Breast cancer cell-secreted miR-199b-5p hijacks neurometabolic coupling to promote brain metastasis

Breast cancer metastasis to the brain is a clinical challenge rising in prevalence. However, the underlying mechanisms, especially how cancer cells adapt a distant brain niche to facilitate colonization, remain poorly understood. A unique metabolic feature of the brain is the coupling between neurons and astrocytes through glutamate, glutamine, and lactate. Here we show that extracellular vesicles from breast cancer cells with a high potential to develop brain metastases carry high levels of miR-199b-5p, which shows higher levels in the blood of breast cancer patients with brain metastases comparing to those with metastatic cancer in other organs. miR-199b-5p targets solute carrier transporters (SLC1A2/EAAT2 in astrocytes and SLC38A2/SNAT2 and SLC16A7/MCT2 in neurons) to hijack the neuron–astrocyte metabolic coupling, leading to extracellular retention of these metabolites and promoting cancer cell growth. Our findings reveal a mechanism through which cancer cells of a non-brain origin reprogram neural metabolism to fuel brain metastases.

1.The conclusions heavily rely on the MDA-MB-231 cell line and its subline.To strengthen the conclusion, the authors should use additional cancer cell lines to demonstrate that EV miR-199b promotes brain metastasis in vivo.2. Glutamine levels are relatively high in the blood and brain.The authors show that EV miR-199b significantly changed glutamine concentration in the brain.How about other common sites of breast cancer metastasis (the lung and bone)? 3. Will the administration of glutamine and lactate in mice promote breast cancer brain metastasis in vivo?How about lung and bone metastasis? 4. Will silencing miR-199b (in breast cancer cells) or its extracellular form block brain metastasis in vivo?
In their manuscript, Ruan and colleagues propose a mechanism by which metastatic Breast Cancer cells alter the metabolic coupling between neurons and astrocytes in the brain to promote the growth of brain metastases.This is an interesting and important question; however, that are several major methodological problems that needs to be addressed: The authors suggest that breast cancer cells release extracellular vesicles (EVs) that contain high levels of miR-199b, which is found at higher levels in the blood of Breast Cancer patients with brain metastases compared to those with metastases in other organs, which is in line with previous that demonstrated the hsa-mir-199b is uniquely up-regulated in metastatic brain tumors, compared to primary brain tumors results (e.g., PMID: 32906592; table 1).Subsequently, the researchers propose that miR-199b-5p targets specific solute carrier transporters in the brain; however, the reasoning leading to this hypothesis is unclear.Additionally, the authors state that the analysis of microRNA.orgdemonstrated an association of miR-199b-5p with EAAT2/GLT1, SNAT2, and MCT2.Nevertheless, an analysis using DIANA tools (TarBase V.8) or rnacentral.orgdid not support this association.Moreover, the authors did not address the regulation of EAAT1 (SLC1A3/GLAST), which is the second main glutamate transporter in astrocytes, by miR-199-5p.Interestingly, DIANA/rnacentral.organalyses suggested that miR-199-5p is linked to the expression of the SLC1A5/ASCT2 transporter, which was suggested to play a role in glutamate reuptake by astrocytes (reviewed in PMID: 30234109).These unresolved issues require further attention.Furthermore, the changes in the rLuc reporter levels (Figure 3) appear minimal, despite being statistically significant, according to the authors.It remains uncertain whether these mutations affected glutamate influx into astrocytes or influxes of glutamine and lactate into neurons.Therefore, the functional effect of these mutations, as with the mimic mir, should be determined.Next, the authors examined the impact of mir-199b encoded EVs on Glu/Gln/Lac uptake and release (Figure 4).Initially, they tested whether NHA or differentiated SH-SY5Y cells took up the fluorescently labeled EVs, but their analysis only suggests a possible association between the EVs and cells.To obtain more accurate results, a z-stack confocal-based analysis with membranal staining, as demonstrated in PMID: 35225262, should be conducted (this should also be performed for 5B).Subsequently, in Figures 4B and 4C, the authors attempted to regulate the levels of EAA2, SNAT2, and MCT2 by overexpressing or inhibiting mir-199b.Although the mRNA data showed statistical significance, the manipulation of mir-199b resulted in only minor changes in the transcript levels of EAA2, SNAT2, and MCT2.The immunoblot analyses seemed more promising, but it is difficult to evaluate them due to the overexposure of the housekeeping protein GAPDH and the lack of statistical data on different biological repeats.Additionally, the statistical analysis of the metabolic measurements in Figure 4E should be redone, particularly with regard to the Gln analyses and the reported P<0.001.Importantly, these experiments lack a clear causal relationship between the observed changes in transcript/protein levels and the metabolic changes.This should be addressed by, for example, knocking down EAA2 in the cells and demonstrating that mir-199b-based manipulation has no effect on Glu uptake levels in those cells.
The measurement of mCherry fluorescent intensity in Figures 5E and 5F does not necessarily correlate with "Enhanced growth of cancer cells."The observed results could be associated with the regulation of mCherry expression itself, cell proliferation, cell cycle arrest, or cell death.To support their hypothesis and obtain more accurate results, the authors should conduct a nontranscriptionally regulated experiment, such as CFSE staining.Of note, using siRNA to support the author's hypothesis (Figure 5F) is a step in the right direction towards establishing a clear causal relationship, but they lack key controls demonstrating the target transcript/protein was indeed downregulated and identifying the cells in which it was downregulated.
The authors should clarify what is the percentage of astrocytes and neurons that up took the EVs, those are found in proximity to the tumor cells.Is the use of anti-CD63 mAbs is the best way to analyse uptke of EVs? Wouldn't it be degarded in the cells the uptook the EVs?The results in Panels B-E are quite unexpted compared to the minimal regulation demonstrated in the in-vitro experiments, and the low "uptake" of the EVs shown in panel A. more over it is clear that those changes are happingn in astrocyte or neurons.According to the immunblot data in panel 6D (that should be revised, similar to previoes coments on the authors immunblots), the autors should prformted immunostainng couple to GFAP and MAP2 staining in shame and tumor-bearing mice.
