Escribir un comentario

This manuscript investigates whether locus-specific expression of LINE-1 elements contributes to hematopoietic dysfunction in VEXAS syndrome. By integrating single-cell transcriptomic analyses with locus-targeted CRISPRa/CRISPRi, pharmacological perturbation, and functional assays, the authors identify L1-10 and L1-15 as candidate regulators of myeloid-biased hematopoiesis in the context of UBA1 mutation.

The study addresses an interesting and potentially important question: whether specific transposable element (TE) loci act as functional contributors to disease-associated hematopoietic abnormalities, rather than simply serving as markers of cellular stress. Prior work has mostly examined TE families or correlations, so functionally testing individual loci is a clear step forward for understanding how TEs intersect with inflammatory and hematopoietic programs in VEXAS. If confirmed, these findings would position individual TE loci as candidate drivers of HSC dysfunction, and could suggest new intervention points for VEXAS and related clonal hematopoietic disorders. The evidence that L1-10/L1-15 modulate myeloid output is supported by concordant effects of CRISPRa in healthy HSPCs and CRISPRi in UBA1-mutant HSPCs, together with improved engraftment and lineage output in xenotransplanted mice. In contrast, the evidence for direct TE–transcription factor regulation, VEXAS specificity, and clonal dominance is currently less direct.

Major Comments

1. The proposed regulatory relationship between L1-10/L1-15 and SPI1/IRF1 requires further clarification. In figure 1, the authors identify potential transcription factor binding motifs and associations between L1-10/L1-15 and SPI1/IRF1. Figure 3 shows that targeted activation of these loci is accompanied by increased expression of the corresponding transcription factors. These findings support a functional relationship, but do not establish direct transcriptional regulation. The context-dependent effects of TE repression are particularly interesting, as SPI1/IRF1 expression appears to respond differently in UBA1-mutant and healthy HSPCs (Figure 4c-d and Extended Data Figure 5e-h). The authors might clarify how they interpret this difference, and whether the current data support direct or indirect regulatory relationship. If feasible, SPI1/IRF1 binding data at these loci, or SPI1/IRF1 knockdown in TE-activated cells, would help distinguish direct regulation from secondary effects. Alternatively, the authors could qualify the mechanistic language, for example, “associated with” rather than “regulates”.

2. The specificity of locus-targeted CRISPRa/CRISPRi would benefit from further clarification. The strategy of targeting unique upstream sequences and assessing the intended loci by ddPCR is a reasonable approach (Figure 3a-e and Figure 4a-d). However, without information on neighboring gene expression and on how the ddPCR primers discriminate among highly similar LINE-1 copies, it remains difficult to attribute the phenotypes specifically to L1-10/L1-15 transcription. I recommend providing the primer design and locus-discrimination rationale, which is important for interpreting the ddPCR data. Examining flanking-gene expression or a second independent gRNA would further strengthen the evidence if feasible, helping exclude effects on adjacent chromatin.

3. The extent of VEXAS specificity should be distinguished from broader age- or disease-associated TE expression. In figure 2, the comparisons with inflammatory, myelodysplastic, and aging-related datasets are valuable and provide important context for the proposed disease association. However, Extended Data Fig. 3i indicates that L1-15 is also differentially expressed in aged versus young HSCs, so its expression is not exclusive to VEXAS. The Unique Score is calculated as the average log2FC difference for 21 TEs relative to three comparisons while no statistical test or threshold is described, so it is unclear what magnitude indicates specificity. The authors might add a statistical assessment or describe the signature as “VEXAS-associated”, which would preserve the significance of the findings without overstating disease specificity.

4. The pharmacological findings would benefit from a clearer distinction between target-associated effects and broader cellular responses. TAK-243 treatment is associated with increased L1-10/L1-15 expression, whereas Auranofin treatment is reported to reduce their expression and improve functional readouts (figure 3m-n and figure 4k-l). Because these compounds can affect broader cellular processes, the results do not necessarily establish that the phenotypic effects are mediated specifically through the two TE loci. The authors might clarify the evidence connecting TE modulation to the functional outcomes. Dose-response and viability data at the doses used would be most informative, as they would show whether TE changes occur at non-toxic exposures. Otherwise, the pharmacological results would be more appropriately framed as supportive of a potential therapeutic connection rather than definitive validation of a TE-specific mechanism.

5. The proposed role of L1-10/L1-15 in clonal dominance remains an incompletely established interpretation. The Abstract states that the study “provides direct evidence that VEXAS-specific TE govern HSC clonal dominance”, and the Discussion proposes a feedback loop that “drives the expansion” of mutant clones. The colony, lineage-output, and xenograft results provide a basis for proposing that L1-10/L1-15 may contribute to mutant-cell fitness. However, functional improvement in these assays is not equivalent to directly demonstrating clonal selection or competitive dominance. The authors might clarify whether the current experiments directly measure clone abundance or competitive fitness, as opposed to changes in population-level function. If direct clonal evidence is unavailable, distinguishing the demonstrated functional effects from the proposed implications would help readers interpret the findings accurately.

6. The relevance to primary VEXAS disease and the scope of the therapeutic implications could be better defined. Much of the functional work uses experimentally manipulated donor HSPCs, including engineered UBA1-mutant cells, which may not fully recapitulate primary VEXAS HSPCs. Validation in primary patient cells would extend the current conclusions. If such experiments are beyond the scope of the current study, the authors could acknowledge this limitation explicitly and distinguish findings in engineered models from evidence in primary disease.

7. The xenotransplantation results support improved engraftment, but long-term HSC restoration would benefit from more cautious interpretation. Since the endpoint is 8 weeks in figure 4i-4j, the authors cannot yet conclude that long-term self-renewal is restored. I recommend describing the findings as improved engraftment and multilineage output, or including serial transplantation if feasible, which would directly assess durable repopulating capacity.

Minor Comments

8. In the Methods, Stellarscope reassigns multimapping reads. Because L1-10 and L1-15 are highly similar to other LINE-1 copies, readers could better judge the reliability of the locus assignments if the authors report the fraction of reads/UMIs uniquely assigned to these loci, and make the analysis code public (currently “available upon request”).

9. The Discussion refers to “Auranofin mediated enhancement of L1-10 and L1-15 expression”, which appears to conflict with the reduced expression in Extended Data Figure 6j-k. The authors might verify this statement so the translational route matches the Results.

Reviewer expertise and competing interests

My expertise is in hematopoiesis, hematopoietic stem cell biology, inflammation, and aging. Xenotransplantation and single-cell TE bioinformatics are outside my primary areas of expertise. I have no competing interests to declare.

I reviewed this preprint as part of the Transparent and Accountable Peer Review training program. This report was written with the support of my research mentor.

Competing interests

The author declares that they have no competing interests.

Use of Artificial Intelligence (AI)

The author declares that they did not use generative AI to come up with new ideas for their review.