Article

ISSN 0006-2979, Biochemistry (Moscow), 2024, Vol. 89, No. 5, pp. 799-816 © Pleiades Publishing, Ltd., 2024.

799

REVIEW

Dysregulation of Immune Tolerance to Autologous iPSCs

and Their Differentiated Derivatives

Margarita E. Bogomiakova

1,2,a

*, Alexandra N. Bogomazova

1,2

and Maria A. Lagarkova

1,3

Federal Research and Clinical Center of Physical-Chemical Medicine, Federal Medical Biological Agency,

119435 Moscow, Russia

Center for Precision Genome Editing and Genetic Technologies for Biomedicine,

Federal Research and Clinical Center of Physical-Chemical Medicine, Federal Medical Biological Agency,

119435 Moscow, Russia

Lomonosov Moscow State University, 119991 Moscow, Russia

e-mail: margbog@rcpcm.ru

Received October 12, 2023

Revised December 21, 2023

Accepted February 13, 2024

Abstract— Induced pluripotent stem cells (iPSCs), capable of differentiating into any cell type, are a promising tool

for solving the problem of donor organ shortage. In addition, reprogramming technology makes it possible to ob-

tain a personalized, i.e., patient-specific, cell product transplantation of which should not cause problems related

to histocompatibility of the transplanted tissues and organs. At the same time, inconsistent information about the

main advantage of autologous iPSC-derivatives– lack of immunogenicity– still casts doubt on the possibility of

using such cells beyond immunosuppressive therapy protocols. This review is devoted to immunogenic properties

of the syngeneic and autologous iPSCs and their derivatives, as well as to the reasons for dysregulation of their

immune tolerance.

DOI: 10.1134/S0006297924050031

Keywords: induced pluripotent stem cells, immune response, immunogenicity, immunotolerance, T-cells, NK-cells,

differentiation

Abbreviations: B2M,beta-2-microglobulin; CD,clusters of differentiation; ESCs, embryonic stem cells; HLA, human leuko-

cyte antigen; iPSCs,induced pluripotent stem cells; KIR,killer-cell immunoglobulin-like receptor; NK,natural killer cells;

PSCs,pluripotent stem cells; RPE,retinal pigment epithelium; SMCs,smooth muscle cells.

* To whom correspondence should be addressed.

INTRODUCTION

Human pluripotent stem cells (PSCs), which in-

clude embryonic stem cells (ESCs) and induced plu-

ripotent stem cells (iPSCs), can unlimitedly proliferate,

and differentiate into almost any type of somatic cells

[1, 2]. These unique properties make them an attrac-

tive and promising tool for modeling various diseases

and drug development [3,4]. Hopes are pinned on dif-

ferentiated derivatives of PSCs as a source of materi-

al for cell therapy, which should solve the problem of

shortage of donor organs and tissues [5].

It is still too early to discuss widespread introduc-

tion of the PSC technology into clinical practice. Only

26  years have passed since the discovery of human

ESCs in 1998 [1], and active work is currently ongoing

to improve protocols for differentiating PSCs into spe-

cialized cells and obtaining three-dimensional struc-

tured tissues in vitro [4]. Another stumbling block to

integrating PSCs into the clinic is high cost of the tech-

nology. According to the recent estimate, generating

a clinical grade iPSCs line under good manufacturing

practice (GMP) costs approximately U$800,000 [4]. An-

other limiting factor is long time required to obtain a

new iPSC line and its subsequent differentiation into

the desired cell type [6]. Additionally, there are cur-

rently no developed standardization parameters that

would be applied to both iPSCs [7] and their differ-

BOGOMIAKOVA et al.800

BIOCHEMISTRY (Moscow) Vol. 89 No. 5 2024

entiated derivatives [8]. Thus, at least in the coming

years, patient-specific iPSC-based therapy is unlikely

to become widespread. Interestingly, almost half of the

clinical trials of PSC-derivatives used cell products de-

rived from only five ESC lines; however, the number

of studies of cells derived from iPSCs has increased

significantly over the past few years [9]. According to

the Federal Law No. 180, using ESCs to develop, pro-

duce, and apply biomedical cell products is prohibited

in Russia. Therefore, only iPSC-derivatives can be used

in clinical practice.

Despite the obvious economic advantages asso-

ciated with production and “scaling-up” of allogeneic

PSC-derivatives, the issue of immune rejection remains

unresolved due to the high polymorphism of the genes

of the major histocompatibility complex (HLA). To pre-

vent immune rejection during allogeneic tissue and

organ transplantation, patients must undergo lifelong

immunosuppressive treatment and its associated side

effects [10]. Initially, it was believed that personalized

therapy based on autologous iPSC-derivatives could cir-

cumvent the problem of histocompatibility [11]. How-

ever, some researchers report that immune response

against the syngeneic and autologous iPSCs is still

possible, which casts doubt on the main advantage

of autologous iPSCs– lack of immunogenicity [12-15].

Thereasons for this phenomenon have yet to be thor-

oughly studied. In this review, we tried to shed light

on the mechanisms of impaired immune tolerance

to autologous iPSCs. It is worth emphasizing that un-

derstanding effects of the significant immune effector

cells– T and NK-cells– on various types of cells and,

in the long-term perspective, tissues derived from

iPSCs will help to find approaches to their suppres-

sion and will be of great importance for successful de-

velopment of translational medicine.

