Induced pluripotent stem cell technology in bone biology

Abstract

Technologies on the development and differentiation of human induced pluripotent stem cells (hiPSCs) are rapidly improving, and have been applied to create cell types relevant to the bone field. Differentiation protocols to form bona fide bone-forming cells from iPSCs are available, and can be used to probe details of differentiation and function in depth. When applied to iPSCs bearing disease-causing mutations, the pathogenetic mechanisms of diseases of the skeleton can be elucidated, along with the development of novel therapeutics. These cells can also be used for development of cell therapies for cell and tissue replacement.

Introduction

Given the technological advances made in biomedical research since the turn of the 20^th century, these are exciting times. We now have at our fingertips some remarkable tools (e.g., cellular reprograming, gene editing, single-cell and spatial transcriptomics, to name just a few) to address not only the cause of human wounds, diseases and disorders, but also potential ways to treat them. But, as is so often the case, new technologies precede our abilities to interpret and understand the information that they provide to us, and also in our ability to use them effectively. Case in point is the discovery of tissue-specific stem and progenitors, and subsequently the concoction of induced pluripotent stem cells (iPSCs), both of which were thought to provide immediate relief to patients with all kinds of afflictions. While there have been few immediate miracles, the fields of tissue engineering and regenerative medicine using iPSCs remain promising, but will require a deeper understanding of stem cell biology in general, and of human iPSCs in particular, using validated assays and appropriate models to make them effective.

What follows below are brief presentations on the history of cell and stem cell biology that lead to development of iPSCs, how iPSCs are made, characterized and modified, with a focus on osteogenic differentiation of iPSCs, and creation of patient-specific iPSCs as models of human disease.

The origins of cell and stem cell biology

Using a microscope, the 1590 invention of two Dutch spectacle-makers, Hans and Zacharias Janssen, the basic structural units that give rise to tissues and organs of the body were first described by Robert Hooke in 1665 as “cellula” in cork (actually, the residual cell wall of plant cells) [1]. Later, in 1674, Anton van Leeuwenhoek first described live microscopic alga (which he termed “animalcules”), which he reported in letters to the Royal Society that were eventually published (reviewed in [2]). From these observations, what became known as the Cell Theory evolved, first proposed by Theodor Schwann and Mattias Jakob Schleiden in 1839 [3], and fully supported by Rudolph Virchow’s authoritative statement in 1855: “Omnis cellula e cellula“; i.e., all cells come from other cells ( published later in [4]). This assertion was in fact a reiteration of Robert Remak’s proposal, first presented in 1841, and then published in 1855, along with a description of three germ layers in developing chick embryos, all of which were derived from cell division [5]. In its current read, Modern Cell Theory states that: 1) all known living things are made up of cells, 2) the cell is the structural and functional unit of all living things, 3) all cells come from pre-existing cells by division (spontaneous generation does not occur), 4) cells contain hereditary information, which is passed from cell to cell during cell division, 5) all cells are basically the same in chemical composition, 6) all energy flow (metabolism and biochemistry) of life occurs within cells (reviewed in [6]). While these tenants seem obvious today, the theory arose during a time in which “spontaneous generation” was considered to be a viable possibility (https://bio.libretexts.org/).

The concept of a “stem” cell did not emerge until 1868, when Ernst Haeckel described a fertilized egg as a “Stammzelle”; that is, an extraordinary single cell that could give rise to all cell types of an organism. In 1886, William Sedgwick also described stem cells in the parts of plants that can regenerate. Additionally, the concept that tissues capable of continuous replacement during the lifetime of an organism such as blood, skin and intestine, rely on the existence of a “stem cell” within that tissue, was gaining traction based on observations around the end of the 19th century by classical histologists (e.g., Regaud working on spermatogenesis [7]) and hematologists (e.g., Weidenreich, Dantschakoff, Maximow, Ferrata, Pappenheim and others, reviewed in [8]). Years later, proof of the ability of HSCs to form all types of blood cells from studies by James Till and Ernest McCulloch in 1963 [9]. Evidence for the existence of post-natal stem cells in skin and intestine followed shortly thereafter (Howard Green, clonal keratinocytes with regenerative power [10]; Christopher Potten, the label-retaining cell in the intestinal crypt, [11]). Friedenstein and Owen provided evidence of a bone marrow stromal stem cell (reviewed in [12]), later termed a skeletal stem cell [13]. Findings since the 1990s have suggested that virtually all postnatal/adult tissues have some form of a tissue-specific stem/progenitor cell, capable of local regeneration, albeit with different potencies and rates of turnover, depending on the demands imposed upon them by their tissue of origin [14-17].

Embryonic stem cells

Perhaps the most significant advance made in cell biology at the end of the 20^th century was establishment of the ability to isolate stem cells from the blastocyst of a mouse embryo [18, 19]. In short order, this technological development was followed by the demonstration that these murine embryonic stem cells (mESCs) could support the formation of every cell in the body when introduced into developing mouse embryos [20, 21] (see Fig. 2B). These findings, along with advances in genetic engineering, revolutionized the field of mouse genetics, leading to the creation of a vast collection of transgenic and knockout mouse strains that are often (but not always) valuable models of human diseases and disorders. These advances also motivated the search for methods to collect equivalent stem cells from human blastocysts, which culminated in the isolation of human embryonic stem cells (hESCs) near the turn of the 20^th century [22, 23]. However, a debate immediately ensued on the ethicality of the use of human embryonic stem cells (as well as fetal tissues), which continues to this day. The use of hESCs in US Government-funded research is tightly regulated as outlined in the NIH Guidelines for Human Stem Cell Research (https://stemcells.nih.gov/research-policy/guidelines-for-human-stem-cell-research), although there is variation in the regulations from one country to another; e.g. in Europe [24].

Fig 2.

Development and stem cell hierarchy and utilization. A. Development starting from the fertilized egg (totipotent), going through the blastocyst stage (pluripotent), and then gastrulation, which establishes the three germ layers (multipotent). B. Pluripotent stem cells can be transplanted into mouse embryos (MOUSE ONLY), and will give rise to all cells of the zygote, embryo and fetus. This was a major advance in mouse genetics. Pluripotent stem cells can also differentiation into all cell types in vitro, and can form embryoid bodies, which can exhibit cells from all three germ layers. However, the gold standard to demonstrate pluripotency is by in vivo transplantation, which leads to formation of a teratoma that contains derivatives of all three germ layers.

