Cerebral infusion of AAV9 vector-encoding non-self proteins can elicit cell-mediated immune responses

Abstract

There is considerable interest in the use of adeno-associated virus serotype 9 (AAV9) for neurological gene therapy partly because of its ability to cross the blood-brain barrier to transduce astrocytes and neurons. This raises the possibility that AAV9 might also transduce antigen-presenting cells (APC) in the brain and provoke an adaptive immune response. We tested this hypothesis by infusing AAV9 vectors encoding foreign antigens, namely human aromatic L-amino acid decarboxylase (hAADC) and green fluorescent protein (GFP), into rat brain parenchyma. Over ensuing weeks, both vectors elicited a prominent inflammation in transduced brain regions associated with upregulation of MHC II in glia and associated lymphocytic infiltration. Transduction of either thalamus or striatum with AAV9-hAADC evinced a significant loss of neurons and induction of anti-hAADC antibodies. We conclude that AAV9 transduces APC in the brain and, depending on the immunogenicity of the transgene, can provoke a full immune response that mediates significant brain pathology. We emphasize, however, that these observations do not preclude the use of AAV serotypes that can transduce APC. However, it does potentially complicate preclinical toxicology studies in which non-self proteins are expressed at a level sufficient to trigger cell-mediated and humoral immune responses.

Figure 1

AAV9-hAADC transfer into striatum. AAV9-hAADC directed robust transduction of rat striatum followed by progressive loss of transduced neurons. Shown is anti-AADC staining of striatal sections (a) 1-, (c) 3-, (e) 5-, and (g) 8 weeks after AAV9-hAADC injection [7 × 10¹⁰ vector genomes (vg)], indicating the progressive decrease in aromatic L-amino acid decarboxylase (AADC) immunostaining in the injected hemisphere. The central panel shows individual hAADC⁺ cells magnified from panels a, c, e, and g with neuronal (short arrow) and glial (long arrow) morphology. The progressive loss of human AADC (hAADC)-expressing cells was associated with a dramatic diminution in the number of NeuN⁺ cells in the transduced striatum. The right-hand panels show NeuN⁺ cells in the area surrounding the injection site (b) 1-, (d) 3-, (f) 5-, and (h) 8 weeks after AAV9-hAADC. This neuronal loss was not observed after injection of an equal dose of AAV2-hAADC (i, j, k). The inserts show a higher magnification of the transduced area in the striatum 1 week, 3 weeks, and 5 weeks after AAV2-hAADC transduction. Short arrows indicate individual hAADC⁺ cells with a neuronal morphology. Bar = 1 mm for a, c, e, g, i, j, k; 100 µm for b, d, e, h and 50 µm for the images in the central panel and inserts i, j, k. AAV9, adeno-associated virus serotype 9.

Figure 2

Cellular tropism of AAV9-hAADC vector in striatum. Representative images show expression of human aromatic L-amino acid decarboxylase (hAADC) in (a) NeuN⁺ cells and in (b) GFAP⁺ cells 1 week after striatal AAV9-hAADC [7 × 10¹⁰ vector genomes (vg)] injection. Bar = 100 µm. AAV9, adeno-associated virus serotype 9; GFAP, glial fibrillary acidic protein.

Figure 3

Neurotoxicity of AAV9-hAADC after injection into thalamus. AADC staining 5 weeks after injection of AAV9-hAADC [7 × 10¹⁰ vector genomes (vg)] into thalamus shows a weakly stained area at the injection site almost completely devoid of (a) AADC immunoreactivity. Higher magnification (central image) reveals only a few hAADC⁺ cells with glial morphology (arrow). (b) NeuN⁺ cells in the transduced area in the thalamus. Striatal injection of AAV9-GFP (7 × 10¹⁰ vg) resulted in widespread expression of (c) green fluorescent protein (GFP) and modest loss of (d) NeuN⁺ cells in the ipsilateral striatum 3 weeks after transduction. The central panel shows individual GFP⁺ cells from panel c at high magnification with neuronal (short arrow) and glial (long arrow) morphology. The table shows the quantification of striatal NeuN⁺ and the cellular tropism of AAV9-GFP 3 weeks after striatal infusion (e). Data are represented as mean number of cells ± SD. ^aP < 0.01 or ^bP < 0.05 compared with noninjected control, Mann–Whitney U-test. Bar = 1 mm for a, c; 100 µm for b, d and 50 µm for central images. AAV9, adeno-associated virus serotype 9; hAADC, human aromatic L-amino acid decarboxylase.

