Volume 6, Issue 3, Pages 394-399 (September 2002) Gene Therapy for Pyruvate Dehydrogenase E1α Deficiency Using Recombinant Adeno- Associated Virus 2 (rAAV2) Vectors  Renius Owen, Ronald J. Mandel, Chandramohan V. Ammini, Thomas J. Conlon, Douglas S. Kerr, Peter W. Stacpoole, Terence R. Flotte  Molecular Therapy  Volume 6, Issue 3, Pages 394-399 (September 2002) DOI: 10.1006/mthe.2002.0683 Copyright © 2002 American Society for Gene Therapy Terms and Conditions

FIG. 1 Design of rAAV-PDH vectors. The salient features of the rAAV-PDH vectors used in this paper are depicted. ITR, AAV2 inverted terminal repeat. PDH, PDH E1α cDNA coding sequence. An, bovine growth hormone poly(A) signal. GFP, enhanced humanized green fluorescent protein cDNA. Molecular Therapy 2002 6, 394-399DOI: (10.1006/mthe.2002.0683) Copyright © 2002 American Society for Gene Therapy Terms and Conditions

FIG. 2 Efficiency of gene transfer into human fibroblasts with rAAV vectors. Flow cytometry, phase contrast microscopy, and confocal microscopy were performed on human fibroblast cultures 48 hours after (A) infection with Ad5 only (MOI = 5 i.u./cell; phase contrast image is shown); (B) infection with rAAV-GFP (MOI =10 i.u./cell; confocal fluorescence image is shown); or (C) infection with rAAV-GFP (MOI =100 i.u./ cell; confocal fluorescence image is shown). For all photomicrographs initial magnification is ×100. Flow cytometry plots are shown as insets, with the GFP-expressing (more highly fluorescent) cells indicated by the R2+R1 window. Corresponding numerical counts are shown beneath each panel with the “% Gated” in each R2+R1 row indicating the proportion of intact cells (gated for size by light scatter) that are also positive for GFP expression. Molecular Therapy 2002 6, 394-399DOI: (10.1006/mthe.2002.0683) Copyright © 2002 American Society for Gene Therapy Terms and Conditions

FIG. 3 Expression of PDHA1 mRNA in human skin fibroblasts transduced with rAAV-PDH at an MOI of 50 i.u./cell. RT-PCR with primers specific for PDHA1 (top half of figure) and human β-actin (bottom half) was performed on total mRNA extracted from skin fibroblasts. Patient, untransduced patient cells. Vector, rAAV-PDH-transduced cells. Normal, skin fibroblast culture from a normal volunteer. PDH, band of the size predicted for PDHA1 primer pair. β-Actin, band of the size predicted for human β-actin control RT-PCR, indicative of the quality of each mRNA sample. Molecular Therapy 2002 6, 394-399DOI: (10.1006/mthe.2002.0683) Copyright © 2002 American Society for Gene Therapy Terms and Conditions

FIG. 4 Partial functional correction of PDH enzyme activity in rAAV-PDHA1 transduced human fibroblasts. Fibroblast cultures were infected with rAAV-PDH vector at an MOI of 50 i.u./cell (under conditions in which approximately 15% of cells would be transduced). After 48 hours, cell extracts were assayed for functional PDH enzyme activity by measuring 14CO2 generation from 14C-pyruvate. Activity was normalized to a purified enzyme extract, and expressed as activity on the y axis. Patient, uncorrected PDH E1α-deficient fibroblasts. Vector, rAAV-PDH-transduced cells. Normal, normal control fibroblasts. HEK, human embryonic kidney 293 cell line control. All differences are significant at P < 0.05. Molecular Therapy 2002 6, 394-399DOI: (10.1006/mthe.2002.0683) Copyright © 2002 American Society for Gene Therapy Terms and Conditions

FIG. 5 In vivo expression of rAAV-PDH-GFP in neurons of the basal ganglia of rat brain. GFP expression in a rat that received rAAV-PDH-GFP in two sites in the left hemisphere and rAAV-GFP in identical sites in the right hemisphere 1 year earlier are shown. (A) Low power photomicrograph showing the region of the posterior transduction site as detected by GFP immunostaining. Even at this low magnification there is a clear qualitative difference between the appearance of the expression pattern obtained from the two different vectors (bar, 2 mm). (B) Higher magnification of the rAAV-PDH-GFP transduction as shown in the left hemisphere in (A). Although, at this magnification, the cellular GFP staining appears lighter than the rAAV-GFP counterpart in (C), there are more positive cells in this hemisphere. (C) Higher magnification of the rAAV-GFP transduction as shown in the right hemisphere in (A). The neuronal staining in this panel appears darker than in (B) due to the homogenous filling of the soma as compared with the more diffuse somal staining seen in (B) (bar, 200 μm and applies to (B)). (D) High power photomicrograph of the neurons shown in (B). The punctate intracellular GFP expression consistent with mitochondrial localization is apparent at this magnification. (E) High power photomicrograph of the neurons shown in (C). The soma and neuropil are filled evenly with GFP in this panel (bar, 50 μm and applies to (D)). (F) Confocal photomicrograph of native fluorescence from a serial section of the same brain shown in (A, B, D). The white arrow points to a neuron containing GFP expressed from rAAV-PDH-GFP. This image, from the hemisphere that received rAAV-PDH-GFP, clearly shows the intracellular segregation of the GFP protein consistent with mitochondria. (G) Confocal image of the rAAV-GFP injected hemisphere of the same section as in (F), demonstrating the virtual complete filling of the neuron by the GFP protein in contrast to the pattern shown in (F) (bar, 4 μm and applies to (F)). Molecular Therapy 2002 6, 394-399DOI: (10.1006/mthe.2002.0683) Copyright © 2002 American Society for Gene Therapy Terms and Conditions