The Role of MicroRNAs Carried by Extracellular Vesicles in Tumorigenesis Through Reprogramming the Mitochondrial Information Processing System

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

In this study, the authors provides a thorough overview of EV-miRNA-Mitochondria Axis and explores the potential of EV-miRNAs as non-invasive biomarkers. This paper presents an interesting discussion and holds the potential to be published in International Journal of Molecular Sciences with some modifications to enhance the clarity and depth of the article.

1. In the part of 1.3.3, it may be better to include cellular origins and target cells of the different types of EVs which deliver miRNAs.
2. It is recommended to include the therapeutic applications of mitochondrial suppression in tumor, especially in pancreatic cancer, in section 4.1.1.

Comments on the Quality of English Language

1. For the title, “reprograming”may be a spelling mistake, while “reprogramming” is more widely used in related researches.
2. The tense in the abstract should be kept consistent: “we summarized”should be followed by “we discussed” and “we highlighted”.
3. On page 6, line 1, section 2.2.1, “Dephosphorylation YBX1 acts as a sorting miRNA 217 into the EVs”should be “YBX1dephosphorylation sorts miRNA 217 into EVs”.

Author Response

Reviewer 1

 

 

1. In the part of 1.3.3, it may be better to include cellular origins and target cells of the different types of EVs which deliver miRNAs.

                                  

Thank you for this valuable suggestion. We have revised Section 1.3.3 to explicitly define the cellular origins and target cells of the EVs involved in apoptotic evasion.

 

Revised test line 100-113 / Part 1.3.3

 

“The intrinsic apoptotic pathway is governed by mitochondrial outer membrane permeabilization (MOMP), regulated by the Bcl-2 family protein. EV-miRNAs confer chemoresistance by disrupting this balance. These apoptosis-modulating EVs typically originate either from drug-resistant tumor cells which transfer their survival phenotype to neighboring, drug-sensitive tumor cells or from stromal cells within the tumor micro-environment. For example, Cancer-associated fibroblast (CAF)-derived EV-miR-21 to neighboring cancer cells, suppressing APAF1, inhibiting apoptosis, and inducing paclitaxel resistance [13]. EVs secreted from breast cancer stem cells (CDCs) and chemoresistant cells express high level of miR-155. miR-155 upregulation in chemo-resistant cells associated with epithelial-to-mesenchymal transition indicates transfer of EV-miR-155 from CDCs and chemoresistant cells [14]. Another example of stromal cells could transfer miRNA to liver cancer cells, when Zhang et al., [15] have revealed a significant downregulation of miR-320a level in CAF-derived EVs and that affect the cancer cell proliferation, migration and metastasis. The authors have also validated that miR-320a in liver cancer cells directly targets 3’-UTR of PBX3 mRNA, which in turns inhibit the activation of the MAPK pathway. It has already been demonstrated that intercellular crosstalk between tumor cells and fibroblasts is mediated by tumor-derived EVs. High-metastatic liver cancer cells secrete EV-encapsulated miR-1247-3p that directly targets B4GALT3, leading to activation of β1-integrin-NF-κB signaling in fibroblasts. Activated CAFs further promote lung metastasis of liver cancer [16].”

 

 

2. It is recommended to include the therapeutic applications of mitochondrial suppression in tumor, especially in pancreatic cancer, in section 4.1.1.

 

We appreciate the reviewer's insightful recommendation to emphasize the clinical relevance of this mechanism. We have expanded Section 4.1.1 to explicitly discuss the therapeutic applications of mitochondrial suppression in pancreatic cancer.

 

Revised test line 598-611 / Part 4.1.1

 

“Targeting mitochondrial metabolism has emerged as a highly attractive strategy for cancer therapy, particularly in pancreatic cancer, where suppressing mitochondrial function is known to severely repress tumor cell proliferation. The mitochondrial suppression directly affects the mitochondrial bioenergetics in the cancer cells. As a promising therapeutic application, human bone marrow mesenchymal stem cell (hBMSC)-derived exosomes can be utilized to deliver tumor-suppressive miRNAs, such as miR-484, directly to pancreatic tumors [8]. This disruption of mitochondrial metabolism downregulates the expression of key ETC genes, such as ND1, ND2, ATP5, Cytb and fatty acid oxidation regulators, like PPAR. Downregulation of ETC genes result in a significant decrease in mitochondrial ATP production and a concurrent trigger in mitochondrial ROS production. This ROS overload activates apoptosis and inhibits tumor growth, in vivo, highlighting how the EV-mitochondria axis can be hijacked therapeutically to induce metabolic exhaustion and mitigate malignant transformation in cancer cells [8]. The OXPHOS can increase survival of pancreatic cancer cells and impair the therapeutic importance of KRAS-ablation therapy. The clinical-grade engineered hBMSC-derived exosomes with the ability to target oncogenic Kras (iExosomes) confirm suppression of oncogenic Kras and an increase in the survival of several mouse models with pancreatic cancer [87].”

