miR‐205 mediates adaptive resistance to MET inhibition via ERRFI1 targeting and raised EGFR signaling

Abstract The onset of secondary resistance represents a major limitation to long‐term efficacy of target therapies in cancer patients. Thus, the identification of mechanisms mediating secondary resistance is the key to the rational design of therapeutic strategies for resistant patients. MiRNA profiling combined with RNA‐Seq in MET‐addicted cancer cell lines led us to identify the miR‐205/ERRFI1 (ERBB receptor feedback inhibitor‐1) axis as a novel mediator of resistance to MET tyrosine kinase inhibitors (TKIs). In cells resistant to MET‐TKIs, epigenetically induced miR‐205 expression determined the downregulation of ERRFI1 which, in turn, caused EGFR activation, sustaining resistance to MET‐TKIs. Anti‐miR‐205 transduction reverted crizotinib resistance in vivo, while miR‐205 over‐expression rendered wt cells refractory to TKI treatment. Importantly, in the absence of EGFR genetic alterations, miR‐205/ERRFI1‐driven EGFR activation rendered MET‐TKI‐resistant cells sensitive to combined MET/EGFR inhibition. As a proof of concept of the clinical relevance of this new mechanism of adaptive resistance, we report that a patient with a MET‐amplified lung adenocarcinoma displayed deregulation of the miR‐205/ERRFI1 axis in concomitance with onset of clinical resistance to anti‐MET therapy.

Thank you for the submission of your manuscript to EMBO Molecular Medicine and please accept my apologies for the unusual delay in getting back to you. We have now finally heard back from the three referees whom we asked to evaluate your manuscript.
You will see from the set of comments pasted below that referees 1 and 2 are supportive and only mention minor points, while referee 3 is much more critical. As EMBO Molecular Medicine focuses on studies of general interest with translational value, we agreed with some of the critical points of referee 3. Furthermore, upon our cross-commenting exercise, referees went back and forth and agreed with the following items that must be addressed for the paper to be further considered: We would therefore welcome the submission of a revised version within three months for further consideration and would like to encourage you to address all the criticisms raised as suggested to improve conclusiveness and clarity. Please note that EMBO Molecular Medicine strongly supports a single round of revision and that, as acceptance or rejection of the manuscript will depend on another round of review, your responses should be as complete as possible.
EMBO Molecular Medicine has a "scooping protection" policy, whereby similar findings that are published by others during review or revision are not a criterion for rejection. Should you decide to submit a revised version, I do ask that you get in touch after three months if you have not completed it, to update us on the status.
Please also contact us as soon as possible if similar work is published elsewhere. If other work is published we may not be able to extend the revision period beyond three months.
Please read below for important editorial formatting and consult our author's guidelines for proper formatting of your revised article for EMBO Molecular Medicine.
I look forward to receiving your revised manuscript. ***** Reviewer's comments ***** Referee #1 (Comments on Novelty/Model System for Author): human data Referee #1 (Remarks for Author): The manuscript by Migliore and colleagues reports on an interesting observation that in a clinical context the miR-205 and ERRFI1 are directly involved in the resistance to MET TKI. The patient data are supported also by in vitro experiments in three cell lines. I think this is a manuscript of interest for the reader of the journal, and I have only minor comments, as this is a well-done study. Also, I consider that the identification of a patient with abnormal miR-205 and ERRFI1 levels and the expected phenotype, explain why in vivo study in mice are not necessary for this study in order to be published in the journal Minor comments: 1 what are the Gene Ontology data for the array results presented in Figure 2H for each of the cell lines? Are these common between the cell lines?
2 what are the exact CpG sites methylated? A clear map of these loci is useful.
3 the figure 2G is nicely done and the Westerns are clear. Was in fact used only one time the normaliser, meaning all the ten hybridizations were done on only one blot? Otherwise, each normaliser foe each membrane has to be presented.
3 the figure 2G is nicely done and the Westerns are clear. Was in fact used only one time the normaliser, meaning all the ten hybridizations were done on only one blot? Otherwise, each normaliser foe each membrane has to be presented.