Development and stem cell definitions

It is now recognized that there is a hierarchy of embryonic, fetal, and post-natal/adult stem cells. That said, there are two definitions that are universally accepted that apply to all stem cells: 1) a single stem cell has the ability to differentiate into functional cells (often with varying phenotypes) that make up a specific tissue (potency); they may be unipotent or multipotent; and 2) they have the ability to self-renew; i.e., they are able to maintain an appropriate number of stem cells within a given tissue through a variety of different types of cell division including asymmetric division [14] (Fig. 1A). However, it is essential to recognize that both properties must be demonstrated by rigorous assays. Potency is often inaccurately estimated by using assays that are prone to artifact, and those that do not demonstrate functionality; self-renewal is not necessarily demonstrated by extensive proliferation and is often more associated with immortalization. Identifying both properties requires the use of clonal populations of cells (i.e., those originating from a single cell). Differentiation assays must be tailored to faithfully recapitulate functional differentiation into a particular tissue, either in vitro or in vivo; e.g., the in vitro pellet culture for cartilage differentiation, and subcutaneous or orthotopic transplantation for bone formation in vivo, both of which are the current gold standards for these two tissues [25]. Self-renewal must be demonstrated by taking the progeny of a single cell (a clone), differentiating it into a tissue, followed by re-isolation of a single cell with the same characteristics as the original cell (e.g., cell surface markers), and subsequent re-differentiation as has been demonstrated for bone marrow-derived skeletal stem cells [25, 26] (Fig. 1B). Potency and self-renewal are the two basic tenets of stem cell biology.

Fig 1.

Stem cell self-renewal – a defining characteristic of a stem cell. A. There are several ways that a stem cell may divide. In one case, a stem cell divides asymmetrically, with one daughter remaining a stem cell, and the other becoming more specified, and may transiently amplify before becoming terminally differentiated. Or a stem cell may divide symmetrically, and in one case, both daughters are committed, which would lead to stem cell loss, and in the other case, both daugters are both stem cells, which may lead to immortalization. The actual process of stem cell maintenance is not yet well understood, and may rely on toggling between the different types of division. B. An assay to demonstrate self-renewal of a skeletal stem cell present in bone marrow. A single cell from bone marrow with a defined phenotype (CD45⁻/CD34⁻/CD146⁺) is isolated by FACS and allowed to form a colony. When the colony is transplanted in vivo, the progeny of that single cell gives rise to a bone/marrow organ (ossicle). A single cell with the same original phenotype (CD146⁺) is isolated from that ossicle, and is able to give rise to a secondary colony, that when transplanted is able to form a secondary ossicle with the same CD146⁺ phenotype (a pericyte) (modified from Sacchetti et al, Cell, 2007).

Stem cell hierarchy

It is now evident that there is a hierarchy of stem cells that evolve during development based on their potency. The fertilized egg (a zygote), often described as the “mother of all stem cells” by the developmental biologist Elizabeth D. Hay, is totipotent; able to form all cells in the body and the fetal components of the placenta. With division of the zygote, the resulting cells remain totipotent for at least two divisions (8 cell stage). With 1-2 more divisions, the morula is formed as a solid clump of cells (up to 64 cells) that goes on to cavitate to establish a blastocyst, which is composed of a hollow sphere of trophoblasts that will form the fetal components of the placenta, and the inner cell mass, composed of epiblasts that will form the fetus, and a layer of hypoblasts that will form the chorion of the placenta. The epiblasts at this point have lost the ability to form trophoblasts and are termed pluripotent, able to make all cell types of the body proper, but not the extraembryonic membranes. As development continues, migration and involution of epiblasts results in reorganization of the embryo (gastrulation), the pattern of which differs between mice and humans [27]. This involution gives rise to the three embryonic germ layers: ectoderm, mesoderm, and endoderm. Cells of the three germ layers become specified by various combinations and gradients of morphogens to form specific organ systems, and are no longer pluripotent, but are multipotent [28, 29] (Fig. 2A)

Naïve vs. primed pluripotent stem cells

As characterization studies progressed on mESCs and hESCs, it became clear that even though all are pluripotent, there are differences between them, dictated by developmental patterns. A deep dive into murine development determined that cells in the morula (created after ~6 divisions of the fertilized egg) must first be specified as extraembryonic cells that will mediate uterine implantation prior to the delineation of cells that will become epiblast cells that will give rise to the entire fetus [30, 31]. It was also determined that cells isolated from the murine pre-implantation inner cell mass could contribute to chimeric mouse blastocysts and are therefore “naive,” and at a basic ground state. On the other hand, post-implantation murine inner cell mass cells, termed epiblast stem cells, mEpiSCs, could not form chimeric blastocysts, and therefore are “primed,” yet they are still pluripotent [32]. The post-implantation inner cell mass was also found to have hypoblasts that contribute to the formation of the chorionic layer of the placenta. Furthermore, prior to the exit from pluripotency, both murine and human ESCs exhibit interconvertible subsets that display a lineage bias based on transcription factor analysis that could impact on subsequent differentiation into cell types from different germ layers [33]. It was also determined that hESCs are more akin to primed mEpiSCs than they are to mESCs [34]. However, it has now been demonstrated that naïve human epiblasts can be isolated de novo, or formed by chemical reset of primed cells (e.g., line HNES1), and can form extraembryonic cell types (so-called “extended pluripotent cells”), but mouse naïve cells cannot [35], another difference between mouse and human development. Taken altogether, not surprisingly, there is a great deal of heterogeneity in embryonic stem cells, depending on the animal species and the stage of embryonic development.

Developmental origins of bone

Of note, bone in different regions of the body is derived from components of two of the three different germ layers: ectoderm that forms neural crest in the proximal region, gives rise to frontal and craniofacial bones, paraxial mesoderm forms somitomeres and somites, which forms the dorsal portion of the skull and the axial skeleton, and lastly, lateral plate mesoderm forms somatic mesoderm that forms the appendicular bone [36-38]. In addition, it has also been suggested that the dorsal root of the aorta (derived from lateral plate somatic mesoderm) gives rise to bone by the budding of mesoangioblasts from the developing aorta into the underlying connective tissue [39, 40]. Thus, there are at least three different embryonic sources of bone, which will be further highlighted below (Fig. 3A).

Fig 3.