In-vivo experiments (Figure 6) -The authors should clarify the percentage of astrocytes and neurons that took up the EVs in proximity to the tumor cells.Additionally, it should be noted whether using anti-CD63 monoclonal antibodies is the most effective way to analyze EV uptake, as it is possible that CD63 may degrade in the cells that have taken up the EVs.The results in Panels B-E are unexpected compared to the minimal regulation demonstrated in the in-vitro experiments and the low "uptake" of EVs shown in Panel A. Moreover, it is unclear whether these changes occur in astrocytes or neurons.However, as the immunoblot data in Panel 6D, which should be revised as per previous comments on the authors' immunoblots, seems to suggest dramatic changes in the protein expression, the authors should conduct immunostaining coupled with GFAP and MAP2 staining in sham and tumor-bearing mice to address this question.It is worth noting that mir-199b has been shown to regulate cancer progression in several cancer types through various mechanisms.However, the current in vivo experiments demonstrates a correlation between mir-199b-carrying EVs, metabolic pathways involving Glu/Gln/Lac, and the astrocytes-neurons axis and tumor cells.There is still no clear causal relationship.This issue should be experimentally addressed, or alternatively, the limitations of the in vivo (and ex-vivo) data should be clearly discussed in the results, discussion, and abstract sections.
Additional comments: 1.The authors ought to present a more comprehensive background on the involvement of astrocytes in primary and secondary brain tumors in the introduction or discussion section.Specifically, they should explore the role of astrocytes in the metabolic reprogramming of the tumor microenvironment (TME).This could be achieved by referencing relevant studies such as PMID 29892069, 35899587, 27225120, and 25639230.2. It would be beneficial for the authors to provide a more extensive context regarding the role of miR-199b in cancer generally, as well as in relation to breast cancer and metastasis.This could be accomplished by citing relevant studies such as PMID 23296799, 35090460, and 32082362, among others.3. The authors need to be more consistent and clearer about the identity of miR-199b, as miR-199b-5p has a different sequence from miR-199b-3p.They should ensure that they are unambiguously referring to miR-199b-5p throughout the manuscript.Doing so will help avoid any confusion and inaccuracies in interpreting their results.4. It would be advantageous to provide one or two lines of summary or conclusion for the section titled "miR-199b is associated with brain metastasis in BC patients and secreted by brain-tropic MBC cells".5.The authors should incorporate an immunoblot analysis of CD63 in the BBM1 extracellular vesicles (EV) shown in Figure 2B.
Reviewer #4 (Remarks to the Author): In this study, the authors demonstrated that EVs derived from brain-metastatic breast cancer cells shuttled miR-199b-5p to neurons and astrocytes, affecting neural metabolism and promoting brain metastases.Their work is great.I guess that experiments in this study needed very tough work.This manuscript is worth considering for publication in this journal, but it can be considered after addressing the following issues in revision.
Major comment 1.It is interesting that breast cancer cell-derived EVs affect neural metabolism, leading to brain metastases.However, as the authors know, this journal has a high impact on this research field, please emphasize how this idea is novel.They should develop the topic in the discussion section.
2. In the introduction section, please discuss the relationship between neural metabolites and brain (metastatic) tumor in more detail.Please describe how critical the level of glutamate, glutamine, and lactate for tumor growth.
3. Please make figure 5(A) understandable.In the experiments using brain slices, it is slightly difficult to understand the time course of the experiments.In some experiments, cancer cells were seeded on the same day (day 2) with EV treatment, but current figure 5(A) doesn't show this time course correctly.Moreover, in figure 5(A), the illustrations of EV treatment and cancer seeding are the same.It is confusing.Please change the illustration.4. In the experiments using brain slices, they demonstrated that the levels of glutamine and lactate were increased by EV treatment to brain slices, leading to significant cancer cell growth.These data are so amazing.They need to carefully verify the validity of these results.Please check the following points.First, EVs could be also taken up by cancer cells on the brain slice.They need to show data on EV-treated cancer cell growth.They showed data on MDA-231/miR-199b in vitro cell growth in figure 7(C).According to this data, I guess that cancer cell growth is down-regulated by EV treatment, but they need to show clearly.Second, they need to confirm whether the cancer cell growth is changed when the levels of glutamine and lactate in CM are up or downregulated.Thirdly, living cell numbers in the brain slices could be reduced by EV treatment.Current results could be obtained when EV treatment causes cell death and reduce living cell numbers.Finally, SLC transporters can be expressed not only in astrocytes and neurons but also in other types of cells in the brain slices.Please examine and discuss these points.
5. The authors need to show the bioluminescent image of the body in in vivo experiments (Figure 6).I am interested in the specificity of EV-miR-199b to the brain metastatic tumor.I wonder if EV-miR-199b have some effect on the tumor in other sites such as lung and lymph nodes.If they confirm that EV-miR-199b does not have any effect on lung or lymph nodes, they succeed in demonstrating that EV-miR199b function is specific to brain.
6.The authors should discuss the relationship between BBB and cancer cell-derived EV microRNAs more in detail.Please cite a following paper.Moreover, does miR-199b-5p have some function on BBB?They should confirm at least the difference of BBB penetration efficiencies between MDA-231/miR-199b EV and MDA-231/ctrl EV (Figure 6A).Tominaga, N., Kosaka, N., Ono, M. et al.Brain metastatic cancer cells release microRNA-181ccontaining extracellular vesicles capable of destructing blood-brain barrier.Nat Commun 6, 6716 (2015).https://doi.org/10.1038/ncomms7716Minor comment 0. Please show how to label cancer cells by mCherry in the methods section. 1. Please show the scale bar on the left picture in figure 5(B).

