Extracellular vesicle-mediated transfer of genetic information between the hematopoietic system and the brain in response to inflammation

Abstract

Mechanisms behind how the immune system signals to the brain in response to systemic inflammation are not fully understood. Transgenic mice expressing Cre recombinase specifically in the hematopoietic lineage in a Cre reporter background display recombination and marker gene expression in Purkinje neurons. Here we show that reportergene expression in neurons is caused by intercellular transfer of functional Cre recombinase messenger RNA from immune cells into neurons in the absence of cell fusion. In vitro purified secreted extracellular vesicles (EVs) from blood cells contain Cre mRNA, which induces recombination in neurons when injected into the brain. Although Cre-mediated recombination events in the brain occur very rarely in healthy animals, their number increases considerably in different injury models, particularly under inflammatory conditions, and extend beyond Purkinje neurons to other neuronal populations in cortex, hippocampus, and substantia nigra. Recombined Purkinje neurons differ in their miRNA profile from their nonrecombined counterparts, indicating physiological significance. These observations reveal the existence of a previously unrecognized mechanism to communicate RNA-based signals between the hematopoietic system and various organs, including the brain, in response to inflammation.

Figure 1. Noninvasive tracing of hematopoietic Cre recombinase activity.

(A) Expressing Cre recombinase specifically in the hematopoietic lineage reveals the contribution of hematopoietic cells to nonhematopoietic tissues by irreversibly switching on reporter gene expression after excision of a floxed stop codon. β-galactosidase expression after recombination can be observed in multiple tissues such as liver (B), lung epithelia (C), and small intestine (D). In the intestine, the labeling of an entire crypt indicates recombination of an intestinal stem cell with concomitant inheritance of the marker to all progeny. (E) Overview of a cerebellar section showing a single recombined Purkinje neuron expressing GFP (green). (F) High magnification of a Purkinje neuron in serial section confocal analysis confirming absence of a second nucleus, excluding cell fusion. Scale bar, 50 µm (B), 20 µm (C), 50 µm (D), 50 µm (E), and 10 µm (F).

Figure 2. Cre mRNA is present in the blood plasma of Vav-iCre mice and contained in EVs including exosomes.

(A) Cre mRNA can be detected in vesicle preparations enriched for exosomes from the blood plasma of Vav-iCre mice by RT-PCR. Each lane represents a result from an individual animal. For detection of Cre mRNA, nested primer PCR was used. The PCR product at 100 bp represents the signal after the second round of amplification. (B) Cre mRNA is localized in the pellet but not in the supernatant after ultracentrifugation of conditioned medium from primary Vav-iCre–positive hematopoietic cells after stimulation by LPS in vitro. Cre mRNA was resistant to RNaseA treatment in all experiments. (C) After treatment with Triton-X to lyse EVs in combination with RNaseA digestion, Cre mRNA is no longer detectable in contrast to RNaseA treatment alone. (D) Vesicular structures between 50 and 100 nm in size were visualized in electron micrographs from Vav-iCre hematopoietic-cell-derived vesicle preparations (scale bar, 50 nm). (E and F) Secreted membrane vesicle subspecies can be separated by density by sucrose gradient ultracentrifugation. Exosomal identity was confirmed by blotting against the specific protein markers ADAM10 and CD9 for all subfractions. (G) Cre protein could not be detected in any of the fractions. Positive controls for all antibodies are shown in boxes to the right. (H) Cre mRNA is present in the exosomal fractions 2–7. The nonexosomal vesicles fractions or apoptotic bodies are characterized by their variability of positive subfractions to complete absence of Cre mRNA. In this experiment, subfractions 9 and 10 are positive, whereas 8, 11, and 12 do not contain any Cre RNA.

Figure 3. EVs containing Cre mRNA are sufficient to induce recombination in Purkinje neurons after intracerebellar injection in vivo.

(A) EV preparations enriched for exosomes prepared from the peripheral blood and bone marrow of Vav-iCre mice were brought into the circulation by tail vein injection or were directly injected into the cerebellum. Injection of Cre RNA-containing EVs into tail veins did not lead to recombination events in the brain (n = 4). (B) β-galactosidase–positive Purkinje neuron in the cerebellum of a reporter mouse 4 d after intracerebellar injection of EVs. (C) Other reporter-gene–positive cells with a shape and size reminiscent of glial cells in proximity to the Purkinje cell layer. (D) Reporter-gene–positive cells displaying a microglia-like morphology. (E) Quantification of reporter-gene–expressing Purkinje neurons after intracerebellar injection of vesicle preparations from Vav-iCre–positive peripheral blood. Control mice (shaded part) were injected with 1 µl purified Cre-recombinase protein at 1 U/µl (light grey) or lysate prepared from Vav-iCre bone marrow (dark grey) and never showed any recombined cells. Scale bar, 50 µm (B and C) and 25 µm (D).