Referee #2 (Remarks for Author):
The manuscript under consideration describes a model by which, miR-205 upregulation drives resistance to MET inhibitors through a by-pass pathway involving activation of EGFR. In addition to identifying that miR-205 is upregulated, the authors determined at least one potential mechanism that miR-205 is elevated, which entails reduced promoter methylation. The authors also determined that the effect is mainly mediated through miR-205 targeting ERRFI, a negative regulator of EGFR. The manuscript concludes with validation that miR-205 and ERRFI are anti-correlated in one human tumor sample and that miR-205 is elevated, and ERRFI reduced, following resistance. Overall the study is well-controlled, the conclusions are justified and the findings are novel. The manuscript will likely be of interest to a broad audience. While the manuscript is generally acceptable, there is one figure/region of text that needs editing and perhaps an additional experiment done to support the conclusions.
The authors state that "deletion of the predicted miR-205 seed sequence ... ...eliminated luciferase downregulation by ectopic miR-205, pointing to a direct regulation of ERRFI1 by miR-205". There are multiple issues with this statement. 1) The authors did not delete the "seed sequence" of miR-205. Seed sequence refers to the miRNA, not the target. They deleted the targeting sequence in ERRFI1. 2) Importantly, they did not delete the sequence in ERRFI1 targeted by the miR-205 seed sequence either. This would have been preferred and would have been more convincing of a direct role for miR-205 in regulating ERRFI. Instead they deleted over 200nt of the 3' UTR of ERRFI1, which contains the miR-205 target sequence, and likely many other regulatory sequences. MiR-205 could alter the expression of other factors that bind this region generating an indirect effect. This is a rather large deletion to make the above conclusion. To conclude that miR-205 is directly targeting ERRFI1 in the predicted targeting sequence, the authors would need to mutate only a portion of the targeting sequence and not delete 200nt.
3) The authors state that they "eliminated" luciferase downregulation. This is somewhat inaccurate. Although not significant, there is still some reduction in luciferase activity, which suggest that there may be another region of the 3' UTR involved. Indeed, the majority of the effect is lost when this region of the 3' UTR is deleted, but not the full effect. The text should be edited. In this manuscript, Migliore et al. describe miR-205 mediated acquired resistance to MET inhibitors. miRNA sequencing showed upregulation of miR-205, and RNA-seq showed downregulation of ERFFI1. Genetic knockdown and expression experiments provide evidence that the upregulation of miR-205 and downregulation of ERRFI1 has a functional impact on EGFR activation, which the authors show has increased activity in the resistant cell lines. The authors suggest this enables resistance to MET-TKIs. In its current form, the data is preliminary and the manuscript lack deep mechanistic insight and has low translational value and is unsuitable for publication.
Major points 1. Consistency in experimental conditions and missing controls. This is an incomplete study because there is a lack of consistency in the compounds used as well as missing controls. Can the authors explain why different drugs are being used for different cell lines? Why is JNJ not being used for the EBC-1 and GTL16 cell lines? Why is PHA and Crizotinib not used in the SG16 cell line? There should be consistency in the drugs and resistant cell lines used for all the figures in the manuscript. Figure 2G makes comparisons between WT without drug and resistant cells with drug. This is not a direct comparison. The authors need to show the resistant cells without drug to confirm that the signalling in the resistant cells is not altered by the presence of the drug. Also, why was HER2 included in this blot, this is not referred to in the text.
Experiments to show that miR-205 levels are altered by overexpression or knockdown for figures 2A-F have not been provided. Figure 2H, the authors only highlight a subset of putative miR205 targets that are downregulated in resistant cells. The entire gene expression dataset needs to be provided in the supplemental data. In particular were there any putative targets that were upregulated in the resistant cells and if so can the authors provide a reason as to why this may be the case.

