<<

Gene Therapy (2005) 12, 1503–1508 & 2005 Nature Publishing Group All rights reserved 0969-7128/05 $30.00 www.nature.com/gt RESEARCH ARTICLE Adeno-associated virus type 4 (AAV4) targets ependyma and in the and RMS

G Liu1,2, IH Martins1,2, JA Chiorini1,4 and BL Davidson1,2,3,5 1Program in Gene Therapy, University of Iowa, Iowa City, IA, USA; 2Department of Internal Medicine, University of Iowa, Iowa City, IA, USA; 3Department of Neurology, University of Iowa, Iowa City, IA, USA; 4National Institute of Dental and Craniofacial Research, National Institutes of Health, Bethesda, MD, USA; and 5Department of Physiology & Biophysics, University of Iowa, Iowa City, IA, USA

The subventricular zone (SVZ) is one of the neurogenic conducted on adult mice. AAV4 vectors show a characteristic niches in the adult mammalian . The SVZ is of interest transduction of the ependyma independent of delivery route. for studies on and stem therapy. Here, we However, AAV4 virus injected into the SVZ targeted GFAP report specific transduction of ependyma and/or astrocytes positive astrocytes forming the glial tube in the SVZ and by recombinant adeno-associated virus type 4 (AAV4) viral (RMS). Our results introduce AAV4 vectors. AAV4 vectors encoding b-galactosidase or eGFP as a new tool by which to manipulate glial cells in the RMS. were injected into the of neonatal and adult Gene Therapy (2005) 12, 1503–1508. doi:10.1038/ C57BL/6 mouse . In addition, SVZ injections were sj.gt.3302554; published online 2 June 2005

Keywords: adeno-associated virus type 4 (AAV4); neurogenesis; ependyma; subventricular zone; astrocytes; radial ; rostral migratory stream

Introduction respectively.6 NPCs have also been successfully labeled with recombinant retroviruses.7,8 The wall of the lateral ventricle is a major site of Adeno-associated virus (AAV) vectors mediate long- neurogenesis in the adult mammalian brain (for reviews term stable expression and induce minimal immune see Gage1, Taupin and Gage2, Aivarez-Buylla3). This responses, which make them attractive vehicles for region contains a distinct organization of different cells in vivo gene transfer to NPCs. While prior work with types. The ependyma lines the ventricles and consists of a AAV type 2 (AAV2) vectors showed poor NPC transduc- single layer of multiciliated cells. The subventricular zone tion following direct SVZ injection, AAV2 did success- (SVZ) is adjacent to the ependyma and contains three fully transduce human NPCs in vitro.9,10 We previously major cell types known as Types A, B and C cells.4 Type B showed that AAVs possessing different capsid structures cells reside in the SVZ and the rostral migratory stream have altered tropism when introduced into the mamma- (RMS) and are glial fibrillary acidic protein (GFAP) lian brain.11 For example, AAV4 transduced ependyma positive. Type C cells are a transient amplifying popula- with high efficiency when injected into the striata or tion derived from type B cells. Type C cells give rise to lateral ventricles of adult mice.11 In the present study, we type A cells, which are ensheathed by an compared the transduction profile of AAV4 introduced astroglial tube. On reaching the subependymal zone of the to the SVZ relative to direct ventricular injection. Our olfactory bulb (OB), Type A cells differentiate into data show that AAV4 transduces Type B astrocytes in the in the granule and periglomerular cell layers. SVZ, and glia overlying the RMS neural tube. Both viral and nonviral strategies have been devel- oped to target distinct cell populations in the lateral ventricular wall, with particular focus on resident neural progenitor cells (NPCs). Falk et al5 found that adenovirus Results and VSV-G pseudotyped retroviruses delivered foreign Newborn mice were injected with AAV4bgal at post- genes into NPCs more efficiently in vitro than nonviral natal day 0 or 1 and killed 4 weeks postinjection. b- methods. When administered into the ventricles or the Galactosidase expression in the ependyma lining the SVZ, adenovirus infected ependymal cells and NPCs lateral ventricles of both hemispheres and the indicated successful intraventricular injections (Figure 1a and b). Double IHC for GFAP and anti-b- Correspondence: Professor BL Davidson, 200 EMRB, University of Iowa, Iowa City, IA 52242, USA galactosidase was performed to investigate if type B cells Received 6 September 2004; accepted 2 April 2005; published online were transduced by AAV4, since SVZ-resident type B 2 June 2005 astrocytes have access to the lateral ventricle via cilia.4 AAV4 targets ependyma and astrocytes in the SVZ and RMS G Liu et al 1504

