Asymmetric Inheritance of Epigenetic States in Asymmetrically Dividing Stem Cells

Abstract

Asymmetric cell division produces two cells that are genetically identical but each have distinctly different cell fates. During this process, epigenetic mechanisms play an important role in allowing the two daughter cells to have unique gene expression profiles that lead to their specific cell identities. While the process of duplicating and segregating the genetic information during the cell cycle has been well studied, the question of how epigenetic information is duplicated and partitioned still remains. In this review, we discuss recent advances in understanding how epigenetic states are established and inherited, with emphasis on the asymmetric inheritance patterns of histones, DNA methylation, non-histone proteins, RNAs, and organelles. We also discuss how mis-regulation of these processes may lead to diseases such as cancer and tissue degeneration.

Introduction

Stem cells are unique in their ability to both self-renew and differentiate into various cell types. Many adult stem cells undergo asymmetric cell division (ACD) to give rise to two daughter cells: one cell that retains its stem cell identity and another cell that enters the differentiation program. This process occurs commonly and is essential for development, tissue homeostasis, and regeneration. Imbalance between self-renewal versus differentiation of stem cells can result in cancer, infertility, or tissue degeneration (reviewed in [1–4]).

Previous studies have described various extrinsic and intrinsic mechanisms that regulate the distinct daughter cell fates after ACD. Extrinsic mechanisms such as signaling molecules from the stem cell-residing microenvironment could act locally to define distinct cell fates [5]. Asymmetric separation of intrinsic factors during stem cell division, such as proteins and RNA molecules, could also help determine distinct cell fates [6]. Recent studies have sought to understand how epigenetic states are inherited during ACD and their roles in defining distinct cell fates.

Epigenetic states are defined by chromatin structure, which play critical roles in regulating transcriptions and cell cycles, as well as DNA damage and repair during the development of metazoans [7,8]. Nucleosomes are the fundamental building blocks of chromatin and are one of the major carriers of epigenetic information in eukaryotic cells. Histones comprise the core component of nucleosomes, which are formed by an octamer of four canonical histones (H2A, H2B, H3, H4) with ~147 bp of DNA wrapping around it [9,10]. Unstructured N-terminal tails extending from the globular C-terminal cores of histones can be modified with covalent posttranslational modifications (PTMs) that dictate epigenetic states. Emerging evidence suggests that mutations to histone genes themselves are linked to tumorigenesis [11].

During ACD, the establishment and maintenance of unique cell identities are regulated by gene expression through epigenetic mechanisms. Classically, epigenetic inheritance has been studied in the context of how an epigenetic state is faithfully maintained through DNA replication and cell division in symmetrically dividing cells. In ACD, the two resulting daughter cells take on different cell fates, which is thought to be defined by their distinct epigenetic states. However, it has remained largely unclear as to how epigenetic states are maintained or changed during ACD. In this review, we focus on factors that are inherited differentially during ACD of stem cells such as the chromatin-associated proteins, highlighting recent studies that provide new insights into this process (Figure 1). We also discuss aberrant epigenetic inheritance that contribute to diseases. These studies shed light on a long-standing question critical to both developmental and chromatin biology.

Figure 1. Summary of Asymmetric Inheritance of Epigenetic States in Different Systems.

Epigenetic mechanisms allow for differential gene expression and the establishment of a cell identity program. During asymmetric cell division, one cell gives rise to two daughter cells that have identical genetic material but take on two different cell fates. Recent work has identified epigenetic factors that are asymmetrically inherited and influence cell fate. At the chromatin level, it has been found that preexisting and newly synthesized histones can be asymmetrically partitioned to different sister chromatids, providing unique epigenetic profiles. Further, DNA methylation can be asymmetric between sister chromatids. These two mechanisms can create epigenetically distinct sister chromatids, which are then recognized and segregated by the mitotic machinery. One set of sister chromatids can have stronger centromeres, which attach to the centrosome with higher microtubule activity, leading to asymmetric segregation of sister chromatids. Beyond the chromatin level, proteins, RNAs, and organelles can also be asymmetrically partitioned during division. Polarity proteins aid in the establishment of the division and help segregate various RNAs, proteins, and RNA-protein complexes asymmetrically. Further, organelles such as the endoplasmic reticulum, mitochondria, and the midbody structure have been observed to be inherited asymmetrically, effecting cell identity.