Lack of mechanistic insights
The manuscript has not explored the molecular mechanisms as to why the resistant cells have higher levels of miR205. They make reference that demethylation may be a reason for this ( figure 1E), but no further mechanistic detail has been provided What is the molecular mechanisms of demethylation? Is it a cause or effect of resistance? Why specifically resistant cells? A sentence "studies suggest a role for de-methylation of miR-205 genomic locus" in the discussion is insufficient. The authors need to perform more experiments establishing the link between methylation and miR205 as well as the mechanism(s) by which resistant cells regulate methylation at this locus.
Fundamental questions as to whether this demethylation is dynamic are not address, e.g. do methylation levels dynamically change in response to drug treatment or is it a function of the final acquired resistant state? Is this specific to MET-TKIs are do other TKIs also induce this effect.

Little translational value
An important element to consider for publication in EMBO Molecular Medicine is whether these findings are of any translational value. The authors have not definitively demonstrated this.
Many of the viability changes in figure 2 and 3 are small and it is unclear if these alterations will actually have an effect on tumour growth. It is essential that in vivo xenograft experiments are performed to demonstrate that these small changes actually translate into a significant decrease in tumour growth. In this day and age, this would be the minimal requirement for oncology studies.
It is unclear as to what is the biomarker that is important for establishing if miR205 is the mechanism of TKI resistance. The authors only use 3 cell lines which originate from different cancer types. Given the limited number of cell lines, the authors have not demonstrated how general the observations are. Secondly the use of a larger panel of cell lines would provide clues as to whether any specific genetic features or biomarkers can indicate sensitivity to the miR205 pathway. The two case studies they provide have contrasting ERRF1 alterations upon acquisition of MET TKI resistance and hence no conclusions can be drawn about how general is this observation and which genetic factors or biomarkers dictate sensitivity to the miR205 pathway. The authors need to increase the number of cell line models and preferably in patient derived models in order to establish biomarker and molecular determinants of miR-205 driven MET TKI resistance.
Other points 1. Fig 2C, B, and D(CRIZ) show significant but not large changes, suggesting miR-205 alone is not sufficient to fully restore resistance. Dose response curves would be more informative.
2. The authors claim in page 12 that "ERRFI1 expression was clearly lower in resistant cells compared to their parental counterpart" However, this is only true in 3 out 5 resistant cell lines shown in Suppl. 4. Page 11 paragraph 2 is largely descriptive and non-specific. It states that "MET largely retained sensitivity to inhibition by MET-TKIs". This statement is non-specific and Fig 2G shows that GTL16 resistant cell lines have higher MET phosphorylation than the WT+TKI. Whilst the level of phosphorylation is reduced compared to WT, this still does not show "sensitivity to inhibition" as claimed by the authors. It also claims that MET inhibition in resistant cells "did not translate into significant suppression of ERK and AKT activation", however Fig 2G shows GTL16 R-CRIZ and R-PHA cells have reduced pAKT. Figure 3D does not show a "sizeable decrease of ERRFI1 protein expression" in GTL16, as claimed by the authors. 6. Figure 3F needs to state is the change from CTRL to miR-205 in the right panela is nonsignificant. 7. Figure 4A -the authors offer no explanation as to why the case study observation in Figure 4 is completely different to the observation in S Fig S4. What is the molecular determinant of miR-205 driven MET-TKI resistance?