Figure 1 b-Galactosidase expression following intraventricular injection of AAV4bgal into newborn mice. (a) Ependymal b-Galactosidase enzyme activity 4 weeks postinjection (arrowheads, ependymal cell layer). (b) b-Galactosidase expression revealed by immunohistochemistry (arrowheads, ependymal cell layer; arrow, ). (c) Dual staining with anti-GFAP (red) and anti-b-galactosidase antibodies did not reveal colocalization of astrocytes and transduced cells. (d) In the olfactory bulb (OB), cells with b-galactosidase activity (arrow; also inset) are localized in the SEL. (e) b-Galactosidase-positive cells (red) in the middle of OB do not express the neuronal cell marker NeuN (green). (f) Triple staining with anti-b-galactosidase (green), anti-NeuN (blue) and anti-GFAP (red) demonstrates that a fraction of b-galactosidase-positive cells in the layer expresses NeuN (arrow) but not GFAP. A cell that does not colocalize with either NeuN or GFAP is also shown (arrowhead). LV, lateral ventricle.

Colocalization of GFAP and b-galactosidase expression examined. b-Galactosidase+ cells were GFAPÀ (Figure was not evident (Figure 1c). Also, the RMS was negative 2b) and NeuN+ (Figure 2c and d). In contrast to the for b-galactosidase activity and anti-b-galactosidase neonatal injection group, no b-galactosidase expressing immunoreactivity (data not shown). However, sparse cells were detected in the SEL of the OB. Thus, AAV4 (o0.1%) positive cells were found in the olfactory bulb transduces ependyma in the lateral ventricular wall (OB; Figure 1d–f). The majority of b-galactosidase+ cells when accessed from the luminal side, with modest to no were restricted to the center of the OB (Figure 1d, inset). transduction of NPCs when introduced into neonates This region corresponds to the subependymal layer or adult mice, respectively. (SEL). Double labeling with NeuN or GFAP and b- To test if AAV4 could transduce other cell types galactosidase antibodies revealed no NeuN+/b-galacto- we introduced virus directly into the SVZ. AAV4 virus sidase+ or GFAP+/b-galactosidase+ cells in the SEL encoding eGFP (AAV4eGFP) was injected into the (Figure 1e and data not shown). Therefore, the majority anterior part of the subventricular zone (SVZa). Previous of b-galactosidase-positive cells in the SEL of the OB studies from our lab showed that striatal injection of were not or glia. Olfactory ventricles are open AAV4 transduced the ependyma with high specificity.11 and connected to the lateral ventricle at early postnatal Surprisingly, 4 weeks after AAV4eGFP injection into the stages.12 Thus, most b-galactosidase+ cells likely resulted SVZ, robust eGFP expression was found in the ependy- from direct ependymal cell transduction following ma, the SVZ and the RMS (Figure 3). The transduced delivery to the postnatal brain. Previous reports on cells in the SVZ were GFAP+/eGFP+ (Figure 3a–c). It is residual ependyma in OB are consistent with this noteworthy that the colocalization of eGFP and GFAP observation.13 was confined to the SVZ and the RMS; transduced Finally, intraventricular injection of AAV4bgal into astrocytes were not found along the injection track or P0/P1 mice resulted in some NeuN+/b-galactosidase+ elsewhere in the brain. Interestingly, none of the cells in the OB in the granule cell layer (Figure 1f, arrow). transduced cells were Tuj-1+, an immature neuronal These cells derive from NPCs in the lateral ventricular marker expressed by migrating neuroblasts (Figure 3d– wall.14 Thus, the presence of b-galactosidase+ granule cell f). However, transduced cells were in close contact with neurons indicates transduction, albeit inefficient, of Tuj-1+ positive neuroblasts, and the arrangement of NPCs by AAV4. GFAP+/eGFP+ cells resembled the glial tube formed by We next injected AAV4bgal into the lateral ventricle of type B astrocytes.4 No eGFP+/NeuN+ cells were found in young adult mice and analyzed brain tissue 4 weeks the SVZ or RMS (Figure 3g–i). Also, we were unable to later. The transduction pattern of AAV4 into adult mouse identify eGFP+ cells in the OB (data not shown). Our ventricles mirrored our studies done in neonatal mice results suggest that the subtype of astrocytes transduced in most respects. The ependyma lining the ventricular by AAV4 differentiate into glial tube cells, or are resident system was b-galactosidase positive (Figure 2a), and glial tube cells at the time of injection. Our results also underlying astrocytes were negative (data not shown). show that AAV4 does not transduce cells with properties Similarly, the RMS was devoid of reporter gene activity. consistent of NPCs. These data also support previous We detected few reporter gene-positive cells in the work by others suggesting that astrocytes in the SVZ and granule cell layer (Figure 2b–d) in three of six mice RMS are distinct from astroglia in other brain areas.4,12,13