 

 

Comments on the Quality of English Language

 

1. For the title, “reprograming”may be a spelling mistake, while “reprogramming” is more widely used in related researches.
2. The tense in the abstract should be kept consistent: “we summarized”should be followed by “we discussed” and “we highlighted”.
3. On page 6, line 1, section 2.2.1, “Dephosphorylation YBX1 acts as a sorting miRNA 217 into the EVs”should be “YBX1dephosphorylation sorts miRNA 217 into EVs”.

 

Thank you for your careful review and for identifying these language and grammatical issues. We have implemented all of the suggested corrections, including correcting the spelling of “reprogramming” in the title, ensuring consistent verb tense throughout the abstract, and revising the sentence in Section 2.2.1 as recommended. In addition, we have thoroughly reviewed the entire manuscript to correct grammatical errors and improve the overall clarity and readability of the text.

 

Reviewer 2 Report

Comments and Suggestions for Authors

This review focuses on the “EV–miRNA–mitochondria axis” and systematically discusses how extracellular vesicles derived from tumor cells or components of the tumor microenvironment selectively package miRNAs and influence tumor initiation and progression by regulating the mitochondrial information-processing system in recipient cells. The manuscript first introduces the classification and biogenesis of EVs, as well as mechanisms of selective miRNA loading, including RNA-binding proteins such as hnRNPA2B1, SYNCRIP, Lupus La, and YBX1, EXOmotifs, 3′-end modifications, and lipid raft/ceramide-dependent pathways. It then highlights how EV-associated miRNAs mediate metabolic reprogramming, dysregulation of mitochondrial fission and fusion, evasion of apoptosis, and therapy resistance. The review further discusses the transfer of mtDNA and intact mitochondria through EVs and their implications for tumor metabolic adaptation and immune regulation. In the clinical section, the authors cover several malignancies, including pancreatic cancer, glioblastoma, breast cancer, hematologic malignancies, ovarian cancer, and renal cancer, emphasizing the potential of EV-miRNAs as non-invasive biomarkers for diagnosis, prognosis, and therapeutic monitoring. Overall, the topic is timely and of considerable interest. The manuscript integrates EV biology, miRNA sorting, and mitochondrial cancer metabolism; however, it would benefit from improved logical focus, clearer differentiation of evidence levels, greater terminological consistency, and more standardized presentation of key mechanistic figures and tables.

1. How do the authors distinguish experimentally validated EV-miRNA–mitochondria interactions from mechanistic associations inferred from correlative or bioinformatic studies?
2. Could the authors clarify whether the proposed “Mitochondrial Information Processing System” represents an established conceptual framework or their own integrative model?
3. How do the authors address the stoichiometric limitation of miRNA abundance per EV when interpreting functional effects in recipient cells?
4. Could the manuscript better differentiate tumor-promoting versus tumor-suppressive EV-miRNAs across different cancer types and cellular sources?
5. What methodological standards should be prioritized to validate EV-miRNAs as clinically reliable biomarkers for mitochondrial reprogramming in cancer?

Author Response

 

1. How do the authors distinguish experimentally validated EV-miRNA-mitochondria interactions from mechanistic associations inferred from correlative or bioinformatic studies?

 

We strongly agree with the reviewer that distinguishing computational/correlative predictions from experimentally validated interactions is critical for scientific rigor. To address this, we have modified our summary table (Table 2) to include a new 'Method' column. This column explicitly defines the evidence base for each cited EV-miRNA–mitochondria interaction, clarifying whether the mechanism was inferred from clinical correlation or confirmed via direct experimental validation.

 

Revised Table 2

 

Cancer Type	Key EV-miRNA	Mitochondrial/Molecular Effect	Clinical Outcome	Method	Citation
Pancreatic	miR-484	Suppresses Wnt/MAPK; Increases ROS; Decreases ATP	Tumor suppression; Metabolic exhaustion	Experimentally validated (in vitro assays & in vivo xenograft models)	[8]
Glioblastoma	miR-451a	Targets LKB1/AMPK; Metabolic stress adaptation	Glucose sensing; Niche remodeling	Experimentally validated (in vitro metabolic assays)	[31]
Breast	miR-122	Targets PKM/GLUT1; Suppresses stromal glucose uptake	Nutrient hijacking (Reverse Warburg)	Experimentally validated (in vivo pre-metastatic niche models)	[18]
Breast	miR-221/222	Targets PUMA/p27; Blocks mitochondrial apoptosis	Tamoxifen/Adriamycin Resistance	Experimentally validated (in vitro drug resistance and apoptosis assays)	[2], [5]
AML	miR-125b	Targets BAK1; Inhibits apoptosis	Chemoresistance; HSC suppression	Clinical correlation & in vitro validation (Patient serum profiling & cell viability assays)	[2]
Ovarian	miR-424	Targets MYB; Downregulates VEGF	Inhibition of Angiogenesis	Experimentally validated (Dual-luciferase reporter assay & in vivo murine models)	[101]
Renal	miR-210	Targets ISCU/COX10; Inhibits ETC	Hypoxic adaptation; Glycolysis	Experimentally validated (in vitro hypoxia models)	[21]

