Ikaros Regulates microRNA Networks in Acute Lymphoblastic Leukemia

Round 1

Reviewer 1 Report (New Reviewer)

In this study, the authors identified tumor suppressive and oncogenic miRNA networks regulated by IKZF1 in B-cell acute lymphoblastic leukemia. Findings from this study provided new insights into how loss of IKZF1 function can promote leukemogenesis and may help drive development of new biomarkers for IKZF1-mutated B-ALL cases. 

Given the critical role of IKZF1 in early B cell development, do the authors think similar miRNA networks would be regulated by IKZF1 in normal early B cells or they are only regulated in leukemia? If so, what might be the molecular determinants that drive the differences in normal early B cells and leukemia? No need to do experiments - the authors can include a discussion in the Discussion section. 

Experiments were done using Ph⁺ B-ALL cells. Do the authors think similar results would be obtained using B-ALL PDXs/cell lines driven by other translocations? In other types, could the miRNA networks regulated by IKZF1 be subtype specific? To address these questions, the authors can include a discussion in the Discussion section. 

Turk A et al. (2022, PMID: 35408829) identified miRNA networks associated with different types of leukemias. Could the authors comment on findings from Turk A et al. (2022)? Any overlap or differences? 

Author Response

We are grateful to the reviewer for their helpful suggestions. We provide a point-by-point response to each of their comments below:

1. Given the critical role of IKZF1 in early B cell development, do the authors think similar miRNA networks would be regulated by IKZF1 in normal early B cells or they are only regulated in leukemia? If so, what might be the molecular determinants that drive the differences in normal early B cells and leukemia? No need to do experiments - the authors can include a discussion in the Discussion section. 

We thank the reviewer for raising this important point. In our revised manuscript we have included a discussion of this in the Discussion (Lines 364-379).

2. Experiments were done using Ph⁺B-ALL cells. Do the authors think similar results would be obtained using B-ALL PDXs/cell lines driven by other translocations? In other types, could the miRNA networks regulated by IKZF1 be subtype specific? To address these questions, the authors can include a discussion in the Discussion section. 

Essentially, we know very little about distinctive miRNA expression patterns in different leukemia subtypes and even less about the biological processes that regulate them. We briefly discuss this point in the revised discussion on Lines 362-373.

3. Turk A et al. (2022, PMID: 35408829) identified miRNA networks associated with different types of leukemias. Could the authors comment on findings from Turk A et al. (2022)? Any overlap or differences? 

We thank the reviewer for referencing this recent publication. We compared our list of differentially regulated miRNAs to those classified as leukemia miRNAs in the Turk 2022 study. We found that 5/31 total IK1-regulated miRNAs overlap with annotated leukemia miRNAs. We have included this information along with the Turk 2022 reference in the revised discussion (lines 197-199).

 

Reviewer 2 Report (New Reviewer)

In this manuscript, Kogut et al show the potential miRNA regulated by IKZF1 by using Ph+ALL cell lines and hypothetize about the consequences of gene deregulation (oncogenes and TS genes). The experiments seem correct although triplicates would be much more appreciated than duplicates. The authors link genetic data with outcome data. I have some minor comments shown below, which could be commented on Discussion section (a bit of limitations of the study):

- RNAseq better than expression array for analyzing differential gene expression 

- Triplicates instead of duplicates in cell cycle experiments (Fig 1C) and expression array (Fig2B)

- Fig 1 C could show the real p-value instead of P<0,05

- Impact on survival made on a low number cohort (n=86)

- Effects investigated only in 2 Ph+ ALL cell lines

- Fig 3 C a gene cannot be seen due to the presence of the scale (5kb)

- Among the 1,081 and 2,829 up and down-regulated miRNA target genes, you selected those which were tumor suppressors (219) or oncogenes (89). But this is a little proportion from the total number of target genes. What about the remaining genes? Gene Ontology was done on 1,81 and 2,829 genes or on 219+89 genes?

- Survival curves (Fig 6) may, in fact, reflect the outcome of IKZF1-deleted vs. IKZF1 WT patients

- “ABL” gene should appear as “ABL1”

Author Response

We thank the reviewer for their helpful comments. We have addressed each point below:

1. “RNAseq better than expression array for analyzing differential gene expression.”

