Induced pluripotent stem cells reprogrammed from primary dendritic cells provide an abundant source of immunostimulatory dendritic cells for use in immunotherapy

Abstract Cell types differentiated from induced pluripotent stem cells (iPSCs) are frequently arrested in their development program, more closely resembling a fetal rather than an adult phenotype, potentially limiting their utility for downstream clinical applications. The fetal phenotype of iPSC‐derived dendritic cells (ipDCs) is evidenced by their low expression of MHC class II and costimulatory molecules, impaired secretion of IL‐12, and poor responsiveness to conventional maturation stimuli, undermining their use for applications such as immune‐oncology. Given that iPSCs display an epigenetic memory of the cell type from which they were originally derived, we investigated the feasibility of reprogramming adult DCs to pluripotency to determine the impact on the phenotype and function of ipDCs differentiated from them. Using murine bone marrow‐derived DCs (bmDCs) as proof of principle, we show here that immature DCs are tractable candidates for reprogramming using non‐integrating Sendai virus for the delivery of Oct4, Sox2, Klf4, and c‐Myc transcription factors. Reprogramming efficiency of DCs was lower than mouse embryonic fibroblasts (MEFs) and highly dependent on their maturation status. Although control iPSCs derived from conventional MEFs yielded DCs that displayed a predictable fetal phenotype and impaired immunostimulatory capacity in vitro and in vivo, DCs differentiated from DC‐derived iPSCs exhibited a surface phenotype, immunostimulatory capacity, and responsiveness to maturation stimuli indistinguishable from the source DCs, a phenotype that was retained for 15 passages of the parent iPSCs. Our results suggest that the epigenetic memory of iPSCs may be productively exploited for the generation of potently immunogenic DCs for immunotherapeutic applications.


| INTRODUCTION
As professional antigen-presenting cells, dendritic cells (DCs) are uniquely responsible for initiating all immune responses and are, therefore, attractive candidates for use in immunotherapy. 1 Accordingly, DCs differentiated from peripheral blood monocytes (moDCs) have been used in more than 200 clinical trials in an attempt to vaccinate cancer patients to defined tumor associated antigens (TAAs). 2 Although such trials have demonstrated only limited efficacy, induced pluripotent stem cells (iPSCs) may provide an unlimited source of previously inaccessible subsets of DCs, such as plasmacytoid DCs 3 and CD141 + DCs displaying superior capacity for the cross-presentation of TAAs. 4 Unlike moDCs, such a novel source is amenable to scale-up, quality control and genome editing, 3,5 making it attractive for downstream therapeutic applications. 6 Nevertheless, iPSC-derived DCs (ipDCs) appear arrested at a stage of development more reminiscent of a fetal rather than an adult phenotype, a finding common to many cell types differentiated from a pluripotent source, including hepatocytes, 7 cardiomyocytes, 8 and forebrain neurons. 9 Human iPSCderived cardiomyocytes, for example, typically display fetal-type ion channel expression, spontaneous contraction, and fetal-type electrophysiology, 8,10 limiting their capacity to integrate into cardiac tissue and electrically couple with adult cardiomyocytes. In the case of DCs, the fetal phenotype is characterized by low MHC class II expression, 5,11,12 excessive anti-inflammatory IL-10 secretion but barely detectable levels of IL-12 upon ligation of pattern recognition receptors (PRRs), a similar inability to secret IL-12 having been described previously among human neonatal DCs. 13,14 Since their advent more than a decade ago, 15 iPSCs have been demonstrated by numerous laboratories to display "epigenetic memory" of the cell type from which they were originally derived, 16,17 attributed to residual DNA methylation patterns and histone modifications within the genome of the source cells. [18][19][20] Indeed, multiple studies have demonstrated retained memories of cell types of origin within iPSCs, including retinal pigmented epithelial cells, 21 pancreatic β islet cells, 22 hepatocytes, 23 blood progenitors, 16 granulocytes, 17 fibroblasts, 16,23,24 skeletal muscle precursors, 17 hepatoblasts, 25 osteoblasts, 26 melanocytes, 23 adipocytes, 24 germline stem cells, 27 keratinocytes, 24 bone marrow mesenchymal stem cells, 28 and cardiac myocytes. 29 The subtle epigenetic signature of the cell type of origin may contribute to an increased tendency for spontaneous redifferentiation back into the donor cell type 30 : indeed, differentiation of cardiac progenitor cell-derived iPSCs into cardiomyocytes was shown to be more efficient than fibroblast-derived iPSCs, 31 whereas iPSCs derived from mesoangioblasts displayed stronger myogenic commitment than conventional, fibroblast-derived iPSCs. 32 This epigenetic memory is, however, eventually lost upon repeated passage of iPSCs, 17 suggesting that continual cell division progressively removes any epigenetic relicts of past incarnations and promotes resolution of epigenetic differences among iPSC lines. 33 In order to address the limitations of the fetal phenotype of ipDCs, we investigated the feasibility of reprogramming existing DCs to pluripotency in order to exploit the epigenetic memory of iPSCs for the cell type of origin to facilitate the downstream differentiation of DCs with a definitive, adult phenotype. Given the sentinel function of DCs which render them highly sensitive to viral pathogens and PRR ligands, we anticipated that reprogramming of existing DCs with viral vectors may prove challenging. However, contrary to expectation, we show here that mouse bone marrow-derived DCs (bmDCs) are fully amenable to reprogramming unless provoked to mature in response to PRR agonists. Furthermore, iPSC lines derived in this way (so-called iPSC DC ) show significant capacity for redifferentiation along the DC lineage pathway, the resulting cells displaying superior immunogenicity in vitro and in vivo compared to ipDCs differentiated from conventional iPSC lines derived from MEFs (iPSC MEF ). Although ultimately lost after passage 15, we conclude that the epigenetic memory retained by the iPSC lines provides an adequate window of opportunity for the production of immunostimulatory DCs for vaccination purposes.