Response to Reviewers
We thank the editor and all reviewers for the constructive comments.In response to these comments, we made substantial revisions to the manuscript and added a considerable amount of new data from in vitro and in vivo experiments.Due to multiple factors that were out of our control in the past year, the revision process took much longer than usual.We appreciate your kind understanding.Below please find our point-by-point responses.

Reviewer #1:
This manuscript presents a fascinating hypothesis, that brain metastasizing tumor cells hijack astrocyte-neuron metabolic communication.If true, this will be an outstanding addition to the literature.A number of technical issues are listed below that require thorough examination before this conclusion can be verified.Overarching issues listed include the use of one high miR199b expressing cell line.We have no idea of how levels of this miR affect the biology.Also, empty EVs at times have intermediate effects-a better control would be to also test EVs with an unrelated miR.The brain metastasis assays are substandard.Images showing putative expression patterns cannot be clearly seen.
Response: We highly appreciate these constructive comments and have generated a new T47D cell line model with miR-199b overexpression as well as a new cell line secreting EVs carrying an unrelated miRNA (miR-211) as an additional control.The new results are summarized below in our responses to individual comments and have been added to the revised manuscript.
1. Can the analysis be repeated with samples that will also discriminate lung and liver metastases in Fig. 1? Response: We were able to obtain a lung-metastasizing variant of MDA-MB-231 (MDA-231-LM2) and a variant of T47D derived from a spontaneous brain metastasis (T47D-BR2).The levels of miR-199b in these additional cell line models and their EVs were measured by qPCR and results are added to new Fig.1d.Unfortunately, for the clinical samples assessed in Fig. 1, we do not have the complete clinical information regarding the status of lung and liver metastases.
2. Do other miRs bind to this sequence shown on Fig. 3? Response: The sequences shown in Fig. 3 (current Fig. 2a) are not predicted as binding sites for other miRNAs.
3. Are SLC1A2 only expressed by astrocytes and not neurons?Are SLC38A2 and SLC16A7 only expressed by neurons and not astrocytes?How about expression in other brain cell types?
Response: To obtain a comprehensive understanding of these genes' expression patterns in different types of brain cells, we re-analyzed a single cell gene expression dataset of human prefrontal cortex samples (GSE168408).Following cell clustering and cell type annotation, five major cell subsets were obtained (new Supplementary Fig. 3a).Expression of EAAT2 (SLC1A2) was the highest in astrocytes, whereas expression of SNAT2 (SLC38A2) and MCT2 (SLC16A7) were the highest in excitatory neurons (new Supplementary Fig. 3b,c).Oligodendrocytes also show some EAAT2 and MCT2 expression; however, in this study we focus on astrocytes and neurons for their metabolic coupling pathway.Response: We created a new MDA-MB-231-derived cell line that overexpressed an unrelated miRNA (miR-211) and secreted EVs carrying a high level of miR-211 as an additional control.This unrelated control miRNA did not alter the expression of SLC genes or nutrient consumptions, as shown in the new data added to new Fig.3b  and 4a.
5. The differences in Gln for neurons in 4E may be statistically significant but are truly minimal.Do the authors have any proof that they are of functional significance?How much miR-199b is used?Do EVs with different levels of this miR have graded effects?

4.
The experiments shown on Fig.4Eare critical.The empty EVs exerted some effect.Pls add an EV with another miR.