Development of bone. A. Following gastrulation and formation of the three different germ layers, the ectoderm gives rise to neural crest cells that migrate away from the neural tube, some of which form the frontal bones of the skull, and the facial bones. The sclerotome (the ventral part of somites made of paraxial mesoderm) forms the axial skeleton, and the lateral plate somatic mesoderm forms the appendicular skeleton. B. There are diseases that specifically affect bones from germ layer specific elements.

Induced pluripotent stem cells

Because of the ethical considerations surrounding the use of hESCs, there understandably was a search for populations of post-natal (and/or adult) cells that could substitute for hESCs. Subsequently, numerous publications flooded the literature claiming “plasticity” of post-natal/adult cells; that is, the ability to transdifferentiate into phenotypes outside of their tissue of origin, which would seem to fit the bill as pluripotent stem cells. Several such populations were initially isolated from bone marrow, such as MAPCs (reviewed in [41], MIAMI cells (reviewed in [42]), and from peripheral blood cells, such as VSELs (reviewed in [43]). These so-called pluripotent cells were identified based on their expression of a few cell surface markers or transcription factors associated with ESCs, and notably, on differentiation assays that often do not reliably predict the functionality of the so-called “transdifferentiated” cells. Many other populations were offered up as being “pluripotent”, including “mesenchymal stem cells.” However, it must be noted that many of these populations were not rigorously shown to be bona fide stem cells; i.e., potent, able to differentiate into functional cell types, or able to self-renew. In the years that followed, it was shown in many tissues, populations of post-natal/adult stem/progenitor populations do exist, but that they do not transdifferentiate outside of their lineage [44], without extreme ex vivo coercion by genetic engineering or chemical treatments [25, 45]. They may display some “flexibility” within their lineage (e.g., conversion of bone marrow stromal cells/skeletal stem cells from an osteogenic fate to an adipogenic fate [46]). In general, it is currently thought that endogenous (in vivo) transdifferentiation rarely occurs under normal circumstances, if at all.

Despite the disappointing performance of various post-natal/adult populations to perform in a pluripotent fashion, the hope of creating such a cell population was fueled by the notion that as long as a cell has an intact nucleus and is in an instructive environment, anything is possible. This thought emanated from the ground-breaking study of Sir John Gurdon, who successfully transferred a nucleus from an intestinal epithelial cell of a tadpole into an enucleated unfertilized frog egg (somatic cell nuclear transfer, SCNT), which resulted in the generation of a tadpole [47, 48]. In order to determine if the developmental stage of the donor nucleus controlled its ability to be successful in SCNT, Wilmut and coworkers introduced ovine nuclei from embryonic or fetal tissue, or adult mammary gland nuclei into enucleated sheep ova, and were successful in generating viable offspring from all three sources [49], one of which was Dolly. It was also determined that when chromosomes are removed from one of two blastomeres at the 2-cell stage of a developing murine zygote, the recipient cell reprograms the donor chromosomes, and the zygote goes on to generate a “mosaic cloned” animal [50] (Fig. 4A).

Fig 4.

Methods of reprogramming somatic cells. A. Somatic nuclear cell and chromosomal transfer. B. Use of viral vectors and other modes of introducing four transcription factors (OSKM) into somatic cells of varying types. Figure modified from Bischoff, SR, NovoHelix https://www.novohelix.com/cell%20line%20models/pluripotent-stem-cells-reprogramming).

From these studies, it was clear that the cytoplasm of the recipient ovum contained factors that were able to induce the reprogramming of the DNA in the donor nucleus back to a primordial state. The search was on to identify the active ingredients for induction of pluripotency in post-natal somatic cells. Reasoning that these factors would be transcription factors expressed in embryonic stem cells, but not in somatic cells, Takahashi and Yamanaka embarked on their identification, first in mESCs [51], and then in human ESCs [52, 53]. From a long list of different transcription factors, they identified four [Octamer-binding transcription factor-3/4 (Oct 3/4); Sex-determining region Y-box 2 (Sox2); Kruppel Like Factor-4 (KLF-4); and Myelocytomatosis viral oncogene homolog (c-Myc)], that bring somatic cells back to a pluripotent state, able to make all cell types of the three embryonic germ layers, based on the formation of teratomas upon in vivo transplantation of the reprogrammed cells. Initially it was thought that the epigenetic memory could not be completely erased from differentiated cells and that reprogramming would not be complete [54]; however, with increased technical proficiency, mounting evidence is that “memory” does not pose a substantial problem in thoroughly reprogramed differentiated somatic cells.

Generation of induced pluripotent stem cells

Cell Source

Virtually any nucleated cell type can be used as a target for reprogramming to form iPSCs, if they are grown under appropriate and optimal cell culture conditions (i.e., healthy cells will be more useful than ones that are struggling ex vivo). However, the ease in which they can be obtained, and the efficacy of reprogramming using optimized methods, are factors that impact on the selection of starting material. Most often skin fibroblasts are used, but skin biopsies are not without issues, and other cell types such as keratinocytes obtained from hair roots and CD34⁺ cells from peripheral blood are increasingly used. Efficiency of reprogramming is impacted by a variety of parameters, including the state of differentiation, and the levels of endogenous expression of the four genes (OKSM) (reviewed in [55]).

Methods of reprogramming

There are a number of ways in which the reprogramming factors can be introduced into the recipient cells (Fig. 4B), all of which have their pluses and minuses with respect to their ease of application and the efficiency in generating reprogrammed cells. Initially, Maloney murine leukemia virus (MMLV)-derived retroviral vectors for each transcription factor were used, which can efficiently transduce dividing cells. After integration into the genome, these constructs are usually spontaneously silenced as expression of the endogenous gene takes over, an essential step in the reprogramming scheme [56]. More recently, lentiviruses capable of infecting dividing and non-dividing cells at high levels have been devised to simplify the transduction process [57]. However, they are not as repressible as MMLV versions and may lead to over-production of the four factors and development of cells that form neoplastic lesions. Lastly, a single picornaviral 2A vector can accommodate all four factors within a single virus [58]. However, a major issue related to the use of integrating viruses is their effect on integrations sites, which may cause activation of proto-oncogenes or inactivation of genes that are a part of essential signaling pathways. A way of reducing the impact of random integration is to remove the transgenes following reprograming, and the Cre/LoxP system was initially used. However, after Cre-mediated excision of these constructs, a single LoxP site remains, which may continue to be disruptive. Consequently, a piggyBac transposase system has been developed which allows for the addition and removal of the transgenes that are bracketed by piggyBac recognition sites. Removal of the polycistronic transgene leaves a clean excision site, with no exogenous DNA remaining [59].

