Transposable Elements in Early Human Embryo Development and Embryo Models

Abstract

Transposable elements (TEs), long discounted as “selfish genomic elements,” are increasingly appreciated as drivers of genomic evolution, genome organization and gene regulation. TEs are particularly important in early embryo development, where advances in stem cell technologies, in tandem with improved computational and next-generation sequencing approaches, have provided an unprecedented opportunity to study the contribution of TEs to early mammalian development. Here, we summarize advances in our understanding of TEs in early human development and expand on how new stem cell-based embryo models can be leveraged to augment this understanding.

INTRODUCTION

A considerable portion of mammalian genomes are composed of Transposable Elements (TEs); in humans, the most recent assemblies identify ~53% of the genome as being composed of TEs [1,2]. There are two major classes of TE’s: retrotransposons, which involve an RNA intermediate that is reverse transcribed during the process of transposition, and DNA-only TEs, which do not use reverse transcriptase for transposition. The most prolific colonization of mammalian genomes has been achieved by retrotransposons, which can broadly be classified into Long Terminal Repeat (LTR) and non-LTR TEs (Figure 1). Non-LTR TEs include families such as Long interspersed nuclear element (LINE) and short interspersed nuclear elements (SINE), while LTR transposons include the Endogenous Retroviruses (ERVs) families. ERV integrants consist of pro-viral sequences, often flanked by regulatory LTRs, as well as solo-LTRs, which have lost their associated pro-virus through recombination or degradation [3]. In humans, most non-LTR and LTR-family TEs are no longer capable of transposition, while some families such as the LINE1 Human Specific (L1Hs), Alu SINE elements, and the composite retrotransposon family SINE-VNTR-Alu (SVA), a SINE derivative, remain mobile in humans [4].

Figure 1. Schematics of Retroviral Class Elements in Humans:

A) (Upper) A typical Human Endogenous Retrovirus (HERV) including (from left to right): Target Site Duplication (TSD), Long Terminal Repeat (LTR), a proviral genome with gag, pro, pol and env genes (Lower) A schematic of the LTR and it’s regulatory regions. Solo LTRs are highly related to those found in the provirus. B) (Top) Long Interspersed Nuclear Element (LINE) containing 2 open reading frames in the sense orientation and one in the anti-sense orientation. (Middle) A Small Interspersed Nuclear Element (SINE), a non-autonomous element which is reliant on LINE activity for mobility. The Alu element is a subfamily of SINEs and is nonautonomous but remains mobile in the human genome. (Bottom) The SINE-VTNR (Variable Number of Tandem Repeat)-Alu (SVA) element, which contains a SINE core repeat, and Alu-like region (antisense to the remainder of the element), a VNTR repetitive region and a SINE-R region, which is derived from HERVK10 and contains homology to LTR5Hs. C) Modes of cis-regulatory TE activity showing (Upper) enhancer activity where H3K27ac-marked TEs are bound by transcription factors (Middle) Promoter-acting TE, where an upstream integrant acts as a promoter for a downstream gene. (Lower) A TE acting as a 3D genome regulatory element.

During early embryogenesis, dynamic changes in RNA expression from TEs is associated with key developmental progressions in totipotent and pluripotent cells including at the time of zygotic genome activation (ZGA), the conversion of totipotent cells in the morula to pluripotent pre-implantation epiblast cells of the blastocyst, as well as during early embryo development including formation of primordial germ cells (PGCs) [5–12]. Studying TE expression and epigenetic regulation in human pre-implantation embryos is possible but limited due to the small number of embryos donated to research following in vitro fertilization (IVF). As described below, most studies on TEs in early human embryos involve tracking RNA expression and chromatin changes. However, a major discovery revealed that pro-viruses originating from ancient HERVK pro-viral genomic integrants are capable of generating viral particles during blastocyst formation when studied in vitro [13].

Evaluating the role of TE’s in post-implantation human embryo development poses even more challenges, especially due to highly limited availability of human embryo samples. To fill this gap in knowledge, human pluripotent stem cells, namely human embryonic stem cells (hESCs) and induced PSCs (hiPSCs), are used to model pre- and post-implantation states of pluripotency, emerging somatic lineages, and PGCs. As described below, the field has used relatively simple two-dimensional (2D) models to generate the foundational knowledge regarding expression, epigenetic state and transcription factor binding to TEs in early human development. However, stem cell-based embryo models (embryo models), which recapitulate various three-dimensional aspects of early human embryo development represent the next frontier for evaluating the functional role of TEs in regulating human biology during this key stage of embryo development.