 

2. Could the authors clarify whether the proposed “Mitochondrial Information Processing System” represents an established conceptual framework or their own integrative model?

 

We appreciate the reviewer's request for clarification regarding the origin of this concept. The 'Mitochondrial Information Processing System' (MIPS) represents a recently established conceptual framework within the broader literature, rather than an integrative model proposed exclusively by our group. It characterizes the mitochondrion as a central signaling organelle that communicates with other cellular compartments to transduce input-to-output biological information. To avoid any ambiguity for the reader, we have already included the appropriate supporting citation (Kuo et al., 2024) where the term is introduced.

 

3. How do the authors address the stoichiometric limitation of miRNA abundance per EV when interpreting functional effects in recipient cells?

 

We sincerely thank the reviewer for raising this critical question regarding the very low miRNA copy number per EV imposes challenging functional consequence in recipient cells.  We already touch upon this debate in Section 2.3.1, but we have modified the paragraph to be more explicit.

 

Added at Lines 369-372 in Section 2.3.1

“Furthermore, the continuous secretion and accumulation of these EVs in the acidic and hypoxic TME concentrates these signals locally, allowing even low-abundance miRNAs to reach the cumulative threshold levels required to repress mitochondrial function in recipient cells [51], [52]. In mammalian cells, approximately 100–1,000 copies of a specific miRNA per cell are needed to measurably reduce target-mRNA levels. Assuming uptake efficiencies of 0.1–1 % of EV cargo and <1 functional miRNA per EV, the number of EVs that must be internalized per cell to reach this threshold. ”

 

 

4. Could the manuscript better differentiate tumor-promoting versus tumor-suppressive EV-miRNAs across different cancer types and cellular sources?

 

We sincerely thank the reviewer for this insightful comment. We agree that explicitly distinguishing between the dual, opposing roles of EV-miRNAs significantly enhances the clarity and mechanistic depth of the manuscript. To address this efficiently, we have introduced a new table (Table 1).

 

Table 1 has been inserted after Line 575.

 

“Table 1: EV-miRNA Functions Based on Tumor-Promoting vs. Tumor-Suppressive Roles

 

Functional Role	EV-miRNA	Cancer Type	Cellular Source (Donor)	Target Cell (Recipient)	Key Mechanism & Biological Outcome	References
Tumor-Promoting	miR-105	Breast	Metastatic breast cancer cells	Endothelial cells	Destroys ZO-1 and tight junctions, promoting vascular leakiness and distant metastasis.	[80]
	miR-155	Pancreatic	Pancreatic cancer cells	Normal fibroblasts	Converts normal fibroblasts into pro-tumorigenic Cancer-Associated Fibroblasts (CAFs).	[81]
	miR-92a-3p	Colorectal (CRC)	Cancer-Associated Fibroblasts (CAFs)	CRC tumor cells	Promotes chemoresistance to 5-fluorouracil (5-FU) and oxaliplatin therapies.	[82]
	miR-21	Liver (HCC) / Breast	Tumor cells, Adipocytes	Hepatic stellate cells, Macrophages	Activates CAFs, drives macrophage M2 polarization, and stimulates angiogenesis.	[83]
Tumor-Suppressive	miR-484	Pancreatic	Human bone marrow MSCs (hBMSCs)	Pancreatic cancer cells	Deactivates the Wnt/MAPK pathway, causing ATP depletion, ROS accumulation, and tumor shrinkage.	[8]
	miR-34a-5p	Colorectal (CRC)	Mesenchymal stem cells (MSCs)	CRC tumor cells	Suppresses c-MYC/DNMT3a/PTEN axis, inhibiting tumor growth and epithelial-mesenchymal transition (EMT).	[84]
	miR-124-5p	Acute Myeloid Leukemia (AML)	Bone marrow MSCs (BMSCs)	Leukemic cells	Inhibits cell-cycle progression and induces apoptosis in AML cells.	[85]
	miR-655-3p	Esophageal (ESCC)	Human umbilical cord MSCs (hUCMSCs)	ESCC tumor cells	Inactivates HIF-1α via the LMO4/HDAC2 axis, significantly inhibiting liver metastasis.	[86]

“

 

 

5. What methodological standards should be prioritized to validate EV-miRNAs as clinically reliable biomarkers for mitochondrial reprogramming in cancer?

