primary human microglia in terms of morphology, the expression of cell surface markers, and immune responses to stimulation with pro-inflammatory cytokines 30,31 . IMhu cells exhibit the same typical microglial phenotype observed in primary and immortalized microglial cultures 31 . IMhu express surface markers specific for human microglia-macrophage lineage such as CD11b, TGFβR, and P2RY12; these markers are also expressed on primary microglia [30][31][32] . IMhu also demonstrate phagocytic and migratory activity characteristic of primary microglia 31,33 . Additionally, it has been demonstrated that IMhu respond to pro-inflammatory stimulation with the activation of an M1 phenotype, involving upregulation of several pro-inflammatory cytokines and chemokines, a response also observed in primary human microglia 30,34 . The characterization of IMhu as a microglial cell model, though, has not yet been extended to an assessment of the cell line's responsiveness to the activation of major PRRs. Thus, before investigating the IMhu immune response to SpCas9 plasmid transfection, we tested the immune-competence of this cell line by stimulating it with ligands of various PRRs. After 24 h, levels of pro-inflammatory cytokines and type I IFNs in the supernatant were determined using HEK-Blue IFN-α/β, HEK-Blue IL-1β, HEK-Blue IL-6, HEK-Blue IL-18, and HEK-Blue TNFα reporter cells (Fig. 1). Stimulation of intracellular DNA receptors with double-stranded DNA (dsDNA) or oligodeoxynucleotides containing unmethylated cytosine-guanine dinucleotides (CpG ODNs) did not induce a detectable production of IFN-α/β or pro-inflammatory IL-6. In contrast, stimulation with the Toll-like receptor (TLR)3 ligand Poly I:C induced significant releases of IFN-α/β, IL-6, IL-18, and TNF-α. Moreover, significantly elevated levels of IL-1β showed that IMhu also responded to inducers of inflammasome signaling (LPS and ATP) (Fig. 1). This suggests that IMhu cells are capable of mounting inflammatory immune responses to ligands of TLR3 and inducers of inflammasome signaling, but have limited immunoreactivity to ligands of intracellular DNA receptors.

Transfection with the SpCas9 plasmid results in Cas9 protein expression in IMhu cells but does not trigger cytokine responses.
To test whether plasmid-mediated gene transfer of Cas9 induces immune responses in cell models of human microglia cells, we designed an experimental plasmid coding for the Streptococcus pyogenes Cas9 (SpCas9) sequence (SpCas9 plasmid) ( Fig. 2A). An identical non-coding plasmid (NC plasmid) without the Cas9 sequence served as negative control for Cas9 staining ( Fig. 2A). To confirm that SpCas9 plasmid transfection results in Cas9 expression, IMhu cells were transfected with either the SpCas9 plasmid or NC plasmid using cationic lipid mediated transfection. At 24 h post-transfection, cells were stained with an anti-SpCas9 antibody and analyzed via fluorescence microscopy (Fig. 2B). We found that approximately 30% of the IMhu cells expressed Cas9 intracellularly. As expected, no Cas9 expression was detected in IMhu cells transfected with the NC plasmid (Fig. 2B).
