Introduction

To maintain long-term tissue homeostasis, it is indispensable to employ a system based on stem cells and to preserve the self-renewal activity and multi-lineage differentiation potential of these stem cells. Proper control of self-renewal ability in hematopoietic stem cells (HSCs) is essential for maintaining functional mature cell production over organismal lifetime. However, it has been reported that DNA damage gradually accumulates through repeated cell divisions over long periods of time, and, along with replicative stress, markedly impairs stem cell function (Fig. 1) [1,2,3,4,5,6,7,8,9]. Ex vivo expansion of HSCs is a major challenge in cellular therapy, made more difficult by the high sensitivity of HSCs to DNA damage. On the other hand, it has been reported that the telomere end is a particularly fragile site in vertebrate chromosomal DNA, and it is highly sensitive to DNA damage. Telomeres consist of specific tandem repeats (5′-TTAGGG-3′), which are oxidation-sensitive G-rich regions [10]. The 3′ single-stranded G-rich overhang can be exposed to the DNA damage machinery, followed by cell cycle arrest and apoptosis. Furthermore, aberrant repair pathways lead to genomic abnormality and instability [11,12,13]. Therefore, prevention of DNA damage response and mechanisms for telomere end protection are required for the maintenance of the hematopoietic system [14,15,16,17,18,19].

Fig. 1
figure 1

The shelterin complex is a gatekeeper that prevents HSC loss through telomere protection. HSCs repeatedly divide for self-renewal and progenitor cell production over organismal lifetime. The telomere is elongated by telomerase, and its activity is regulated by shelterin molecules. It is also the role of the shelterin complex to inhibit DNA damage response (DDR) caused by replicative stress. Since accumulation of excessive DNA damage leads to cell death, stem cell aging, and hematological malignancies, proper protection by shelterin is indispensable for maintenance of the normal HSC pool

Full size image

The shelterin complex is composed of six protein subunits, TRF1, TRF2, RAP1, TIN2, TPP1, and POT1; they bind to the telomeric DNA region and are linked to the maintenance and regulation of telomere length and terminal loop structure [20, 21]. Shelterin is known to play an important role in protecting telomeres from signaling via ataxia telangiectasia-mutated (ATM) and ataxia telangiectasia and RAD3-related (ATR)-mediated DNA damage response (DDR) [22,23,24,25,26]. Here, we focus on the shelterin component proteins and describe the molecular mechanisms underlying maintenance of the self-renewal activity of hematopoiesis and controlling senescence and telomere dysfunction-associated disease.

The role of shelterin proteins in telomere protection

Several shelterin component proteins directly bind to telomeric DNA in maintaining telomere homeostasis. The Pot1/Tpp1 heterodimer binds to telomeric single-stranded DNA (ssDNA), and Trf1/Trf2 binds to the double-stranded region [27,28,29]. The main role of Trf1/Trf2 in the telomere is the maintenance of the loop structure and the suppression of DDR. Trf1 controls telomere length negatively and promotes telomere duplication [30, 31]. It has been suggested that Trf1 separates from telomeres when phosphorylated by ATM-dependent on the activity of the MRE11-RAD50-NBS1 complex, and this dissociation promotes telomere elongation by telomerase [32]. Therefore, appropriate activation of ATM is necessary for maintaining telomere homeostasis. On the other hand, Trf2 suppresses ATM signaling and downstream telomeric DNA resection [33], non-homologous end joining (NHEJ) and homology-directed repair (HDR) [34,35,36]. In mammals, the Trf2 interacting protein (Trf2ip), Rap1 forms a heterodimer with Trf2 and protects telomeres from inappropriate HDR pathway activation [37, 38]. Rap1 is the only shelterin molecule that inactivated mice do not result in embryonic lethality. The phenotype of Rap1 conditional knockout in epidermis shows telomere shortening, increased DNA damage, and skin hyperpigmentation [39]. In its canonical telomere-associate role, Tin2 is a central component of shelterin, contributing to the entry of the Tpp1/Pot1 heterodimer into the nucleus and tethering it to Trf1/Trf2 [22, 40, 41]. Tpp1, which is encoded by the gene Acd, forms a Pot1/Tpp1 heterodimer with Pot1, another shelterin member; this heterodimer caps telomeres by binding to the telomere G-rich overhang [27, 28]. On the surface of the oligonucleotide/oligosaccharide binding fold domain (OB-fold) domain of the Tpp1, small clusters of amino acids called TEL patches are essential for both the recruitment of telomerase and promotion of telomere elongation activity [42,43,44,45]. Pot1 is a member of the shelterin complex, which binds to telomeric single-stranded DNA (ssDNA) G-rich overhangs via N-terminal OB-fold domains and caps telomeres to preserve genomic integrity [46, 47]. POT1 cooperates with heterogeneous nuclear ribonucleoprotein A1 (hnRNPA1) and telomeric repeat-containing RNA (TERRA) to suppress aberrant DDR by inhibiting the binding of replication protein A (RPA) to ssDNA [48]. Although human shelterin contains a single POT1 gene, there are two Pot1 orthologs, Pot1a and Pot1b in the mouse genome, which have different functions in telomeres [49,50,51,52]. Pot1a is essential for suppressing cytogenetic abnormalities, which occur due to the activation of DDR in telomeres and the subsequent activation in NHEJ and HDR-dependent repair pathways [34, 49, 50, 53]. Deficiency of Pot1a in mice leads to early embryonic lethality. On the other hand, Pot1b is involved in maintaining G-tail length through the regulation of the Apollo nuclease and Exo1, or DNA polymerase [51, 52]. Although, Pot1b-deficient mice are viable and fertile, Pot1b-deficient cells show long single-stranded telomeric overhangs and telomere shortening [49, 50, 54].