It is now appreciated that some TEs can serve as cis-regulatory elements, and therefore have the potential to drive cell-type specific gene expression. In both mice and humans, extensive binding of transcription factors to ERV-family elements, especially the key regulators of pluripotent stem cell self-renewal (NANOG, OCT4 and KLF4), have driven the hypothesis that TEs contribute to gene regulation in the pluripotent state of mammals [14]. Further supporting this view, many ERV LTRs harbor multiple pluripotency factor binding sites, and Chromatin immunoprecipitation followed by sequencing (ChIP-seq) has shown that ERV LTRs are, in many cases, co-bound by multiple pluripotent transcription factors [14,15]. In addition to being transcription factor-bound, specific families of ERVs are also enriched with chromatin and histone modifications associated with active states of gene expression [11,16–18]. Evidence for a role of TEs in 3D genome organization exists as well, as CTCF binding sites are found in TEs of the LINE, SINE and LTR superfamilies. These sites are thought to be functional, as alterations in 3D genome organization are observed when TEs are altered via CRISPR/cas9 editing [19,20]. These observations support a role for TEs as potential cis-regulatory enhancer/promoter elements in regulating states of pluripotency in the early human embryo.

Epigenetic silencing of ERVs, and therefore the decommissioning of active ERV LTR-family enhancer/promoter elements during early embryo development likely involves Kruppel Associated box Zinc-finger proteins (KRAB-ZFPs), a broad family of DNA binding proteins that bind TEs and work in tandem with TRIM28 (KAP1) to facilitate the deposition of repressive H3K9me3 heterochromatin at the targeted ERVs [21–24]. Enrichment of H3K9me3 at ERVs is highly dynamic during early human embryo development, beginning at the 4-cell stage before ZGA, and through blastocyst formation [7,8]. This is important as the human genome is largely demethylated during these embryonic stages [25], suggesting that H3K9me3 likely serves as the predominant repressive epigenetic modification responsible for dynamically silencing TEs in the early embryo. Supporting this hypothesis, ultra-low input ChIP-seq for H3K9me3 in pre-implantation embryos has revealed that intergenic ERV LTRs are progressively enriched with H3K9me3 from the 4-cell stage to formation of the blastocyst [7,8]. Notably, some primate and hominoid-specific TEs including LTR5Hs, LTR7B and SVA_D are not marked with H3K9me3 by the 8-cell stage and are enriched for pluripotency factor motifs. In particular, SVA_D elements are highly enriched for DUX motifs, suggesting a cis-regulatory role for these elements [7,8]. To evaluate the necessity of TE remodeling at the 8-cell stage, Yu et al. performed CRISPRi-mediated repression of SVA_D elements in human embryos at the 4-cell stage, which resulted in a developmental block at the 8-cell stage, [8] indicating SVA_D expression, or the chromatin state at SVA_D is involved in human ZGA.

Similarly, there is evidence that TEs function in the morula as totipotency is extinguished in order to give rise to trophectoderm and ICM of blastocysts. Specifically, chromatin accessibility of morula cells reveals two distinct chromatin states; 1) cells where regions of high chromatin accessibility are enriched for transcription factor motifs involved in trophectoderm formation (GATA and TEAD families), and 2) cells where regions of high chromatin accessibility are enriched for transcription factor motifs associate with pluripotency (NANOG, SOX2 and KLF4). Interestingly, cells of the putative outer morula (the ones fated to become trophectoderm), acquire H3K9me3 at hominoid-specific ERVs, including those of the ERVK family (MER11B, MER11C LTRs and HERK9-int) and ERV1 family (LTR12). Footprinting analysis of the putative ICM reveals that these ERV LTRs are bound by pluripotency factors OCT4 and SOX2 [8]. Taken together, these observations support the hypothesis that targeted silencing of the aforementioned ERVs safeguards specification of extraembryonic lineages by precluding ectopic activation of the pluripotency program [8]. Collectively, these results suggest that active epigenetic remodeling of TEs is important for early human embryonic development.