IMMUNOGENICITY OF SYNGENEIC

AND AUTOLOGOUS iPSCs

AND THEIR DERIVATIVES TOWARD T-CELLS

The possibility that cells differentiated from au-

tologous iPSCs can provoke an immune response was

widely considered only after the publication of Zhao

etal. [12]. In this study, the authors showed that sub-

cutaneous administration of iPSCs to syngeneic, i.e.,

linear, mice led to formation of teratomas, where T-cell

infiltration zones were found. Moreover, the process

was accompanied by subsequent necrosis and regres-

sion of the resulting teratomas. At the same time, ter-

atomas formed after administration of the syngeneic

ESCs with the same genetic background caused an

immune response much less often. The authors not-

ed that the teratomas from iPSCs were rarely rejected

when episomal reprogramming was used. However,

in this case, the formed teratomas were rejected with

the frequency of 10-20%, and most of them were also

infiltrated by T-cells. Nevertheless, these results were

met with some skepticism by the scientific community,

mainly because undifferentiated iPSCs are not consid-

ered as a source of cells for clinical use [11].

More recent studies have been somewhat contra-

dictory, although most still indicated lack of immuno-

genicity of the syngeneic iPSCs derivatives. The oppo-

site findings were reported only in a few studies. Thus,

additional evidence of immunogenicity of the cells

derived from syngeneic iPSCs was presented in 2013

by Araki etal. [16]. The authors reported similar fre-

quency of teratoma rejection formed by both syngene-

ic iPSCs and ESCs. It was suggested that the immune

response to teratomas is potentially related to the ex-

pression of genes regulating pluripotency. In particu-

lar, the authors relied on the previously obtained data

that the transcription factor OCT4 may have immuno-

genic properties [17]. In addition, Araki et al., for the

first time, found signs of immune response to the ter-

minally differentiated derivatives of syngeneic iPSCs.

Transplantation of cardiomyocytes differentiated from

the iPSCs led to the significant T-cell infiltration of the

graft in the syngeneic mice [16].

Another study reported complete survival of ter-

atomas formed by the syngeneic ESCs, although some

teratomas still showed areas of T-cell infiltration [18].

Furthermore, the authors compared immunogenicity

of endothelial cells, hepatocytes, and neuron precur-

sors differentiated from the syngeneic ESCs and iPSCs.

Differentiated iPSCs derivatives did not induce signs

of specific T-cell response either in the in  vitro model

or after transplantation into the syngeneic mice. Thus,

Guha etal. [18] showed that the degree of immunoge-

nicity of iPSCs can decrease in the process of differen-

tiation. The authors of another work [19] came to the

same conclusion. Analysis of the functional state of im-

mune cells found in the transplantation area showed

that teratomas were infiltrated predominantly by the

cytotoxic T-cells, while endothelial cells differentiat-

ed from iPSCs – by the regulatory T-cells and macro-

phages [19]. Another study reported complete survival

of teratomas formed by the syngeneic ESCs, although

some teratomas still showed areas of T-cell infiltration

[18]. Moreover, the authors compared immunogenicity

of endothelial cells, hepatocytes, and neuron precur-

sors differentiated from the syngeneic ESCs and iPSCs.

Differentiated iPSCs derivatives did not induce signs of

specific T-cell response either in the invitro model or

after transplantation into syngeneic mice. Thus, Guha

et al. [18] showed that the degree of immunogenicity

of iPSCs can decrease in the differentiation process.

The authors of another work [19] came to the same

conclusion. Analysis of the functional state of immune

cells found in the transplantation area showed that

DYSREGULATION OF IMMUNE TOLERANCE TO AUTOLOGOUS iPSCs 801

BIOCHEMISTRY (Moscow) Vol. 89 No. 5 2024

teratomas were infiltrated predominantly by the cyto-

toxic T-cells, but endothelial cells differentiated from

iPSCs– by the regulatory T-cells and macrophages [19].

Another study with the non-human primate mod-

el was published in 2013 [20]. The authors compared

immune response to autologous and allogeneic trans-

plantation of the iPSC-differentiated midbrain dopa-

minergic neurons into Macaca fascicularis brains [20].

It was found that the significant amounts of microglia

and T-cells infiltrated the allografts, while autologous

neurons elicited minimal immune cell response. Simi-

lar work was done with the neural precursors of iPSCs,

where no significant infiltration was observed in the

autologous cell transplantation areas [21,22]. Interest-

ingly, Morizane etal. detected a limited T-cell response

to the autologous dopaminergic neurons in some ex-

perimental groups, if they were differentiated from

the iPSCs derived by retroviral transfection [20]. How-

ever, when the iPSCs were obtained by the episomal

reprogramming system, their derivatives did not lead

to the immune response of autologous T-cells. These

data indicate that viral integration of pluripotency fac-

tors during the iPSCs production may affect immuno-

genicity of the cell products.

Some limitations of the work of Morizane et al.

[20] are worth noting. First, the authors were able to

analyze immune response only after euthanasia of the

animals, approximately 3-4 months after transplan-

tation. Attempts were made to track dynamics of the

immune response using positron emission tomography

and measuring cytokine content in blood and cerebro-

spinal fluid. However, the results were highly variable

and correlated poorly with the postmortem histologic

data. Additional time points could more accurately de-

termine likelihood of the immune response to trans-

plantation of autologous iPSC-derivatives. Moreover,

although the authors demonstrated that the degree

of immune response was higher when the allogeneic

dopaminergic neurons were transplanted, rejection

was not observed even without immunosuppressive

therapy. This phenomenon could be explained by the

brain being an immune-privileged organ. Moreover,

these data are consistent with the clinical observations

of long-term survival of the allogeneic dopaminergic

neurons derived from the fetal material in the patients

with Parkinson’s disease who received only short-term

or no immunosuppression [23,24].

