Development of an intraventricular adeno-associated virus-based labeling strategy for glioblastoma cells nested in the subventricular zone

Abstract

Background: Glioblastoma (GBM) is a dreadful brain tumor, with a particular relationship to the adult subventricular zone (SVZ) that has been described as relevant to disease initiation, progression, and recurrence.

Methods: We propose a novel strategy for the detection and tracking of xenografted GBM cells that are located in the SVZ, based on an intracerebroventricular (icv) recombinant adeno-associated virus (AAV)-mediated color conversion method. We used different patient-derived GBM stem-like cells (GSCs), which we transduced first with a retroviral vector (LRLG) that included a lox-dsRed-STOP-lox cassette, upstream of the eGFP gene, then with rAAVs expressing the Cre-recombinase. Red and green fluorescence is analyzed in vitro and in vivo using flow cytometry and fluorescence microscopy.

Results: After comparing the efficiency of diverse rAAV serotypes, we confirmed that the in vitro transduction of GSC-LRLG with rAAV-Cre induced a switch from red to green fluorescence. In parallel, we verified that rAAV transduction was confined to the walls of the lateral ventricles. We, therefore, applied this conversion approach in 2 patient-derived orthotopic GSC xenograft models and showed that the icv injection of an rAAV-DJ-Cre after GSC-LRLG tumor implantation triggered the conversion of red GSCs to green, in the periventricular region. Green GSCs were also found at distant places, including the migratory tract and the tumor core.

Conclusions: This study not only sheds light on the putative outcome of SVZ-nested GBM cells but also shows that icv injection of rAAV vectors allows to transduce and potentially modulate gene expression in hard-to-reach GBM cells of the periventricular area.

Keywords: Glioblastoma; SVZ-nested GBM cells; adeno-associated virus; intracerebroventricular; subventricular zone.

Figure 1.

rAAV-1, rAAV-5, and rAAV-DJ efficiently transduce patient-derived GBM stem-like cell (GSC) cultures. (A) Flowchart representing the rAAV-red fluorescent protein (RFP) transduction of GSCs for the screening of most efficient serotypes (created with BioRender). (B) % GB1 cells expressing RFP in 2D-culture after a 3-day exposition to rAAVs (n = 3 independent experiments; one-way ANOVA, P-values for each comparison in Supplementary Figure S1, A). (C) Immunofluorescent staining of the RFP protein after rAAV-DJ-RFP transduction of GB1 cells (Scale bar = 50 μm). (D) % GB1 cells expressing RFP in 3D-culture after a 3-day exposition to rAAVs (n = 3 independent experiments; one-way ANOVA, P-values for each comparison in Supplementary Figure S1, B).

Figure 2.

rAAV-1, rAAV-5, and rAAV-DJ efficiently convert GSC-LRLG cells in culture from dsRed⁺ to eGFP⁺. (A) Flowchart representing dsRed gene excision and fluorescent eGFP gene expression in GSCs induced by rAAV-mediated Cre recombinase (created with BioRender). (B) Representative images of the dsRed and eGFP signals in T033-LRLG cultures transduced with rAAV-DJ-Cre, over 14 days. (C-D-E) % T033-LRLG cells that are found dsRed⁺, dsRed⁺/eGFP⁺, or eGFP⁺ only at 3, 7, and 14 days post-transduction with rAAV-1-Cre (C), rAAV-5-Cre (D) and rAAV-DJ-Cre (E). (NT = non transduced, at day 0). (F) % eGFP⁺ T033-LRLG cells over time, with the 3 rAAV serotypes (n = 3 independent experiments). (G) Results from the 2-way ANOVA, provided the statistical effect of the serotype and the effect of time. (H-I-J) % T049-LRLG cells that are found dsRed⁺, dsRed⁺/eGFP⁺, or eGFP⁺ only at 3, 7, and 14 days post-transduction with rAAV-1-Cre (H), rAAV-5-Cre (I) and rAAV-DJ-Cre (J) (NT = non transduced, at day 0). (K) % eGFP⁺ T049-LRLG cells over time, with the 3 rAAV serotypes (n = 3 independent experiments). (L) Results from the 2way ANOVA, provided the statistical effect of the serotype and the effect of time.

Figure 3.

The rAAV-DJ-Cre-eYFP has lateral ventricles-restricted spreading 1 week after icv injection and disseminates to adjacent neuron-rich structures after 4 weeks. (A) rAAV-1-eYFP, rAAV-5-eYFP, and rAAV-DJ-eYFP remain confined to the lateral ventricle walls one week after icv injection. Images are representative of n = 3 mice per serotype tested (scale bars = 100 μm). (B) Sagittal views at the VL medial wall (400 μm laterally to Bregma), through the VL (1mm laterally to Bregma), and at the VL lateral wall (1500 μm laterally to Bregma) to visualize YFP⁺ cells (scale bars = 400 μm). (C) Comparison of the volume (in μm³) of YFP fluorescence in transduced areas in the lateral ventricles neighboring (n = 5 mice; one-way ANOVA, P-values for each comparison in Supplementary Figure S1, A). (D) Longest distance (μm) between the further YFP⁺ cells in the mCP and dlCP and the lateral wall of the VL (n = 5). LSc, lateral septal nucleus; mCP, medial caudoputamen; dlCP, dorsolateral caudoputamen; Fi, fimbria; CC, corpus callosum; HPF, hippocampus.

Figure 4.

rAAV-DJ-Cre icv injection induces red-to-green conversion of GB1-LRLG that reached the subventricular zone (SVZ) in an orthotopic xenograft model. (A) Flowchart representing the experimental workflow. GB1-LRLG cells were implanted in nude mice and rAAV-DJ-Cre-YFP was icv injected at week 6 (group I) or week 12 (group II). Then mice were sacrificed at week 10 or week 16, respectively. (B) Analysis of right hemispheres with a lightsheet microscope, highlighting dsRed and eGFP-positive tumor areas. (C) Quantification of the dsRed-positive tumor area in mice from groups I and II (Mann–Whitney U-test, n = 8 mice per group). (D) Quantification of the eGFP-positive tumor area in mice from groups I and II (n = 8 mice per group). (E) Correspondence between dsRed + and eGFP + volumes in group II animals (linear regression).

Figure 5.

rAAV-DJ-Cre icv injection induces red-to-green conversion of T033-LRLG that reached the subventricular zone (SVZ) in an orthotopic xenograft model. (A) Flowchart representing the experimental workflow. T033-LRLG cells were implanted in nude mice and rAAV-DJ-Cre-YFP was icv injected intraventricularly at week 5, and mice were sacrificed at week 6. (B-C) Immunofluorescence stainings of eGFP⁺ and dsRed⁺ cells in different regions of the mouse brain (right striatum, right lateral ventricle, left lateral ventricle). Sections of 3 different animals are displayed (scale bar = 50 or 100 μm).

Figure 6.

Characterization of eGFP + T033-LRLG cells. Immunofluorescence costainings of SOX2 and Ki67 with eGFP⁺ and dsRED⁺ cells in different regions of the mouse brain implanted with T033-LRLG and one week after rAAV-DJ-Cre icv injection. Sections of 3 different animals are displayed (scale bar = 100 μm).