EDITOR COMMENTS
You will see from the set of comments pasted below that referees 1 and 2 are supportive and only mention minor points, while referee 3 is much more critical. As EMBO Molecular Medicine focuses on studies of general interest with translational value, we agreed with some of the critical points of referee 3. Furthermore, upon our cross-commenting exercise, referees went back and forth and agreed with the following items that must be addressed for the paper to be further considered: 1-One way of showing if the drug is hitting the same target is demonstrating cross-resistance across the 3 compounds: therefore including the extra chemicals in all the cell lines is required (ref3, point 1) As suggested, we performed experiments assessing cell viability after exposing both wt and resistant derivatives to escalating doses of the three anti-MET drugs used in our work (i.e. Crizotinib, PHA-665752 and JNJ-38877605). Figure 1 A,B,C in the revised manuscript shows that all the resistant derivatives were strongly cross-resistant, independently from the drug to which they were initially exposed to generate resistance.

2-Ref. 3 wants to see the western blot of the drug-selected cells in the absence of drug to check whether the resistant cells have not undergone signalling rewiring such that at baseline there is elevated phosphorylation of AKT and ERK (ref. 3 point 1 Fig. 2G)
To address this issue, resistant cells were grown in the absence of the drug to which they had been rendered resistant, and lysates analyzed by WB. As shown in Figure 2I, ERK and AKT activation was not, or only marginally affected, by drug withdrawal.

3-authors have to show that methylation of the promoter is altering the expression of miR-205, it's only correlative now (ref. 3 point 2)
We performed additional experiments in which we evaluated changes in the methylation status of the miR-205 genomic region in parallel with variations in the expression levels of miR-205. As shown in Figure 1D,E of the revised manuscript, demethylation of miR-205 genomic sequences was paralleled by a sizeable increase of miR-205 expression. This datum indicates that dynamic changes of CpG methylation in the genomic region of the miR-205 locus are mirrored by variations of miR-205 expression. This provides additional support to our model of epigenetic regulation of miR-205.

4-More cells should be included to validate that the effect is wide spread (or not) (ref. 3 point 3)
In the original work we performed our experiments in three cell lines (two established cell lines and a primary cell line), all of which addicted to MET. In the revised version we added two more METaddicted cell lines, namely KATO II and SNU-5, both of which showed anti-correlated changes of miR-205 and ERRFI1 expression upon the acquisition of resistance to MET TKIs (Figures EV1E,F, EV2 B,C,D,E). However, no increase of miR-205 was observed in two other MET-addicted cell lines (H1993 and Hs746T) rendered resistant to both crizotinib and PHA-665752 (these negative data are not included in the revised version but only presented to the reviewer. If requested, we can add them as an appendix figure). Interestingly, when we expressed miR-205 in a not-MET-addicted lung cancer cell line (A549) we observed a decrease of ERRFI1 expression and concomitant upregulation of EGFR expression/activation (Appendix Figure S8). In aggregate, these experiments suggest that the described mechanism of resistance to MET TKIs, while not ubiquitous, does play a role in several cellular models of different histological origin (i.e. lung and gastric cancer).

5-The full dataset for the gene expression analysis needs to be provided.
The full dataset has now been deposited at NCBI Sequence Read Archive, SRA (accession number SUB3868037) and at NCBI Gene Expression Omnibus, GEO (accession number GSE114406).

6-Evidence to support the findings in vivo. 2 referees out 3 support this.
As requested, we performed xenograft experiments by injecting GTL16 resistant cells (R-CRIZ) transduced with lentiviral (LV) stocks generated from pCDH-anti-miR-205 or the empty vector pCDH (Appendix Figure S4A). Mice were treated with crizotinib (25 mg/kg) and tumor volume was monitored for 18 days. As shown in Figure 2G, anti-miR-205 transduced cells reacquired sensitivity to crizotinib treatment, while tumors generated by control cells remained insensitive to the drug. In the mirror experiment, GTL16 cells infected with either control or miR-205 LV stocks (Appendix Figure S4B) were injected in mice. When tumors reached an approximate volume of 150 mm 3 , treatment with crizotinib or with vehicle was started. Figure 2H shows that tumors generated by miR-205 overexpressing cells were refractory to Crizotinib treatment, while tumors generated by control cells where highly sensitive.
We performed similar in vivo experiments with control and ERRFI1 over-expressing GTL16 R-CRIZ cells. As shown in Figure EV 3B, tumors generated by ERRFI1 derivatives of GTL16 resistant cells were significantly smaller compared to control cells, implying that forced ERRFI1 expression was able to partially revert resistance. Interestingly, the effect on tumor growth was less dramatic than that observed in anti-miR-205 derivatives ( Figure 2H). This result suggests that, although ERRFI1 appears to be the most critical miR-205 target in this context, other miRNA targets are likely to contribute to the development of the resistant phenotype. This point has been discussed in the text (page 8 of the manuscript).
In aggregate, the above data validate in tumor xenotransplants the model emerged from molecular genetics and pharmacological approaches in in vitro model systems.
We thank the editor and the referees for having encouraged us to pursue in vivo experiments that further strengthen the relevance of our findings.