Extra-telomeric function of shelterin proteins

Recently, functions of shelterin molecules other than protecting and maintaining telomere called extra-telomeric function have been reported. TRF2 also has functions as a transcriptional regulator. In the tumor vasculature of human cancers, TRF2 transcriptionally activates PDGF receptor β and increases angiogenesis [55]. Conversely, the Wilms’ tumor suppressor WT1 and the Wnt/β-catenin signaling pathway regulate the expression level of TRF2 [55, 56]. Rap1 activates the NF-κB-mediated pathway through binding to IKK and also binds to the sub-telomeric region as a transcriptional repressor, thereby contributing to the regulation of metabolic programs, including fatty acid and glucose metabolism, and PPARα signaling [57, 58]. Indeed, the absence of Rap 1 leads to the negative control of gene expression of metabolism and cell proliferation. Based on this change, Rap1 deficient mice show adult-onset obesity due to metabolic defects. This phenotype is related to glucose intolerance and insulin resistance. TIN2 is involved in metabolic regulation. TIN2 can also be located in mitochondria and negatively regulates energy production by suppressing oxidative phosphorylation [59]. TPP1-TIN2 binding inhibits the transition of TIN2 to mitochondria and promotes entry into the nucleus. Together, these findings suggest that changes in the abundance of shelterin protein affect cellular metabolic regulation. TIN2 Study shows the concept that observation of quantitative changes in shelterin protein is necessary to promote understanding of extra-telomeric functions in shelterin molecules.

Role of shelterin genes in the maintenance of normal hematopoiesis

Acute Trf1 deficiency in mouse bone marrow hematopoietic cells leads to aplastic anemia and bone marrow failure syndrome (BMF) (Table 1) [18, 60, 61]. Interestingly, although this increases telomeric DDR and p53-p21-mediated cellular senescence in hematopoietic progenitor cells, telomere shortening does not occur. On the other hand, chronic Trf1 inactivation leads to severe telomere shortening, similar to human dyskeratosis congenita (DC). HSC pools in Trf1-deficient mice are depleted through aberrant cell cycle progression, leading to decreased bone marrow reconstitution ability. These effects underlie the pathogenic effects of Trf1-mediated shelterin mutation in human DC, through loss of telomere protection, followed by the rapid depletion of HSPCs.

Table 1 Role of shelterin components in the maintenance of HSCs
Full size table

Tpp1/Acd deficiency in mice leads to embryonic or perinatal lethality, but analysis of embryonic hematopoiesis of mice which are homozygous for the spontaneously occurring hypomorphic Acd allele has revealed that Acd/Tpp1 is essential for embryonic development (Table 1) [16, 62, 63]. In the hypomorphic Acd mice, HSPCs show loss of cell cycle quiescence and G2/M arrest, resulting in loss of long-term bone marrow reconstitution ability. This phenotype is similar to that of HSPCs in the fetal liver of Acd (flox/flox); Vav-Cre + mice, in which the HSPC pool is severely depleted. In adult hematopoiesis, as a result of conditional inactivation of Acd, acute chromosome instability, up-regulation of p53 target genes and G2/M arrest are observed. Moreover, HSPCs are immediately exhausted after Acd deletion. In the case of competitive bone marrow transplantation, Acd-deficient HSPCs do not support bone marrow repopulation activity. Interestingly, p53 deletion does not rescue Acd-deficient HSPC function. Thus, it is suggested that Acd/Tpp1 maintains the homeostasis of the hematopoietic system via p53-dependent or independent pathways.