Based on the random integration of the retroviral approaches, the use of non-integrating methods is attractive, and include the use of adenoviral vectors, in particular for clinical applications. Adenoviral methods result in transient expression of the transgenes but are not as efficient in inducing reprogramming as integrative viruses [60]. Sendai viruses are also attractive based on their lack of pathogenicity in humans. They have been reported to efficiently infect dermal fibroblasts, as well as highly differentiated cell types such as T cells [61]. However, it takes several rounds of passage to remove the virus, the presence of which limits the differentiation capacity of the recipient cells. This has been circumvented using temperature sensitive mutants [62]. Recently, an improved procedure utilizing temperature sensitivity along with Sendai vectors for OCT4, SOX2, KLF4 and LMYC resulted in generation of naïve human iPSCs, without the use of feeder cells (described below), that had a better differentiation capacity in comparison with other methods [63]. Other methods using plasmids, miniCircles, proteins, mRNAs (as depicted in Fig. 4B) have also been tested, with varying degrees of efficiency.

Ex vivo expansion

There are numerous ways to ex vivo expand iPSCs, too numerous to explain in detail here. The majority of publications suggest that the initial reprogramming and subsequent expansion in the immediate period after reprogramming be done on irradiated murine embryonic fibroblast (MEF) feeder cells layers, and that the reprogrammed cells be transitioned onto Matrigel^™-, or fibronectin-, or vitronectin-, or laminin-511-coated substrates (reviewed in [64]). Each laboratory has its preference. However, the major point is that it must be ensured that the resulting population has maintained its pluripotency as judged by the assays described below. This checklist should be performed any time the lines have undergone extensive ex vivo expansion, and at the beginning of a series of experiments.

Conventionally, iPSCs are propagated as colonies in both feeder and feeder-free culture systems (e.g.; [65]). There are some limitations to passaging fragmented colonies, including low cell yields and generation of heterogeneously differentiated cells. In addition, the resulting colonies are not clonal (i.e., do not originate from a single cell). To improve hPSC culture methods, a method was developed based on non-colony type monolayer (NCM) culture of dissociated single cells and use of the Rho-associated kinase (ROCK) inhibitor, Y-27632. Cultivation of hPSCs on defined extracellular proteins such as the laminin isoform 521 (LN-521) without the use of ROCK inhibitors has also been optimized. Cells grown in this fashion have been efficiently transfected or transduced with plasmid DNAs, lentiviral particles, and oligonucleotide based microRNAs into hPSCs such that genetically modified cells can be used for molecular analyses and drug discovery. The NCM-based methods are less time-consuming and technically easier to perform than colony-type cultures and can be suitably scaled up to produce high numbers of homogeneous hPSCs for potential cell therapies, stem cell research, and drug discovery [66, 67].

Characterization of pluripotency

Once established and expanded in culture, reprogrammed cells must be verified as pluripotent through a number of assays. It must be demonstrated that they express levels of OCT4 and NANOG (and other pluripotency transcription factors such as TDGF1, DNMT3B, GABRB3 and GDF3) at levels equivalent to that expressed by hESCs (the gold standard for comparison) by qRT-PCR and immunocytochemistry. Other pluripotency markers such as SSEA3 and SSEA4 (glycolipid antigens), and TRA-1-60 and TRA-1-81 (keratan sulfate antigens) are evaluated by FACS [68, 69]. Subsequently, in vitro differentiation (Fig. 2) can be used to initially demonstrate formation of cell types that are representative of the three different germ layers; e.g. cells identified by TUBB3, NES and MAP2 expression (ectoderm); cells that are DES, ACTA2 positive or beating cardiomyocytes are formed (mesoderm); and liver-associated genes identified such as HNF4a, AFP and ALB expression (endoderm) [66, 69]. Pluripotency can also sometimes be determined by looking for derivatives of each germ layer in embryoid bodies (Fig. 2B); however, EB formation is often not consistent and is affected by the composition of the culture medium [70]. Consequently, the gold standard for determination of the pluripotent potential of the reprogrammed cells is by teratoma formation in vivo, along with identification of derivatives of all three germlines (Fig. 2B). This can be done by injecting the cells underneath the kidney capsule or subcutaneously in a mixture containing Matrigel^™ [69, 71].

It is advantageous to generate a considerable number of frozen cell aliquots at a similar passage prior to starting a series of experiments (a working bank, so to speak). Lastly, it is essential that the karyotype of the cells be evaluated at this point as well, to ensure that there are no detectable chromosomal defects. There are a number of assays that are informative [72], but not fool proof. It would be desirable to have assays that are more sensitive (and affordable) to detect small defects, which hopefully, will be developed in the future.

Formation of osteoprogenitors by hiPSCs

hiPSCs are powerful tools for studying developmental biology and mechanisms underlying pathological processes. An attractive and informative application of hiPSCs is in recapitulating the organogenesis of human bone. As described above, the human skeleton originates from three distinct embryonic lineages: the frontal skull and facial bones arise from the neural crest (NC), the axial skeleton from the paraxial mesoderm (PM), and the appendicular skeleton from the somatic lateral plate mesoderm (LPM) [73]. These differences in the developmental origins of bone are also reflected by diseases that affect specific bone types (Fig. 3b). Examples include Muenke syndrome (OIMM # 134934) and cherubism (OIMM # 118400) that affect frontal and facial bones, respectively; Robinow syndrome (OMIM # 268310), which affects the vertebrae of the axial skeleton, and acheiropodia (OMIM # 200500), which causes truncation of the upper and lower extremities of the appendicular skeleton [74] (Fig. 3B). In an in vivo study utilizing a Muenke syndrome mouse model, authors reported significant shortening of pre-sphenoid and basisphenoid bones, whereas basioccipital bones were unaffected. They demonstrated that the sphenoid and ethmoid bones are neural crest-derived, while the basioccipital bone is paraxial mesoderm in origin [75]. Further investigation of disorders affecting particular bone types would be greatly informative to bone development and disease mechanisms, making the derivation of lineage-specific osteogenic progenitors (OPs) from hiPSCs highly beneficial.

