Cover Page
The following handle holds various files of this Leiden University dissertation:
http://hdl.handle.net/1887/62204
Author: Chen, X.
Title: Determinants of genome editing outcomes: the impact of target and donor DNA structures
Issue Date: 2018-05-16
Summary and
Concluding Marks
C
linical treatments are transitioning from traditional “one-fits-all” chemical drugs, to personalized or precision medicine based on the genetic informa- tion of each patient. Either in biomedicine as a whole or in precision medi- cine exclusively, genome editing technologies are going to play an ever-increasing role by helping decipher the function of genomic sequences and by providing the tools for permanently correcting the underlying cause of genetic disorders - the ide- al situation one can envision in the future. Related to this, engineered RNA-guid- ed CRISPR-Cas9 nucleases are becoming one of the main driving forces to bring these ideas into reality. However, there are still crucial pitfalls associated with CRIS- PR-Cas9 nucleases that need to be tackled before their implementation as agents in standard-of-care therapies. These bottlenecks include ineffective cellular delivery of the genome editing reagents, their off-target DNA cleaving activities and the ineffi- cient and/or imprecise the exogenous DNA incorporation into the human genome.To tackle these issues, researchers are developing improved delivery agents, scal- ing-up their production and, critically, enhancing the specificity of double-stranded DNA break (DSB) formation. Finally, researchers are also dissecting the cellular fac- tors that can affect genome editing outcomes, notably among these, the chromatin status of target chromosomal sequences. It is under this context that, this thesis has investigated, on one end, a DSB-free genome editing strategy coined in trans paired nicking, and on the other, the impact of higher-order chromatin conformations on the activity of genome editing reagents and the choice of DNA repair pathway, i.e., non-homologous end joining (NHEJ) versus homology-directed repair (HDR).
Chapter 1 reviews the expanding genome-editing tool armamentarium and associ- ated DNA modification strategies, with an emphasis on programmable nucleases based on CRISPR systems. Chapter 2 demonstrates that in trans paired nicking ge- nome editing can precisely incorporate small and large DNA segments at different loci in human cells, including the AAVS1 “safe harbor” locus in pluripotent stem cells.
Different HDR genome editing strategies involving either DSBs or single-stranded DNA breaks (SSBs), were tested in parallel and assessed in terms of their relative efficiency, specificity and accuracy. Conventional genome editing based on DSB formation readily yielded mutation byproducts presumably derived from NHEJ;
genome editing based on SSB formation was, in turn, inefficient. Critically, in trans paired nicking, resorting to the generation of SSBs at chromosomal and plasmid donor DNA using CRISPR-Cas9 “nickases”, led to efficient and seamless genome editing without attendant detection of NHEJ-derived mutations at target alleles.
Consistent with the initial homologous recombination model proposed by Holliday in 1964, it is possible that the coordinated generation of SSBs at target and donor DNA substrates facilitates key rate-limiting HDR steps, e.g., reciprocal single-stranded DNA invasion by both interacting partners. In this regard, further investigations are warranted to tease apart the players and mechanisms governing nick-induced HDR
Summary and Concluding marks / 164
pathways through, for instance, unbiased library screens or knocking-out candidate DNA repair genes. From a practical point of view, it will be interesting to assess the potential of in trans paired nicking for endogenous gene tagging, and editing multi-copy genes or genes with overlapping sequences. Next to this, it will be worth determining whether in trans paired nicking is compatible with Cas9 orthologs and high-specificity Cas9 variants to further improve its versatility and accuracy, respec- tively. Admittedly, a possible drawback of the in trans paired nicking strategy might be its unsuitability for the genetic modification of quiescent cells due to their lack of an active, canonical, HDR machinery. Hence, alternative strategies that can achieve efficient and non-mutagenic genome editing in post-mitotic cells are in demand and should thus be enthusiastically pursued.
Chapter 3 probes, in a quantitative manner, the impact of alternate chromatin states on the performance of genome editing tools, in particular, programmable nucleas- es based on S. pyogenes CRISPR-Cas9 nucleases and Xanthomonas sp. transcription activator-like effector (TALE) proteins. For these experiments, complementary gain- of-function and loss-of-function cellular models were set-up to track targeted gene knockout levels by different CRISPR-Cas9 nucleases and TALE nucleases (TALENs) at euchromatin versus heterochromatin. In these human-cell systems, reporter tar- get alleles transit from compact heterochromatin to relaxed euchromatin through the doxycycline-dependent release of a tTR-KRAB fusion protein whose effector do- main Krüppel-associated box (KRAB) recruits, amongst other heterochromatin-as- sembling factors, KAP-1 and HP-1. The data generated in these reporter systems demonstrates that TALENs and CRISPR-Cas9 nucleases are both significantly hin- dered by KRAB-induced heterochromatin in living cells with the former more so than the latter. This finding is intriguing in view of the fact that, in contrast to TALE proteins, S. pyogenes CRISPR-Cas9 nucleases do not evolve to assess and cleave genomic DNA in the nuclei of eukaryotic cells. It is of note, however, Chapter 4 in- dicate that S. aureus CRISPR-Cas9 nucleases are significantly more affected by closed chromatin than their S. pyogenes counterparts. Whether this differential “chromatin barrier” results from these nucleases having different protospacer adjacent motifs, conformational energy thresholds for ATP-independent helicase activation, or other downstream processes, warrants further investigation.