Although RNA expression and epigenetic modifications can be mapped to TE families and subfamilies, a caveat exists when studying the role of individual TE integrants and regulation of neighboring gene expression. The repetitive nature of TEs, combined with the short-read lengths acquired using traditional Illumina sequencing precludes accurate mapping to unique genomic sites [26]. Furthermore, in the case of mobile elements, such as L1Hs and Alu, precise mapping of new integrants across individuals using a reference genome is impossible as the reference would likely not include the position of person-specific integrations. Still, computational advances and long-read sequencing (discussed below) are beginning to address these shortfalls, promising better platforms for interrogating the gene regulatory role of TEs in human biology.

In addition to TE regulation at the chromatin level, it is now appreciated the N6-methyladenosine modification to endogenous viral mRNA is also a potent modality of ERV (and LINE1) regulation. Evidence in mouse pluripotent stem cells has shown that ERV RNAs transcribed from the retrotransposition-active IAP family are heavily marked by m6A due to activity of the METTL3/MEETL14 heterodimer. m6A-marked ERV RNAs are then bound by YTHDF1/2/3, m6A readers, which in turn results in clearance of viral mRNAs [27]. While little is known about the role of m6A in regulating ERVs, SVAs or LINE elements in the early human embryo, it is apparent that analysis of m6A at TEs during early human embryo development and human pluripotent stem cells is needed.

Evaluating TEs using Human Pluripotent States in culture

Evaluating TE expression, epigenetic regulation and function in pre-implantation and early post-implantation human embryos is challenging. Therefore, TE analysis in hESCs and hiPSCs has emerged as an important in vitro model to understand the role of TEs in pluripotency. Pluripotency is a spectrum of states from naïve to primed, with three major states of pluripotency (naïve, formative and primed) successfully captured as self-renewing stem cells in vitro. A fourth state of pluripotency, called latent pluripotency, occurs in PGCs and is captured in vitro with the differentiation of PGC-Like cells (PGCLCs) [28–36]. (Figure 2) Culture of human pluripotent stem cells in naïve conditions recapitulates the transcriptome and key epigenomic features of cells from the morula at Carnegie Stage 2 (CS2) and pre-implantation epiblasts at CS3-CS4 (Figure 2). Primed and formative culture conditions recapitulate the post-implantation epiblast at CS4-CS5 [37,38]. The differentiation of hPGCLCs in vitro results in the formation of germ cells equivalent to those found in post-implantation embryos between CS5-CS7 (Figure 2) [10,39,40]. In the following section, we will highlight key studies evaluating TE’s in the four states of human pluripotency.

Figure 2: In Vitro and Stem-Cell Based Embryo Models:

Top) 2D-culture in vitro models representing various states in embryonic development paired with their embryonic time-matched counterparts below (In vivo). From left to right: 8-Cell Like Cells (8CLCs), non-renewing transient subpopulations in naïve human pluripotent stem cell (PSCs) cultures. Naïve PSCs, which represent the inner cell mass of the pre-implantation epiblast. Naïve cells are marked by high expression of HERVK and functional reliance on LTRT5Hs. Formative cells, which can be stably cultured but have not yet been identified in in vivo human embryos, thus the post-implantation time point they may correspond to is unknown (?) . Primed PSCs represent cells of the post-implantation epiblast, and are marked by high expression of HERVH and functional activity of LTR7(up1/2) integrants. Far right is the PGCLC aggregate model, in which PGCLCs occupy a state of latent pluripotency. CS3-CS7 embryos are shown as a lateral cross-section.

Bottom) Stem Cell-based Embryo models (Left to right): Blastoids represent the pre-implantation embryo. The Posteriorized Embryonic-like Sac (P-ELS) model involves gradients of BMP4 exposure driving formation of the amnion, hPGCLCs and expression of T in the pre-streak Epiblast. Two variation of a 3D-printed embryo model and a gastruloid are shown. 1) lineages derived by Warmflash 2014, from outer to inner are trophectoderm-like, Endoderm like, Primitive-streak like, and ectoderm like. 2) Model of Jo et al. 2022, where they observe (from outer to inner) amnion-like cells, PGCLCs, and ectoderm-like cells. 3) A gastruloid model representative of BMP4-treated gastruloids in Minn et al. 2020 (from outer to inner): amnion-Like and trophectoderm-like cells, endoderm and PGCLCs, mesoderm and primitive streak like, epiblast-like and ectoderm-like cells. Pluripotent pre-implantation lineages are noted in the key as well as primitive steak. Other cells represented in the models are shown in grey.