Differentiation of hiPSCs directly into OPs has been reported by several studies that employ various differentiation conditions and selection criteria for purifying cell populations [69, 76-78]. However, the embryonic germ layer from which they originated is often not described, or even considered. Instead, two-step differentiation protocols broadly refer to osteogenic precursors as "mesenchymal stem cell-like" populations. There is a major flaw to this train of thought in that “mesenchyme” is an undifferentiated connective tissue that has not yet been specified, and consequently, differentiation-specifying protocols need to be applied [25]. Furthermore, it relies on use of non-specific cell surface markers that select cells of fibroblastic morphology, without specifically selecting those that are related to bone development. Using these types of starting cell populations is a limitation in recapitulating the early developmental stages of bone and in reducing heterogeneity in cell populations. Previous studies that attempted to address this shortcoming performed stepwise differentiation of hiPSCs into osteoprogenitors in serum- and feeder-free conditions [73, 79, 80], but still did not generate all three OP lineages, or demonstrate the true osteogenic capacity by using rigorous assays, and/or describe transcriptomic patterns. As a result, the characterization of the differences among hiPSC-derived OPs remained incomplete. To address these gaps, the Robey lab reported a method of stepwise differentiation of lineage-specific OPs from hiPSC-derived PM, LPM, and NC cells using chemically defined and serum-free culture conditions [65].

Stepwise differentiation of hiPSCs into OP cells and their characterization

To recapitulate the development of lineage-specific osteogenic sub-populations, it is important to establish the stepwise differentiation steps of hiPSCs (Fig. 5) and identify their signature transcriptomic patterns. From primitive streak to paraxial and lateral plate mesoderm, and neural crest (which does not develop via the primitive streak), the success of the differentiation system is supported by the upregulated expression of corresponding lineage markers. Kidwai et al. characterized differences among transcriptomic patterns and highlighted key markers of the three cell populations [65].

Fig 5.

Differentiation of hiPSCs into lineage-specific osteoprogenitor cells from three different embryonic origns (paraxial mesoderm, lateral plate mesoderm and neural crest (Kidwai et al, Stem Cells, 2020).

Differentiation of hiPSCs into primitive streak

To mimic gastrulation during which the primitive streak (PS) forms, others have previously differentiated hiPSCs into PS using a GSK inhibitor [65, 81, 82]. Previous studies have also shown the expression of T-box Transcription Factor T (T in mouse, Brachyury in humans), Mesoderm Posterior BHLH Transcription factor 1 (MESP1), Mix Paired-Like Homeobox (MIXL1), and Forkhead Box F1 (FOXF1) in hiPSC-derived PS [65, 81, 82]. T, MIXL1, and MESP1 are early, reliable markers of gastrulation [83, 84]; FOXF1 expression signifies the development of mesoderm during the late primitive streak stage [85].

Differentiation of primitive streak into paraxial mesoderm

Given that the PS gives rise to PM, LPM, and definitive endoderm [86, 87], PS cells differentiated into PM using the BMP inhibitor, LDN193189, and the TGF-β inhibitor, SB431542, is well established [65]. Platelet-derived Growth Factor Receptor A (PDGFRα) has been used as a key marker of paraxial mesoderm [88-90], and in vivo studies have shown that T-Box Transcription Factor t (TBX6 )is expressed in nascent and maturing paraxial mesoderm [91]. Moreover, Paired Box 3 (PAX3) and Mesogenin 1 (MSGN1) are known master regulators of paraxial mesoderm [92]. Kidwai et al. reported forming PS-derived PM by using SB431542, a TGF-β-mediated SMAD 2/4 inhibitor, and BMP-mediated SMAD 1/5 inhibitor, LDN193189 [65].

Differentiation of primitive streak into lateral plate mesoderm

Eomesodermin (EOMES), Hematopoietically Expressed Homeobox (HHEX), NK2 Homeobox 5 (NKX2-5), and Islet-LIM Homeobox 1 (ISL) are established markers for lateral plate mesoderm [65, 88, 92, 93]. EOMES plays a crucial role in early gastrulation, and its deficiency results in loss of LP formation [94]. Activated by HHEX [95], NKX2-5 is reported to be a key modulator of LPM maturation [96]. Additionally, ISL1 is reported to be upstream of the Sonic Hedgehog pathway for LPM differentiation [97]. Previously, PS differentiation into lateral plate mesoderm was reported by Tan and colleagues [80].

Differentiation of hiPSCs into neural crest

NC differentiation follows a different route than PM and LPM (NC is not derived from the primitive streak). NC cells have been reported to differentiate directly from hiPSCs [90, 98]. Aristaless-like Homeobox 4 (ALX4), Goosecoid Homeobox (GSC) (cranial neural crest markers); Sry-box Transcription Factor 9 and 10 (SOX10, SOX9) (neural crest specifier genes), and Forkhead Box C2 (FOXC2) and Heart And Neural Crest Derivatives Expressed 1 (HAND1) (neural plate border genes) are the best known markers for neural crest cells [99, 100]. ALX4 is upregulated in NC, and its mutation is associated with craniofacial disorders [100]. SOX10 appears during neural crest migration and regulates both neural crest survival and differentiation [101]. GSC is required during embryogenesis and normal formation of craniofacial structures [102, 103]. SOX9 precedes markers of migratory neural crest [104]. FOXC2 [105] and HAND1 expression have been reported in nerual crest cells, and are also involved in RUNX2-IHH-regulated endochondral ossification [106].

Osteogenic differentiation of lineage-specific progenitors

hiPSC-derived progenitors were then differentiated into three OP groups in serum-free medium as has been previously reported [65, 107]. Kidwai et al. reported high expression of the osteogenic factors, Runt-Related Transcription Factor 2 (RUNX2), Distal-Less Homeobox 5 (DLX5), Secreted Phosphoprotein 1 (SPP1) and Bone Gamma-Carboxyglutamate Protein (BGLAP) in PM-OPs, LPM-OPs, and NC-OPs [65]. Most importantly, the bona fide differentiation of OPs into mature bone-forming cells was validated by subcutaneous transplantation (Fig. 6). Here, it must be noted that the vast majority of studies studying osteogenic differentiation of any population of cells, including so-called “mesenchymal stromal/stem cells” and pluripotent stem cells (PSCs) have relied on the use of an in vitro differentiation assay that is known to result in artifactual results. Osteogenic media includes β-glycerophosphate (BGP) or other sources of phosphate (along with varying concentrations of dexamethasone, with or without a BMP and ascorbic acid). However, if the test population of cells expresses alkaline phosphatase (ALP), which undifferentiated PSCs do, as well as many types of stromal cell populations (including adipogenic precursors), it cleaves BGP to form inorganic phosphate, and when the phosphate concentration becomes high enough in the medium, calcium phosphate spontaneously precipitates, and stains very well with von Kossa or with Alizarin Red, but it is not hydroxyapatite that is characteristic of mineralized matrix (e.g., [108]). The gold standard for determining osteogenic differentiation is by in vivo transplantation of the test population of cells in combination with an appropriate scaffold underneath the skin [109], under the kidney capsule or orthotopically. Histologically, one must see osteoblasts on the bone-forming surface, and osteocytes in a matrix that is fluorescent when illuminated with UV light [110]. The use of polarized light can also determine orientation of collagen fibers (woven bone vs. lamellar bone) [111] (e.g., [69]). This can distinguish between bone (collagen fibers are visualized) versus dystrophic calcification (devoid of collagen). In addition. It must be determined that the bone is of donor origin, either by in situ hybridization probes for human-specific ALU DNA sequences [65], or by staining with a donor species-specific antibody [112].