As aforementioned, a preeminent shortcoming of RNA-guided nucleases is their off-target activity. To address this issue head-on, researchers are developing new reagents and strategies to enhance the specificity of genome editing protocols. For instance, target site discrimination by RNA-guided nucleases can be improved through rationally engineered high-specificity Cas9 variants, offset RNA-guided
“nickase” pairs, and truncated gRNAs (Tru-gRNAs) whose shorter spacer sequenc- es often bias DNA binding towards fully complementary sequences. Chapter 3 and Chapter 4 demonstrate that, albeit to different extents, all these high-specificity
RNA-guided nuclease platforms suffer from having their target sequences embed- ded in KRAB-induced heterochromatin. In fact, with the exception of Tru-gRNAs, heterochromatin seems to impinge a higher activity barrier to these high-specificity platforms than to conventional RNA-guided nucleases. Future studies should seek to determine whether, not only on-target, but also off-target profiles of program- mable nucleases change due to different epigenomes, such as those characteristic of specific cell types and cell differentiation stages.
Using the same tTR-KRAB reporter cells, Chapter 5 studys the impact of alternate chromatin conformations on the relationship between NHEJ versus HDR in genome editing. In these experiments, donor DNA templates were provided in the form of single-stranded oligodeoxyribonucleotides, standard plasmids or Integrase-defec- tive lentiviral vector (IDLV) genomes. In contrast to the observation that hetero- chromatin partially impairs NHEJ-mediated gene editing (Chapters 3 and 4), the absolute levels of HDR-mediated gene editing are similar at euchromatin and het- erochromatin. Hence, the balance between NHEJ and HDR shifts towards HDR when target sequences transit from euchromatin to heterochromatin. As corollary, when compared to euchromatic sites, at heterochromatic sites, there are fewer muta- genic NHEJ-derived indels for each HDR-derived gene editing event.
Although the “downstream” chromatin transition in tTR-KRAB reporter cells entails recruiting endogenous chromatin-remodeling complexes and establishing “physio- logical” KRAB-dependent chromatin states, the “upstream” epigenetic regulation of target sequences is surrogate in character. Thus, it should be insightful to com- plement the data from tTR-KRAB-regulated reporter alleles with those from endog- enous loci harboring different epigenetic marks and chromatin organizations. This line of inquiry should profit from the expanding set of epigenetic-remodeling tools which, in a targeted manner, can modulate DNA methylation and/or post-transla- tionally alter nucleosome compositions, e.g., histone acetylation/deacetylation. Col- lectively, the data presented in Chapters 3, 4 and 5 raise the possibility for directing genome editing outcomes by controlling the higher-order epigenetic states of target sequences. In particular, “opening-up” a heterochromatic target site might increase gene knockout efficiencies, while transiently “closing” a euchromatic target site might reduce NHEJ-derived indels upon HDR-based genome editing. Such ancillary approaches might become feasible with the aid of the aforementioned toolbox con- sisting of epigenetic-remodeling programmable factors. Of note, besides higher-or- der chromatin structures, recent in vitro studies have demonstrated that nucleosome positioning can, per se, influence target DNA cleavage by CRISPR-Cas9 nucleases.
Interestingly, modulating the dynamics of nucleosome positioning in vitro through ATP-dependent chromatin remodelers greatly enhanced target DNA cleavage by CRISPR-Cas9 nucleases.
Summary and Concluding marks / 166
Chapter 6 provides an overview of the main characteristics, pros and cons of com- monly used viral vector systems that are being converted into delivery vehicles for genome editing tools. In particular, IDLVs, adeno-associated viral vectors and ad- enoviral vectors. As a case-in-point, Chapter 7 covers the emerging investigations using viral vectors for the ex vivo or in vivo correction of defective dystrophin-encod- ing gene causing the lethal muscle-wasting X-linked disorder Duchenne muscular dystrophy (DMD). Regarding the in vivo delivery of programmable nucleases, it is important to consider the potential immunogenicity of genome editing reagents and delivery vehicle components, e.g., bacterial Cas9 proteins and viral vector capsids, respectively. In fact, recent experiments have identified pre-existing humoral and cellular immune responses to S. pyogenes and S. aureus Cas9 nucleases in the human population. These aspects are especially relevant in in vivo settings where genome editing components directly encounter the recipient’s immune system.
In conclusion, the interplay between different donor DNA structures and target chromatin environments modulates the outcomes of genome editing in terms of their efficiency, specificity and accuracy. Hence, the findings of this thesis provide a new path for improving these three key parameters that underlie robust and seam- less genetic modification of human cells.
