Immunological thresholds in neurological gene therapy: highly efficient elimination of transduced cells might be related to the specific formation of immunological synapses between T cells and virus-infected brain cells

Abstract

First-generation adenovirus can be engineered with powerful promoters to drive expression of therapeutic transgenes. Numerous clinical trials for glioblastoma multiforme using first generation adenoviral vectors have either been performed or are ongoing, including an ongoing, Phase III, multicenter trial in Europe and Israel (Ark Therapeutics, Inc.). Although in the absence of anti-adenovirus immune responses expression in the brain lasts 6-18 months, systemic infection with adenovirus induces immune responses that inhibit dramatically therapeutic transgene expression from first generation adenoviral vectors, thus, potentially compromising therapeutic efficacy. Here, we show evidence of an immunization threshold for the dose that generates an immune response strong enough to eliminate transgene expression from the CNS. For the systemic immunization to eliminate transgene expression from the brain, > or = 1 x 10(7) infectious units (iu) of adenovirus need to be used as immunogen. Furthermore, this immune response eliminates >90% of transgene expression from 1 x 10(7)-1 x 10(3) iu of vector injected into the striatum 60 days earlier. Importantly, elimination of transgene expression is independent of the nature of the promoter that drives transgene expression and is accompanied by brain infiltration of CD8(+) T cells and macrophages. In conclusion, once the threshold for systemic immunization (i.e. 1 x 10(7) iu) is crossed, the immune response eliminates transgene expression by >90% even from brains that receive as little as 1000 iu of adenoviral vectors, independently of the type of promoter that drives expression.

Keywords: Gene therapy; adenovirus; neuroimmunology; promoter; threshold.

Fig. 1.. (A) RAd36 induces expression of β-galactosidase in astrocytes

Rat brain sections injected with RAd36 were stained for β-galactosidase (green) and GFAP, a marker of astrocytes (magenta) and analyzed with confocal microscope. Confocal analysis demonstrated that β-galactosidase is expressed in astrocytes. Note the colocalization of β-galactosidase and GFAP in the merged image. Scale bar, 30 µm. (B) Determination of the dose of systemic immunization needed to silence transgene expression in the CNS. Panels a–h show β-gal immunohistochemistry in rats injected with 10⁷ iu Ad-mCMV-βgal in the striatum and immunized 30 days later with 10¹–10⁸ iu Ad-hCMV-HPRT. Only doses 10⁵ to 10⁷ iu are illustrated. Peripheral immunization with ≤10⁵ iu did not reduce brain transgene expression (a,b,i) or increase serum neutralizing antibodies (j). Immunization with 10⁶ iu induced loss of transgene expression in some animals and one animal from each group is illustrated (c–f,i,j). Following peripheral immunization of ≥10⁷ iu transgene expression was lost in all animals (g,h,i), with all animals showing significant titers of neutralizing antibodies (j). Each dot in i and j represents one animal.

Fig. 2.. β-gal immunohistochemistry of rat brain injected with increasing doses of Ad-mCMV-βgal in the striatum

Animals were injected with 10¹–10⁷ iu Ad-mCMV-βgal in the striatum and 30 days later systemically with either saline (left) or 10⁸ iu Ad-hCMV-HPRT (right). Peripheral immunization with 10⁸ iu Ad-hCMV-HPRT reduced brain transgene expression independently of the dose of adenovirus injected into the brain.

Fig. 3.. Quantification of β-gal-positive cells after injection with Ad-mCMV-βgal in the striatum and Ad-hCMV-HPRT systemically

Rats were injected with increasing doses of Ad-mCMV-βgal in the striatum (10¹–10⁷ iu) and injected systemically 30 days later with either saline or 10⁸ iu Ad-hCMV-HPRT. Expression of β-gal was always reduced following immunization with 10⁸ iu Ad-hCMV-HPRT despite the different doses injected intracranially.

Fig. 4.. Immunohistochemical detection of β-gal, CD8 and ED1

β-gal, CD8 and ED1 in brain sections from rats injected in the striatum with different virus (Ad-hCMV-βgal, Ad-mCMV-βgal, RAd122, RAdsyn and RAdβact) containing different promoters (hCMV, mCMV, RSV, syn and βact, respectively) to drive expression of β-gal. Thirty days after injection animals were injected intradermally with either Ad-hCMV-HPRT (immunized animals) or with saline (control group). Sixty days after intradermal injection, animals were sacrificed and perfused. The expression of β-gal, CD8 and ED1 is shown in immunized (Ad-hCMV-HPRT) and in control animals (saline). (A,B) 1st, 3rd and 5th columns show low magnification of β-gal, CD8 and ED1, respectively, and the boxed areas are shown at higher magnification (right columns). (C) The area occupied by β-gal is reduced significantly in immunized animals (Ad-hCMV-HPRT) compared to controls (C1). Stereological estimation of the number of CD8 cells increased significantly in animals immunized with Ad-hCMV-HPRT compared to controls (C2). The area occupied by ED1 also increased significantly in immunized animals compared to controls (C3). *P<0.05, **P<0.01 Student’s t-test.

Fig. 4.. Immunohistochemical detection of β-gal, CD8 and ED1

β-gal, CD8 and ED1 in brain sections from rats injected in the striatum with different virus (Ad-hCMV-βgal, Ad-mCMV-βgal, RAd122, RAdsyn and RAdβact) containing different promoters (hCMV, mCMV, RSV, syn and βact, respectively) to drive expression of β-gal. Thirty days after injection animals were injected intradermally with either Ad-hCMV-HPRT (immunized animals) or with saline (control group). Sixty days after intradermal injection, animals were sacrificed and perfused. The expression of β-gal, CD8 and ED1 is shown in immunized (Ad-hCMV-HPRT) and in control animals (saline). (A,B) 1st, 3rd and 5th columns show low magnification of β-gal, CD8 and ED1, respectively, and the boxed areas are shown at higher magnification (right columns). (C) The area occupied by β-gal is reduced significantly in immunized animals (Ad-hCMV-HPRT) compared to controls (C1). Stereological estimation of the number of CD8 cells increased significantly in animals immunized with Ad-hCMV-HPRT compared to controls (C2). The area occupied by ED1 also increased significantly in immunized animals compared to controls (C3). *P<0.05, **P<0.01 Student’s t-test.

Fig. 5.. Activated T cells that express tyrosine kinases form immunological synapses with virally infected cells

Confocal images show three immunological synapses formed between T cells and virally infected astrocytes in the striatum of rats injected with Ad-hCMV-TK and immunized systemically 30 days later with Ad-hCMV-HPRT. Activated T cells show expression, phosphorylation and polarization of tyrosine kinases when in close apposition with virally infected astrocytes, indicating the activation and signaling of TCR in response to antigen encounter. Top row: a T cell (red, stained for TCR) that expresses phosphorylated ZAP-70 (green) forms an immunological synapse with a virally infected cell (white, stained for TK). Middle row: a T cell (red, stained for TCR) expressing phosphorylated Lck (green) polarized to the area of contact with the target cell (white, stained for TK). Bottom row: a CD8⁺ T cell that expresses LAT (green) forms an immunological synapse with a virally infected cell (white, stained for TK). White arrows indicate the areas of close apposition between T cells and infected brain cells. Scale bar, 20 µm.