Fig 6.

In vivo transplantation of lineage-specific osteoprogenitor cells into immunocompromised mice (NSG) in combination with a ceramic scaffold (Attrax^™) after 16 weeks of transplantation. Bone formation was verified to be of donor origin based on identification of human cells by human ALU in situ hybridization (Kidwai et al, Stem Cells, 2020).

The formation of bone was verified using the assays as described above [65]. Interestingly, it was found that both PM-OPs and LP-OPs formed small areas of chondrocytes after eight weeks of in vivo transplantation, which may represent the cartilage intermediate in endochondral ossification that occurs in the axial and appendicular skeleton [113]. On the other hand, evidence of chondrogenic ossification in NC-OPs was not found [65]. Instead, the flat bones of the ventral skull forms from neural crest-derived OPs in the frontal bones that proliferate and condense into compact nodules containing osteoblasts, which deposit an osteoid matrix that later mineralizes [113].

Characterization of lineage-specific osteoprogenitors (OPs)

Paraxial mesoderm osteoprogenitors (PM-OPs)

It was also reported that expression of Cadherin 6 (CDH6), Fibrinogen Beta Chain (FGB), C-X-C Motif Chemokine Ligand 6 (CXCL6), and Insulin-like Growth Factors (IGFs) and related genes were exclusively found in PM-OPs [65]. It has been shown that CDH6 is a target of TGF-β and is regulated by RUNX2 [114]. FGB induces RUNX2 activity through the SMAD1/5/8 signaling pathway [65]. CXCL6 has been reported to play an important role in bone formation during fetal development and in response to hormonal and mechanical stimuli [115]. Similarly, IGFBP1 and its ligands, play key roles in bone metabolism [116].

Lateral plate mesoderm osteoprogenitors (LPM-OPs)

For LPM-OPs, Fibroblast Growth Factor 7 (FGF7) and Homeobox (HOX) family genes (HOXA13, HOXC10, and HOXC11) were found to be exclusively present [65]. In LP-OPs, FGF7 is expressed in connective tissues and plays an essential role in regulating long bone development [117]. The cluster of HOX genes is known to provide cells with specific positional identities on the anterior-posterior axis [117]. They also have important roles in long bone embryogenesis [52].

Neural crest osteoprogenitors (NC-OPs)

It was reported that there are high levels of Fibroblast Growth Factor 1 (FGF1) in NC-derived OPs (NC-OPs), an important signaling molecule in bone-related processes. The 23 members of the fibroblast growth factor (FGF) family bind to fibroblast growth factor receptors (FGFRs), leading to receptor dimerization and trans-autophosphorylation of the kinase domain [118]. Together, FGF-FGFRs play essential developmental and homeostatic roles in the skeleton by regulating chondrocyte and osteoblast differentiation and proliferation [119]. Indeed, aberrancy in their signaling cascades causes various well-established skeletal diseases, such as achondroplasia from FGFR3 gain-of-function mutations [120]. FGFR signal transduction is comprised of four major pathways: phosphoinositide-3-kinase/AKT (PI3K/AKT), phospholipase Cγ (PLCγ), signal transducer and activator of transcription (STAT), and the RAS/mitogen-activated protein kinase (MAPK) pathways [121]. MAPK is the predominant downstream pathway of activated FGFRs, modulating cell proliferation and, in certain contexts, differentiation [121]. Kidwai et al. reported that FGF1 regulates RUNX2 at multiple levels, with evidence implicating MAPK involvement, specifically extracellular signal-regulated kinases 1 and 2 (ERK1/2 or MAPK1/3) [65]. Taken together, these studies validate hiPSCs as a powerful tool to model bone development and draw attention to FGF1 as a protein of interest for future studies on disorders of neural crest-derived structures.

iPSCs bearing disease-causing mutations – development of new models of human diseases

Since Yamanaka's studies first documented human hiPSC generation, there have been an explosion of interest in the utilization of hiPSCs [52]. Human-derived iPSCs overcomes both ethical and translational barriers in hES-based human therapeutics. The potential of a patient-specific cell-based treatment to manufacture several types of critical tissues and cells, such as OPs, hematopoietic stem cells, cardiomyocytes, and motor neurons, has attracted international interest [122]. Researchers anticipate using hiPSCs in at least two ways: as a source for autologous cell-based therapy without immune rejection [122] and as a tool for disease modeling to discover patient-specific treatments [123].

Consistently, animal models and immortalized cell systems have been utilized in conventional disease modeling research to investigate disease pathophysiology and treatment development [124]. The screening of therapeutics for monogenic disorders, such as rare skeletal anomalies, cancer, neurological diseases, and congenital heart diseases, has utilized cells that have been artificially altered [125]. Multiple mutations, as observed in compound-mutation disorders, have not been explored for drug screening, whereas monogenic diseases have been intensively studied to seek therapies. It has proven difficult to construct disease models using several genes that mimic the majority of clinical disorders [124]. 95% of novel medications that have been evaluated using artificially modified cells were subsequently abandoned due to off-target effects [126].

To mitigate the concerns associated with off-target effects, it has been proposed to conduct research on human tissues. However, research involving human samples present significant obstacles. First, obtaining fresh affected human patient samples is challenging. Since the majority of human samples are collected at the latter stages of the disease or postmortem, it is not possible to critically evaluate the evolution of pathological characteristics. Consequently, hiPSC disease modeling has developed as an alternate way for gathering both essential human pathology data and a continuous supply of diseased cells. The rapid advancement of hiPSC-mediated disease models and innovative treatment strategies [51] has occurred. Comparing human hiPSC disease models with their parental lines reveals significant advantages [127]. For instance, it gives a complex genetic signature of a human tissue and an inexhaustible resource for determining the mechanism behind the development of disease and creating innovative treatments.