The role of chromatin in epigenetic inheritance

Epigenetic inheritance must occur in the context of chromatin. Chromatin structure, consisting of DNA wrapped by histones and other associated proteins, directly influence gene expression which consequently determines cell identity. The chromatin architecture can be modulated at the DNA level through methylation and at the protein level through PTMs. Before a cell divides, the genetic information is replicated and the chromatin structure is established on the duplicated sister chromatids, which are then recognized and segregated to each daughter cell during mitosis. Recent studies have advanced our current knowledge on how histone and DNA modifications are established and inherited during ACD of stem cells.

The role of histones in epigenetic inheritance

It has been studied how histones are inherited during ACD of Drosophila male germline stem cells (GSCs), which gives rise to a self-renewed stem cell and a differentiating daughter cell [12,13]. By differential label, pre-existing (old) histones and newly synthesized histones display asymmetric inheritance patterns, with the old histones being preferentially maintained in the self-renewed stem cell and the newly synthesized histones being inherited by the differentiating daughter cell. This asymmetry is specific to the H3-H4 tetramer, as H2A and H2B were inherited symmetrically between the two daughter cells [13]. It has been hypothesized that the old histones maintain an epigenetic memory that is passed to the self-renewed stem cell, while the new histones can be modified to induce a differentiation program.

Global asymmetric histone inheritance in male GSCs would require that 1) old and new histones are asymmetrically incorporated into sister chromatids and 2) the epigenetically distinct sister chromatids are recognized and segregated differentially [14] (Figure 2A). Both of these processes have been studied, giving insight into how this asymmetry is established and partitioned.

Figure 2. Asymmetric Histone Inheritance during Asymmetric Cell Division.

A. Two step model for asymmetric histone inheritance. In order to observe global asymmetric histone inheritance as seen during the ACD of male Drosophila GSCs, two steps must occur: first, old and new histones must be asymmetrically deposited onto different sister chromatids during DNA replication (Establishment), and second, these epigenetically distinct sister chromatids must be recognized and segregated asymmetrically during division (Recognition). B. Asymmetric inheritance of CENP-A during division. In Drosophila ISCs, it was found that old CENP-A was retained in the renewed stem cell, while new CENP-A was segregated toward the differentiating daughter cells. Further, CENP-A was found to be inherited asymmetrically in male and female Drosophila GSCs, where the renewed stem cell inherited more CENP-A than the differentiating daughter cells [gonialblast (GB) in male and cystoblast (CB) in female].

Establishment of distinct epigenetic states

There are three unique ways in which pre-existing and newly synthesized histones can be incorporated into newly replicated DNA: 1) The H3-H4 tetramer may split into two H3-H4 dimers, which then tetramerize and are incorporated into chromatin, 2) Pre-existing H3-H4 tetramers stay unsplit and are shuttled randomly onto each sister chromatid along with newly synthesized histones, or 3) Pre-existing H3-H4 tetramers are specifically shuttled toward the leading or lagging strand as newly synthesized histones populate the other strand. Current research supports the inheritance of the old (H3-H4)₂ tetramer without splitting [15]. Thus, the question becomes whether pre-existing and newly synthesized H3-H4 tetramers are symmetrically or asymmetrically incorporated into sister chromatids in different biological contexts.