We agree that the sequencing technology is better for detection and discovery of different small RNA species including miRNAs in samples. In fact, we used both sequencing and miRNA arrays for this project and found that our results with the miRNA arrays were very consistent, whereas our sequencing will require more optimization. We therefore proceeded here with the analysis of the array data. For this we included a statement of the study limitations in the discussion section (lines 457-461).

 

2. “Triplicates instead of duplicates in cell cycle experiments (Fig 1C) and expression array (Fig2B).”

The results of the growth suppression for Fig 1B have been reproduced numerous times and the cell cycle results in Fig 1C were consistent with previously published studies using the other cell line (PDX2) used in this study. Taken together, we feel quite confident in these results. For the expression array analysis (Fig 2B) use of technical triplicates might have improved our results by identifying additional miRNAs. However, we chose to evaluate IK1-regulated miRNAs using two biological replicates (separate cell lines) each with technical duplicates. Thus, we have four individual data points providing a stringent analysis of Ikaros-regulated miRNAs.

 

3. “Fig 1 C could show the real p-value instead of P<0,05.”

We have updated the p-value to the Figure 1 legend in the revised manuscript.

4. “Impact on survival made on a low number cohort (n=86).”

As far as we know, the TARGET ALL project is the largest cohort of B-ALL patients with comprehensive miRNA profiles.

 

5. “Effects investigated only in 2 Ph+ ALL cell lines.”

We acknowledge that 2 cell lines is a limitation of the study and we have included a revised the description of the study limitations in the discussion section (line 463). Indeed, having more cell lines that correspond to this unique subtype (Ph⁺ and IZKF1-mutated) would be powerful, but we unfortunately have a limited number of patient-derived cell lines to work with.

5. “Fig 3 C a gene cannot be seen due to the presence of the scale (5kb)”.

We thank the reviewer for pointing this out, we have now corrected this error in the revised manuscript.

5. “Among the 1,081 and 2,829 up and down-regulated miRNA target genes, you selected those which were tumor suppressors (219) or oncogenes (89). But this is a little proportion from the total number of target genes. What about the remaining genes? Gene Ontology was done on 1,81 and 2,829 genes or on 219+89 genes?”

We apologize for the confusion. The gene ontology analysis (Fig. 5) was performed on all the target genes for all of the 21 up-regulated or all of the 10 down-regulated miRNAs (1,542 and 3,290 target genes, respectively). There are many pathways associated with the targets of either group of up- or down-regulated miRNAs, and many genes were targets of both up- and down-regulated miRNAs. Therefore, to focus on the unique target for either up or down regulated miRNAs, we curated a list of unique non-overlapping target genes for the up- and down-regulated DE miRNAs (1,081 and 2,829 genes, respectively). We have clarified the results text in the revised manuscript (lines 241-242). \

8. “Survival curves (Fig 6) may, in fact, reflect the outcome of IKZF1-deleted vs. IKZF1 WT patients”.

This is an interesting point raised by the reviewer. We have analyzed the TARGET data to identify samples with IKZF1 deletion or point mutations in this cohort that have both matched miRNA expression and whole genome sequencing data. We have found that there are only seven samples harbor IKZF1 deletions and the survival analysis, albeit low n, shows no difference in outcome between these 7 patients and the other group of wild-type patients.

8. “ABL” gene should appear as “ABL1”.

We thank the reviewer for pointing this error (line 427).

Reviewer 3 Report (New Reviewer)

Kogut et al characterized the microRNA expression change upon Ikaros induction in two established B-ALL cell lines using Microarray analysis. The authors further investigated the physiological relevance and prognostic significance of the altered microRNA expression change. Overall, the manuscript uses a coherent language, references are appropriately cited, and sufficient background is provided. However, the study requires improvement on certain fronts before it can be accepted for publication.

1. In Figure 1A, it would be beneficial if the authors could comment on the expression change of IK6 upon IK1 induction, especially in the case of the MXP5 cell line.

2. In Figure 1B, stats should be provided for each time point.

3. In figure 3C, t-test should also be performed for Phase S and G2/M.

4. The authors need to provide more clarity on the Y-axis of figure 2B. To my understanding, the numbers on the Y-axis stood for the baseMean reads after DE analysis in R. However, these numbers do not match those as reported in the Volcano plot. For example, the fold change of hsa-miR-3178 was around 2⁴, but the bar graph in figure 2B showed about a 2-fold increase. Also, it would be beneficial if the authors could expand a bit more on how fold change was calculated for miRNAs that read 0 in the Microarray analysis in the methods section.  