 

We sincerely appreciate the reviewer highlighting this critical translational hurdle. We completely agree that establishing stringent methodological standards is paramount for validating EV-miRNAs as clinically reliable biomarkers. To address this, we have added a new subsection titled 'Methodological Standards for Clinical Validation of EV-miRNA Biomarkers' to the manuscript.

 

Added at Lines 729

 

“5. Methodological Standards for Clinical Validation of EV-miRNA Biomarkers

While EV-miRNAs hold immense promise as non-invasive biomarkers for tracking tumor progression and metabolic shifts, their clinical validation requires overcoming significant technical challenges and establishing rigorous methodological standards [4]. A primary hurdle is the lack of standardized isolation methods, which can lead to the co-isolation of non-vesicular proteins and lipoproteins [37]. Unfortunately, most current miRNA biomarkers are unlikely to have much clinical utility. As we and others have shown, strategies for identifying miRNA biomarkers in the first wave of discovery were inadequate and challenged by variable isolation methods, non-standard analysis designs, and an underappreciation of the cellular source of a miRNA [107]. To ensure clinical reliability, accurate pre-analytical processing is essential. For instance, utilizing plasma rather than serum is often preferred to avoid contamination from coagulation-associated EVs released by platelets and lysed blood cells [37]. Furthermore, because EVs can associate with protein complexes bound to free-circulating miRNAs, treatments utilizing proteinase K and RNase A prior to miRNA extraction are heavily recommended during the biomarker discovery and clinical validation phases to remove extravesicular contamination [37]. While these enzymatic treatments may be too cumbersome for routine, high-throughput clinical diagnostics, they are essential for establishing initial biomarker accuracy; failure to implement these steps during validation can result in up to a 70% over-estimation of EV-miRNA quantities in patient plasma due to externally bound contaminants. ensuring the analysed signature is genuinely encapsulated within the EV [4]. Analytical variability poses another major challenge, as current RNA quality control standards were originally optimized for cellular RNA and do not accurately represent the unique RNA cargo typically found in small EVs. Furthermore, achieving effective RNA quantification requires the establishment of robust biological normalization controls (such as stable small non-coding transcripts) to overcome technical variability before downstream computational processing [4]. Therefore, rigorous normalization strategies must be implemented. This includes establishing optimal endogenous reference miRNAs (such as miR-16-5p, miR-423-3p, or miR-191-5p) or utilizing exogenous normalizers (e.g., spike-in cel-miR-39) to reduce technical noise and inter-individual variability across clinical samples [37]. Finally, to ensure reproducibility and facilitate successful clinical translation, studies must adhere to universally established reporting frameworks, such as the MISEV 2023 (Minimal Information for Studies of Extracellular Vesicles) guidelines, which provide foundational criteria for EV isolation and characterization [37].”

Reviewer 3 Report

Comments and Suggestions for Authors

This review article presents a comprehensive overview of the emerging role of extracellular vesicle (EV)-associated microRNAs in regulating mitochondrial function during tumorigenesis. The manuscript integrates multiple aspects of EV biology, including biogenesis, miRNA sorting, mitochondrial dynamics, metabolic reprogramming, apoptosis, and intercellular communication across various malignancies. The topic is timely and potentially relevant to the fields of cancer metabolism, extracellular vesicles, and translational oncology. However, several conceptual, structural, and scientific issues should be addressed before the manuscript can be considered for publication. My comments are described as follows:

 

Comments:

