Visualizing HIF-1α mRNA in a Subpopulation of Bone Marrow-Derived Cells to Predict Retinal Neovascularization

Abstract

Bone marrow-derived progenitor cells and macrophages are known to migrate into the retina in response to inflammation and neovascularization. These migratory cells might play important regulatory roles in the pathogenesis of neovascularization, a common complication observed in diabetic retinopathy, retinopathy of prematurity, and retinal vein occlusion. Hypoxia-inducible factor 1α (HIF-1α) has been shown to contribute to the pathogenesis of retinal inflammation and neovascularization. However, contributions of monocyte-derived macrophages to neovascularization are largely unknown. We hypothesized that selective visualization of these microglia/macrophages could be a powerful method for predicting the onset of neovascularization and its progression at the molecular level. In this report, we describe the synthesis of a new hybrid nanoparticle to visualize HIF-1α mRNA selectively in microglia/macrophages in a mouse model of neovascularization. HIF-1α expression was confirmed in MRC-1 positive monocytes/macrophages as well as in CD4 positive T-cells and CD19 positive B-cells using single-cell RNA sequencing data analysis. The imaging probes (AS- or NS-shRNA-lipid) were synthesized by conjugating diacyl-lipids to short hairpin RNA with an antisense sequence complementary to HIF-1α mRNA and a fluorophore that is quenched by a black hole quencher. We believe that imaging mRNA selectively in tissue specific microglia/macrophages could be a powerful method for predicting the onset of neovascularization, its progression, and its response to therapy.

Figure 1:

Schematic drawing and hybridization motif of shRNA-lipid conjugates showing the design incorporating anti-sense sequence complementary to the HIF-1α mRNA. The shRNA-lipid conjugates are stabilized using 2’-OMe nucleotides. A fluorescence dye (FL Dye) and a black hole quencher (BHQ) are incorporated onto opposite ends of the shRNA construct as shown. Upon hybridization to the target sequence, the intramolecular quenching is compromised, and fluorescence occurs, thus reporting the hybridization event. Since the shRNA-lipid conjugates contain a hydrophilic and a lipophilic components, it is likely that they form lipid-micelle structures. From the dynamic light scattering (DLS) measurements and transmission electron microscopy (TEM) imaging, we observed that shRNA-lipids from spherical nanoparticles of around 10 nm. The mRNA recognition moiety resides on the surface of the nanoparticles and the lipid-core remains inside.

Figure 2:

In vitro localization of HIF-1α mRNA targeted AS-shRNA-lipid in mouse Müller cells (MMC) using fluorescence microscopy. Expression of HIF-1α mRNA was induced in MMCs by exposing the Cells under hypoxia. DAPI was used to counterstain the nucleolus (blue). (A-D) AS-shRNA-lipid derived fluorescence was minimally observed in normoxic MMCs and the fluorescence was significantly increased in hypoxic MMCs, supporting the expression of HIF-1α mRNA in hypoxic MMCs. (E) Comparative assessment of fluorescence intensities in normoxic and hypoxic MMCs, showing increased fluorescence in hypoxic cells. Fluorescence intensities were measured computationally using ImageJ software (n = 6). Statistical significance *** P<0.0001.

Figure 3:

Characterization of CD4 positive T cells, CD19 positive B cells, MRC-1 positive macrophages and expression of HIF-1α using single cell RNAseq data analysis in isolated white blood cells from P17 mouse OIR and RA control mice. (A-C) Representative aggregated t-SNE plot showing higher numbers of CD4 positive T cells in room air (RA) control group (A); higher numbers of CD19 positive B cells in oxygen-induced retinopathy (OIR) group (B); macrophage mannose receptor C-1 (MRC1) positive cells mostly in OIR group (C). (D) HIF-1α expression was observed in T cells, B cells and MRC1 positive macrophages. t-SNE plot was created using Loupe Cell Browser program.