At the RNA level it is now appreciated that the different states of pluripotency in vitro express differing repertoires of TEs, with corresponding ChIP-seq and chromatin accessibility data revealing the epigenetic status of TEs in each state [17,31,32]. For example, naïve pluripotent stem cells exhibit high expression of the HERVK provirus (HERVK-int) and LTR5Hs, a Hominoidea -specific LTR of the HERVK(HML2) family [31]. RNA expression of LTR5Hs in naïve human pluripotent stem cells is coupled with hypomethylation and enrichment of H3K27ac at these TEs. Interestingly, LTR5A and 5B, two older members of the HERVK(HML2) family first identified in Hominoidea (Apes) and Catarrhini (New World Monkeys), respectively, are not highly expressed or marked with H3K27ac in the naïve state. Members of the SVA family, specifically SVA_D, are also expressed by naïve human pluripotent stem cells, and these integrants are correspondingly enriched in H3K27ac [16,31]. This is of particular interest, as the HERVK10-derived SINE-R region of the composite SVA_D element shares significant homology with LTR5Hs (Fig. 1), meaning this region of SVA_D likely recruits the same transcription factors and KRAB-ZFPs as LTR5Hs [16,41]. Functional studies in naïve human pluripotent stem cells using CRISPRi to target LTR5Hs and SVA_D results in wide-spread gene deregulation, especially of genes of involved in 3D genome organization, cell polarity and lineage restriction [16,42]. Thus, naïve human pluripotent stem cells do not simply express LTR5Hs and SVA_D, but are also reliant on chromatin remodeling and/or RNA expression of these TEs for proper maintenance of the naïve state.

Naïve hESCs also have high accessibility and expression of LTR7Y, a regulatory element of the HERVH provirus. While some LTR7B and some LTR7 elements exhibit H3K27ac enrichment in the naïve state, it is the LTR7Y elements that are highly enriched in H3K27ac, and are bound by KLF4, KLF5 and NANOG [16–18]. Notably, different naïve culture conditions lead to differences in LTR7 family expression, with 5i/L/A conditions resulting in higher LTR7Y expression and H3K27ac, while 3i/L naïve conditions lead to higher expression of LTR7. This is likely a result of 5i/L/A conditions and 3i/L naïve conditions capturing different states on the pluripotency spectrum. Indeed, the responsiveness of LTR7, LTR7Y and LTR7B to different naïve culture conditions likely suggests they are highly dynamic and sensitive to developmental progression through the different states of pluripotency in the early embryo.

As the human blastocyst implants during CS4 to CS5, the naïve pluripotent cells of the pre-implantation epiblast convert to a state of pluripotency referred to as formative. This is a state competent to differentiate into somatic cells and PGCs following exposure to appropriate differentiation cues [32]. In the mouse embryo, formative pluripotency corresponds to cells of the post-implantation epiblast at the early egg-cylinder stage. In humans, the location of formative cells in the post-implantation embryo is not yet known, but likely corresponds to post-implantation pluripotent cells somewhere between day 8–12 post-fertilization. Despite this unknown, formative cells have been reported in culture [32]. Formative cells share some TE overlap with naïve human pluripotent stem cells, including expression of HERVK, however formative cells uniquely express LTR6A [32]. Curiously, formative cells share expression of a number of KRAB-ZFPs with naïve cells, including ZNF676, ZNF560, ZNF528. However, much of the formative gene transcriptional profile is shared with the primed state [32]. This similarity in gene expression profiles elevates LTR6A as a key marker of the formative state, and underscores the utility of TEs as markers of developmental populations which are hard to capture in vivo.

Primed hPSCs, which recapitulate post-implantation epiblast cells, express pro-viral integrants of the HERVH and HERVK families [17,43]. Furthermore, the epigenome of HERVH associated LTR regulatory sequence LTR7 are hypomethylated and enriched with H3K27ac in the primed state [17,31,44–47]. In a recently phylogenic analysis of HERVH and LTR7, new subfamilies were identified, and their evolutionary trajectories defined. Of particular interest are Homininae-specific LTR7up1 and LTR7up2, which were previously clustered with LTR7. Re-analysis of LTR7Y, LTR7up1 and LTR7up2 using these updated phylogenies revealed that LTR7Y elements are expressed by naïve human pluripotent stem cells, whereas it is the LTR7up families that are highly expressed in primed cells, with LTR7up1 and LTR7up2 accounting for most of the observed LTR7 transcription in this state. Consistent with these observations, LTR7up1 and LTR7up2 are also enriched with H3K27ac, and are bound by pluripotency factors NANOG, SOX2 and FOXP1 in primed cells [48]. Altogether, this supports the idea that young TEs of the LTR7up families likely acting as cis-regulatory regions in primed state pluripotency.