Referee #1 (Remarks for Author):
The manuscript by Migliore  We thank the reviewer for the positive evaluation of our work.

Minor comments:
1 what are the Gene Ontology data for the array results presented in Figure 2H for each of the cell lines? Are these common between the cell lines?
We performed the GO analysis for the genes presented in the new Figure 2L. The results show that the pathways in which downregulated miR-205 target genes are involved are not shared between the two models (see the graphs below for reviewer only). A hypothesis to explain this result is that in our NGS analysis we identified a small number of mir-205 downregulated targets and thus the GO analysis was quite restricted. Indeed, only three targets are shared by the two models, namely ERRFI1, Epiregulin and ZEB1. Epiregulin downregulation could seem unexpected; however, as we observed a general rewiring of EGFR ligands, the observed Epiregulin decrease could be biologically irrelevant due to the compensatory presence of other EGFR ligands.

what are the exact CpG sites methylated? A clear map of these loci is useful.
We apologize for not having included the map of the analyzed CpGs in the previous version. The map is now shown in Appendix Figure S2.
3 the figure 2G is nicely done and the Westerns are clear. Was in fact used only one time the normaliser, meaning all the ten hybridizations were done on only one blot? Otherwise, each normaliser foe each membrane has to be presented.
We apologize for not showing the normalizer for each of the two blots in Figure 2G of the original submission. In addressing a point raised by Reviewer 3, we modified the blots shown in the actual Figure 2I and the actin normalizer is now shown for both of them.

Referee #2 (Remarks for Author):
The 3) The authors state that they "eliminated" luciferase downregulation. This is somewhat inaccurate. Although not significant, there is still some reduction in luciferase activity, which suggest that there may be another region of the 3' UTR involved.

Indeed, the majority of the effect is lost when this region of the 3' UTR is deleted, but not the full effect. The text should be edited.
We thank the Reviewer for pointing out our mistake. The text has been edited to comply with the Reviewer's objections. As suggested by the reviewer, we have mutagenized the miR-205 targeted sequence in the ERRFI1 3' UTR. Figure 3F shows that the luciferase activity of the mutated reporter is not affected by ectopic miR-205, which provides cogent proof that this sequence is indeed targeted by miR-205 in our cellular models. As pointfully suggested by the editor, we tested cross-resistance among the different derivatives and found that they display extensive cross-resistance to all the MET kinase inhibitors used in our study ( Figure 1A,B,C). This datum indicates that our cell lines developed resistance to structurally different MET TKIs which share a common mechanism of action. This datum is also compatible with a model whereby the same mechanism of resistance (i.e. signal rewiring via the miR-205-ERRFI1-EGFR circuit) allows MET addicted cells to escape from pharmacological inhibition of the MET kinase by different TKIs.

