Keep your distance or lose control Free

In this issue of Blood, Felder et al¹ demonstrate that, for normal developmental control of globin gene expression, the enhancer (locus control region [LCR]) must be located remotely from the promoters of the HBG (fetal) and HBB (adult) globin genes.

Enhancers can regulate genes across substantial genomic distances (up to thousands of kilobases), often activating distal rather than more proximal genes. Chromatin looping brings enhancers into close proximity to their cognate promoters.² At the β-globin gene cluster, the LCR loops to the fetal-type HBG2 and HBG1 globin genes in fetal erythroid precursor cells, whereas in adult erythroid cells it skips over these genes and contacts the more distal HBB gene.³ The LCR promotes transcription of all β-like globin genes but does not possess any intrinsic developmental stage specificity; this is encoded within the promoter-proximal sequences.⁴ Prior work showed that forcing a chromatin loop between the LCR and the β-globin genes can lead to their activation,⁵ which helped to establish that physical proximity is a key factor in enhancer-mediated gene activation.

A vexing question in the field has been why so often evolution has resulted in far genomic separation of enhancers from their cognate promoters, requiring complicated mechanisms to establish specific physical connections. Would it not be much simpler to position the enhancer right next to its target as seen in lower organisms? For the β-globin locus, Felder et al now show that the answer is more complicated. They demonstrate that remoteness of the LCR is not merely an obstacle to be surmounted, but it is a necessity for proper developmental gene control. Using a strategy termed Del2Rec (delete-to-recruit) the authors removed ∼25 kb of genomic sequence to position the LCR within ∼500 base pairs of the HBG2 gene in adult erythroid cells. Even though the binding sites for all known HBG repressors were left intact, HBG2 transcription was strongly induced, accompanied by increased chromatin accessibility and gained “active” histone marks. Expression of HBB was diminished, consistent with gene competition for LCR activity.⁶ In the future, it will be important to understand how LCR proximity can override HBG silencing mechanisms. For example, are key HBG repressors, such as BCL11A, LRF, or NF1A, displaced by LCR proximity, and is HBG DNA methylation lost?

The authors excluded the possibility that potential LCR-initiated read-through enhancer RNAs, which in some cases produce full-length mRNAs when close to a gene,⁷ may extend to the now adjacent HBG2 gene and activate it. Moreover, they ruled out the possibility that the deleted region contains HBG2 repressor activity by showing that inversions (instead of deletions) that appose the LCR with the HBG2 gene but retain the entire genomic interval also activate HBG2 expression. Hence, strong linear separation between the LCR and the HBG2 gene is required to allow repression of HBG2 in adult cells. To determine whether enhancer distance is also essential in a different context, similar experiments were carried out at the α-globin locus. Minimizing the distance between a major α-globin enhancer and the embryonic HBZ promoter activated HBZ expression in adult erythroblasts. Together, these findings assign a previously unrecognized function to genomic sequences with no intrinsic regulatory function: by providing linear separation between enhancers and genes, they enable proper developmental control of gene expression.

An unexpected finding of this study is that the LCR only activated the HBG2 gene but not the neighboring, just slightly more distal HBG1 gene. In normal fetal erythroblasts, the LCR activates both genes. The authors provide a plausible explanation for this discrepancy: chromatin loops that are established via protein-protein interactions are short-lived and dynamic, allowing for oscillating interactions. In contrast, contacts established by permanently placing an enhancer next to a promoter may lack such flexibility and constrain the enhancer’s range of action.

Previous studies in transgenic mice showed that linking the LCR to the HBG gene still allowed for “autonomous” silencing of the HBG gene in adult erythroblasts.⁸ Although HBG silencing was variable between LCR-HBG mouse strains and subject to position effects, these findings suggest that, in vivo, LCR-HBG proximity fails to override the silencing forces at the HBG promoter. To reconcile these seemingly contradictory findings, Felder et al reason that, in the transgenic experiments, the 11-kb spacing between the LCR and HBG gene was greater and thus less capable of surmounting the inherent HBG repression mechanisms. In a related study, Vyas et al⁹ showed that placing part of the LCR (5'HS2) within ∼2 kb of the embryonic HBE gene leads to partial activation upon introduction of the construct into a murine adult stage erythroid cell line. Thus, the ability of enhancers to overcome gene repression depends on both their genomic proximity to target promoters and their intrinsic strength. More generally, this idea is consistent with findings that the effects of enhancers diminish as a function of genomic distance from the promoter.¹⁰ 

Considering the plethora of mechanistic studies on hemoglobin switching, it is likely that multiple forces are at play, including direct promoter-proximal repression, differential LCR usage, and gene competition. Felder et al show that a proper genomic distance of the LCR is needed for these mechanisms to function correctly. More generally, because overpowering the gene-silencing mechanism can occur simply by linear juxtaposition or forced chromatin looping,¹¹ this suggests that, during evolution, enhancer strength, distance, and looping forces have been balanced to enable dynamic gene expression changes during development at multigene loci such as the α- and β-globin gene clusters.

Lastly, the authors raise the possibility of exploiting Del2Rec for the treatment of hemoglobinopathies. Elevating HBG levels is beneficial to patients with sickle cell disease or β-thalassemia. Deletion of the intervening sequence between the LCR and HBG induced HBG transcription to levels even exceeding those obtained by interference with BCL11A, a proven therapeutic target. Clinical application would require that the Del2Rec genome edits can be carried out without detrimental off-target genomic alterations. Other disease-causing genomic rearrangements may similarly be corrected by altering enhancer-promoter arrangements.

In conclusion, Felder et al reveal that DNA sequences lacking recognized regulatory elements are important for precise gene expression by providing genomic distance between enhancers and promoters.

Conflict-of-interest disclosure: The author declares no competing financial interests.