5. In figure 3B and 3C, better annotation of the genome browser tracks would be beneficial, such as the exact location of mir-551. Also, the scale should be fixed in panel C.

 

Author Response

We would like to thank the reviewer for their helpful comments and critical evaluation of the manuscript. We have addressed the comments point-by-point below:

1. In Figure 1A, it would be beneficial if the authors could comment on the expression change of IK6 upon IK1 induction, especially in the case of the MXP5 cell line.”

We thank the reviewer for their insightful observation. This is something we have observed in other experiments as well. We are currently investigating the mechanism, but our current hypothesis is that Ikaros autoregulates its own expression and IK1 induction also represses the endogenous IK6 expression.

 

2. “In Figure 1B, stats should be provided for each time point.”

We have added the appropriate statistical analysis results in the Figure 1 legend.

3. “In figure 3C, t-test should also be performed for Phase S and G2/M.”

We have performed the statistical tests to show the significance differences and have modified the Figure 1 accordingly. Thank you.

 

4. The authors need to provide more clarity on the Y-axis of figure 2B. To my understanding, the numbers on the Y-axis stood for the baseMean reads after DE analysis in R. However, these numbers do not match those as reported in the Volcano plot. For example, the fold change of hsa-miR-3178 was around 2⁴, but the bar graph in figure 2B showed about a 2-fold increase. Also, it would be beneficial if the authors could expand a bit more on how fold change was calculated for miRNAs that read 0 in the Microarray analysis in the methods section.  

We apologize for the confusion. In the revised manuscript, we have re-represented Fig. 2A as a MA plot by transforming the data onto M (log ratio) and A (mean average) scales, then highlighting differentially regulated miRNAs. The statistical details were clarified in Material and Methods section.

 

5. “In figure 3B and 3C, better annotation of the genome browser tracks would be beneficial, such as the exact location of mir-551. Also, the scale should be fixed in panel C.”

We thank the reviewer for their suggestions and we have corrected this Figure in the revised manuscript.

 

 

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

To the Authors:

The manuscript submitted aims to determine the role of Ikaros in the regulation of miRNA expression in B-ALL. They first set out to determine the miRNA targeted by Ikaros with potential tumor suppressor or oncogenic roles in B-ALL. Next, they used pathway enrichment analysis to identify target genes for those miRNAs, and examined the miRNA regulated by Ikaros associated with B-ALL outcome.

Abstract: The abstract reflects the content of the manuscript.

Keywords: The keywords are appropriate.

Introduction: The introduction is consistent with the subject.

Line 55 – Ikaros (IKZF1) (first time it is mentioned in the Introduction)

Materials and Methods: The methods used were selected based on their ability to address the objectives of the study and are well described in this section.

Results:

As a general consideration, the experimental model is well suited although the n is small.

For clarity, the authors should consider placing the figures closer to the paragraphs/text where the results illustrated are explained, instead of being presented all together (3.6. Figures).

The legend of Figure 3 should include the indication that the B-ALL cell lines referred to are ICN1 and LAX2 (line 284).

Author Response

We thank the reviewer for their positive evaluation of our study, as well as the critical feedback. We have addressed each of the points below.

1. Line 55 – Ikaros (IKZF1) (first time it is mentioned in the Introduction).

Author’s response: thank you for pointing this out. We have added this correction.

2. For clarity, the authors should consider placing the figures closer to the paragraphs/text where the results illustrated are explained, instead of being presented all together (3.6. Figures).

Author’s response: We agree. We were following the journal’s template and will ask the journal to place the figures closer to the text where they are described in the final layout.

3. The legend of Figure 3 should include the indication that the B-ALL cell lines referred to are ICN1 and LAX2 (line 284).

Author’s response: Thanks for pointing this out. We have modified the Figure legend to provide this information.

Reviewer 2 Report

 The authors studied the role of IKZF1 in Ph+ ALL cell lines and identified some miRNAs could participate in the pathomechanisms of the disease.

1. The authors should have more cell lines for a clear conclusion, especially Ph- cell lines.

2. Though the cell lines contributed to the results, how about real samples from patients?

3. The exact mechanisms of potential miRNAs should be studied for conclusion.

Author Response

We thank the reviewer for their favorable feedback and constructive suggestions for our manuscript. We have addressed their comments point-by-point below.