Figure 2G makes comparisons between WT without drug and resistant cells with drug. This is not a direct comparison. The authors need to show the resistant cells without drug to confirm that the signalling in the resistant cells is not altered by the presence of the drug. Also, why was HER2 included in this blot, this is not referred to in the text.
According to the referee's suggestion, we performed a WB analysis of both resistant and WT cells, either in presence or absence of MET TKIs. As shown in the new Figure 2I we did not observe major differences in the activation of downstream signaling molecules when resistant derivatives were grown in the absence of the drugs. We agree that the HER2 expression/activation status is not informative and this panel was removed from the current manuscript version.

Experiments to show that miR-205 levels are altered by overexpression or knockdown for figures 2A-F have not been provided.
We regret that we did not include the requested controls in the previous version of the paper. These data are now shown in Appendix Figure S3. Figure 2H, the authors only highlight a subset of putative miR205 targets that are downregulated in resistant cells. The entire gene expression dataset needs to be provided in the supplemental data. In particular were there any putative targets that were upregulated in the resistant cells and if so can the authors provide a reason as to why this may be the case.
The main focus of our study was to investigate whether upregulated miRNA expression could underpin resistance. This focus led us to analyze genes showing an expression pattern opposite to that of the miRNAs of interest. However, we did find a subset of putative miR-205 targets that was upregulated, as shown in the list provided in Appendix Figure S6. In the context of resistance to MET TKIs we suspect that some putative miR-205 targets could be upregulated despite increased miR-205 expression because they are either poor miR-205 targets or liable to regulatory cues capable of overruling miR-205-dependent control.

Lack of mechanistic insights
The manuscript has not explored the molecular mechanisms as to why the resistant cells have higher levels of miR205. They make reference that demethylation may be a reason for this ( figure 1E) We aimed at identifying mechanism/s involved in miR-205 upregulation in our cellular models, which led us to investigate the methylation status of the miR-205 genomic region. To address some of the reviewer's criticisms, we performed additional experiments to seek correlations between dynamic changes of the miR-205 locus methylation status and miR-205 levels. As shown in Figure  1D,E, we observed that miR-205 expression was significantly increased in wt (i.e. TKI-sensitive) cell lines upon 5-Aza-2'-deoxycytidine-induced de-methylation of CpG islands mapping to the miR-205 locus. We think that these novel data strengthen our model that epigenetic modifications may underpin miR-205 regulation. We do agree that additional experimentation would be required to further solidify this model. Yet, we feel that additional in-depth work would go beyond the scope of our work and would add limited value to the prevalently translational angle of this study. We acknowledge that this is a truly critical issue and we have strived to address the Reviewer's point. We performed xenograft experiments by injecting GTL16 resistant cells (R-CRIZ) transduced with control or anti-miR-205 lentivirus stocks, (Appendix Figure S4A). Mice were treated with crizotinib and tumor volume was monitored for 18 days. As shown in Figure 2G, anti-miR-205 expression reverted resistance, while tumors generated by control cells remained insensitive to the drug. In the mirror experiment ( Figure 2H), tumors generated by wt GTL16 miR-205 overexpressing cells became refractory to crizotinib treatment, while tumors generated by control cells remained sensitive.

Little translational value
All together these results support the hypothesis that miR-205 upregulation causes MET-addicted cancer cells to acquire resistance to MET-TKIs.
We also performed in vivo experiments using GTL16 R-CRIZ resistant cells expressing ectopic ERRFI1. As shown in Figure EV 3B, tumors generated by GTL16 R-CRIZ-ERRFI1 cells were significantly smaller compared to controls, implying that re-expression of ERRFI1 was able to partially revert resistance. Interestingly, the effect on tumor growth was less dramatic that what observed in tumors generated by GTL16 pCDH-anti-miR-205 cells ( Figure 2H). This result suggests that while ERRFI1 is likely the most critical miR-205 target in this context, other miRNA target molecules can contribute to the resistant phenotype.
We believe that the above in vivo data strengthen our model and enhance its translational value. We thank the Reviewer for encouraging us to pursue this experimentation. We agree that the diagnostic/prognostic value of a biomarker gains robustness when the biomarker is validated across different models. It is also true, however, that different mechanisms of resistance can develop within each tumor type, an undeniable complexity that poses a serious, and yet ineludible, challenge to biomarker discovery/development. This is indeed the case when resistance does not have a genetic basis. Here we provide evidence that in at least three different cellular contexts of MET addiction (lung adenocarcinoma -EBC-1; gastric carcinoma -GTL16; primary gastric cells -SG16) resistance to MET TKIs may be caused by deregulation of the miR-205-ERRFI1-EGFR axis.