Figure 4:

Imaging HIF-1α mRNA selectively in neovascularization in mouse P17 OIR retina. Isolectin B4 was used to visualize the retinal vasculatures. The OIR mice received intraperitonel injections of AS-shRNA-lipid conjugates. Eighteen hours post-injection, retinal tissues were analyzed ex vivo. (A-B) The shRNA-lipid conjugates incorporating anti-sense sequence complementary to HIF-1α mRNA was localized in cells that are also associated with neovascular tufts (arrows). (C-D) Higher magnification of A-B showing strong fluorescence punctate patterns presumably due to hybridization of AS-shRNA-lipid with HIF-1α mRNA in cells in the OIR retina. (E-F) AS-shRNA-lipid derived fluorescence was also observed in cells that are localized around the neovascular tufts (arrows) in the P17 OIR retinas. Scale bar 30 µm in A-B and 20 µm in C-F.

Figure 5:

Co-localization of AS-shRNA-lipid derived fluorescence with IBA1 positive cells in the P17 mouse OIR retina. (A-C) Isolectin B4 was used to visualize the retinal vasculatures (green); IBA1 was used to stain the microglial cells (magenta). (D-E) AS-shRNA-lipid derived fluorescence was localized in cells that are also associated with IBA1 positive cells (white arrows), suggesting that IBA1 positive microglia/macrophages (magenta color) could express HIF-1α mRNA that are also associated with neovascularization. (F-G) AS-shRNA-lipid derived fluorescence were not observed in ramified IBA1-positive cells (yellow arrows) in the OIR retina.

Figure 6:

The AS-shRNA-lipid conjugates were minimally detectable in IBA1 positive ramified/resting microglia (magenta color) in room air (RA) raised healthy control retinas. Isolectin B4 was used to visualize the retinal vasculatures. Microglia were observed on the surface of superficial capillary plexus (SCP), middle capillary plexus (MCP) and deep capillary plexus (DCP) and are ramified. AS-shRNA-lipid conjugates were not detectable in SCP (A-D); and were not detectable in MCP or DCP (E-L). Scale bar 50 µm.

Figure 7:

Depletion of macrophages using clodronate-liposome reduces the AS-shRNA-Lipid derived fluorescence in P17 OIR mouse retinas compared to control PBS-injected OIR retinas. AS-shRNA-lipid derived fluorescence was monitored in mice with intraperitoneal injection of AS-shRNA-lipid. IBA1 was used to visualize the microglia/macrophages in the retina and IB4 was used to visualize the vasculatures. (A) A large number of IBA1 positive cells were observed in the control PBS-injected OIR retinas. (B) Although a large number of IBA1 positive cells were observed in the clodronate-liposome injected retina, AS-shRNA-lipid derived fluorescence significantly decreased in the clodronate-liposome injected retinas as shown in C.

Figure 8:

Effect of inhibitors/stimulator on shRNA-lipid internalization into the primary retinal microvascular endothelial cells (MRMEC). shRNA-lipid conjugates (both AS- and NS-) were synthesized using only the fluorescence dye (Cy5) at the 5’ end without using the black hole quencher (BHQ). Thus, the probes are not quenched and are always ON. Cells were exposed to different inhibitors or stimulators for 30 min prior to addition of the shRNA-lipid conjugates. Plate based fluorescence assays and fluorescence imaging were performed two hour post-incubation to monitor probe internalization. (A) shRNA-lipid internalization is inhibited in presence of clathrin-mediated endocytosis inhibitor (sucrose), suggesting the possibility of clathrin-mediated endocytosis of shRNA-lipid, * P<0.05 (n=4). (B) Inhibition of endocytosis (missing punctate structures) could be visualized using confocal imaging (yellow arrows). Scale bar 25 μm. (C) In addition, shRNA-lipid internalization is independent of the macropinocytosis pathways as observed from phorbol esters treatment; and also internalization is independent of filipin-sensitive caveolae-mediated transport.