Exosomes to the Rescue

Acute kidney injury, AKI, is relatively common in the hospital setting, and beyond the identification of the underlying cause and supportive measures, little can be done to facilitate healing of the compromised kidney. Administered bone marrow–derived mesenchymal stem cells (MSCs), however, have been shown to accelerate recovery and repair from AKI, in both humans and mouse models of disease.^2,3 There is some debate regarding the mechanism by which this occurs, with two competing ideas prevailing; the first is that MSCs home in on the injured kidney, engraft, and differentiate, and the second is that some humoral factor is responsible for the proregenerative effect. Engraftment and differentiation, when observed, occur at a low frequency and thereby contribute marginally to renal tissue restoration. Engraftment-independent mechanisms appear to constitute the major repair mechanism for renal tissue after AKI.⁴ Indeed, the injection of cell-free supernatant derived for MSCs facilitates healing after AKI-induced injury to the same extent as the MSCs themselves.⁵ Candidates mediating this observed protective effect range from small peptide mediators to cytokines to extracellular vesicles (ECVs). Careful work by Bruno et al. clearly shows that the majority of the protective effect was due to ECVs.^6,7 In this issue of the Journal of the American Society of Nephrology, Collino, et al. show that most of this proregenerative effect is due to the microRNA (miRNA) content of ECVs produced by MSCs.⁸

ECV is a “grab-bag” term used to describe any vesicle made and released from a cell, including exosomes, ectosomes, and apoptotic bodies, as reviewed in greater detail elsewhere.⁹ Classic exosomes are produced within the intracellular endosome by the multivesicular body. This body is a large vacuole into which small vesicular packages are budded in an outside out membrane orientation to form the intraluminal vesicles, which will in turn become exosomes when released from the apical aspect of the cell.¹⁰ Ectosomes are derived from plasma membrane budding and contain membrane markers as well as a small sample of the cytoplasm. Apoptotic bodies arise from the extensive plasma membrane blebbing that occurs during apoptosis, and contain cytoplasm and densely packed organelles.¹¹ ECVs contain miRNAs and miRNA transfer via ECVs has been well established as a signaling mechanism in immunology and cancer biology.^12–14

In this issue, Collino et al. demonstrate that both MSCs and the MSC-derived ECVs are able to protect mice from AKI induced by intramuscular injection of glycerol, a model of AKI induced by rhabdomyolysis. In this model, MSCs or MSC-derived ECVs were given well after the onset of AKI, 3 days after glycerol injection, and strongly suppressed AKI-induced injury. Two days after ECV injection (5 days after AKI induction), the kidneys looked almost normal histologically and BUN levels were close to those seen in uninjured mice. Using RNA interference against the miRNA processing enzyme Drosha, the investigators were able to produce matched batches of MSCs or ECVs globally deficient in miRNAs (control-MSCs, Drosha-MSCs, ECV-control, and ECV-Drosha). Silencing of Drosha, the nuclear dsRNA ribonuclease responsible for the first step in miRNA maturation, led to a 55% reduction of all miRNAs in ECVs. Importantly and slightly surprisingly, knockdown of Drosha did not influence the expression of the classic MSC surface markers, the expression of ECV markers on the ECVs, or their ability to bind to and integrate with murine tubular epithelial cells. When MSCs were injected into mice with glycerol-induced AKI, the difference in outcome was stunning. In the case of control-MSC–treated AKI mice, BUN was reduced nearly to normal and the number of hyaline casts and the amount of tubular necrosis were also reduced. Mice injected with Drosha-MSCs had an outcome more or less identical to untreated AKI mice, suggesting that miRNAs control the well documented proregenerative effects of MSCs. When ECVs were introduced into AKI mice, the results were identical to those seen in mice treated with MSCs. Importantly, ECV-Drosha treatment did not ameliorate AKI-induced damage. These data suggest that the miRNA content of ECVs can profoundly reprogram an injured kidney, reduce inflammation and necrosis, and radically improve renal function, as measured by BUN.

Since miRNAs regulate mRNA stability, the investigators used RNA-seq to examine the poly-A transcriptome of normal, AKI, and experimentally treated mice. The control treatments, either ECVs or MSCs, had a large effect on the transcriptome of the kidneys compared with AKI alone; whereas the Drosha-knockdown treatment groups, either ECVs or MSCs, had a significantly lesser effect. For example, ECV-control reproducibly increased 610 transcripts and decreased 706 whereas their ECV-Drosha only increased 111 and decreased 129 transcripts.

Collino et al. utilized gene ontology to visualize the signaling pathways being modulated by MSC/ECV treatment. The authors showed that the kidney mRNAs increased in the pooled control-MSC/ECV treated group compared with mice with induced AKI were associated with amino acid, butanoate, and propanoate metabolism, the peroxisome proliferator-activated receptor signaling pathway, and the complement and coagulation cascades. Conversely, kidney mRNAs decreased in the control-MSC/EV treated groups were associated with inflammation, the extracellular matrix–receptor interactions, chemokine/cytokine signaling, and cell cycle.

Focusing on miRNAs found in ECVs and matching these to the 3′ untranslated regions of all mRNAs made in AKI kidney, the investigators showed that there were 209 potential gene targets regulated by miRNAs, of which 165 (77.8%) were still misregulated in Drosha-knockdown treatment groups. Among the 165 miRNA regulated targets, there is an over-representation of genes involved in extracellular matrix–receptor interaction, focal adhesion, Wnt signaling, and p53 pathways. Indeed, both the Wnt and p53 pathways have been implicated in the genesis of fibrosis and chronic kidney disease.^15,16

The authors further analyzed two miRNA regulated targets, Lipocalin-2 (Lcn2) and fibrinogen (Fg) subunits α, β, and γ, both markers of inflammation and tubular injury. Lcn and Fg subunits α, β, and γ were substantially increased in AKI. Both MSC-control and EV-control treatment decreased the level of Lcn2 transcripts and Fg β subunit protein, while MSC-Drosha and EV-Drosha did not. These data imply that Lcn2 and Fg subunit β may be useful markers for response to ECV therapy.

Clinically, an ECV-based strategy for ameliorating AKI-induced damage would be extremely useful, perhaps allowing some individuals to avoid dialysis and transplantation. However, ECVs from donor MSCs will have HLA antigens on their surface that may sensitize the patient to allo-HLA. Indeed, the authors used HLA class 1 as a FACS marker for ECVs. Allo-HLA recognition might be a drawback to an ECV-based therapy, but could be circumvented by producing ECVs that are HLA null or fully synthetic. Production of synthetic ECVs with a customized miRNA payload that are not recognized by the patient’s immune system will be the ultimate goal.

In summary, the MSCs are themselves potently proregeneratative in murine models of AKI, and this article demonstrates that the effect is mediated in large part by the miRNA content of the MSC ECVs. However, there are still some unanswered questions that will have to be addressed before therapies derived from these observations can be used routinely in the hospital setting: Which vesicular component of the ECV pool is responsible for the proregenerative effect, classic exosomes, ectosomes, or apoptotic bodies? How do the ECVs home in on injured tissue and what are the adhesins and ligands involved in the cell–ECV interaction? How do ECVs get to tubular epithelial cells? Are the ECVs trafficked through the glomerular basement membrane or are they transported across the vascular endothelium and the two basement membranes? Finally, which miRNAs are responsible for the proregenerative effect? These questions will have to be answered before ECV therapy can be used for general clinical use.

Disclosures

None.