1. The review introduces the concept of the “Mitochondrial Information Processing System (MIPS)” as a central framework in tumorigenesis; however, could the authors clarify how this concept differs from previously established models of mitochondrial signaling and metabolic reprogramming in cancer?
2. Although the manuscript extensively discusses EV-miRNA-mediated mitochondrial regulation, could the authors provide a more critical evaluation of whether these interactions are cancer-type specific or represent universal oncogenic mechanisms across malignancies?
3. The manuscript describes selective miRNA sorting through RBPs such as hnRNPA2B1, SYNCRIP, and Lupus La; however, are there quantitative or comparative studies demonstrating which sorting pathway predominates under hypoxic or therapeutic stress conditions?
4. Several sections describe EV-miRNAs targeting mitochondrial dynamics and apoptosis, but could the authors discuss whether these mitochondrial alterations are reversible and how long-lasting the phenotypic effects remain in recipient cells?
5. The review highlights the role of EV-miR-210 in regulating ETC components and glycolytic adaptation; however, could the authors clarify whether miR-210 exerts similar mitochondrial effects in both tumor cells and stromal cells, or whether the downstream consequences differ according to cellular context?
6. While the manuscript discusses mitochondrial transfer and “mitovesicles,” the mechanistic distinction between EV-mediated mtDNA transfer and whole mitochondrial transfer remains somewhat unclear. Could the authors further clarify the biological significance and relative contribution of these two processes in tumor progression?
7. The authors propose EV-miRNAs as promising non-invasive biomarkers; however, could they discuss the current limitations regarding standardization of EV isolation methods, miRNA normalization strategies, and inter-study reproducibility that may hinder clinical translation?
8. Several studies cited in the review rely heavily on in vitro or preclinical evidence. Could the authors provide a clearer distinction between findings validated in clinical patient cohorts and those that remain largely experimental?
9. The manuscript emphasizes the oncogenic role of EV-miRNAs, but could the authors expand the discussion regarding tumor-suppressive EV-miRNAs and whether these molecules may also contribute to anti-tumor mitochondrial reprogramming?
10. The conclusion suggests that EV-based delivery of anti-miRs or miRNA mimics may represent future therapeutic strategies; however, could the authors discuss the major translational barriers, including EV targeting specificity, biodistribution, immunogenicity, and large-scale manufacturing challenges?

Author Response

 

1. The review introduces the concept of the “Mitochondrial Information Processing System (MIPS)” as a central framework in tumorigenesis; however, could the authors clarify how this concept differs from previously established models of mitochondrial signaling and metabolic reprogramming in cancer?

 

We appreciate the reviewer’s insightful question regarding the theoretical distinctions between these frameworks. The 'Mitochondrial Information Processing System' (MIPS) represents a recently established conceptual framework within the broader literature, rather than an integrative model proposed exclusively by our group. It characterizes the mitochondrion as a central signalling organelle that communicates with other cellular compartments to transduce input-to-output biological information. To avoid any ambiguity for the reader, we have already included the appropriate supporting citation (Kuo et al., 2024) where the term is introduced.

To clarify, traditional models of metabolic reprogramming such as the classical Warburg effect have historically viewed mitochondria primarily as passive bioenergetic factories. Those established models focus heavily on localized, intracellular shifts in energy production (e.g., transitioning from oxidative phosphorylation to aerobic glycolysis) to support the immediate demands of tumor proliferation.

In contrast, the Mitochondrial Information Processing System (MIPS) represents a more recent and integrated paradigm. It conceptualizes mitochondria not just as powerhouses, but as dynamic, central signalling organelles. Because the primary scope of our manuscript is to detail the specific mechanisms of EV-miRNA-mediated communication rather than to provide a historical comparison of broader metabolic models, we felt that delving into this theoretical distinction within the text might distract from the main narrative. However, if the reviewer feels that explicitly including this distinction would benefit the readers, we would be happy to incorporate a brief explanation into the revised manuscript.

 

1. Although the manuscript extensively discusses EV-miRNA-mediated mitochondrial regulation, could the authors provide a more critical evaluation of whether these interactions are cancer-type specific or represent universal oncogenic mechanisms across malignancies?

 

We sincerely appreciate the reviewer’s insightful comment. We agree that critically evaluating functional dichotomy provides an important framework for the reader.

 

In Section 3 ('The EV-miRNA-Mitochondria Axis of Oncogenesis'), we outline the universal strategies shared across malignancies. And conversely, to emphasize how the precise molecular execution of these universal strategies is highly cancer-type specific, we dedicated Section 4 ('Clinical Implications of EV-Mediated Mitochondrial Reprogramming in Malignancies') to evaluating distinct tumor microenvironments. Within this section, we explicitly contrast how unique malignancies adapt their EV-crosstalk. Furthermore, We have now added summary tables (Table 1) of these distinct EV-miRNA networks, systematically categorizing them by their unique cellular sources and specific cancer types.