Figure 4. Transfer of Cre mRNA in blood chimeras.

(A) Schematic drawings of the experimental strategies to test for recombination events in blood chimeras. Lethally irradiated ROSA26-LacZ mice receive BM cells from Vav-iCre-ROSA-GFP mice. The bone marrow of recipient mice was tested for engraftment by flow cytometry analysis of GFP expression (representative analysis in right panel; wild-type bone marrow, red line; Vav-iCre-ROSA-GFP donor bone marrow, green line; bone marrow of ROSA26-LacZ recipient mouse after engraftment, blue line). For adoptive transfer experiments, the same combination of transgenes was used with spleen and lymph node cells as donor organs. Representative flow cytometry analysis of GFP expression of donor (green line) compared to wild-type cells (red line). At the time of analysis, GFP-positive cells were undetectable in recipient spleens (blue line). Two months after bone marrow transplantation, recombined cells can be detected in the liver (B), granular cell layer (C), and Purkinje cell layer (D), as well as associated with blood vessels (E). None of the recombined LacZ-/X-Gal–positive cells were positive for GFP, excluding cell fusion. Scale bar, 10 µm (B–E).

Figure 5. Peripheral inflammation increases the number of recombined Purkinje neurons.

The number of recombined Purkinje neurons is low in healthy animals (A) but increases dramatically after peripheral inflammatory conditions (B). Inflammatory injuries were induced by subcutaneous injection of LLC2s or peritonitis. Animals were analyzed 12 d after injection when tumors were formed. Peritonitis was induced by a single i.p. injection of thioglycolate broth (1 ml in 3% PBS). Mice with peritonitis and ECL were analyzed 4 d after injection. (C) Filled bars represent results from Vav-iCre and empty bars from Tie2-Cre reporter mice. The p values were calculated by two-tailed t test for groups with unequal variance. (D and E) We did not observe any recombined Purkinje neurons that were binucleated in either transgenic mouse line after induction of an inflammation. (F and G) Microglia (white arrows) were always negative for the marker gene in healthy animals as well as after an inflammation. (H and I) Transendothelial electrical resistance (TEER, top panel) decreases and the corresponding capacitance (Ccl, bottom panel) of the bEnd5 endothelial monolayers increases significantly 24 h and 48 h after addition of bone-marrow-derived EVs compared to conditioned medium supernatant after ultracentrifugation. Vertical line at 0 h indicates media exchange. Scale bar, 100 µm (A and B), 50 µm (D and F), 10 µm (E), and 5 µm (G).

Figure 6. Reporter-gene–positive neurons can be found in multiple areas of the brain.

(A and B) β-galactosidase–positive cells in the SN/VTA that can be TH-positive dopaminergic neurons as well as TH-negative cells with neuronal morphology (C). (D–F) β-galactosidase–positive cells in the cortex are also positive for NeuN. (G–I) Overview of the dentate gyrus in the hippocampus with recombined neurons in the granular cell layer that are positive for the neuronal marker NeuN. (J and K) Recombined neurons after ECL in hippocampal areas CA2 and CA3 and nonneuronal GFAP-negative recombined cells at the lesion site (L). Scale bar, 100 µm (D, G, H, and J), 50 µm (A, E, K, and L), and 10 µm (B, C, F, and I).

Figure 7. Recombined Purkinje neurons display a different miRNA profile compared to nonrecombined neighboring neurons.

(A) Example for an X-Gal–positive neuron and corresponding nonrecombined Purkinje neuron that were collected by LCM for miRNA analysis. Cells were identified based on morphology and location in the Purkinje cell layer. Nonrecombined Purkinje neurons in the direct vicinity were cut out as references, depicted below. (B) Pairwise comparison of nonrecombined versus recombined Purkinje neurons from the same animal. Each symbol represents an individual miRNA; colored symbols represent miRNAs that could only be detected in one pool. Symbols representing higher expressed miRNAs are at the top; those representing lower expressed miRNAs are at the bottom of each column. The comparison of recombined versus nonrecombined PKNs from the same animals reveals significant differences in their miRNA profile. (C) Venn diagram depicting miRNAs specific for recombined PKNs. The three miRNAs displayed in the box are those that can be found in both samples of the recombined PKNs but not in any sample of the nonrecombined PKN. (D) EVs isolated from the serum of mice with peritonitis display a different miRNA content than those of healthy control animals.

Figure 8. Model for EV transfer of RNA from immune cells to the brain.

Possible modes of RNA transfer from blood to brain: (A) EVs are released from blood cells into the blood stream, where they can cross the BBB and fuse with neurons. (B) Alternatively, leukocytes enter the brain and only EVs released in short distance to the target cell are able to bind and release their content. This direct signaling of immune cells to the brain is independent of microglia.