Next, we analyzed the immune responses of IMhu cells to SpCas9 plasmid transfection. To this end, IMhu cells were transfected via cationic lipid mediated transfection with either the SpCas9 plasmid or the NC plasmid, or treated with the transfection reagent Lipofectamine 3000 (L3000) alone. At 24 h post-transfection supernatant was harvested and the release of 105 cytokines was assessed using Proteome Profiler™ Antibody Arrays. Heat map analysis of cytokine ratios of SpCas9 plasmid-transfected versus L3000 treated cells, or NC plasmid-transfected versus L3000 treated cells respectively showed that neither transfection with the SpCas9 plasmid nor with the NC plasmid triggered a detectable change in the release of any of the measured immune mediators (Fig. 2C). To verify these results, we repeated this experiment including IMhu cells which were either treated only with the SpCas9 plasmid (but no L3000) or not stimulated, as additional controls. At 6 h, 12 h, 24 h, and 48 h after stimulation, cytokine levels of IL-1β, IL-6, IL-18, and TNF-α were determined using HEK Blue™ cells. No differences in the levels of any of these cytokines were observed between SpCas9 plasmid-or NC plasmid-transfected cells, Figure 1. IMhu cytokine response to stimulation of major PRRs. Concentrations of TNF-α, IFN-α/β, IL-1β, IL-6, and IL-18 in the supernatant of IMhu cells 24 h following stimulation with 5 μg/ml dsDNA and 1.5 μl/ ml L3000; 1 µM CpG ODN and 5 µl/ml LTX; 10 μg/ml Poly I:C; or 5 μg/ml LPS and 5 mM ATP. Cytokine concentrations were measured using HEK-Blue IFN α/β, IL-6, TFNα, IL-1β, and IL-18 reporter cells. Error bars represent means + SD of n = 3 per group. Normally distributed data was assessed using One-way ANOVA with post hoc Tukey's tests. Non-normally distributed data was assessed using non-parametric Mann-Whitney tests or Kruskal-Wallis tests with post hoc assessment using Dunn's tests with control for joint ranks and Bonferroni adjustment. *p < 0.05, **p < 0.01. Asterisks indicate significant differences in comparison to unstimulated cells.  Figure S1A). To evaluate whether IMhu cells responded to other plasmids of similar or larger size encoding for SpCas9 and/ or different fluorescent proteins, IMhu cells were transfected with two additional plasmids (mKate: 4.8 kbp and the EGFP SpCas9 plasmid: 9.3 kbp) ( Figure S1B). Again, measurements of cytokine production from supernatant samples using HEK Blue™ IFN α/β-, IL-1β-, IL-6-, IL-18-, or TFNα-cells revealed no significant changes in cytokine releases of the plasmid-stimulated cells compared to the non-stimulated controls ( Figure S1C). These results indicate that neither plasmid transfection in general, nor plasmid-mediated gene transfer of Cas9 triggers immune responses in IMhu microglia cells.

SpCas9 plasmid transfection induces a vector-related release of IL-6 and IL-8 in ARPE-19 cells.
To evaluate whether plasmid-mediated gene transfer of Cas9 induces immune responses in human RPE cells, ARPE-19 cells were transfected with the SpCas9 plasmid and a non-coding control plasmid. As transfection efficacy 35 and nuclear delivery of plasmids 36,37 have been shown vary in relation to vector size [35][36][37] , the Cas9 sequence was replaced by a non-coding stuffer sequence enlarging the plasmid to the size of the 8 kbp SpCas9 plasmid (stuffer plasmid) (Fig. 3A, upper panel) to create the control plasmid. Two additional EGFPencoding plasmids: an EGFP-SpCas9 plasmid (9.3 kbp) expressing Cas9 and EGFP under the same promoter and a respective equally sized non-coding EGFP-stuffer plasmid were used (Fig. 3A, lower panel) to facilitate precise flow cytometric comparison of transfection rates. To assess Cas9 and/or EGFP protein expression following plasmid transfection, ARPE-19 cells were transfected with equal concentrations of the SpCas9 plasmid, the stuffer plasmid, the EGFP-SpCas9 plasmid, or the EGFP-stuffer plasmid. As the SpCas9 plasmids and their  www.nature.com/scientificreports/ respective stuffer control plasmids were of equal size and almost identical molar mass, the transfection at equal mass corresponded to an almost equimolar transfection of the SpCas9 plasmids and their respective control plasmids. After 24 h cells were stained with SpCas9 antibodies as described. Microscopic evaluation confirmed intracellular Cas9 expression exclusively in SpCas9 plasmid and EGFP-SpCas9 plasmid-transfected cells and a comparable EGFP expression exclusively in EGFP-SpCas9 plasmid and EGFP-stuffer plasmid-transfected cells (Fig. 3B) and demonstrated Cas9 and EGFP co-expression following transfection with the EGFP-SpCas9 plasmid as expected (Fig. 3B, third panel). For flow cytometric quantification of the transfection rates, EGFP-SpCas9 plasmid or the EGFP-stuffer plasmid-transfected cells were harvested at 24 h post treatment, stained with the cell death marker 7-AAD, and subsequently analyzed. Living ARPE-19 cells were gated and the percentage of EGFP positive cells was determined as shown in Fig. 3C,D. This analysis confirmed comparable transfection rates of the EGFP-SpCas9 plasmid [16.13 ± 3.8% (mean ± SD)] and the EGFP-stuffer plasmid (20.92 ± 6.43%) ( Fig. 3E) with no significant differences seen between groups (p = 0.557).