All previous studies have described immunogenic-

ity of the animal iPSCs– mice and primates. However,

studying immunogenicity of the autologous human

iPSC-derivatives is essential for clinical application.

In 2015, Zhao etal. studied this issue in a humanized

mouse model with the reconstructed human immune

system [13]. They found T-cell infiltration and tissue

necrosis areas in the most teratomas formed from

iPSCs. However, the degree of immune response to

the autologous iPSCs was weaker than to the allogene-

ic ESCs. Therefore, the authors hypothesized that only

certain derivatives of iPSCs could induce rejection.

In addition, deep sequencing of the T-cell receptor

(TCR) repertoire of the infiltrating lymphocytes re-

vealed their oligoclonal character, pointing to the an-

tigen-specific response of T-lymphocytes to the autol-

ogous iPSCs. For allogeneic ESCs, the polyclonal TCR

repertoire was revealed.

Furthermore, histological sections of the terato-

mas infiltrated with T-cells were analyzed to identify

potentially immunogenic tissues. The authors found

that the desmin-positive smooth muscle cells (SMCs)

were significantly more frequently surrounded by

the infiltrating T-cells. In contrast, the retinal pigment

epithelial cells (RPE) were almost never infiltrated by

T-cells. Next, the authors compared immunogenicity of

the two cell types, SMCs and RPE. It turned out that the

autologous SMCs were more immunogenic due to dys-

regulated expression of the tumor-associated genes,

particularly HORMAD1 and ZG16. Ectopic expression

of ZG16 in the RPE cells resulted in the significant

T-cell response in autologous recipients.

Thus, the data on immunogenicity of syngeneic

and autologous iPSCs derivatives against T-cells are

contradictory. Nevertheless, most of them are encour-

aging, such as, for example, relatively recent works

performed with the pigs [25], monkeys [26], humanized

mice [27] models as well as invitro studies [15,28], where

immune tolerance to the cellular products derived

from iPSCs was demonstrated.

IMMUNOGENICITY OF SYNGENEIC

AND AUTOLOGOUS iPSCs

AND THEIR DERIVATIVES TOWARD NK-CELLS

While primary function of T-lymphocytes is to

recognize foreign molecules, including neoepitopes,

NK-cells have a different principle of immunological

recognition. The classical “missing self” hypothesis

assumes that NK-cells recognize and destroy all cells

lacking HLA class I molecules [29]. Based on the cur-

rent viewpoint, activation of NK-cells is a more com-

plex concept and is determined by interaction of the

signals from two types of receptors on their surface:

activating and inhibitory [30]. Predominance of the

inhibitory signals upon interaction with the target cell

does not disturb the anergy of NK-cells, while predom-

inance of the activating signals triggers their cytotoxic

program. In turn, predominance of the activating sig-

nals can be caused by increase in the amount (level) of

the ligands for activating NK-cell receptors on the tar-

get cells and decrease in the inhibitory ligands, mainly

HLA class I molecules. Such imbalance in physiologi-

cal conditions is determined by various pathological

BOGOMIAKOVA et al.802

BIOCHEMISTRY (Moscow) Vol. 89 No. 5 2024

processes, including oncogenesis, viral and bacterial

infections, and stress [31].

Activity of NK-cells against undifferentiated PSCs,

including syngeneic and autologous ones, was previ-

ously highlighted in several in  vitro studies [32,  33].

Generally, low expression of HLA class I molecules is

a characteristic feature of PSCs [34], so high activity of

NK-cells is primarily due to the absence of inhibitory

signals. However, some studies indirectly noted contri-

bution of the activating ligands. For example, Frenzel

et  al. found that the preliminary blocking of activat-

ing receptor NKG2D significantly reduces cytotoxicity

of the NK-cells co-cultured with the syngeneic mouse

ESCs [35]. Another study showed that the NK-cells with

knockout of the Klrk1

–/–

gene encoding the NKG2D re-

ceptor lysed a significantly lower percentage of the

ESCs than the wild-type NK-cells [36]. Interestingly,

experiments with blocking antibodies have shown a

role for another activating receptor, DNAM-1, for hu-

man PSCs [33]. High sensitivity of PSCs to NK-cells is

due to two factors simultaneously: low expression of

HLA-I molecules and increased expression of activat-

ing ligands.

As for the in  vivo immune response, NK-cells are

known to limit teratoma formation after subcutaneous

injection of both syngeneic [37] and autologous iPSCs

[38]. These findings suggest that the residual PSCs,

which could potentially remain in the graft among

the differentiated cells, would fail to form teratomas

and would be rejected by the NK-cells. Melendez etal.

reported that NK-cells can act as an internal barrier

during reprogramming in  vitro and in  vivo [39]. It was

demonstrated that NK-cells can recognize and destroy

the partially reprogrammed cells shown to express the

ligands for activating NKG2D and DNAM-1 receptors.