Figure 4

Time-dependent glia reaction to striatal injection of AAV9-hAADC, AAV2-hAADC, and AAV9-GFP. Micrographs (a–e) are representative of 1- and 3-week survival for equal doses of AAV2-hAADC, AAV9-hAADC, and AAV9-GFP, and of an 8-week survival for a lower dose of AAV9-hAADC (f). Glial activation in transduced striatum is indicated by Iba1 (a–f) and glial fibrillary acidic protein (GFAP) staining (g–l). Immunostaining for Iba1 and GFAP shows a modest reaction of glial cells to AAV2-hAADC injection (a,b,g,h). In contrast, a massive glial activation to AAV9-hAADC (higher dose) was observed between 3 and 8 weeks after transduction (c,d,i,j,m). Injection of a lower dose of AAV9-hAADC also induced prominent but delayed glial activation 8 weeks after transduction (f,l,m). Significant microglial and astroglial cell was also detected 3 weeks after AAV9-GFP injection into striatum (e,k,m). Data are represented as the mean percentage of Iba1 or GFAP optical density (OD) in the striatal region of the contralateral side ± SD. The values obtained from the contralateral striatum are scaled to 100%. Statistical intervals are indicated by P values as follows: ^aP < 0.01 or ^bP < 0.05 compared with noninjected control, Mann–Whitney U-test. Bar = 1 mm. AAV9, adeno-associated virus serotype 9; GFP, green fluorescent protein; hAADC, human aromatic L-amino acid decarboxylase.

Figure 5

Glial cells adopt an antigen-presenting phenotype after adeno-associated virus serotype 9 (AAV9) injection. (a–c) Representative images depict Iba1⁺ microglia (a) that are MHC II⁺ (b). Iba1 and MHC II colocalize (c) and are mainly associated with blood vessels 7 days after AAV9-hAADC transduction (7 × 10¹⁰ vg). (d–f) Representative images show (d) GFAP⁺ astrocytes, immunopositive for (e) MHC II. (f) GFAP and MHC II colocalize and were detected in the transduced striatum 3 weeks after AAV9-hAADC vector injection. Representative images show amoeboid microglia that are MHC II⁺ (g–i) 3 weeks after AAV9-GFP transduction. The table shows the percentage of MHC II/Iba1 and MHC II⁺/GFAP⁺ cells in the striatum 1, 3, 5, and 8 weeks after AAV9-hAADC injection and 3 weeks after AAV9-GFP injection ± SD (j). ^aP < 0.05 compared with noninjected control, Mann–Whitney U-test. Bar = 50 µm. GFP, green fluorescent protein; GFAP, glial fibrillary acidic protein; hAADC, human aromatic L-amino acid decarboxylase.

Figure 6

Striatal influx of armed T-cells after AAV9-hAADC and AAV9-GFP transduction. (a) Hematoxylin–eosin (H&E) staining of brain 3 weeks after AAV9-hAADC [7 × 10¹⁰ vector genomes (vg)] shows a multiple scattered perivascular inflammatory foci (arrows) localized in the transduced striatum. (b) H&E staining of contralateral (noninjected) striatum shows the vessel lumen without any pathology. The higher magnification of the inset from a (a1). (c) Inflammatory foci after AAV9-hAADC infusion were also detected in other transduced areas, e.g., ipsilateral entopeduncular nucleus and thalamus. (d) Numerous CD8⁺ T cells in parenchyma were found in close contact with blood vessels. Quantification of the density of inflammatory foci and infiltrated CD8⁺ T cells into transduced striatum at various time points after AAV9-hAADC (low- and high-dose) and 3 weeks after AAV9-GFP injection (e). Data are the mean number of foci per section and mean number of CD8⁺ T cells per mm² ± SD Statistical significance is indicated as follows: ^aP < 0.01 or ^bP < 0.05 compared with noninjected control, Mann–Whitney U-test. Bar = 500 µm for a, c; 100 µm for a1, b, and 50 µm for d. AAV9, adeno-associated virus serotype 9; GFP, green fluorescent protein; hAADC, human aromatic L-amino acid decarboxylase.

Figure 7

Dot–blot assay for circulating anti-hAADC antibodies. (a) Positive control (dilutions of monoclonal antibodies against hAADC; www.millipore.com; Cat. AB136). (b) Negative control serum from naive rat. Sera samples obtained from rats 8 weeks after (c,d) AAV9-hAADC transduction revealed a much higher titer of anti-hAADC antibodies (two representative rats) than in rats that received AAV2-hAADC [two representative rats (e,f)]. AAV9, adeno-associated virus serotype 9; hAADC, human aromatic L-amino acid decarboxylase.