To analyze immune responses of ARPE-19 cells to plasmid-mediated gene transfer of Cas9, ARPE-19 cells were transfected with either the SpCas9 plasmid or treated with the transfection reagent Lipofectamine LTX (LTX) alone. At 24 h following transfection, cytokine levels in the supernatant were determined using Proteome Profiler™ Antibody Arrays. Heat map analysis demonstrates that transfection of ARPE-19 cells with the SpCas9 plasmid triggered an IL-8 response (Fig. 4A). IL-8 was increased 14-fold following SpCas9 plasmid-transfection when compared to LTX treatment only.
To strengthen the results of this semiquantitative analysis with a quantitative method, IL-8 concentrations in the supernatant of ARPE-19 cells transfected with the SpCas9 plasmid or the EGFP-SpCas9 plasmid were measured at five time points after treatment (3 h, 12 h, 18 h, 24 h, and 48 h) using a sandwich ELISA. To determine whether the observed IL-8 response was Cas9-related or plasmid-vector-related, this experiment also included ARPE-19 cells transfected with the stuffer plasmid and the EGFP-stuffer plasmid. ARPE-19 cells which were either treated with only the SpCas9 plasmid, with LTX, or left unstimulated served as additional controls. Transfection with both SpCas9-encoding plasmids as well as with both stuffer plasmids induced a strong IL-8 release, significant for all plasmid-transfected groups at 48 h after stimulation (48 h: SpCas9 plasmid + LTX vs unstim. control p < 0.001; stuffer plasmid + LTX vs unstim. control p < 0.001; EGFP-SpCas9 plasmid vs unstim. control p = 0.014; EGFP-stuffer-plasmid vs unstim. control p < 0.001), whereas no significant differences in IL-8 levels were observed between the additional controls ( Fig. 4B). Comparison between plasmid-transfected ARPE-19 cells revealed no significant differences in the IL-8 response between cells treated with SpCas9 encoding plasmids and stuffer plasmids (SpCas9 plasmid vs stuffer plasmid, p = 0.368; EGFP-SpCas9 plasmid vs EGFP-stufferplasmid, p = 0.354) (Fig. 4B).
Chen et al. have shown that IL-8 secretion by ARPE-19 cells is mediated by the induction of NF-κB and MAPK signaling, and activation of these signaling pathways leads to the additional release of IL-6 38 . To test whether plasmid-mediated gene transfer of Cas9 in ARPE-19 cells triggers IL-6 secretion, we measured IL-6 in the supernatant of SpCas9 plasmid-transfected cells and the corresponding control groups using a sandwich ELISA. This more sensitive cytokine analysis showed that transfection with both SpCas9-encoding plasmids as well as with both stuffer plasmids resulted in an increase in IL-6 production which was significant for all groups at 48 h after stimulation (48 h: SpCas9 plasmid + LTX vs unstim. control p < 0.001; stuffer plasmid + LTX vs unstim. control p < 0.001; EGFP-SpCas9 plasmid vs unstim. control p < 0.001; EGFP-stuffer-plasmid vs unstim. control p < 0.001) (Fig. 4C). No differences in IL-6 concentrations were observed between cells treated only with either the SpCas9 plasmid or LTX and unstimulated controls. Moreover, there were no significant differences in the IL-6 response between cells treated with SpCas9 encoding plasmids and stuffer plasmids (SpCas9 plasmid vs stuffer plasmid, p = 0.424; EGFP-SpCas9 plasmid vs EGFP-stuffer-plasmid, p = 0.755) (Fig. 4C).