Recent work has yielded different answers to this question, depending on the system. For example, two recent papers investigated the incorporation pattern of old histone H3 into the newly synthesized DNA in mouse embryonic stem cells and budding yeast, respectively [16,17]. Both studies identified that old H3 are incorporated symmetrically on the leading versus lagging strand; however, in the mouse embryonic stem cells, there was a slight bias toward the leading strand, whereas in yeast, there was a slight bias toward the lagging strand. It is important to note that yeast H3 is an ancient form that is closely related to the histone variant H3.3 in higher eukaryotic systems. Furthermore, both groups found that by impairing certain components of the replication machinery, the respective biases could be boosted, thus resulting in an enhanced asymmetric pattern. This indicates that replication components have an underappreciated role as histone chaperones for balancing old histone incorporation between the two strands. These studies also suggest that there is a molecular control at the replication fork in order to counteract biased parental H3-H4 incorporation during DNA replication in symmetrically expanding cells. These results also provoke the question of how replication-dependent nucleosome reassembly is regulated in asymmetrically dividing cells.

Investigations of histone incorporation during DNA replication in the asymmetrically dividing male Drosophila GSCs use a chromatin fiber technique to directly visualize old versus new histone distribution in replicating or newly replicated sister chromatids [13,18]. These studies have revealed that old H3 histones are preferentially incorporated by the leading strand whereas the new H3 populate the lagging strand (Figure 2A). Additionally, an increase in unidirectional DNA synthesis was detected in these cells, indicating a possible coordinated replication fork movement [13]. It is possible that the strand-specific incorporation and coordinated replication fork movement act together to establish distinct epigenetic states between sister chromatids prior to mitosis; however, the molecules at the replication fork responsible for these processes remain to be studied.

Recognition of epigenetically distinct sister chromatids

To achieve the asymmetric histone inheritance pattern observed during ACD of male Drosophila GSCs, epigenetically distinct sister chromatids must be recognized by the mitotic machinery. Several hypotheses propose that epigenetic differences at chromosomal domains known as centromeres contribute to biased sister chromatid attachment and segregation during mitosis (reviewed in [19]). Recent work has identified a “mitotic drive” in male Drosophila GSCs [20]. This study demonstrates that cis-factors from each sister centromere and trans-acting factors from the mitotic machinery work in concert to ensure the proper attachment and segregation of old versus new histone-enriched sister chromatids during ACD of GSCs (Figure 2A).

Notably, the centromere, defined by a histone H3 variant called centromere protein-A (CENP-A), is where kinetochore protein complexes that are essential to the faithful segregation of chromosomes during cell divisions are nucleated [21,22]. Although each pair of Drosophila sister chromatids show similar levels of CENP-A proteins in symmetrically dividing progenitor germ cells, asymmetrically dividing stem cells show approximately 1.4-fold more CENP-A inherited by the self-renewing GSC than the differentiating daughter cell (Figure 2B). Different amounts of CENP-A between individual sister centromeres can be visualized as early as prometaphase, suggesting this asymmetry is likely established prior to mitosis and becomes detectable when sister centromeres resolve in mitosis [20]. Evidence for trans-acting factors that recognize the CENP-A asymmetry between sister chromatids comes from the observation of temporally asymmetric microtubule activity during mitosis. The microtubules emanating from the mother centrosome located toward the stem cell side first interact with the nuclear envelope, resulting in a polarized nuclear envelope breakdown followed by the preferential attachment of these microtubules to the stronger centromeres, ensuring that the set of sister chromatids with more CENP-A is segregated to the self-renewing GSC. These asymmetries in cis- and trans-acting factors are specific to the asymmetrically dividing GSCs, as symmetrically dividing spermatogonial cells did not show asymmetric sister centromere or asymmetric microtubule activity [20].

Interestingly, a similar phenomenon in asymmetric sister centromeres has previously been reported in asymmetrically dividing Drosophila female GSCs [23]. In addition, in asymmetrically dividing Drosophila intestinal stem cells (ISCs), old CENP-A is preferentially retained in ISCs [24]. Collectively, these reported studies demonstrate that CENP-A protein asymmetry at sister centromeres and the sequential activity of mitotic machinery could drive asymmetric epigenetic inheritance in multiple stem cells (Figure 2B).