| Preparation of source DCs
Bone marrow from the dissected femurs of adult male CBA/Ca mice was depleted of erythrocytes and cultured at 7.5 × 10 5 cells/mL in

Significance statement
Induced pluripotent stem cells (iPSCs) offer a scalable source of rare dendritic cell (DC) subsets, not otherwise accessible from patients, that are amenable to manufacture and quality control. However, in common with other cell types differentiated from a pluripotent source, DCs become arrested in their developmental pathway, significantly limiting their immunogenicity. This study showed that a simple modification to protocols for the derivation of iPSC lines fully resolves this issue and demonstrated the feasibility of reprogramming existing DCs to pluripotency, thereby exploiting the epigenetic memory of iPSCs for the cell type of origin.
By capturing the epigenetic profile of existing DCs, the resulting iPSCs spawn populations of DCs that are highly immunostimulatory and suitable for use in immunotherapy. 90 mm diameter tissue culture plates in R10 medium (RPMI 1640 supplemented with 10% fetal calf serum [FCS], 2 mM L-glutamine, 1% sodium pyruvate and 50 μM 2-mercaptoethanol) further supplemented with~5 ng/mL murine GM-CSF harvested from the supernatant of an X6310 cell line stably transfected with the murine Gmcsf gene. Nonadherent cells were removed on days 3 and 6 of culture when the medium was replaced and cells were harvested on day 7. DCs were purified using anti-CD11c-APC monoclonal antibodies (mAb) followed by anti-APC magnetic beads, according to the manufacturer's instructions (Miltenyi Biotec, Bisley, Surrey, UK).  iPSC lines were routinely passaged every 3 days.   counter-stained with anti-CD4-APC mAb as described above. CD4 + T cell proliferation was analyzed using a FACSCalibur cytometer.

| Statistical analysis
All presented data are representative of multiple independent experiments. Data are reported as mean ± SD. Results were analyzed using GraphPad Prism (Version 7.0). Where appropriate, results between groups were compared using a two-tailed unpaired t test. Differences between groups were considered statistically significant at P < .05. Asterisks indicate the level of statistical significance (*P < .05; **P < .001).