1. The authors should have more cell lines for a clear conclusion, especially Ph- cell lines

Author’s response: The reviewer suggests that we include Ph-neg cell lines in the analysis. We agree that it is of great interest to evaluate miRNA expression in different subtypes of B-ALL. We have now included this in the Discussion (see page 14, lines 428-431). However, in this project, we specifically investigate the role of Ikaros in the regulation of miRNAs, with the hypothesis that mutations in the IKZF1 gene may lead to altered expression of cancer-associated miRNAs. IKZF1-mutations are highly associated with Ph+ B-ALL, and we therefore specifically chose to use two biological replicates of IKZF1-mutated Ph+ B-ALL cell lines to address this specific question. We have included a statement to clarify this in the start of the results (page 4, lines 174-175) and discussion sections (page 13, lines 341-343).

2. Though the cell lines contributed to the results, how about real samples from patients

Author’s response: We agree this is of high interest, and we sought to address this with the analysis of TARGET dataset (Figure 6). One challenge, is that there are very few available miRNA datasets on primary patient samples for B-ALL, and while the TARGET dataset we analyzed had miRNA profiling from ~175 primary B-ALL patient samples, only a small number of those were Ph+ or harbored IKZF1 mutations (only 7 IK6-DN+ samples), and none of those were Ph+. Thus it was not feasible to investigate specific questions pertaining to miRNA patterns in patient with IKZF1 mutations. We anticipate that this will be an area of expansion in the future, with future primary patient samples being analyzed for their miRNA expression, and this will be an interesting aspect to study once datasets are available.

3. The exact mechanisms of potential miRNAs should be studied for conclusion

Author’s response: Our study has uncovered multiple miRNAs that change expression with IK1 induction in IKZF1-mutated Ph⁺ B-ALL. Our analysis shows that many different miRNAs are regulated by Ikaros, and each miRNA can potentially regulate multiple mRNAs together forming a complex network. This is an important first step in understanding the role of miRNAs in B-ALL and we agree that it is important to study the mechanistic details of each of these miRNAs in B-ALL pathogenesis in future investigations. However, this is beyond the scope of the current study.

Reviewer 3 Report

Here authors test a hypothesis that hematopoietic transcription factor Ikaros (IKZF1) isoform IK1 (wild-type Ikaros) promotes growth suppression of the Philadelphia chromosome (Ph+) B cell acute lymphoblastic leukemia (B-ALL) cells. IKZF1 mutations or deletions are frequently noted in high number of Ph+ B-ALL patients and are predictor of high-risk pre-B-ALL. Authors have previously demonstrated that ectopic expression of IK1 inhibited growth of Ph+ pre B-ALL PDX2 cells. In this report, authors investigate molecular transducers of IK1-dependent inhibition of pre B-ALL cells involving microRNAs (miRs). Ph+ pre B-ALL cell lines were generated that express inducible IK1 to demonstrate that IK1 induction caused G1 phase cell cycle arrest. A high-throughput analysis of miRs following induction of IK1 was carried out and miRs with significant (log2 fold) change were identified. Expression of a subset of IK1-induced tumor suppressive or IK1-repressed oncogenic miRs was confirmed in fig 2B, and IK1 direct binding to select number of miRs was determined by ChIP analyses (Table S3). Further a KEGG pathway enrichment analysis was carried out for IK1 upregulated differentially expressed miRs and IK1-down-regulated differentially expressed miRs (Fig 4) to identify p53 tumor suppression, cell cycle and cancer pathways as the predominant pathways altered by IK1 expression in Ph+ pre B-ALL cells. Database analyses of key miR-targeted mRNAs was carried out in fig 5. Importantly, data in figure 6 show potential prognostic values of three, IK1-regulated miRs for pre B-ALL. Overall the studies are well performed with appropriate controls and statistical analyses. As many miRs are known to function as tumor suppressive or oncogenic in other cancer models, a large part of the data essentially validates tumor suppressor or oncogenic functions of many miRs in B-ALL. The concern is that miRs are well known to be pleiotropic in transducing their tumor growth regulatory functions to the extent that they can also signal by targeting expression of other miRs. Although in fig 2B, it seems that IK1 directly binds with two miRs (4674 and 551a), it is unclear whether changes in levels of other miRs noted in fig 2B involved direct IK1 binding. This will be important to clarify whether the miRs indicated for prognostic value in fig 6 are direct targets of IK1 or are down-stream targets of miRs that are direct targets of IK1.    Perhaps, anti-miR sequences (antagomirs) to miR-4674 or miR-551a (if available) could be utilized to determine whether and to the extent expression of other miRs listed in Fig 2B could be further altered in cells expressing inducible IK1 and respective anti-miR. Authors will also need to demonstrate/validate experimentally that key gene/mRNA of miR-4674 (such as JAK1, STAT5, or SOX) are down-regulated following IK1 induction. As the cell model is Ph+ pre B-ALL cells, it will be important to also demonstrate whether miR-663a is a direct target of IK1 and that IK1 induction inhibits growth of these cells in part by suppressing expression of ABL1 mRNA as BCR-ABL translocation is a feature of Ph+ B-ALL. A minor concern also pertains to data in Fig 3B, C. Authors need to present additional details/explanation in results and legend to clarify the implications of the data analyzed from two additional B-ALL cell lines.