Taken together, plasmid transfection of ARPE-19 cells with either the SpCas9-encoding plasmids or the stuffer control plasmids triggered significant IL-8 and IL-6 secretion. These effects were comparable between plasmids both coding and non-coding for Cas9, suggesting that in ARPE-19 cells, the immunogenic effect of Cas9 transfection is triggered by the transfected plasmid DNA rather than by the presence or expression of the Cas9 transgene.

Plasmid transfection upregulates immunological surface markers on ARPE-19 cells. It has
been shown that stimulated ARPE-19 cells upregulate immunological surface markers indicative of cell activation, including HLA-ABC and HLA-DR major histocompatibility (MHC) antigens, and CD54 (ICAM-1) 25 . To analyze whether plasmid-mediated gene transfer of Cas9 leads to cell activation, ARPE-19 cells were stimulated with the SpCas9 plasmid and the stuffer plasmid. At 24 h post-stimulation, cells were harvested, stained with fluorescent antibodies against HLA-ABC, HLA-DR and CD54, and the cell death marker 7-AAD, and subsequently analyzed via flow cytometry. Living ARPE-19 cells were gated as shown in Fig. 3C and surface expression of HLA-DR, HLA-ABC, and CD54 was determined (Fig. 5A). We found that neither SpCas9 plasmid-and stuffer plasmid-transfection nor stimulation of ARPE-19 cells with the Cas9 plasmid or LTX alone induced changes in the expression of HLA-DR (Fig. 5A,B; lower panels). However, both SpCas9 plasmid transfection and stuffer plasmid transfection led to a significant upregulation of HLA-ABC (SpCas9 plasmid + LTX vs unstim. control p < 0.001; stuffer plasmid + LTX vs unstim. control p < 0.001) and CD54 (SpCas9 plasmid + LTX vs unstim. control p < 0.001; stuffer plasmid + LTX vs unstim. control p < 0.001), with no significant differences in the expression levels of these markers seen between the two groups. No changes in HLA-ABC and CD54 expression were observed in cells treated only with either the SpCas9 plasmid or LTX (Fig. 5A,B; upper panels). Similar results were observed following transfection with the EGFP-SpCas9 plasmid and the EGFP-stuffer-plasmid: here, transfection also did not lead to changes in the expression of HLA-DR (Fig. 5C,D; lower panels), but caused increases in the expression of HLA-ABC and CD54. These were significant for HLA-ABC (EGFP-SpCas9 plasmid + LTX vs unstim. control p < 0.001; EGFP-stuffer plasmid + LTX vs unstim. control p < 0.001). A similar trend was  (Fig. 5A,B; upper panels).
Overall, this suggests that the upregulation of HLA-ABC and CD54 expression on ARPE-19 cells was triggered by the transfection of the DNA plasmid rather than by the presence or expression of the Cas9 transgene.

Plasmid transfection decreases the viability rate of ARPE-19 cells.
It has been shown that cationic lipid transfection of DNA plasmids can induce cell death 12 . To determine whether cationic lipid-mediated transfection of the SpCas9 plasmid influences cell viability, the transfected IMhu cells and ARPE-19 cells were first examined microscopically. There was no evidence of increased cell death in SpCas9 plasmid-transfected and NC plasmid-transfected IMhu microglia up to 48 h after treatment ( Figure S2A). In contrast numerous floating spherical cells in the supernatant of ARPE-19 cultures suggested reduced cell viability from 24 h on after transfection with the SpCas9-encoding plasmids or the corresponding stuffer plasmids ( Figure S2B). To investigate the viability of transfected ARPE-19 cells in more detail, these cells were either transfected with the SpCas9 plasmid or stuffer plasmid, or treated with LTX or SpCas9 plasmid only and compared to unstimulated cells. After 24 h, 7-AAD negative living ARPE-19 cells were quantified by flow cytometry (compare Fig. 3C). The analysis showed a minimal reduction in the viability percentages of SpCas9 plasmid-transfected and stuffer plasmid-transfected ARPE-19 cells when compared to control groups (SpCas9 plasmid + LTX vs unstim. con- (C) Transfection with both the SpCas9-encoding plasmids and the stuffer control plasmids also triggered a significant production of IL-6. Data was analyzed via one-way ANOVA followed by Bonferroni's comparison tests for selected pairs of columns. Asterisks indicate significant differences in comparison to unstimulated ARPE-19 cells. Bars represent means + SD of pooled mean values from three independent experiments measured in duplicates. *p < 0.05, **p < 0.01, ***p < 0.001.  (Fig. 6A). Similar results were observed for EGFP-SpCas9 plasmid-transfected and EGFP-stuffer plasmid-transfected cells (EGFP-SpCas9 plasmid + LTX vs unstim. control p = 0.009; EGFP-stuffer plasmid + LTX vs unstim. control p = 0.02) (Fig. 6B). The percentage of living cells did not differ between SpCas9 encoding plasmids or the corresponding stuffer plasmids, suggesting weak cytotoxic effects induced by plasmid transfection rather than the presence or expression of the Cas9. Interestingly, the transfer of DNA plasmids can trigger strong innate immune responses in muscle cells 11 , demonstrating the potential immunogenicity of plasmid DNA. Moreover, Cas9 is a peptide of bacterial origin and therefore carries the potential for non-self-recognition and immunogenicity. Such an effect could lead to adverse reactions and negate therapeutic efficacy in the target tissue. Indeed, recent publications demonstrate that Cas9 can induce humoral and cell-specific immune responses in human blood cells [13][14][15][16][17] . On the other hand, no Cas9-specific immune responses were observed in a pre-clinical study of CRISPR/Cas9 retinal gene editing 45 , suggesting a potential site-and cell-specific reaction to Cas9. We therefore investigated the innate immune response to Cas9 in model systems relevant to ocular gene therapy. Specifically, we evaluated immune responses to plasmid-transfection of SpCas9 in human retinal cell models and demonstrated cell type-specific immunogenicity of the DNA plasmid vector, but not of SpCas9, in a model of human RPE cells and a complete absence of plasmid-related or SpCas9-related immunogenicity in a human microglia model. In plasmid-mediated gene transfer of Cas9, potential immunogenic components that could trigger innate immune responses consist of plasmid DNA, the transgene mRNA, and the transgene protein. Additionally, immune responses observed following transfection of experimental plasmids lacking the Cas9 sequence might www.nature.com/scientificreports/ be induced by the expression of vector encoded fluorescence proteins such as GFP 46 . It has been shown that transfection of non-coding plasmid DNA into muscle cells induced upregulation of the endosomally expressed DNA sensor TLR9 and of other PRRs responding to cytoplasmic DNA and triggered pro-inflammatory cell infiltration 11 . Moreover, electrotransfer of SpCas9 transgene RNA stimulated an immune response in human CD34+ hematopoietic stem cells 4 and extracellular application of SpCas9 proteins induced a pro-inflammatory cytokine release in human monocytes 14 . Human RPE cells have been shown to express intracellular receptors for DNA including TLR9 and the cytosolic DNA receptor cGAS 27,47 , and also exhibit PRRs capable of detecting RNA 26,48 and extracellular proteins 48,49 . Moreover, ARPE-19 cells were found to react to the TLR9 ligand CpG DNA and to intracellular mitochondrial DNA 27,50 . Additionally, human primary RPE cells responded to ligands of the RNA sensing TLR3 48 . Similarly, human microglia cells have been shown to express cGAS 51 and innate PRRs responding to RNA and proteins 52 , but in contrast to RPE cells, the expression of TLR9 is low or absent in primary human microglia 52 . In line with this we observed that although IMhu microglia cells reacted to ligands of a RNA sensing TLR (TLR3), they did not respond to the TLR9 ligand CpG ODN. Additionally, we also found that they did not react to dsDNA. Collectively, this suggests that differences may exist in the capability of microglia and RPE cells to sense intracellular DNA. Differences in the immune responsiveness to DNA provide a possible explanation for the discrepancy in the reactivity of IMhu cells and ARPE-19 cells to SpCas9 plasmid transfection. While SpCas9 plasmid transfection did not elicit a cytokine response in IMhu cells, in ARPE-19 cells it caused a release of pro-inflammatory IL-8 and IL-6 and an upregulation of the cellular activation markers HLA-ABC and CD54 (ICAM-1) in response to the SpCas9 plasmid as well as to the EGFP-SpCas9 plasmid. Similar responses were also induced in cells transfected with the corresponding stuffer control plasmids. In the latter groups, the observed immune reaction could neither be attributed to the presence of Cas9 mRNA nor to Cas9 protein. Moreover, no responses were detected in ARPE-19 cells when the Cas9 plasmid was applied extracellularly. Taken together, this suggests that intracellular plasmid DNA vectors triggered inflammatory immune responses in ARPE-19 cells. In contrast, in IMhu cells, the absence of an immune response to plasmid transfection might have been related to their reduced reactivity to DNA antigens and/or intracellularly expressed proteins.