It is unclear as to what is the biomarker that is important for
As requested, we investigated the other MET-addicted cellular models in our hands to verify if this mechanism could be even more general. We generated KATO-II and SNU-5 cells which showed increased levels of miR-205 and decreased expression of ERRFI1 ( Figure EV1E,F, Figure  EV2B,C,D,E). These results provide further support that miR-205 is one of the mediators of acquired resistance to MET TKIs.
On the other hand, when we rendered H1993 (lung cancer) and Hs-746t (gastric cancer) cells resistant to MET TKIs, we found that miR-205 was not expressed in either wt or resistant derivatives (see table below reporting qPCR analysis), implying that other mechanism/s underpin MET TKIs resistance in these cells.
As already stated in the main text, samples from tumors addicted to MET activation which develop resistance to MET TKIs are extremely rare because MET targeted treatments have not been approved yet. Thus, validation of our model on a larger scale of clinical samples is not feasible for the time being. That said, we believe that the validation in a patient of the mechanism of resistance we discovered in vitro provides an initial but highly encouraging proof of concept of our model. While we are not in a position to comment on how general this mechanism might be, we see that the combination of in vivo experimentation reported in Figure 2 G,H of the revised manuscript and clinical data in Figure 4 provides strong translational value. Finally, the fact that two patients experienced different mechanisms of resistance is in agreement with the consolidated notion that targeted treatment may generate different mechanisms of resistance.
Other points 1. Fig 2C, B, and D(CRIZ) show significant but not large changes, suggesting miR-205 alone is not sufficient to fully restore resistance. Dose response curves would be more informative.
We agree with the reviewer that the differences are not always impressive. We note that the largest differences were observed when miR-205 expression was inhibited in cells with the highest miR-205 expression (SG16 and EBC-1 resistant derivatives) or when miR-205 was overexpressed in cells with a low endogenous amount (wt GTL16 cells). Fig.S2.

The authors claim in page 12 that "ERRFI1 expression was clearly lower in resistant cells compared to their parental counterpart" However, this is only true in 3 out 5 resistant cell lines shown in Suppl.
We have quantified the intensity of the WB bands shown in Figure EV 2A. The graph clearly demonstrates that 4/5 resistant derivatives show decreased ERRFI1 expression. This was not the case of GTL16 R-PHA that, as stated in the text, relied for resistance on KRAS amplification and not on ERRFI1 downregulation/EGFR activation. We also note that relatively minor changes in ERRFI1 expression have been reported to have a strong impact on EGFR activation in in vitro and in vivo model systems (Anastasi et al., 2016). Fig 2G, whereas Fig 2G uses the term "MAPK" -needs to be consistent.

Page 11 uses the term "ERK" in the main text in reference to
We apologize for that discrepancy. In the revised version we consistently used the term ERK all over the manuscript. Fig 2G shows  We agree with the reviewer that the level of MET activation in resistant cells is higher than that observed in wt cells exposed to TKIs. However, the partial sensitivity of resistant cells to MET TKIs is testified by the observed increase in MET phosphorylation upon TKI removal (see Figure 2I). Furthermore, because GTL16 are MET amplified, it is possible that residual MET Tyr phosphorylation upon MET TKI treatment reflects cross phosphorylation by EGFR. Concerning GTL16 resistant cells, it is true that at baseline they show pAKT levels lower than wt cells; however AKT activation is clearly higher than in wt cells exposed to the MET TKI, which, in several models, has been shown to be sufficient to generate resistance. We have edited the text to take into account the above details and offer a more precise and balanced presentation of data in Figure 2I. Figure 3D does not show a "sizeable decrease of ERRFI1 protein expression" in GTL16, as claimed by the authors.