Deletion of Pot1b results in the formation of a long G-rich overhang by nucleolytic processing of the 5′ end of the C-rich strand by Apollo and Exo1, with progressive loss of total telomere length [52, 54, 64]. It has been suggested that the extension of the overhang recruits RPA and results in activation of an ATR-p53-dependent checkpoint response that impairs HSC function. In Pot1b-deficient mice, several premature aging-related phenotypes appear by 15 months of age (Table 1) [54, 65]. The mice exhibit low reproductive ability with a reduction in testes size and increased numbers of apoptotic cells in the testes and intestine. In the peripheral blood, leucopenia, anemia, and thrombocytopenia are observed. In adult hematopoiesis, severe depletion of the HSPC component, loss of cell cycle quiescence and decreasing bone marrow reconstitution ability are observed. Increased DDR, chromosomal fusion, and apoptosis were observed in HSCs of 4-month-old mice, as telomere protection declined. Pot1b deletion in the background of telomerase inactivation presents with a more severe phenotype, including bone marrow failure and premature death (Table 1) [17, 54, 64, 65]. In Pot1b-telomerase-p53 multi-deficient mice, apoptosis and depletion of HSPCs were partially restored [65]. Moreover, the simultaneous loss of Pot1b and the cellular senescence marker p16 (INK4a), which restricts stem cell function [66, 67], activates ATR-dependent DDR, resulting in increased p53-p21-dependent cell cycle arrest and apoptosis in HSPCs [17]. In Pot1b-p53 double-deficient mice, activation of DDR-based ATR and subsequent p53-independent apoptosis increases, suggesting that p73 is involved in this cascade [68]. Therefore, the function of Pot1b in HSPCs is to maintain telomere end stability and suppress p53-dependent and independent apoptosis.

Pot1a/POT1 maintains HSC activity with age

Recently, we demonstrated that the expression of Pot1a in HSCs is important for suppressing telomeric DDR and maintaining stem cell activity, which decreases with aging (Fig. 2) [14]. Knockdown of Pot1a increases telomeric DDR and reduces long-term bone marrow reconstitution activity (Table 1). Also, the balance in the supply of mature cells changed, as myeloid cells increased, and B cells decreased. Decreased expression levels of Pot1a in HSCs resulted in reduced cell proliferation and increased apoptosis. Similarly, overexpression of a dominant-negative Pot1a form lacking the OB-fold domain resulted in the reduction of the long-term repopulation capacity of HSCs. Pot1a binding to ssDNA via the OB-fold region contributed to telomere stability and maintenance of functional HSCs. In contrast, exogenous POT1a protein treatment suppressed the telomeric DDR and maintained the reconstitution activity of HSCs (Table 1, Fig. 2). Furthermore, the exogenous Pot1a treatment partially restored the reduced reconstitution ability of aged HSCs and normalized the differentiation of donor cells. Furthermore, we observed that exogenous POT1a protein treatment or overexpression of POT1a altered metabolism via mTOR signaling and oxidative phosphorylation in HSCs. Exogenous POT1 protein treatment can also be applied to human HSCs. The treatment of human umbilical cord blood-derived HSCs with exogenous human POT1 protein maintained human HSC activity in culture (Table 1). It has recently been reported that the POT1 can bind not only ssDNA, but also sub-telomeric and non-telomeric DNA through its OB-fold domain [69]. As extra-telomeric roles, POT1 may be involved in gene transcription, replication, and repair. Our findings may implicate POT1 in these functions, but further research is needed for the elucidation of detailed mechanisms.

Fig. 2
figure 2

Maintenance of HSC activity by exogenous POT1a. In the steady state, HSCs slowly but repeatedly divide and self-renew. With progressive aging, telomere DDR accumulation and BM reconstitution ability decrease in HSCs. Similar defects occur during HSC expansion in in vitro culture. Pot1a levels in HSCs decrease in these processes, but they recover by exogenous POT1a addition and protect from the loss of HSC activity

Full size image

The single deficiency of most shelterin genes leads to acute injury and severe phenotypes. However, in the study of aging, mouse models with such acute disorders may not strictly represent models of aging, as the phenotype resulting from physiological aging is milder [70]. Therefore, to further clarify the aging of HSPCs, it may be necessary to construct stepwise expression control models in vivo.

Abnormalities in shelterin gene related to hematological malfunction

It has recently been reported that expression deficiency and mutation of shelterin genes are one of the causes of hematological disease (Table 2). In maintaining the stability of telomeres, the amount of each shelterin protein is strictly controlled, and once the balance is lost, the fragility of telomeres may be unmasked, which may cause chromosomal abnormalities by unwarranted repair activity.