Latent pluripotency is a feature of PGCs. During embryo implantation and progression of pluripotent cells from the naïve to primed state, hPGCs are specified [49,50]. Based on primate models, this is speculated to occur at CS5, coincident with amnion and extraembryonic mesoderm formation but initiated before the formation of the primitive streak [39,51]. Studying hPGC specification in vivo is extremely challenging due to the rarity of early human post-conceptus tissues donated to research. A recent paper profiling single cells of a CS7 post-implantation human embryo demonstrated that hPGCs exhibit a suite of TEs with similarities to naïve cells [9,10,40]. Specifically, this data set revealed that CS7 hPGCs in vivo express HERVK, LTR5Hs and SVA_D [40].

In order to model the expression and function of TE’s during hPGC specification, the differentiation of hPGCLCs from primed human pluripotent stem cells is used (Figure 2) [33–35,52]. Induction of hPGCLCs in vitro leads to a dynamic change in TE expression relative to the primed state, with one of the most significant changes involving up-regulation of LTR5Hs. At the epigenetic level in hPGCLCs, LTR5Hs sequences become hypomethylated, enriched in H3K27ac, and bound by NANOG, SOX15, SOX17 and TFAP2C transcription factors all of which are necessary for hPGC specification and maintenance [10,49,53,54]. Using CRISPRi to dampen LTR5Hs accessibility before hPGCLC induction results in a significant reduction in the competency human pluripotent stem cells to differentiate into hPGCLCs [10]. This suggests that LTR5Hs is necessary either for acquisition or maintenance of latent pluripotency upon hPGCLC formation.

TE Expression in Totipotent-like cells in culture

Following fertilization (CS1), the maternal and paternal genomes undergo epigenetic reprogramming to create a totipotent state competent for ZGA, a process where embryonic transcription is activated in the diploid embryo [5,55,56]. In humans, ZGA occurs in the 8-cell embryo, while in the mouse it occurs in the 2-cell embryo. In both mouse and human embryos, ERVL (MERVL/HERVL) family elements are broadly derepressed with ZGA, coincident with high expression of ZSCAN4 and the TRIM family of genes, as well as expression of Dux/DUXA [5,6,57,58]. Using these characteristics, human 8-cell like cells (8CLCs) have been identified sporadically in hPSCs cultured under a variety of naïve culture conditions, including 4 inhibitor (4i), 5i, NHSM (Naïve human stem cell medium), PXGL (PD0325901, Gö6983, XAV939, LIF ), t2iLGö [6] or from naïve hPSCs reverted to a defined media called e4CL (enhanced 4 chemicals + LIF: PD0325901, IWR1, DZNep and TSA), which increases the subpopulation of 8CLCs [5]. Notably, the transcriptomes of 5i and PXGL 8CLCs are closest to those of e4CL, although e4CL was much more potent at generating 8CLC cells [5]. 8CLCs have also been identified in populations of pre-Epiblast-like PSCs (prEpiSCs), which are cultured in media containing inhibitors of MEK (GSK1120212), WNT (XAV939), PKC (Go6983), Src (A419259) and EED/H3K27me3 (DZNep) [12]. In all conditions, 8CLCs are transient and metastable, interconverting between 8CLCs and naïve PSCs in vitro. Therefore, like totipotent cells of the embryo, 8CLCs lack the capacity for self-renewal [56]. Despite these challenges, reliable isolation of 8CLCs and the ability to identify them in single-cell transcriptomic experiments have enabled characterization of their unique TE-repertoire. TE expression in 8CLCs compared to the naïve cells reveals a general enrichment of ERVs including HERVK, but with a marked increase in expression of HERVL and its associated LTRs, MLT2A1 and MLT2A2 [5,6,12]. This is in line with TE expression in the 8-cell human embryo [5,6,12]. Importantly, neither expression nor chromatin accessibility of HERVL, MLT2A1 and MLT2A2 were observed in any other naïve hPSCs in culture, suggesting expression of these LTRs is tightly associated with the 8CLC state and not naïve pluripotency [5,6,12,45,59,60].