Further, using the transgenic mouse line [40] express-

ing four reprogramming factors from the Yamanaka

cocktail (OSKM) under doxycycline treatment, the au-

thors showed that partial reprogramming in  vivo oc-

curs more efficiently when NK-cells are depleted and,

on the contrary, is significantly reduced when they are

adoptively transferred [39].

Response of the syngeneic and autologous NK-cells

to differentiated iPSC-derivatives needs to be better

understood. For example, increased sensitivity of the

hepatocyte-like cells differentiated from the murine

iPSCs (iPS-HLC) to syngeneic NK-cells in  vitro has been

reported [41]. At the same time, syngeneic somatic

cells– hepatocytes– practically did not induce the NK-

cell response. Interestingly, this work also determined

immune response of NK-cells to the hepatocyte-like

cells derived from ESCs (ES-HCs). The authors found

that the lysed ES-HCs percentage was almost twice

as high as that of the lysed iPS-HCs. ES-HCs, but not

iPS-HCs, appeared to have ligands for activating the

NKG2D receptor. In addition, knockout of the NKG2D

receptors in the NK-cells significantly reduced percent-

age of the lysed ES-HSCs but not iPS-HSCs. Thus, this

work confirmed previous findings that elimination of

the murine PSCs and their differentiated derivatives is

mainly due to interaction of the NKG2D receptor with

its ligands [35, 36, 42].

In another work, the NK-cell response to trans-

plantation of cardiomyocytes differentiated from the

syngeneic iPSCs (miPSC-CMs) was studied in a mouse

model [14]. It was shown that survival rate of the sub-

cutaneously transplanted miPSC-CMs was significant-

ly higher in the NK-cell-depleted mice. In the control

mice, in addition to infiltration of the graft by NK-cells,

there were signs of the NK-cell degranulation and re-

jection of the miPSC-CMs. Analysis of the NK-cell li-

gand expression showed that the miPSC-CMs weakly

expressed the MHC class I molecules and were stained

with antibodies to the ligands for NKG2D and DNAM-1

receptors. Blocking of the NKG2D and DNAM-1 recep-

tors or increasing the MHC-I expression by the IFNγ

(interferon gamma) pretreatment mitigated cytotoxic

properties of the NK-cells in  vitro. Also, it decreased

the NK-cell infiltration into the transplanted areas and

necrosis of miPSC-CMs in  vivo.

At the same time, immunogenicity of the human

iPSC-derivatives to NK-cells is poorly investigated.

It was studied mainly upon engineering of the im-

mune-evasive or “universal” PSCs. This approach is

believed to be an alternative to the traditional immu-

nosuppressive therapy, since derivatives of such cells

will be suitable for any recipient [43, 44]. The most

commonly used strategy to create hypoimmunogenic

cells is to completely “turn off” expression of HLA mole-

cules, both classes  I and II. To suppress the HLA class  I

expression, the beta-2-microglobulin (B2M) gene, which

encodes the light subunit required for stable heterodi-

mer formation, is usually knocked out [45-47]. To sup-

press the HLA class  II expression, the CIITA gene, tran-

scription factor required for the HLA-II expression, is

usually knocked out [48-50]. Cells devoid of HLA mol-

ecules should become completely invisible to the re-

cipient’s T-cells, both CD8

and CD4

[51]. On the other

hand, elimination of the HLA class  I molecules makes

the PSC-derivatives sensitive to cytotoxic properties

of NK-cells [46, 52, 53]. Therefore, obtaining PSC lines

with the reduced immunogenicity usually involves

two steps: first, HLA expression should be suppressed,

and then additional factors should be added to avoid

NK-cell response [53-57].

A  priori studies of low-immunogenic iPSC-de-

rivatives were performed with an allogeneic model.

Since NK-cells can not recognize foreign molecules,

immune tolerance to the iPSC-derivatives can be eval-

uated using NK-cells of allogeneic origin but with one

stipulation. It is known that the NK-cells alloreactiv-

ity can theoretically be caused by mismatch of the

DYSREGULATION OF IMMUNE TOLERANCE TO AUTOLOGOUS iPSCs 803

BIOCHEMISTRY (Moscow) Vol. 89 No. 5 2024

Fig. 1. Mechanism of NK-cell alloreactivity exemplified by the ligand mismatch for KIR2DL receptors. HLA-C1/C1 NK-cells are

activated by interaction with HLA-C2/C2 target cells (no inhibitory signal via KIR2DL3-receptor). HLA-C2/C2 NK-cells are acti-

vated by interaction with HLA-C1/C1 target cells (no inhibitory signaling through KIR2DL1 receptor). HLA-C1/C2 NK-cells are

activated by interaction with HLA-C1/C1 target cells (no inhibitory signaling through KIR2DL1-receptor) and by interaction with

HLA-C2/C2 target cells (no inhibitory signaling through KIR2DL3-receptor).

KIR family (killer-cell immunoglobulin-like receptor)

ligands via the mechanism of “missing-self” recogni-

tion. All HLA-C alleles are divided into two groups –

HLA-C1 and HLA-C2 – depending on the sequence of

amino acids at positions 77 and 80 of the alpha chain,

which determines their ability to bind to the NK-cell

receptors KIR2DL3 and KIR2DL1, respectively [58].