The potential of hiPSCs for disease modeling and drug screening is being utilized largely in the areas of neurology, cardiology and hematology. However, there are only a handful of studies using hiPSCs for the study of skeletal diseases. One example is the use of hiPSCs to characterize pathogenetic mechanisms brought about by changes in Notch signaling, as described below.

Development of iPSC-derived models of human disease in Notch signaling

Notch plays a central role in cell fate decisions and in the differentiation and function of cells of the osteoblast and osteoclast lineages [128, 129]. The extracellular domain of Notch interacts with ligands of the Jagged and Delta-like families and as a result of the interaction, Notch is cleaved and its intracellular domain (NICD) is released [130, 131]. The NICD forms a complex with recombination signal-binding protein for Ig of κ region (Rbpjκ) and mastermind-like (Maml) to regulate transcription of target genes [132-134]. [135-137].

Notch1, 2 and 3 and low levels of Notch4 transcripts are expressed in skeletal cells [129]. NOTCH1 and NOTCH2 are structurally similar, but their functions are not redundant [138-140]. NOTCH1 inhibits osteoblast and osteoclast differentiation, whereas NOTCH2 induces osteoclastogenesis directly and indirectly [141, 142]. NOTCH3 is preferentially expressed by vascular smooth muscle cells and osteocytes and is not expressed by the myeloid lineage [143]. NOTCH3 induces Receptor activator of NF-κB (RANKL) and as a consequence, osteoclast differentiation. Mutations in the extracellular domain of NOTCH3 cause cerebral autosomal-dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL, OMIM 125310) [144-146].

Selected pathogenic variants of Notch signaling associated with skeletal disorders

Hajdu Cheney Syndrome (HCS) (OMIM # 102500) is a rare and devastating disorder characterized by numerous skeletal manifestations, including craniofacial developmental defects, short stature, bone loss with fractures and acroosteolysis associated with inflammation of the distal phalanges [129]. Iliac crest biopsies of subjects afflicted by HCS reveal decreased trabecular and cortical bone, and enhanced bone resorption without impairment of bone formation [147-150]. HCS is associated with mutations or short deletions in exon 34 of NOTCH2 upstream of the PEST domain [129, 151-154]. The mutations lead to the premature termination of a protein product lacking sequences necessary for the proteasomal degradation of the NOTCH2 NICD so that the protein is stable and a gain-of-NOTCH2 function ensues.

Lateral Meningocele (LMS), or Lehman Syndrome (OMIM # 130720), is a rare disorder characterized by craniofacial anomalies, hypotonia and meningocele [155]. Skeletal manifestations are numerous and include craniofacial developmental defects, short stature, scoliosis and osteopenia [156]. Exome sequencing of subjects afflicted by LMS has demonstrated the presence of point mutations or short deletions in exon 33 of NOTCH3, upstream of the PEST domain. This leads to the translation of a truncated, but stable protein, devoid of the PEST domain, and a gain-of-NOTCH3 function [157]. Autosomal dominant inheritance as well as de novo heterozygous mutations have been reported for HCS and LMS..

Alagille Syndrome (AS, OMIM # 118450) is an autosomal dominant disorder associated with pathogenic variants of the Notch ligand, JAG1, and less frequently, NOTCH2. The variants result in loss of function. AS is characterized by paucity in the development of the bile ducts causing cholestasis[158, 159]. Skeletal abnormalities include butterfly shaped vertebrae, shortened interpedicular distance and short stature. Osteoporosis is most likely the consequence of hepatic insufficiency and multiorgan involvement since the inactivation of either Jag1 or Notch2 result in increased, and not decreased, bone mass [160, 161].

Models of monogenic disorders associated with Notch signal activation

We created and rigorously validated a knock-in mouse model harboring a HCS mutation and termed Notch^tm1.1Ecan [128]. The homozygous mutation is associated with craniofacial developmental abnormalities and is lethal, and heterozygous Notch2^tm1.1Ecan mutant mice exhibit profound cancellous and cortical bone osteopenia, short limbs and splemenogaly reproducing functional outcomes of the human disease [128, 162, 163]. Notch2^tm1.1Ecan mice are sensitized to the osteolytic actions of the inflammatory cytokine tumor, necrosis factor α (TNFα) [164]. Notch2^tm1.1Ecan mice, like humans with HCS, have increased bone resorption, secondary to a direct effect of the NOTCH2 gain-of-function on osteoclastogenesis as well as the increased expression of RANKL by cells of the osteoblast lineage [128].

We created a mouse model harboring a Notch3LMS mutation termed Notch3^em1.1Ecan and heterozygous mutant mice exhibit profound cancellous and cortical bone osteopenia, reproducing aspects of the human disease [129]. . The cause of the osteopenia is increased bone resorption secondary to enhanced expression of RANKL by cells of the osteoblast lineage [129, 143, 165].

A mouse model heterozygous for a Jag1 null allele and a Notch2 hypomorphic allele exhibits impaired differentiation of intrahepatic bile ducts, jaundice, growth retardation and defects in heart, eye, and kidney development [166]. The phenotype recapitulates many of the features of Alagille Syndrome. A mutagenesis-induced mouse model harboring a Jag1 (JAGH268Q) hypomorphic mutation termed Nodder (Jag1^Ndr), was found to be Notch signaling incompetent and recapitulated the phenotype of Alagille Syndrome [167]. The model exhibits defects in bile duct differentiation and function and reduced expression of insulin-like growth factor 1 (IGF1). This is also found in humans with Alagille Syndrome and possibly explains the bone loss and short stature observed in the syndrome [167, 168].

iPSCs and monogenic disorders associated with Notch signal alterations

The mouse has been used as a valuable model of human disease. However, humans and mice have developmental, genetic and physiological differences making the validation of cellular events and the testing of therapeutic approaches in humans or human cells a necessity [169]. In addition, not every pathogenic variant harbors the same mutation. Therefore, treatments specific to each mutation often need to be designed and tested for effectiveness in appropriate cellular systems for the approach to be applicable to humans. This requires ample supply of cells from afflicted individuals so that specific treatments can be tested in vitro to document efficacy prior to their possible use in vivo. An attractive application of iPSC technology is that it allows the isolation of patient-derived cells, or the introduction of the genetic alterations associated with specific disorders into well characterized cell lines, providing an experimental system to study the pathogenesis of the disease and also to devise therapeutic strategies. The establishment of iPSCs (those with relevant mutations introduced into normal iPSCs, or those derived from affected individuals and mutation-corrected isogenic controls) can be utilized as a model for testing the cellular mechanisms responsible for the disease and the effectiveness of a therapeutic strategy. We believe that pathogenic variants affecting Notch signaling are ideal for the study of iPSCs as a model of human disease since they are monogenic, have an early onset during development and are associated with a distinct phenotype [170].