 

“Table 1: EV-miRNA Functions Based on Tumor-Promoting vs. Tumor-Suppressive Roles

 

Functional Role	EV-miRNA	Cancer Type	Cellular Source (Donor)	Target Cell (Recipient)	Key Mechanism & Biological Outcome	References
Tumor-Promoting	miR-105	Breast	Metastatic breast cancer cells	Endothelial cells	Destroys ZO-1 and tight junctions, promoting vascular leakiness and distant metastasis.	[80]
	miR-155	Pancreatic	Pancreatic cancer cells	Normal fibroblasts	Converts normal fibroblasts into pro-tumorigenic Cancer-Associated Fibroblasts (CAFs).	[81]
	miR-92a-3p	Colorectal (CRC)	Cancer-Associated Fibroblasts (CAFs)	CRC tumor cells	Promotes chemoresistance to 5-fluorouracil (5-FU) and oxaliplatin therapies.	[82]
	miR-21	Liver (HCC) / Breast	Tumor cells, Adipocytes	Hepatic stellate cells, Macrophages	Activates CAFs, drives macrophage M2 polarization, and stimulates angiogenesis.	[83]
Tumor-Suppressive	miR-484	Pancreatic	Human bone marrow MSCs (hBMSCs)	Pancreatic cancer cells	Deactivates the Wnt/MAPK pathway, causing ATP depletion, ROS accumulation, and tumor shrinkage.	[8]
	miR-34a-5p	Colorectal (CRC)	Mesenchymal stem cells (MSCs)	CRC tumor cells	Suppresses c-MYC/DNMT3a/PTEN axis, inhibiting tumor growth and epithelial-mesenchymal transition (EMT).	[84]
	miR-124-5p	Acute Myeloid Leukemia (AML)	Bone marrow MSCs (BMSCs)	Leukemic cells	Inhibits cell-cycle progression and induces apoptosis in AML cells.	[85]
	miR-655-3p	Esophageal (ESCC)	Human umbilical cord MSCs (hUCMSCs)	ESCC tumor cells	Inactivates HIF-1α via the LMO4/HDAC2 axis, significantly inhibiting liver metastasis.	[86]

“

 

 

3. The manuscript describes selective miRNA sorting through RBPs such as hnRNPA2B1, SYNCRIP, and Lupus La; however, are there quantitative or comparative studies demonstrating which sorting pathway predominates under hypoxic or therapeutic stress conditions?

 

We appreciate the reviewer raising this point. Understanding which sorting pathways quantitatively dominate under stress is a critical frontier in the field. However, based on our review of the current literature, there are no comprehensive comparative studies that definitively establish a single 'predominant' pathway under generalized hypoxia or therapeutic stress. To address this in the text, we have added a brief paragraph ('Dynamic Regulation of miRNA Sorting Under Stress') to Section 2.2.

 

Added at Lines 246-261

 

“There is currently a lack of comprehensive, quantitative head-to-head studies establishing whether any single cargo sorting pathway predominates under specific conditions such as hypoxia or therapeutic stress. Instead, the available evidence supports a highly dynamic and context-dependent process in which distinct stressors engage multiple, partially overlapping sorting mechanisms. For example, while classical RNA-binding proteins like hnRNPA2B1 and SYNCRIP are well documented, cells exposed to noxious stimuli can alternatively utilize Caveolin-1 (CAV1) to selectively sort miRNAs into vesicles [8]. Similarly, cellular stress can trigger reversible binding between the RNA-binding protein HuR and specific miRNAs (such as miR-122), actively augmenting their extracellular export to manage the stress response. Other alternative mechanisms, such as ceramide-dependent secretion and Ago2-associated loading, also reflect the complex regulation of EV cargo under varying microenvironmental pressures [8]. Ultimately, EV-miRNA sorting under stress does not appear to rely on one universal pathway, but rather on a flexible, stimulus-specific network of sorting machineries.”

 

 

4. Several sections describe EV-miRNAs targeting mitochondrial dynamics and apoptosis, but could the authors discuss whether these mitochondrial alterations are reversible and how long-lasting the phenotypic effects remain in recipient cells?

 

We agree with the reviewer that the duration of these phenotypic effects is an important biological consideration. Because miRNAs function post-transcriptionally, the mitochondrial alterations they induce are transient; kinetic studies show that EV-RNAs progressively degrade after uptake (Chakrabortty et al., 2015). We have added a brief clarification regarding the reversible nature of this mechanism to Section 3.

 

Added at Lines 541

“Because EV-miRNAs act as post-transcriptional regulators rather than inducing permanent genetic changes, their effects on recipient mitochondrial dynamics and apoptosis are inherently transient. Kinetic studies demonstrate that internalized EV-RNAs progressively degrade over time, indicating that the maintenance of this metabolic reprogramming relies entirely on a continuous supply of tumor-derived EVs [79].”

 

5. The review highlights the role of EV-miR-210 in regulating ETC components and glycolytic adaptation; however, could the authors clarify whether miR-210 exerts similar mitochondrial effects in both tumor cells and stromal cells, or whether the downstream consequences differ according to cellular context?

 

We thank the reviewer for this excellent question. We agree that clarifying the context-dependent nature of EV-miR-210 is important. To address this, we have added a brief clarification to Section 3.1.1

 

Added at Lines 387

 

“Importantly, the downstream consequences of EV-miR-210 are highly dependent on the cellular context. While it targets ETC components in stromal fibroblasts to force a glycolytic shift, the same miRNA alternatively targets inhibitors like Ephrin-A3 (EFNA3) when internalized by endothelial cells, thereby promoting angiogenesis rather than purely altering bioenergetics [3].”