According to this principle, all donors and recipients

can be categorized into the following groups: HLA-C1/

C1, HLA-C1/C2, and HLA-C2/C2. NK-cells, in the process

of licensing or “learning” during maturation, acquire

tolerance to the specific set of HLA-C alleles on their

own cells. The conditions of NK-cell response to HLA-C

allele mismatch by interaction with the KIR2DL recep-

tors are presented in Fig.  1. Thus, if derivatives dif-

ferentiated from the HLA-homozygous iPSC lines are

used as target cells, heterozygous HLA-C1/C2 NK-cells

will respond to the absence of any of the KIR-ligands.

It was confirmed in the work by Ichise et al. [59].

The authors showed [59] that the HLA-C1/C2 NK-cells

isolated from the blood of healthy donors lysed T-cells

and endothelial cells differentiated from the HLA-C1/

C1 iPSCs. In turn, ectopic expression of HLA-C2 in the

differentiated HLA-C1/C1 derivatives diminished the

NK-cell response. In addition to alloreactivity to the

mismatched ligands of the KIR2DL receptors, the NK-

cell responses to the absence of ligands for KIR3DL1

(epitope Bw4) and KIR3DL2 (HLA-A3, A11 alleles) re-

ceptors have also been reported [60]. Thus, the alloge-

neic model can be used to assess immunogenicity of

the iPSCs derivatives toward NK-cells; however, only

under condition that the donors participating in the

study are typed.

Interestingly, in some studies with hypoimmuno-

genic PSC-derivatives, no significant difference was

found in the response of NK-cells to PSC knockout

derivatives and wild-type PSC-derivatives. However,

it should be noted that in all these studies, the target

cells and the donors involved were not typed. Dif-

ferences in cytotoxicity and number of CD107a

, i.e.,

degranulated NK-cells cocultured with the wild-type

SMCs or SMCs not expressing HLA classI, were statisti-

cally insignificant [55]. In turn, the wild-type RPE cells

induced extremely high NK-cell cytotoxicity, compa-

rable to the knockout PSC-derivatives [50]. Moreover,

percentage of the degranulated NK-cells also did not

differ between the wild-type and knockout RPE cells,

although it varied greatly depending on the donor.

High cytotoxicity of NK-cells was also observed against

the cardiomyocytes differentiated from ESCs [57].

BOGOMIAKOVA et al.804

BIOCHEMISTRY (Moscow) Vol. 89 No. 5 2024

Endothelial cells induced the same level of NK-cell

degranulation regardless of the HLA class I expres-

sion, although NK-cell cytotoxicity was higher against

the ESC-derivatives with the B2M gene knockout [61].

Itshould be noted that in the studies mentioned above,

the authors did not focus on the NK-cell response to

the wild-type PSC-derivatives. Instead, they described

absence of hypersensitivity of the HLA-negative cells

to the NK-cells.

In our recent work, we also found that the fibro-

blast-like iPSC-derivatives with the B2M gene knock-

out (ΔiPS-fibro) show the same degree of sensitivity

to allogeneic and autologous NK-cells as the wild-type

fibroblast-like cells – iPS-fibro [15]. Unlike other au-

thors, we used parental fibroblasts as a negative con-

trol for the NK-cell reaction. This comparison enabled

us to detect absence of complete immunologic tol-

erance to the differentiated iPSC-derivatives from

the autologous NK-cells. Transcriptome analysis re-

vealed a significant imbalance of the NK-cell ligands

in the iPS- fibro. Compared to the parental fibroblasts,

iPS- fibro simultaneously showed significant decrease

in the expression of HLA-I molecules and increase in

the expression of ligands for activating DNAM-1 and

NKG2D receptors. Further in this work, it was shown

that the NK-cell ligands in the differentiated iPSC-de-

rivatives can be balanced by pretreatment of the cell

cultures with IFNγ [15].

Another study noted sensitivity of the renal iPSC-

derivatives to the autologous NK-cells [28]. The au-

thors found that percentage of the activated NK-cells

cocultured with the proximal epithelial cell precursors

was lower than with the more “mature” iPSC-deriva-

tives. However, these differences were not discussed in

detail in the article. According to the RNA-sequencing

data, the level of HLA class  I transcripts increased in

the proximal epithelial cells during prolonged cultur-

ing. These data are in agreement with the results ob-

tained with iPS-fibro [15] as well as with some activat-

ing NK-cell ligands, particularly MICA and NECTIN2.

Itexplains predominance of the activating signals and

initiation of cytotoxic program in the NK-cells. It is also

interesting to note that, in contrast to the earlier work

[33], Rossbach etal. [28] did not observe high activity

of NK-cells against the undifferentiated human iPSCs.

It should be noted that the role of NK-cells in the

solid organ transplantation remains quite controver-

sial [60,  62]. There is evidence that some subsets of

NK-cells may play a role in regulation of allograft tol-

erance, and that NK-cells are, nevertheless, involved

in the T-cell-mediated and antibody-mediated allograft

rejection [63]. Without immunosuppressive therapy,

which affects cytotoxic activity and adjusts degranu-

lation properties, the activated NK-cells produce IFNγ

that could contribute to the development of chronic in-

flammation and enhance the predominantly T-cell-me-

diated immune response [64]. Thus, the iPSC-based cell

therapies should also consider immunogenicity of the

iPSC-derivatives toward NK-cells.

POSSIBLE REASONS OF THE IMMUNE RESPONSE

TO AUTOLOGOUS IPSC DERIVATIVES

It is still unclear what the crucial factor for some-

times-observed immunogenicity of the autologous iPSC-

derivatives is. Generally, the T-cell response can be

explained by formation of neoantigens and aberrant

gene expression (Fig.2a), and the NK-cell response can

be explained by imbalance between the activating and

inhibitory ligands in the target cells (Fig.2b).