Current studies in our laboratory use the iPSC lines, NCRM1 and NCRM5, generated by the NIH Regenerative Medicine Program and obtained through RUCDR Infinite Biologics. NCRM1 and NCRM5 cells are derived from CD34⁺ cord blood by episomal nucleofection and are well-characterized pluripotent cells devoid of chromosomal structural aberrations. NCRM5 cells can differentiate into the myeloid lineage and into osteoprogenitors, making them ideal for studies related to skeletal cells. CRISPR/Cas9 technology was used to create mutations duplicating those found in subjects afflicted by HCS and Lehman Syndrome and replicated in Notch2^tm1.1Ecan and Notch3^em1.1Ecan mice.

We have developed protocols to induce the differentiation of iPSCs into the osteoclast and osteoblast lineages. iPSCs developed into multinucleated TRAP-positive cells that expressed CTSK, CALCR, ACP5 and NFATC1. Moreover, when cultured on bovine bone slices, the cells resorbed bone confirming their osteoclastic nature. We documented that NOTCH1, 2 and 3 receptors are expressed and that NOTCH2 is the prevalent receptor in osteoclasts, whereas NOTCH4 is not detected reproducing the expression pattern of murine osteoclasts.

Future studies will determine whether Notch mutations associated with specific mutogenic disorders alter cellular behavior and replicate the one observed in primary cell cultures of murine cells isolated from mutant mice. iPSCs from humans with Alagille Syndrome have been created and could be used to determine the role of JAG1 mutations in the differentiation of cells of the osteoblast and osteoclast lineages [171].

Other iPSC models of human skeletal diseases

As mentioned above, there are a limited number of reports on hiPSCs derived from patients with a genetic skeletal disease. Recently, an iPSC line from a patient with a rare craniofacial disease, Muenke Syndrome (MS, OMIM 602849), was established. MS is a disease caused by the p.Pro250Arg variant in FGFR3. MS is the leading genetic cause of craniosynostosis and results in a variety of disabling clinical phenotypes. If is characterized by uni- or bilateral coronal suture synostosis, macrocephaly, dysmorphic craniofacial features, and dental malocclusion [65, 172-174]. To model the disease and study the pathogenic mechanisms, an hiPSC line was generated from a patient diagnosed with MS. Successful reprogramming was validated by cellular morphological features, karyotyping, loss of exogenous reprogramming factors, expression of pluripotency markers, mutation analysis, and teratoma formation [173, 174]. Future studies will focus on identifying the defects in iPSCs differentiated into a suture-like phenotype to determine the pathogenetic mechanisms by which craniosynostosis develops in MS. hiPSCs have also been derived from patients with other skeletal diseases, such as Craniometaphyseal Dysplasia [175], Osteogenesis Imperfecta, Osteopetrosis, Fibrodysplasia Ossificans Progressiva and Marfan’s Syndrome (reviewed in [176]). Using these various cell lines, work is on-going to determine the impact of the disease-causing mutations on osteogenic and osteoclastic differentiation and function.

Genome editing

In addition to advances in cellular reprogramming, considerable advances have been made in our ability to edit DNA. The three major techniques that are currently in wide use rely on zinc-finger nucleases (ZFNs), transcription activator-like effector nucleases (TALENs) and clustered regularly interspaced short palindromic repeat (CRISPR)-Cas-associated nucleases (reviewed in [177]). All rely on creating double stranded DNA breaks (DSBs), followed by promoting the repair pathways: 1) non-homologous end-joining (NHEJ); or 2) homologuous recombination (HR), otherwise called homology-directed repair (HDR). All three types of genome editing enzyme systems have their pluses and their minuses, and the choice of which to use is determined by the goal to be achieved [177]. Genome editing is improving, but there are a number of problems relating to the formation of indels caused by DSB-induced site modifications, non-specific cuts, large deletions, and large rearrangements [178, 179].

Development of genome editing technology is particularly pertinent to the use of iPSCs for several reasons. First, correcting the gene mutation in a patient-derived iPSC line generates an isogeneic line that is the best possible control for experimental studies. When using the patient’s cells directly, the only possible controls are cells derived from normal, but non-relative individuals, a parent, or a sibling, none of which are genetically matched. Genetic variability is unavoidable in these cases, and can have a major impact on cell differentiation, and studies related to specific pathogenetic disease mechanisms. Second, correcting the mutation in the patient’s own cells results in an autologous cell product for potential cell therapy. Lastly, being able to create a disease-causing mutation within a normal iPSC line allows for the creation of a cell model of disease that might not necessarily be possible due to the paucity of patient-derived material and/or the rarity of the disease.

Summary and conclusions

Human iPSCs are proving to be useful to the biomedical community in many ways. Their use avoids ethical concerns due to the fact that they can be easily generated from a number of sources in relatively non-invasive ways, which also adds the advantage of generating autologous cells. Based on in vitro assays, and their in vivo performance in teratoma assays, hiPSCs can form virtually every cell type in the body. There have been substantial advances in differentiation protocols of hiPSCS in the last few years, although there are still challenges in achieving appropriate maturity levels. hiPSCs bearing disease-causing mutations can be used to model specific diseases both in vitro and in vivo. High throughput drug screening can be applied to hiPSCs to not only determine factors that regulate cell differentiation, but also to identify potential drug therapies for specific genetic disorders. Genome editing has been successfully applied to hiPSCs ex vivo, and while not without problems, the techniques are rapidly improving, thereby providing cells for cell and tissue replacement (or even gene correction in vivo in the future). Advances in hiPSC biology are currently be applied for the generate cell relevant to bone research, and undoubtedly lead to the advancement in our understanding of pathogenetic mechanisms and development of novel therapeutics and cell and tissue replacement strategies.

Highlights.

• Creation of iPSCs marks a major advance in the cell biology

• Functional OCs and OBs can be differentiated from iPSCs

• Mutated iPSCs can model human skeletal diseases