Author Response

We thank the reviewer for their favorable and constructive evaluation of our manuscript. We appreciate that they specifically pointed out that our studies were overall “well performed with appropriate controls and statistical analysis”. We address their comments point-by-point below.

1. The reviewer pointed out that miRNAs are known to be pleiotropic, a point we agree with. It is known form literature that several miRNAs can have both oncogenic and tumor suppressor functions, depending on the context. We have included this point with a relevant citation in our Discussion (see page 14, lines 360-361).
2. The reviewer points out that “Although in fig 2B, it seems that IK1 directly binds with two miRs (4674 and 551a), it is unclear whether changes in levels of other miRs noted in fig 2B involved direct IK1 binding.” We found that Ikaros binds within 50 kb of 9 of the 31 DE miRNAs (Figure 3A), with two representative examples represented in Fig 3b and c, and the 9 listed in Table S3. We have changed text in the results section and figure 3 legend to clarify this (see page 5, line 209-210). While we did not see evidence of direct Ikaros binding within 50 kb of the remaining 21 miRNAs, their expression was altered upon expression of IK1. This could be due to long-range regulatory interactions or other mechanisms beyond transcriptional regulation, including miRNA processing and stability.
3. The Reviewer pointed out that it is unclear whether Ikaros binds to the 3 miRNA in Figure 6 with prognostic significance: For these 3 (hsa-miR-130b, hsa-miR-4649 and hsa-miR-26b) we did not see evidence of Ikaros binding within 50kb of these loci. We have added text in the results section to clarify this (see page 15, lines 426-428).
4. The reviewer suggests several interesting aspects for future follow-up, such as specific analysis of miR-4674, 551a or 663a and analysis of the effect on candidate target genes (JAK1, STAT5, SOX, ABL1). In this study we show that many different miRNAs are regulated by Ikaros, and that each miRNA can potentially regulate multiple mRNAs together forming a complex network. This is an important first step in understanding the role of miRNAs in B-ALL. While we agree with the reviewer that these are highly interesting research questions to pursue, this is beyond the scope of this study.
5. Lastly, this reviewer requests a clarification in regards to the data in Figure 3 from the two additional B-ALL cell lines. This was Ikaros ChIP-seq performed in two biological replicates (with technical replicates) of Ph+ B-ALL with wild type Ikaros expression. We thank the reviewer for pointing out the need for this clarification, and we have now added text to clarify this in Figure 3 legend (see lines 292-293).

Round 2

Reviewer 2 Report

Though the authors responded and modified the manuscript as reviewers' suggestion accordingly, lack of other cell lines and/or real patient samples for further analysis could hamper the final conclusions.

Reviewer 3 Report

Thank you for responses to my prior concerns. The responses noted in points 3 and 4 do not seem to robustly address the concerns.

For point 3 response, as the focus is the IK1 alterations in pre-B-ALL cancers and the noted miRs are not direct targets they may not necessarily be the predictors of IK1-dependent effects. This fact needs to be stated in the results section instead of discussion.

The concern in point 4 needs to be addressed for robust validation of the hypothesis. Authors will need to conduct the requested studies to address the signaling pathway(s) modulated by IK1-miR axes. Respectfully, the requested studies are felt as within the scope of this study.