Scientific Reports
A detailed analysis of the immune responses to Cas9 plasmid transfection in ARPE-19 cells revealed no significant differences in the IL-8 and IL-6 responses between cells treated with SpCas9 encoding plasmids and stuffer plasmids. There were also no significant differences seen in the expression of HLA-ABC and CD54 between ARPE-19 cells transfected with SpCas9 encoding plasmids and stuffer plasmids. Collectively, this indicates that the inflammatory immune responses were triggered by the intracellular plasmid DNA vector but not by the presence or the expression of the SpCas9 transgene.
The cytokines and surface molecules upregulated in ARPE-19 cells following plasmid transfection have been shown to play important roles in ocular inflammation. Pro-inflammatory IL-6 is a critical mediator of uveitis 53 and induces the disruption of tight junction complexes between RPE cells, leads to the VEGF-induced recruitment of retinal microglia to the RPE layer and thereby compromises the barrier function of the RPE 54 . IL-8 is also associated with human retinal inflammatory diseases [55][56][57] and has a chemotactic effect on neutrophil granulocytes 58 as well as highly cytotoxic CD8+ T cells 59 . Moreover, ICAM-1, which is also expressed by human primary RPE cells, is critically involved in the cross-migration of leukocytes across the blood-retinal barrier 60 , while antigen presentation via HLA-ABC molecules is required for the activation of CD8+ cytotoxic T cells. This suggests that an upregulation of these immune molecules in the retina induced by cell or gene therapies might promote ocular inflammatory processes.
We observed that cationic lipid transfection of Cas9 plasmids not only induced pro-inflammatory immune responses in ARPE-19 cells but also resulted in a minor decrease in cell viability. Interestingly, cell viability was also reduced in ARPE-19 cells transfected with the stuffer control plasmids, but not in cells that were only treated with either transfection reagent or plasmid DNA. This suggests that increased cell death was not caused by the transfection reagent alone or by extracellular plasmids or the Cas9 transgene. Increased cell death in cells transfected with DNA-cationic lipid complexes, but not in cells treated with uncomplexed material, was also observed by Nguyen et al. 12 In this study the effect of cationic lipid transfection of bacterial plasmids was studied in human HeLa cells. A detailed analysis revealed that cell death following transfection with plasmid DNAcationic lipid complexes was caused by apoptosis as shown by substantial DNA fragmentation and upregulation of genes involved in the ER stress-mediated apoptosis pathway 12 . Thus, it is possible that the minor decrease in cell viability seen in our plasmid-transfected ARPE-19 cells was caused by apoptosis induced by plasmid DNAcationic lipid complexes. Overall, this suggests that in vivo approaches of plasmid-mediated retinal gene transfer should be monitored for the occurrence of inflammatory immune responses and cell death.