5.
We agree that ERRFI1 decrease is not homogeneous in all the cell lines (ranging from 50 to 80%, as shown by the added quantitation - Figure 3D). We have changed the sentence by removing "sizeable".
6. Figure 3F needs to state is the change from CTRL to miR-205 in the right panela is nonsignificant.
We apologize for having forgotten to show that the difference was not significant. This has now been added in the figure.
7. Figure 4A -the authors offer no explanation as to why the case study observation in Figure 4 is completely different to the observation in S Fig S4. What is the molecular determinant of miR-205 driven MET-TKI resistance?
As stated, patient #1 became resistant to MET TKI as a consequence of further MET amplification (Oddo et al, BJC, 117:347-352, 2017); no change in the miR-205/ERRFI1/EGFR axis was observed ( Figure EV4). Patient's #2 tumor, instead, showed increased miR-205 expression and ERRFI1 downregulation (Figure 4), which we pinpointed as a mechanism of clinical resistance to MET TKIs by analogy with our in vitro and in vivo experimental models. While we acknowledge that additional molecular analyses would be required to provide further support to our model in the clinical setting, we note that the scarcity of pathological material obtained at biopsy precluded further studies. Thank you for the submission of your revised manuscript to EMBO Molecular Medicine. We have now received the enclosed reports from the referees that were asked to re-assess it. As you will see the reviewers are now supportive and I am pleased to inform you that we will be able to accept your manuscript pending final editorial amendments.
Please submit your revised manuscript within two weeks. I look forward to seeing a revised form of your manuscript as soon as possible.
***** Reviewer's comments ***** Do the data meet the assumptions of the tests (e.g., normal distribution)? Describe any methods used to assess it.
Is there an estimate of variation within each group of data?
Is the variance similar between the groups that are being statistically compared?

YES
We tested for normality all the group of data used as input for T--Test evaluation, by means of the Shapiro--Wilk test implemented in Graph Pad 7.02.
no, see next point given the small size of each group of data the estimate of within group variation is not reliable. We therefore employed the more conservative eteroschedastic t--tes or ANOVA.

Data
the data were obtained and processed according to the field's best practice and are presented to reflect the results of the experiments in an accurate and unbiased manner. figure panels include only data points, measurements or observations that can be compared to each other in a scientifically meaningful way. graphs include clearly labeled error bars for independent experiments and sample sizes. Unless justified, error bars should not be shown for technical replicates. if n< 5, the individual data points from each experiment should be plotted and any statistical test employed should be justified the exact sample size (n) for each experimental group/condition, given as a number, not a range; Each figure caption should contain the following information, for each panel where they are relevant:

Captions
The data shown in figures should satisfy the following conditions: Source Data should be included to report the data underlying graphs. Please follow the guidelines set out in the author ship guidelines on Data Presentation.
Please fill out these boxes ê (Do not worry if you cannot see all your text once you press return) a specification of the experimental system investigated (eg cell line, species name).

B--Statistics and general methods
the assay(s) and method(s) used to carry out the reported observations and measurements an explicit mention of the biological and chemical entity(ies) that are being measured. an explicit mention of the biological and chemical entity(ies) that are altered/varied/perturbed in a controlled manner.
a statement of how many times the experiment shown was independently replicated in the laboratory.
Any descriptions too long for the figure legend should be included in the methods section and/or with the source data.
In the pink boxes below, please ensure that the answers to the following questions are reported in the manuscript itself. Every question should be answered. If the question is not relevant to your research, please write NA (non applicable). We encourage you to include a specific subsection in the methods section for statistics, reagents, animal models and human subjects.