The role of DNA methylation in epigenetic inheritance

Adding a methyl group to a cytosine base results in 5mC DNA methylation, a relatively stable and heritable epigenetic modification known for gene regulation and transposable element silencing [25]. This mark can be established by de novo DNA methyltransferases, maintained through maintenance methyltransferases, and removed by demethylases [25]. During DNA replication, the 5mC mark would need an active mechanism to be copied onto the newly synthesized strand, and the methylation pattern may be diluted or reestablished without such a mechanism. Intriguingly, differential DNA methylation has been reported between sister chromatids in stem cells, with the highest asymmetry in pluripotent blastocysts and the lowest asymmetry in cultured HEK293 cells [26]. It has been speculated that this asymmetry is established by passive dilution or active demethylation during DNA replication. Using bisulfite sequencing, it has been shown that DNA replication can lead to strand biased DNA demethylation, suggesting a possibility to create asymmetry between sister chromatids [27]. Furthermore, it has been found that DNA methyl transferase plays a role in the nonrandom segregation of chromosomes during ACD of mouse and human embryonic stem cells, which introduces the possibility that DNA methylation could lead to specific epigenetic inheritance patterns [28].

In addition to cytosine methylation, adenine can be methylated, leading to a mark known as 6mA. Much of the research regarding DNA methylation has been focused on 5mC, but recent studies have started to elucidate the roles of 6mA [29]. It has been demonstrated that 6mA interacts with the H3K4me2 histone modification during intergenerational epigenetic inheritance in C. elegans [30,31]. Furthermore, studies suggest that high levels of 6mA may promote stemness and that removal of this mark could allow for differentiation (reviewed in [32]). These findings present 6mA as an epigenetic mark that influences the determination of cell identities and can be specifically inherited.

Asymmetric inheritance of other factors

While chromatin structure plays a large role in establishing epigenetic states, other factors can also regulate gene expression and function. It has been shown that intrinsic cell fate determinants, such as certain RNAs and proteins, are asymmetrically inherited during ACD [6,33]. Here we will discuss recent studies that show that RNAs, non-histone proteins, and organelles are inherited asymmetrically during ACD to influence the fate of the resulting daughter cells (Figure 1).

Asymmetrically inherited non-histone proteins and RNAs

Many cytoplasmic proteins have been found to segregate asymmetrically during ACD. This has been well studied in Drosophila stem cell systems in which polarity proteins, such as the Par complex proteins, influence the mitotic spindle orientation and affect the inheritance of cell fate determinants like Numb, Prospero, and Brat (reviewed in [6]). The Notch ligand Dill has also been shown to be asymmetrically inherited by one daughter cell during neural stem cell division in the adult mouse brain. This asymmetric inheritance allows the neural stem cell to produce its own niche, as this activates the Notch pathway and allows for stem cell quiescence [34]. Additionally, many proteins are asymmetrically inherited in dividing budding yeast, where mother cell-enriched proteins contribute to aging and daughter cell-enriched proteins aid in bud construction and genome maintenance [35].

RNA binding proteins, such as Staufen identified in both Drosophila and mammalian neural progenitor cells, are found to be asymmetrically segregated during ACD, along with coinherited cargo mRNAs [36,37]. In Drosophila, the inheritance of Staufen and the associated mRNAs promote differentiation and suppress the neural stem cell fate [37]. Many investigations have shown that asymmetrically inherited mRNAs can influence cell fate [25]. For example, asymmetrically inherited ASH1 mRNA represses mating-type switching in yeast [38–40]. It has also been shown that mRNAs are inherited asymmetrically through germ granules in Drosophila and C. elegans, which directly influence germline fates [41,42]. A long non-coding RNA, cherub, has been found to segregate asymmetrically in asymmetrically dividing Drosophila neuroblasts [43].

