Generation of universal and hypoimmunogenic human pluripotent stem cells

Abstract There is a need to store very large numbers of conventional human pluripotent stem cell (hPSC) lines for their off‐the‐shelf usage in stem cell therapy. Therefore, it is valuable to generate “universal” or “hypoimmunogenic” hPSCs with gene‐editing technology by knocking out or in immune‐related genes. A few universal or hypoimmunogenic hPSC lines should be enough to store for their off‐the‐shelf usage. Here, we overview and discuss how to prepare universal or hypoimmunogenic hPSCs and their disadvantages. β2‐Microglobulin‐knockout hPSCs did not harbour human leukocyte antigen (HLA)‐expressing class I cells but rather activated natural killer (NK) cells. To avoid NK cell and macrophage activities, homozygous hPSCs expressing a single allele of an HLA class I molecule, such as HLA‐C, were developed. Major HLA class I molecules were knocked out, and PD‐L1, HLA‐G and CD47 were knocked in hPSCs using CRISPR/Cas9 gene editing. These cells escaped activation of not only T cells but also NK cells and macrophages, generating universal hPSCs.

human body. [6][7][8][9][10] Subsequently, the differentiated cells from hPSCs can be utilized as the cell source for patient therapies (Figure 1 Furthermore, class II expression can be induced in other cell types upon stimulation by interferons. 11 Each polymorphic HLA gene harbours very large numbers of multiple alleles (Table 1), 12 and therefore, it is extremely difficult to find a cell, organ or tissue donor who matches a specific pair of HLA alleles. Typically, cells, tissues and organs with mismatched HLA haplotypes are transplanted into patients treated with immunosuppressive therapy. However, prolonged treatment with immunosuppressive medicine, which is required to prevent the rejection of mismatched cells and grafts, often leads to dangerous side effects.
The disadvantage of hPSC therapies is that many types of hPSCs with specific HLA class I and class II molecules need to be stocked for their banking for hPSC therapies ( Figure 1A(a)). Taylor and colleagues suggested that the development of approximately 150 hESC lines with different HLA types would be necessary to provide sufficient HLA types to match hESC derivatives for most patients in the UK. 13 Nakajima and colleagues suggested that 170 hESC lines with different HLA types would support HLA matching for 80% of patients in Japan (but not 100% of patients). 14 However, more than 150 000 donors must be screened to generate 140 HLA-A, -B and -DR homozygous hPSC lines. 11 The extremely high cost needed to store large numbers of hPSC lines, even 150-170 cell lines, for their transplantation will lead to a deficit in the national insurance budget if hPSC therapies are approved for usage by national insurance in every country.
Recently, hPSCs, which have approximately the same characteristics as hESCs, were generated from adult somatic cells through the "forced" expression of (a) certain pluripotent genes, 15 such as Sox2, Oct3/4, klf-4 and l-myc or c-myc, (b) proteins (protein-induced F I G U R E 1 Human ESC and hiPSC therapy. A, (a) A single cell line matched to a specific patient's HLA type is selected among many hESC and hiPSC lines, which will be utilized for cell therapy in a specific patient. (b) Patient cells are collected and reprogrammed into patientspecific hiPSCs, which will be utilized for cell therapy in a specific patient. (c) A few lines of hypoimmunogenic or universal hESCs and hiPSCs are stored, which will be utilized for cell therapy in any patient. B, Overview of stem cell therapy using universal hESCs and hiPSCs. Wild-type hESCs and hiPSCs are prepared to be hypoimmunogenic or universal hESCs and hiPSCs by knocking out certain HLA genes and/ or knocking in immune-related genes. Then, universal hESCs and hiPSCs are differentiated into specific cell types and subsequently used for cell therapy to treat patients PSCs) 16,17 or (c) miRNAs, 16,18 which can be prepared from patient cells. 16 These cells were named hiPSCs. However, one of the difficulties associated with the application of hiPSCs for patient-specific therapy is that much time is required to generate mature and clinical-grade hiPSCs because they require an evaluation to prove that there is no contamination with viruses or other pathogens and no genetic abnormality, contributing to the high cost associated with hiPSC therapies and the time-consuming and laborious procedures ( Figure 1A(b)). However, patient-specific hiPSCs have no immunogenicity-related problems after the transplantation of differentiated cells from these hiPSCs because hiPSCs have identical HLA haplotypes.