 

 

6. While the manuscript discusses mitochondrial transfer and “mitovesicles,” the mechanistic distinction between EV-mediated mtDNA transfer and whole mitochondrial transfer remains somewhat unclear. Could the authors further clarify the biological significance and relative contribution of these two processes in tumor progression?

 

We thank the reviewer for highlighting the need to distinguish between these two mechanisms. To address this, we have added a brief clarification to Section 3.4, explicitly contrasting their roles.

 

Added at Lines 486

 

“Mechanistically, these processes serve distinct biological roles in tumor progression. While the transfer of whole, functional mitochondria primarily provides a direct physical rescue to replace damaged machinery and restore OXPHOS, the transfer of mtDNA fragments frequently acts as a remote signaling mechanism, functioning as damage-associated molecular patterns (DAMPs) that activate pathways like TLR9 to polarize the immune microenvironment toward an immunosuppressive state [56].”

 

 

7. The authors propose EV-miRNAs as promising non-invasive biomarkers; however, could they discuss the current limitations regarding standardization of EV isolation methods, miRNA normalization strategies, and inter-study reproducibility that may hinder clinical translation?

 

We sincerely appreciate the reviewer raising this important point regarding the hurdles of clinical translation. We completely agree that the lack of standardization across isolation and analytical methods is currently the most significant bottleneck in the field. To address this (and in accordance with a similar comment raised during the review process), we have added a dedicated subsection titled 'Methodological Standards for Clinical Validation of EV-miRNA Biomarkers' to the manuscript. In this section, we explicitly discuss the current limitations hindering clinical translation, specifically addressing the points raised by the reviewer:

 

Added at Lines 729

 

“5. Methodological Standards for Clinical Validation of EV-miRNA Biomarkers

While EV-miRNAs hold immense promise as non-invasive biomarkers for tracking tumor progression and metabolic shifts, their clinical validation requires overcoming significant technical challenges and establishing rigorous methodological standards [4]. A primary hurdle is the lack of standardized isolation methods, which can lead to the co-isolation of non-vesicular proteins and lipoproteins [37]. Unfortunately, most current miRNA biomarkers are unlikely to have much clinical utility. As we and others have shown, strategies for identifying miRNA biomarkers in the first wave of discovery were inadequate and challenged by variable isolation methods, non-standard analysis designs, and an underappreciation of the cellular source of a miRNA [107]. To ensure clinical reliability, accurate pre-analytical processing is essential. For instance, utilizing plasma rather than serum is often preferred to avoid contamination from coagulation-associated EVs released by platelets and lysed blood cells [37]. Furthermore, because EVs can associate with protein complexes bound to free-circulating miRNAs, treatments utilizing proteinase K and RNase A prior to miRNA extraction are heavily recommended during the biomarker discovery and clinical validation phases to remove extravesicular contamination [37]. While these enzymatic treatments may be too cumbersome for routine, high-throughput clinical diagnostics, they are essential for establishing initial biomarker accuracy; failure to implement these steps during validation can result in up to a 70% over-estimation of EV-miRNA quantities in patient plasma due to externally bound contaminants. ensuring the analysed signature is genuinely encapsulated within the EV [4]. Analytical variability poses another major challenge, as current RNA quality control standards were originally optimized for cellular RNA and do not accurately represent the unique RNA cargo typically found in small EVs. Furthermore, achieving effective RNA quantification requires the establishment of robust biological normalization controls (such as stable small non-coding transcripts) to overcome technical variability before downstream computational processing [4]. Therefore, rigorous normalization strategies must be implemented. This includes establishing optimal endogenous reference miRNAs (such as miR-16-5p, miR-423-3p, or miR-191-5p) or utilizing exogenous normalizers (e.g., spike-in cel-miR-39) to reduce technical noise and inter-individual variability across clinical samples [37]. Finally, to ensure reproducibility and facilitate successful clinical translation, studies must adhere to universally established reporting frameworks, such as the MISEV 2023 (Minimal Information for Studies of Extracellular Vesicles) guidelines, which provide foundational criteria for EV isolation and characterization [37].”

 

 

8. Several studies cited in the review rely heavily on in vitro or preclinical evidence. Could the authors provide a clearer distinction between findings validated in clinical patient cohorts and those that remain largely experimental?