It was initially assumed that using different repro-

gramming vectors could be the reason for iPSC immu-

nogenicity. Even in the first work, it was shown that

the teratomas formed by retroviral iPSCs were more

often rejected in the syngeneic hosts [12]. In turn, us-

ing episomes as a reprogramming vector significantly

reduced percentage of the rejected teratomas. Similar

results were obtained during transplantation of dopa-

minergic neurons into the primate brain [20]. It is well-

known that the retroviral and lentiviral constructs are

predominantly integrated into the transcriptionally ac-

tive sites that could cause mutations, genome instabili-

ty, and chromosomal aberrations. In addition, there is

evidence that the subsequent activation of transgenes

correlates with the aberrant production of the immu-

nogenic protein OCT4 [17].

It is assumed that immunogenicity of the “inte-

gration-free” iPSCs should be lower than of the iPSCs

obtained by retro- and lentiviral transfection. However,

to the best of our knowledge, there are no detailed

studies on this topic. The data comparing genomic in-

stability in the iPSC lines obtained by various repro-

gramming methods are contradictory. Thus, some re-

searchers report a similar number of point mutations

[65], as well as copy number variations (CNVs) [66].

On the contrary, others showed 2-fold lower number

of mutations in the integration-free” iPSCs [67, 68],

which means lower probability of neoepitope forma-

tion. The frequency of genetic variations was also low

in the human iPSCs obtained with episomal constructs

[69]. In the recent years, other methods of reprogram-

ming without integration have been suggested, partic-

ularly involving endogenous pluripotent genes using

the CRISPR/Cas9 system [70]. However, as far as we

know, no additional information about their genomic

instability has been provided. In any case, non-integra-

tive reprogramming methods are currently the safest

and most effective for further clinical use [4].

Commonly, differences in immunogenicity are ex-

plained by mutations and, consequently, by formation

of neoepitopes [71]. First, these may be mutations that

DYSREGULATION OF IMMUNE TOLERANCE TO AUTOLOGOUS iPSCs 805

BIOCHEMISTRY (Moscow) Vol. 89 No. 5 2024

Fig. 2. Immunogenicity of autologous iPSCs and their derivatives. a)Immune response of T-cells to autologous iPSCs and their

differentiated derivatives could be due to recognition of immunogenic neoepitopes formed as a result of mutations or aberrant

expression of immunogenic genes. b) Immune response of NK-cells to autologous iPSCs and their differentiated derivatives

could be due to imbalance of the NK-cell ligands in the target cells. Predominance of activating signals triggers cytotoxic pro-

gram in NK-cells.

existed in the parental somatic cells. Thus, fibroblasts

are one of the most frequently used sources of cells

for reprogramming [72]. It has been reported that mu-

tations acquired due to UV-induced mutagenesis are

present in ~50% of human iPSCs reprogrammed from

the skin fibroblasts [73]. Such mutations are charac-

terized by the C-to-T or CC-to-TT substitutions and are

often observed in melanomas [74]. Other somatic cells

can act as an alternative to fibroblasts. For example, it

has been reported that the iPSCs obtained from hema-

topoietic stem cells contain significantly fewer point

mutations, insertions, and deletions than the iPSCs

obtained from the skin fibroblasts  [75]. Peripheral

blood cells are also very often used as a source of cells

for reprogramming. Rouhani et al. revealed that the

blood-derived iPSCs contained fewer mutations than

the fibroblast-derived iPSCs  [76]. At the same time,

there is evidence that mutations in the blood cells

also accumulate with age [77]. Thus, the work using 16

iPSC lines obtained from the blood cells of donors aged

21-100 years demonstrated that the frequency of muta-

tions in the iPSC increases linearly with the donor age

[78]. In addition, frequency of the mitochondrial DNA

mutations in human iPSCs also increases with the do-

nor age, and this can lead to metabolic defects in iPSCs

[79]. Although denovo mutations in the mitochondrial

DNA are usually rare for iPSCs [80, 81], they can lead

to formation of immunogenic neoepitopes and pro-

voke immune response even during the autologous

transplantation, as was shown earlier [82].

Thus, somatic cells of the young donors may have

a comparative advantage for deriving iPSCs. This is

also confirmed by the results of recent work, where

iPSCs were obtained from the umbilical cord blood

erythroblasts and did not contain mutations in the

protein-coding regions [83]. In addition, age of the do-

nors could affect cultural properties of iPSCs. Thus,

it was shown that the iPSCs obtained from the older

mice proliferated not as good as the iPSCs obtained

from the young mice [84].

Mutations in iPSCs could also occur during re-

programming. Such mutations are usually similar to

the mutations caused by oxidative stress (C-to-A sub-

stitutions) and are predominantly found in the lami-

na-associated domains – condensed regions of heter-

ochromatin positioned at the nuclear periphery [85].

BOGOMIAKOVA et al.806

BIOCHEMISTRY (Moscow) Vol. 89 No. 5 2024

Earlier studies showed that up to 75% of point muta-

tions in iPSCs occur during reprogramming [86, 87].

Presence of various mutations in the isogenic iPSC

clones and lower frequency of point mutations in the

isogenic ESC presumably indicates that such mutations

were not obtained from the parental somatic cells [67].