Asymmetrically inherited organelles

It has been known that the asymmetric inheritance of organelles, such as mitochondria and peroxisomes, influences longevity and survival in yeast [44–46], but recent studies have suggested that organelle inheritance also influences cell fate. For example, in human mammary stem-like cell divisions, daughter cells that inherit fewer mitochondria sustain stem cell properties, but the loss of this asymmetry results in the loss of the correlated stemness traits [47]. In Drosophila proneuronal cells, the Endoplasmic Reticulum (ER) is found to asymmetrically partition during mitosis 14 of the embryo, just before the onset of lineage-specific cell fate determination [48]. Similarly, the ER is partitioned asymmetrically in Drosophila neuroblast divisions, where the neuroblast inherits a larger portion of the ER compared to the differentiating ganglion mother cell [49]. Furthermore, the midbody, the intracellular bridge formed during cytokinesis, can be asymmetrically inherited (reviewed in [50]), as seen in Drosophila GSCs where it is inherited by the self-renewed stem cell in males but by the differentiating daughter cells in females [51]. Loss of this asymmetric inheritance is observed in overpopulated stem celllike cells when overexpressing stem cell niche factors, indicating that this inheritance pattern is correlated with the asymmetric outcome of GSC ACD [51]. Taken together, these results demonstrate that asymmetric organelle inheritance may influence or correlate with the distinct cellular identities during ACD of stem cells.

Mis-regulation of epigenetic states in diseases

The earliest connection between epigenetics and disease was described in the early 1980s, where abnormal changes to DNA methylation was linked to cancer (reviewed in [52]). Since then, many studies further validated this by identifying more connections between abnormal histone PTMs and cancer [53]. Many histone modifying enzymes responsible for placing (“writer”), removing (“eraser”), and reading (“readers”) PTMs are also found to be mutated in cancers such as lung cancer, liver cancer, and lymphomas [54].

Emerging studies also suggest a relationship between the dysregulation of stem cell ACD and tumorigenesis. Genetic studies in the 1970s screening for mutations that lead to brain tumor formation in Drosophila identified Lgl, Dlg, and Brat, all key regulators of ACD (reviewed in [1]). Mutations of these genes result in failure of neuroblast differentiation into normal neurons or glial cells, instead neuroblast overproliferate and cause brain tumor. Subsequent findings demonstrate that mutations in mitotic kinases such as Aurora-A kinase and Polo-like kinase 1(PLK1) also lead to an overproliferation of neuroblasts [55–57]. Moreover, the long non-coding RNA cherub, as discussed above, is normally segregated asymmetrically to the differentiating daughter cell during Drosophila neuroblast ACD. However, cherub has an increased amount of expression in de-differentiated cells, leading to the overproliferation of neuroblasts and brain tumor formation [43]. Collectively, these findings suggest that defects in ACD could be a causal reason of tumorigenesis.

Recently, mutations in histone genes themselves have been reported in several human cancers, including gliomas, sarcomas, head and neck cancers, and various carcinosarcomas [11]. These ‘oncohistone’ mutations can occur in both the N-terminal tail and the globular C-terminal histone fold domains of all four canonical histones (H3, H4, H2A, and H2B), which often occur at or near residues carrying important PTMs and in turn might impact reading or writing of the corresponding modifications. For example, several well-documented oncohistones are the dominant mutations at the Lys27 (H3K27M) and Gly34 (H3G34V/R) residues of histone H3, which are found in 78% of diffuse intrinsic pontine gliomas (reviewed in [58]). Additionally, the histone variant H3.3 is mutated at Lys25 (H3.3K25M) in 95% of chondroblastomas [11] (Figure 3A).

Figure 3. Oncohistone Mutations Contribute to Tumor Formation.