 

We sincerely appreciate the reviewer raising this important point. We completely agree that clearly distinguishing between purely experimental findings (pre-clinical models) and those validated in clinical patient cohorts is crucial for providing a realistic translational outlook. To address this distinction comprehensively without disrupting the narrative flow of the manuscript, we have explicitly detailed these validation parameters in our summary table (Table 2: Cancer-Specific EV-miRNA-Mitochondria Interactions). Specifically, we included a dedicated 'Method' column that systematically categorizes the level of evidence for each cited mechanism.

 

Revised Table 2 / Line 696

 

Cancer Type	Key EV-miRNA	Mitochondrial/Molecular Effect	Clinical Outcome	Method	Citation
Pancreatic	miR-484	Suppresses Wnt/MAPK; Increases ROS; Decreases ATP	Tumor suppression; Metabolic exhaustion	Experimentally validated (in vitro assays & in vivo xenograft models)	[8]
Glioblastoma	miR-451a	Targets LKB1/AMPK; Metabolic stress adaptation	Glucose sensing; Niche remodeling	Experimentally validated (in vitro metabolic assays)	[31]
Breast	miR-122	Targets PKM/GLUT1; Suppresses stromal glucose uptake	Nutrient hijacking (Reverse Warburg)	Experimentally validated (in vivo pre-metastatic niche models)	[18]
Breast	miR-221/222	Targets PUMA/p27; Blocks mitochondrial apoptosis	Tamoxifen/Adriamycin Resistance	Experimentally validated (in vitro drug resistance and apoptosis assays)	[2], [5]
AML	miR-125b	Targets BAK1; Inhibits apoptosis	Chemoresistance; HSC suppression	Clinical correlation & in vitro validation (Patient serum profiling & cell viability assays)	[2]
Ovarian	miR-424	Targets MYB; Downregulates VEGF	Inhibition of Angiogenesis	Experimentally validated (Dual-luciferase reporter assay & in vivo murine models)	[101]
Renal	miR-210	Targets ISCU/COX10; Inhibits ETC	Hypoxic adaptation; Glycolysis	Experimentally validated (in vitro hypoxia models)	[21]

 

 

9. The manuscript emphasizes the oncogenic role of EV-miRNAs, but could the authors expand the discussion regarding tumor-suppressive EV-miRNAs and whether these molecules may also contribute to anti-tumor mitochondrial reprogramming?

 

We sincerely thank the reviewer for this suggestion. To address this without disrupting the manuscript's narrative flow, we have added a brief clarification to Section 3.1 emphasizing that the EV-miRNA-mitochondria axis is not exclusively oncogenic.

 

Added at Lines 369

 

“Importantly, the EV-miRNA-mitochondria axis is not exclusively oncogenic; it exhibits a functional duality where specific EVs can also deliver tumor-suppressive miRNAs to drive anti-tumor mitochondrial reprogramming [58]. For instance, the EV-mediated transfer of miRNAs such as let-7a and miR-126 can actively suppress OXPHOS and alter cellular metabolism to halt tumor progression [59]. Furthermore, EVs delivering miR-27b or mesenchymal stem cell-derived miR-484 can severely disrupt mitochondrial activity in target cancer cells, leading to lethal ROS accumulation, bioenergetic exhaustion, and mitophagy-mediated apoptosis [8], [60].”

 

 

1. The conclusion suggests that EV-based delivery of anti-miRs or miRNA mimics may represent future therapeutic strategies; however, could the authors discuss the major translational barriers, including EV targeting specificity, biodistribution, immunogenicity, and large-scale manufacturing challenges?

 

We sincerely thank the reviewer for highlighting these critical translational hurdles. We completely agree that while EV-based therapies hold immense promise, providing a realistic discussion of the current clinical bottlenecks is essential for a balanced conclusion. To address this, we have expanded our Conclusion section to explicitly discuss the four major translational barriers identified by the reviewer.

 

 

Added at Lines 750

“The clinical translation of EV-based miRNA therapeutics must overcome several formidable barriers. Systemically administered EVs are frequently subjected to rapid clearance by the mononuclear phagocyte system, leading to poor tumor biodistribution and potential off-target effects that necessitate advanced surface engineering to improve tumor tropism [3], [108]. Furthermore, while native EVs generally exhibit low immunogenicity, potential immune and inflammatory reactions to repeated administrations of engineered or allogeneic EVs remain a clinical safety concern [5]). Finally, large-scale clinical deployment is currently hindered by significant manufacturing challenges, including the lack of standardized isolation protocols, low vesicle yields, batch-to-batch inconsistency, and suboptimal miRNA loading efficiencies [4].”

Round 2

Reviewer 3 Report

Comments and Suggestions for Authors

Dear Authors,

Thank you for your careful revision and responses to the reviewers' comments. The manuscript has been substantially improved, and I have no further questions or concerns.