Additional information on the mutations that occur

during reprogramming was provided by Rouhani etal.

[88]. The authors identified unique mutations in the

isogenic iPSC lines derived from the monoclonal-origin

endothelial precursors.

It is believed that mutations occur at the earliest

stages of reprogramming before the first cell division

after the introduction of reprogramming factors or

immediately after the first or second division [67,  83].

The authors of the latest work also found that tempo-

rary deficit of the control components of the G1/S cell

cycle at the initial stage of the reprogramming process

leads to accumulation of mutations [83]. At the same

time, the data on mutations that occur during repro-

gramming are pretty ambiguous. Recent studies have

shown that up to 90% of the various SNPs and indels in

iPSCs originate from somatic cells [73,  89]. Moreover,

Kosanke et al. showed that only 2% of the mutations

detected in iPSCs were not detected in the parental en-

dothelial cells used for reprogramming [90].

The third reason for mutations in iPSCs is long-

term cultivation. Such mutations are formed stochasti-

cally and are much less common than the pre-existing

somatic mutations or reprogramming-induced muta-

tions. Mutations caused by prolonged cultivation are

believed to provide proliferative benefits [91, 92]. For

example, one iPSCs line carried four additional point

mutations in the late passages compared with the cells

of early passages [86]. Nevertheless, according to the

recent data, frequency and spectrum of the mutations

induced by long-term cultivation of iPSCs do not differ

from the mutations occurring at the pre-gastrulation

stage of embryogenesis [93]. Another study showed

that the rate of accumulated mutations in the long-

term cultured iPSCs is lower than in the intestinal and

liver stem cells [94]. The authors found that more than

a third of mutations were caused by the C-to-A substi-

tutions associated with oxidative stress. Cell cultiva-

tion under hypoxic conditions (3% oxygen) for three

months significantly reduced the number of single

nucleotide substitutions. Similar results were obtained

for the ESCs: the authors observed a more than two-

fold decrease in the frequency of mutations under hy-

poxia [95]. The obtained data can be used to optimize

conditions of IPSC cultivation.

For clinical use, it is crucial to understand how

mutations can affect the iPSC phenotype, including

whether they can trigger carcinogenesis. There are

studies indicating that point mutations are predom-

inantly found in the cancer-associated genes [86].

Repeated mutations in the TP53 tumor suppressor

gene were detected both by the whole exome sequenc-

ing of ESC and by analyzing the publicly available RNA

sequencing data from 120 PSC lines [92]. In addition to

mutations in the TP53 gene, repeated mutations were

found in other tumor-associated genes, such as CDK12,

EGFR, and PATZ1 [96,  97]. A recent study revealed mu-

tations in the BCOR gene, often found in hematological

diseases, in more than 25% of the analyzed iPSC lines

[76]. In contrast, no association with the tumor-associ-

ated genes has been found in other studies [65, 69, 73,

94]. Point mutations were mostly found in the areas

of inactive chromatin, so they were unlikely to cause

undesirable effects. However, unique mutations in var-

ious isogenic iPSC clones were usually found in active

promoters and could alter gene expression [73]. This,

in turn, could lead to formation of immunogenic deter-

minants or affect effectiveness of differentiation into

the desired cell type [98], which is also essential for re-

generative medicine.

In addition to mutations induced by reprogram-

ming or cultivation, epigenetic changes that regulate

expression of various proteins should also be consid-

ered. First of all, it applies to disruption in DNA meth-

ylation in the PSC lines. Some studies noted that ab-

errant methylation pattern in iPSCs may be similar to

the tumor cells [99-101]. Moreover, it was shown that

methylation deregulation can persist in differentiated

cells [102]. However, it is worth noting that DNA meth-

ylation in the PSCs can be dynamic, respond to culture

conditions, and vary depending on the cell line [95].

It is also believed that some cells are not fully

undergoing the reprogramming process, and iPSCs

could to a large extent retain transcriptional and epi-

genetic memory of their origin [103]. Nevertheless, re-

sults of the studies on this topic are quite

contradictory,

but according to the modern concepts, molecular and

functional differences in different iPSC lines are lost

during prolonged cultivation [104]. At the same time,

it was shown in the recent study that reprogramming

through the stage of naive iPSCs (TNT-reprogramming)

completely erases epigenetic memory and corrects

epigenetic aberrations that have arisen [105]. Such

TNT-iPSCs turned out to be more similar to ESCs from

the molecular and functional point of view than the

iPSCs obtained by the standard method.

Epigenetic peculiarities could explain abnormal

expression of immunogenic proteins. Thus, at least two

studies have demonstrated that the “somatic memory”

phenomenon could influence immunogenicity of the

iPSCs [106,  107]. Mouse iPSCs obtained from the Sertoli

cells, anatomically related to the immune-privileged

regions, formed teratomas more efficiently than the

iPSCs obtained from the embryonic fibroblasts [106].

Moreover, differentiated derivatives of the syngene-

ic ESCs demonstrated a reduced in  vitro activation

DYSREGULATION OF IMMUNE TOLERANCE TO AUTOLOGOUS iPSCs 807

BIOCHEMISTRY (Moscow) Vol. 89 No. 5 2024

of allogeneic T-cells compared to the iPSCs obtained

from the embryonic fibroblasts. However, it is worth

noting that in the later passages, the authors did not

observe differences in immunogenicity of the iPSCs

obtained from different somatic cells. These results

confirm that “somatic memory” in the iPSCs may be

present only in early passages [106]. Another study

showed that the umbilical cord mesenchymal cells are

a less immunogenic source of cells for reprogramming

than the skin fibroblasts [107].