Table 2 Mutation and expression abnormalities of shelterin components in the hematopoietic malignancies
Full size table

Abnormal expression of shelterin molecules is observed in various hematopoietic malignancies. Impaired expression of TRF1, TRF2, and RAP1 is detected in chronic myeloid leukemia (CML), acute lymphoblastic leukemia (ALL) and acute myeloid leukemia (AML) [71,72,73,74,75,76]. Also, abnormal expression of TIN2, ACD/TPP1, and POT1 is involved in chronic lymphocytic leukemia (CLL), aplastic anemia, and multiple myeloma [72, 77,78,79,80]. The attenuation of telomere maintenance is involved in the genetic syndrome dyskeratosis congenita (DC) and its severe form Hoyeraal-Hreidarsson (HH) syndrome [81, 82]. DC is a congenital hematopoietic deficiency syndrome characterized by bone marrow failure, abnormalities of the skin, nails and mucous membranes and premature hair graying and loss, and in most cases involves the shortening of telomeres. The genes responsible for DC are mostly involved in telomere control: DKC1 (Dyskerin), NOP10, NHP2, TINF2 (encoding TIN2), TCAB1, TERC, TERT [83]. TIN2 mutations were identified in DC patients [84,85,86]. Although TIN2 regulates the expression levels of TRF1 and TPP1, the TINF2 mutation found in DC patients reduces the interaction between TIN2 and telomerase without significantly affecting their levels [87]. Also, the HP1 binding site of TIN2 is necessary for sister telomere cohesion in cell division, but this site is located in the DC-related TINF2 mutation cluster [88]. The TINF2 mutation suppresses the interaction between TIN2 and HP1γ and may impair the maintenance of telomere length by inhibiting sister telomere cohesion in the telomere region in S phase [86]. A mouse model with a knock-in mutation in a Tinf2 allele shows a human DC-like phenotype, including mild pancytopenia and telomere shortening [25]. As a result of the mutation of Tinf2, telomeres are broken down and shortened, through a telomerase-independent mechanism. This result may point to the function of Tin2 in sister telomere cohesion in the division phase [86], or to the role played by crosslinking Trf1–Trf2 to regulate telomere compaction and to protect chromosomal ends [89]. In one family affected by HH syndrome, a clinically serious DC mutant, a mutation was found in ACD. It was found that there is a one amino-acid deletion in the TEL patch of TPP1 (ΔK170), which is important for promoting recruitment and elongation of telomerase within the OB-fold domain [90]. This mutation has also been found in aplastic anemia and related BMF disorders [78]. Consequently, recruitment of telomerase via the TEL patch on the surface of TPP1 is necessary for proper maintenance of the telomere length of HSCs.

It has been reported that changes in the expression levels of several shelterin components are involved in CLL [72]. Compared with normal lymphocytes, expression of TRF1, RAP1, and POT1 decreased and expression of ACD/TPP1 increased in CLL cells. Telomere dysfunction-induced foci (TIFs), an indicator of telomere defects, increased in CLL-derived cells [77]. Therefore, telomere instability may be caused in CLL-derived cells. In addition, it was reported that POT1 somatic mutations occurred in approximately 5% of CLL cases [91, 92]. In particular, clusters of mutations were observed in the OB-fold domain, which is a DNA binding motif. It has been suggested that POT1 mutation leads to telomere instability and chromosomal abnormality, which is favorable for the acquisition of malignant features of CLL. Overexpression of Y223C, one of the POT1 mutations found in CLL, in hematopoietic cells derived from p53 heterozygotes, promotes the development of transplantable myeloid dysplasia [93]. Furthermore, double deficiency of Pot1a and p53 in mouse common lymphoid progenitors (CLP) leads to aggressive thymic lymphoma [94]. These findings suggest that tumor formation is suppressed by the ATR-p53 pathway, even if the function of Pot1a declines, but if this pathway is also impaired at the same time, tumorigenesis becomes extremely aggressive.

Conclusions and future directions

Much progress has been made in the understanding of the maintenance of telomere homeostasis, especially regarding how terminal elongation and end protection are cooperatively regulated. Based on these developments, many causative genetic mutations have been identified in hematological malignancy and human hematopoietic failure syndromes. The genetic mutations found in TINF2, ACD and POT1 impair the function of the normal shelterin complex, leading to failure of proper telomere protection and subsequent chromosomal instability. Clarifying the decline of shelterin function caused by such point mutations is an important theme in clinical research centered on maintaining the integrity of the telomere. The deeper understanding of the maintenance of telomere integrity will also lead to the development of ex vivo expansion strategies for HSCs. On the other hand, it is also necessary to understand the extra-telomeric functions of shelterin molecules, which are now being uncovered. For example, Tin2 is involved in the control of metabolic activities. Rap1 shows transcriptional activity in the sub-telomeric and non-telomeric regions. In addition, Pot1 may be involved in both transcriptional regulation and metabolic control in HSPCs. In the future, clarifying the molecular function of shelterin proteins in both telomeric and extra-telomeric roles may be particularly beneficial for the full understanding of hematopoietic homeostasis and hematopoietic disorders.