| PREPAR ATI ON ME THODS OF UNIVER SAL (HYP OIMMUNOG ENIC) HPSC S
It is necessary to develop hPSCs that do not express HLA class I and class II molecules (universal hPSCs) even after differentiation into specific cell lineages. These universal or hypoimmunogenic hPSCs can be used for clinical therapy in patients with different types of HLA class I and class II molecules ( Figure 1A(c) and B). There are several strategies that can be used to reduce the number of banked hPSC lines: (a) generation of universal hPSCs by knocking out β2microglobulin (B2M), (b) generation of HLA-homozygous hPSCs by gene editing and (c) generation of universal or hypoimmunogenic hPSC by knocking in specific genes. We will discuss these methods, and we will focus on the advantages and disadvantages of these methods in preparing universal and hypoimmunogenic hPSCs.
One million megakaryocytes derived from B2M-knockout hiP-SCs were intravenously transfused into NSG mice. Two hours and one day after megakaryocyte transfusion, peripheral blood mononuclear cells were collected from the mice, and the expression of human CD41 and CD42b (megakaryocyte and platelet markers) was evaluated by flow cytometry. The results indicated that their engineered platelets could survive in vivo 24 hours after the injection of B2M-knockout megakaryocytes into mice.
However, there is a concern that cells without HLA class I molecules are targeted by NK cell-mediated cell lysis because they lack the missing-self response. Furthermore, only 6-30 platelets are generated from a hPSC-derived megakaryocyte, whereas 2000-10 000 platelets are generated from a natural HSC-derived megakaryocyte.
It is necessary to develop a more sophisticated differentiation medium and method for megakaryocyte induction from hPSCs.
Several other researchers have also designed and developed HLA-universal platelets by disrupting the B2M gene using transcription activator-like effector nuclease (TALEN) engineering in hiPSCs 22,25 or RNA interference (RNAi) engineering in hiPSCs and CD34 + HSC progenitor cells. 24,25,36,37 Petrus-Reurer et al 20 generated B2M-knockout hESCs, class II major histocompatibility complex transactivator (CIITA)-knockout hESCs, and both B2M-and CIITA-knockout hESCs using CRISPR/ Cas9 engineering as well as retinal pigment epithelium (RPE) cells differentiated from these gene-edited cells. All of their gene-edited cells suppressed both CD4 + and CD8 + T-cell activation, but these edited cells enhanced NK cell activation, although they did not show a stronger cytotoxic activity than wild-type (WT) hESCs.   HLA-B-knockout hiPSCs exhibited a weaker immune response than the mother hiPSCs according to HLA-targeted complement-dependent cytotoxicity assays. 30 These results indicate that HLA-edited hiPSCs represent a source of immunocompatible and off-the-shelf hiPSCs as well as their differentiated cells for patient therapies.
However, we still need to prepare many hPSC lines with only par-

| Generation of HLA-homozygous cell lines by gene editing
HLA-homozygous cell lines can reduce the number of hPSC cell banks. Therefore, several researchers aimed to generate HLAhomozygous hPSCs from HLA-heterozygous hPSCs through a gene editing process ( Figure 2B(b)). 11

| Universal hPSC generation by knocking in some specific genes
Han et al 31 generated hypoimmunogenic hESCs using CRISPR/Cas9 gene editing in which HLA-A, -B and -C as well as CIITA (a HLA class II molecule) were knocked out ( Figure 3B). These cells were named KO cells ( Figure 3B(b)). Subsequently, PD-L1 ( Figure 2B(f) is necessary for immune protection, and the expression of one of these genes is insufficient for immune protection. 33 Notably, knocking out only PD-L1 and CTLA4-Ig suppressed the T-cell response but did not suppress the NK cell killing response nor the macrophage phagocytosis response in this study. However, these results suggest that knocking in specific genes related to the inhibition of the immune response in hPSCs can also be used to generate universal or hypoimmunogenic hPSCs without inducing systemic immune suppression.
It is known that maternal immune tolerance to the foetus is

| CON CLUS I ON AND FUTURE PER S PEC TIVE S
Hypoimmunogenic hPSCs in which only a single HLA-C allele is expressed but HLA-A, HLA-B and CIITA are knocked out, prepared using CRISPR/Cas9 gene editing by Xu et al, 11

CO N FLI C T S O F I NTE R E S T
The authors declare no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data sharing is not applicable to this article as no new data were created or analysed in this study.