Impaired expression of the genes associated with

the NK-cell response is another reason for immunoge-

nicity of iPSCs and their derivatives [15]. Both increase

in the signals from activating receptors and decrease

in the signals from inhibitory receptors can trigger

cytotoxic program of NK-cells. In other words, proper

balance between the inhibitory and activating ligands

could make the target cell invisible to NK-cells [108].

In contrast, an impaired balance of the NK-cell ligands

in the iPSC-derivatives could cause excessive activa-

tion of NK-cells. Thus, intensity of the HLA class  I mol-

ecule expression and activating ligands and adhesion

molecules would influence the degree of immune re-

sponse. Previously, we showed that all these factors

were responsible for the increased NK-cell response

to iPSC-derivatives [15]. First, we observed a relatively

low gene expression of the HLA-I molecules, major in-

hibitory ligands in the fibroblast-like iPSC-derivatives

(iPS-fibro). Second, the genes coding for the main ac-

tivating NK-cell ligands were upregulated in the iPS-

fibro. Expression of the stress-induced molecule MICA

(NKG2D ligand) gene was more than 1.5 times higher

in the iPS-fibro than in their parental fibroblasts. The

DNAM-1 ligands, NECTIN2 (CD112) and PVR (CD155),

and the NKp30 ligand, NCR3LG1 (B7-H6), underwent

a more noticeable increase in the gene expression

with more than 3-fold-change in the iPS-fibro. Third,

the genes of some adhesion molecules were also over-

expressed in iPS-fibro. Interaction of the adhesion

molecules with their receptors on NK-cells facilitates

formation of tight junctions between the NK-cell and

the target cell and leads to assembly of immunologi-

cal synapses essential for the target cell killing [109].

The ICAM-1 (LFA-1 ligand) and VCAM-1 (VLA-4 or α4β1

integrin ligand) genes were upregulated in the iPSC-

derivatives. Hence, imbalance between the NK-cell

ligands in iPSC-derivatives was determined simultane-

ously by low intensity of the inhibitory signals and ele-

vated intensity of the activating signals [15].

Vulnerability to the action of NK-cells can be ex-

plained by insufficient maturity of the differentiated

iPSC-derivatives and low level of the HLA-I class mol-

ecules compared to the parental somatic cells. Thus,

increase in the HLA-I expression was shown during

prolonged cultivation or passaging, at least for the RPE

cells [50], proximal renal epithelium cells [28], and

iPS-fibro [15]. Another risk of immature phenotype is

expression of embryonic or fetal proteins, which are

also typical for some cancers (for example, alpha-fe-

toprotein) [110]. Despite the active development of

differentiation protocols, several cell types can be dif-

ferentiated in  vitro only to an immature phenotype, in

particular, cardiomyocytes [111], hepatocytes [112], or

beta-cells [113].

Increased expression of the activating NK-cell li-

gands is worth noting separately. Analysis of the pub-

licly available RNA-seq datasets [114-116] showed that

expression of the NECTIN2, PVR, CADM1, and CD70

genes was upregulated in the independently derived

fibroblast-like cells compared to the isogenic fibro-

blasts used for reprogramming [15]. Imperfect micro-

environment during in  vitro differentiation may affect

proper balance between the ligands for the NK-cell

receptors in this type of iPSC-derivatives. In addition,

high levels of the MICA and NECTIN2 gene expression

were observed in the proximal epithelial cells of the

kidney [28]. Since each cell type expresses its own set

of proteins, it will be necessary to determine expres-

sion pattern of the ligands of the NK-cell receptors for

clinical use. It is worth noting that the cells that belong

to immune-privileged tissues could have immunomod-

ulatory functions to suppress immune response. It was

shown that some types of the differentiated PSC-deriv-

atives, in particular RPE cells [117,  118], retinal gan-

glion cells [119], neuron precursors [120-122], neural

crest cells [123, 124], and chondrocytes [125] demon-

strate reduced immunogenicity even to allogeneic lym-

phocytes.

Different cultivation conditions could affect im-

munogenicity of iPSCs and their derivatives. As men-

tioned earlier, prolonged cultivation could lead to ac-

cumulation of mutations in the cells at later passages

[86,  94]. The cryo-pause method, i.e., storing iPSCs as

ready-to-use aliquots from one passage, can reduce

frequency of genomic aberrations caused by passaging

and prolonged cultivation of iPSCs [126]. Considering

that oxidative process during reprogramming and pro-

longed cultivation could lead to C-to-A substitutions

[85, 88, 94], antioxidants could reduce mutagenic load

in the iPSCs. In particular, antioxidants were report-

ed to reduce CNVs in the iPSCs [127]. A recent study

also indicated that introduction of antioxidant trans-

genes, such as superoxide dismutase  1 (SOD1) and 2

(SOD2), glutathione peroxidase  1 (GPX1), and N-acetyl-

cysteine (NAC), reduced the number of transversions

iniPSCs[83].

Selection of the reagents used for cultivation and

differentiation could affect immunogenic properties

of iPSCs and their derivatives. Using xenogeneic

materials for PSC cultivation could complicate fur-

ther clinical use of the PSC-derivatives. For example,

ESCs and embryoid bodies were shown to absorb