In summary, we have shown that plasmid-mediated gene transfer of SpCas9 induced an increased release of pro-inflammatory cytokines and an upregulation of cellular activation markers in a model of human RPE cells, but not in human microglia. Importantly, this immune response was only induced by plasmid transfection but not specifically by Cas9. As the immune reactivity of retinal cells has been shown to depend on their tissue context 61 , immune responses to retinal gene therapy might also depend on the ocular or cellular microenvironment of the transfected cells, indicating that in vitro immune responses may differ from in vivo responses. Nevertheless, our results strongly suggest a generally low immunogenicity of Cas9 in microglia and RPE cells. Results demonstrating no or low immunogenicity of Cas9 are also seen in vivo studies of CRISPR/Cas9 retinal gene editing. Thus, no peripheral adaptive immune responses to SaCas9 were detected in non-human primates subretinally injected with AAV5-encoded CRISPR-SaCas9 to correct the common deep-intronic mutation in CEP290 associated with LCA10 (EDIT-101; Editas Medicine, Inc.) 45

Plasmid transfection of IMhu and ARPE-19. Cationic lipid mediated transfection of IMhu cells with
SpCas9 plasmids and NC plasmids and of ARPE-19 cells with EGFP-SpCas9 plasmids, EGFP-stuffer plasmids, SpCas9 plasmids and stuffer plasmids was performed using Lipofectamine 3000 or Lipofectamine LTX with Plus reagent respectively (both Thermo Fisher Scientific, Waltham, MA, USA). IMhu cells were seeded at 2 × 10 5 cells per well of a 24-well plate. ARPE-19 cells were seeded at 9 × 10 5 cells per well of a 24-well plate and at 2.5 × 10 4 cells per well of a 96-well plate, respectively. Both cell types were incubated in cell culture medium overnight prior to transfection with Cas9 or control plasmids at optimized plasmid-reagent ratios. Additionally, IMhu were transfected with an mKate fluorescent plasmid and the EGFP-SpCas9 plasmid. IMhu were transfected with 1.5 μl/ml Lipofectamine 3000, 1 μl/ml Plus reagent and 1.5 μg/ml plasmid DNA. ARPE-19 cells were transfected with 1.5 µl/ml Lipofectamine LTX, 0.375 µl/ml Plus reagent and 375 ng/ml plasmid DNA. Stimulation with 10 µg/ml Poly (I:C) served as positive control. Cells were incubated for 24 h at 37 °C prior to collection of supernatant and fixation of cells for immunohistochemistry staining. To assess the time point after plasmid transfection at which the Cas9 protein started to be expressed by IMhu cells and ARPE-19 cells, the cells were transfected with the EGFP-SpCas9 plasmid and EGFP fluorescence was analyzed using fluorescence microscopy. www.nature.com/scientificreports/ fluorescence signal corresponding to the amount of cytokine bound was generated utilizing IRDye 800CW followed by LI-COR detection. A LI-COR Odyssey ® Infrared Imaging System (LI-COR, Bad Homburg, Germany) was used to detect near-infrared fluorescence. The Odyssey scan was run with a resolution of 84 μm, an intensity of 5, 800 nm, an absorbance of 774 nm, and an emission of 789 nm. Signal intensity, measured in pixel density, was analyzed using ImageStudioLite Software (LI-COR Biosciences, Lincoln, NE, USA). The average signal of the duplicate spots was measured for each analyte and determined and normalized to the average signal of the reference spots after correction with the background signal.

Measurement of cytokines.
Concentrations of IL-8 and IL-6 in the supernatants of ARPE-19 cells were determined by sandwich ELISA using the Human IL-8 DuoSet ELISA and the Human IL-6 DuoSet ELISA (both R&D systems, Minneapolis, MN, USA) according to the manufacturer's protocol. Cytokine concentrations in the supernatants of IMhu cells were measured using HEK-Blue™ IFN α/β, IL-6, TFNα, IL-1β, and IL-18 reporter cells (all InvivoGen, San Diego, CA, USA). These cells allow the detection of the respective cytokine through the activation of an NF-κBinducible promoter and the production of secreted embryonic alkaline phosphatase (SEAP). Concentrations of SEAP in the supernatant can be assessed by a SEAP detection assay (InvivoGen, San Diego, CA, USA) using QUANTI-Blue™, a reagent that turns blue in the presence of SEAP. HEK-Blue™ cell measurement of cytokines was performed as per manufacturer's protocol. In brief, HEK-Blue™ cells were incubated with supernatant from stimulation experiments at 37 °C for 24 h. Induced HEK-Blue™ supernatant was then collected and incubated with QUANTI-Blue™ at 37 °C and SEAP concentrations were assessed using Infinite M200 microplate reader (Tecan, Männedorf, CH) at 640 nm.