A. Association of oncohistone mutations with cancerous tumors. Mutations in histone genes have been associated with human cancers, and many of these mutations directly affect residues that receive different PTMs. Well documented onchohistone mutations in H3 and variant H3.3 include K27 (K→M), G34 (G→R/V/W/L), and K36 (K→M), that are associated with various cancers including gliomas, chondroblastomas, osteosarcomas, and carcinomas. B. Oncohistone mutation causes tumor formation in Drosophila GSCs. It was found that a mitotic phosphorylation at Thr3 on the H3 tail distinguished old and new histone-enriched sister chromatids during ACD of GSCs. When the Thr residue was mutated to an unphosphorylatable Ala, asymmetric histone inheritance was disrupted, leading to symmetric inheritance of old and new histones and caused tumor formation or stem cell loss.

Another oncohistone mutation has been identified in the Drosophila male germline [14]. Xie et al. reported that histone H3 threonine 3 phosphorylation distinguishes old versus new H3-enriched sister chromatids during ACD of GSCs. Expressing a mutated histone that disrupts this phosphorylation (Thr3 to Ala mutation or H3T3A) leads to progenitor germ cell overproliferation as well as germ cell loss, likely due to disruptions of asymmetric H3 segregation and mis-inheritance of proper epigenetic memory in the two daughter cells resulting from ACD of GSCs (Figure 3B). Altogether, these results demonstrate the importance of asymmetric inheritance of epigenetic states in stem cell asymmetric division.

Conclusions and perspectives

A long-standing question in developmental biology is how different cell fates are established, maintained, or changed, when all cells are derived from one zygote. However, studying this question in developing multicellular organisms is challenging, given that most of these decisions are transient in vivo. Since many stem cells divide asymmetrically over a long period of time during development and homeostasis, these systems provide a unique opportunity to trace epigenetic inheritance under physiological conditions and examine the pathological consequences that occur from disrupting these processes.

In this review, we discuss findings on asymmetric inheritance of epigenetic states and the underlying mechanisms during ACD of stem cells. The emerging evidence has shown that many intrinsic factors are asymmetrically inherited to regulate the identity of the two daughter cells following ACD. Many of these features are specific to ACD and not observed during SCD. In the future, more studies will be needed to address how extrinsic cues may take part in this process and better understand the biological significance of inheriting unique chromatin landscapes in the two cells arising from ACD. Presumably, changes in chromatin states lead to changes in gene expression. Therefore, inheriting differential epigenetic states during ACD could be an elegant mechanism in establishing the unique transcriptional programs that define each daughter cell’s distinct fate. One of the two cells from ACD undergoes dramatic transcriptional changes required for differentiation and the other cell maintains the transcriptional identity as a stem cell.

Newly developed high sensitive genomic techniques such as single cell (sc) RNA-seq [59], scATAC-seq [60,61], sc chromatin immunocleavage sequencing (scChIC-seq) [62,63] or CUT&RUN [64,65], and imaging based methods such as Super Resolution Chromatin Fiber [66], fluorescence in situ hybridization (FISH), ATAC-see [67], in combination with ever improving methods to purify or obtain read-out only from cells of interest, will provide a better understanding of the distinct epigenetic states being transmitted and the downstream effects on gene expression patterns in ACD. In addition to these new genomic techniques, new advances in high spatial and temporal imaging technology methods as well as single-molecule biophysics will further contribute to our understanding of how distinct epigenetic states are inherited and how they function in cells in vivo. Additionally, new advances will lead to a better understanding of how misregulation of asymmetric epigenetic states contribute to cancer and how to reverse it. The mechanisms of oncohistone mutations have the potential to open an entirely new avenue of research focused on understanding how epigenetic mis-regulation and cancer progression or tissue dystrophy are mechanistically linked. Using model systems to examine histone mutations will allow more investigations into their molecular and cellular mechanisms and will shed new light on the treatment of many related human diseases.