Emerging Technologies for Understanding TEs in early Human Embryo Development

Although human pre-implantation embryo development can be studied using IVF embryos, there are limitations in the number of embryos that can be used and, in some jurisdictions, research with human embryos is not allowed [61]. This has led to the emergence of embryo models as critical avatars of human embryo development [62]. New embryo models recapitulate some key aspects of early human development, but not all. The closest representation of an intact embryo are blastoids, resembling pre-implantation blastocysts consisting of spatially organized pre-implantation epiblast, primitive endoderm and trophoblast [63–66]. These emerging models represent a powerful tool in which the function of TE’s can be evaluated using approaches such as CRISRi of SVA_D elements as in [7]. Although powerful, the blastoid model cannot recapitulate the embryonic events in an 8-cell embryo prior to blastocyst formation.

For post implantation stages, embryo models including gastruloids [67,68], post-implantation amniotic sac embryoids (PASE), posteriorized-embryonic-like sacs (P-ELS) [69,70], and micropatterned self-organizing discs [36,65,71] have emerged. These tools enable more thorough interrogation of TEs in lineage specification after embryo implantation, which is of particular importance as the divergence and origin of some post-implantation lineages, such a hPGCs, remains unclear. These models, coupled with advances in sequencing technologies hold promise in understanding the role of TEs throughout development, including a more complex understanding of TE epigenetic reprogramming and its role in properly regulating early embryonic lineage restriction.

In tandem with advancing in vitro models, there has also been a rapid expansion of low-input and single-cell sequencing technologies. Single-cell RNA-seq has become important in examining both rare populations, such as 8CLCs in [5,6], as well as in-depth profiling of the transcriptome of rare clinical samples, such as SMART-seq of the human Carnegie Stage (CS) 7 embryo [40]. Third-generation long-read sequencing technologies also hold promise for better characterization of the RNA expression and mapping of TEs, helping overcome long-standing challenges in mapping to specific genomic sites. For instance, Single-Cell Long-Read RNA-seq (CELLO-seq) [72] can be used to map TE transcripts to their original loci, helping to resolve questions about the cis- and trans-acting roles of TEs. While the higher error rate of third-generation sequencing still imposes some technical limitations on its use, it also harbors distinct advantages, such as the ability to read DNA 5-methycytosine and RNA N6-methyladenosine natively without enzymatic modification to nucleic acids during library preparation [73,74].

Another significant advance in sequencing is the advent of single-cell multi-omics, which are powerful tools to understand how transcriptomes, chromatin accessibility, and epigenetic landscapes are remodeled at the single-cell level. Multi-omic approaches, such as Low-Input Chromatin Accessibility and Transcriptome (Li-CAT-seq) sequencing as in [59], have helped link in vitro observations with in vivo data. These data are of particular importance when studying the expression and epigenetic state of TEs in early development. Here, multi-omic approaches which can map transcriptome, chromatin accessibility and DNA methylome at the single cell level such as scNMT-seq [75] and snmCAT-seq [76], hold promise in better understanding the potential roles of DNA methylation in early human development. These advances are particularly powerful when coupled with low-input chromatin profiling technologies such as CUTnTag and Ultra-Low Input ChIP which have enabled high-quality profiling of chromatin with as few as 100 cells [8]. These advances help understand how epigenetic regulation of TEs, including their activity as enhancers and the dynamics of their silencing during differentiation, contribute to mammalian development.

Conclusions

Here, we summarized major changes in the RNA expression and epigenetic state of TEs across early human embryo development and in pluripotent stem cells in vitro. Importantly, we argue here that distinct states of pluripotency can be defined by a unique repertoire of TEs, their expression profile and epigenetic state forming a molecular fingerprint that can be used to discriminate major transitions in embryo development including ZGA, different states of pluripotency and PGCs. While similar epigenetic profiling of other embryonic stages has not been examined in the same way, such an approach holds promise for improving our understanding of how expression and epigenetic reprogramming at TEs could shape other early embryonic lineage decisions in humans. Disrupting TE expression and chromatin accessibility with CRISPRi provides a glimpse into the functional role of TE’s in early human development, and this functional work is likely to expand when coupled with embryo models. It is important to note, however that confirmation of key observations with embryo models will benefit from access to equivalent stages of human embryo development where possible, and with appropriate regulatory oversight and public engagement as to the use of embryos in research.