| Conventional ipDCs display a fetal phenotype
To confirm previous observations of the fetal phenotype expressed by DCs differentiated from conventional iPSCs, MEFs were transduced with SeV vectors encoding Oct4, Sox2, Klf4, and cMyc and multiple monoclonal iPSC lines were established. The representative cell line, iPSC MEF SV 2 , displayed characteristic cell surface (SSEA-1) and intracellular (Oct4, Nanog) markers of pluripotency (Supporting Information Figure S1A) and gave rise to EBs which, when engrafted under the kidney capsule of syngeneic recipients, formed teratomas containing tissues derived from each of the embryonic germ layers, confirming their pluripotency (Supporting Information Figure S1B). Culture of EBs in vitro with rmIL-3 and GM-CSF, as described previously, 35 Figure 2E). Prior to reprogramming, source cells therefore displayed the desired phenotype and function of highly immunogenic DCs.

| Fully differentiated DCs can be reprogrammed to pluripotency
Purified bmDCs were reprogrammed to pluripotency with the use of CytoTune 2.0 (SeV), forming iPSC-like colonies as early as day 8 following transduction ( Figure 3A). The greatest reprogramming efficiency was achieved using high levels of Klf4 (a ratio of KOS:cMyc: Klf4 of 5:5:5). Control MEFs, reprogrammed in parallel, likewise formed classical, domed iPSC colonies when plated onto monolayers of feeder cells in 6 well plates ( Figure 3B). To quantify reprogramming efficiency, wells were stained with methylene blue to facilitate the enumeration of colonies. Reprogramming efficiency of control MEFs, a cell type particularly amenable to reprogramming, ranged between 0.18% and 1.68% ( Figure 3C and Table 1). In contrast, the efficiency of DC reprogramming was significantly lower, ranging between 0.03% and 0.11% of cells successfully forming iPSCs ( Figure 3D and Table 1).
No colonies were observed in wells cultured in the absence of SeV ( Figure 3E). Likewise, CD11c + bmDCs induced to mature by prior exposure to LPS, failed to yield any iPSC colonies ( Figure 3F), suggesting that the tractability of DCs for reprogramming is highly dependent on their maturation status.
Although reprogramming efficiency of bmDC was comparatively low, numerous colonies were harvested and monoclonal iPSC lines  Figure 5B). Co-stimulatory molecules and MHC class II were consistently upregulated upon exposure to 1 μg/mL LPS to induce maturation ( Figure 5B), which also gave rise to cells with highly dendritic morphology ( Figure 5A). Importantly, levels of IL-12 secretion in response to LPS were not significantly different from those obtained from bmDCs, even at the highest concentration of LPS used (P = .06) ( Figure 5C), which contrasted strongly with the secretion profile of conventional ipDCs differentiated from iPSC MEF ( Figure 1D). Levels of IL-10 were likewise largely indistinguishable from those of bmDCs cultured in parallel ( Figure 5D).
In order to determine the capacity of iPSC DC -DCs to process and present exogenous protein antigen in the context of MHC class II,   Table S1). Together, these data support the notion that early passage iPSCs reprogrammed from primary DCs readily redifferentiate into DCs that are not subject to the fetal phenotype that typifies those differentiated from conventional iPSC MEF and therefore display an immunostimulatory phenotype conducive to their use in immunotherapy.

| DISCUSSION
The importance of DCs in the initiation of immune responses among antigen-naïve T cells has fueled interest in iPSCs as a renewable and scalable source of rare DC subsets with desirable properties not normally vested in moDCs, 3    Importantly, this therapeutic window may be significantly extended through the use of recently described SeV vectors targeted by microRNA-302, known to be rapidly upregulated following the reprogramming of cells to pluripotency. This elegant system results in the auto-erasure of vectors and transgenes shortly after the establishment of iPSCs of either mouse or human origin. 49 In this context, we and others 50 have recently demonstrated the feasibility of reprogramming human moDCs to pluripotency suggesting the potential for the future clinical translation of our findings.
In terms of cancer immunotherapy, our approach may permit the derivation of iPSC lines from small numbers of moDCs differentiated in vitro from a patient's peripheral blood monocytes. This approach carries many advantages over the conventional use of dermal fibroblasts. First, the isolation of monocytes and their subsequent differentiation may be achieved over a 2 day period, 51 comparing favorably with the duration of time required to expand dermal fibroblasts from a skin biopsy to sufficient numbers required for reprogramming. 52 In addition, simple venipuncture required for monocyte isolation is significantly less invasive, thereby carrying a lower risk of infection com-