Gene Therapy AAV4 targets ependyma and astrocytes in the SVZ and RMS G Liu et al 1505

Figure 2 b-Galactosidase expression after AAV4bgal injection into the ventricles of adult mice. (a) b-Galactosidase enzyme activity was limited to the ependymal cell layer 4 weeks postinjection. Arrow, choroid plexus; arrowhead, ependyma. (b–d) Evidence for b-galactosidase+ in the granule cell layer of the OB. Dual immunohistochemical staining showed that b-galactosidase+ cells express NeuN (c and d, arrows) but not GFAP (b, arrow).

Discussion consistent with NPCs.3,18,19 Thus, it is possible that the rare b-galactosidase+/NeuN+ cells in the granule cell It has been established that massive migration of NPCs layer derived from similarly rare transduced radial glia. from the lateral ventricular wall to the OB occurs in A second possibility is that the b-galactosidase+ cells in neonatal and adult rodents.14,15 Although the exact the granule cell layer seen 4 weeks after AAV4 delivery number of cells added to the OB postnatally at a given to neonatal ventricles arose from transduced immature time has not been determined, it is reported that over ependyma in the VZ. This possibility requires that 10 000 labeled cells were observed 15 days after [3H]- ependyma have NPC potential,20 a controversial suppo- thymidine injection in adult mice.15 The number is sition that has little supportive data. estimated to be even higher at early postnatal stages We found only rare, isolated reporter gene-positive cells during peak migration. However, as illustrated in Figure in the OB granule cell layer following AAV4 delivery to 2, only sparse transduced cells were found in the OB adult mice ventricles. Those cells expressed the neuronal following introduction of AAV4 reporter vectors into the cell marker NeuN, indicating transduction of NPCs by lateral ventricles of neonatal mice. Moreover, the majority AAV4. It has been suggested by Johansson et al20 that of OB-resident transduced cells were in the SEL and were ependymal cells function as multipotent neural stem cells regionally and antigenically consistent with ependyma. in adult mice. Most studies, however, suggest that Type B The cellular composition and gross structure of the astrocytes in the SVZ are the bona fide adult neural stem lateral ventricular wall in early postnatal stages and cell.21–24 Type B cells functioning as neural stem cells adulthood differ. At P0, the lateral ventricular wall is contact the ventricular lumen with a single . In composed of two compartments, the ventricular zone either scenario of NPC lineage, AAV4 vectors would have (VZ) contacting the lateral ventricle and the adjacent physical access. Our data show that transduction of NPCs SVZ. It is reported that radial glia account for approxi- by AAV4 is rare following intraventricular injection into mately 95% of all cells in the VZ, whereas ependymal adult or neonatal mice brains. cells, which are immature at P0, comprise less than 2% The transduction profile after AAV4 introduction to of the population.16 Spassky et al17 propose a lineage the SVZ contrasts starkly the near exclusive ependymal relationship between radial glia and ependymal cells. cell transduction following ventricular delivery. Astro- Thus, the reporter gene-positive ependyma we identified cytes in the SVZ and the RMS form a tube-like structure 4 weeks after the introduction of AAV4 to P0/P1 brains around neuroblasts.4,12,13 A study by Doetsch et al4 in our study may derive from directly transduced radial suggests that there are two subtypes of astrocytes in glia, immature ependyma or both. Embryonic and the SVZ. Type B1 cells are larger and stain lighter perinatal VZ resident radial glia possess characteristics (morphologically by EM) compared to type B2 cells, and

Gene Therapy AAV4 targets ependyma and astrocytes in the SVZ and RMS G Liu et al 1506

Figure 3 AAV4eGFP transduction of astrocytes in adult SVZ and RMS. (a–c) Immunohistochemistry for eGFP and GFAP reveals transduced astrocytes in the SVZa. (d–f) Tuj-1+ neuroblasts in the RMS do not express eGFP, but remain in close contact with transduced cells. (g–i) Transduced cells in the RMS do not express NeuN, the mature neuronal marker. Bars, a and c, 10 mm (applies to a–f); g, 50 mm (applies to g–i).

ultra-structually resemble the tube-forming astrocytes in for growth factor production. Similarly, enzymes defi- the RMS. Type B2 cells contact the ventricular lumen cient in lysosomal storage diseases could be supplied via with a single cilium and are the presumptive NPC; the CSF from AAV4 transduced ependyma. injection of [3H]-thymidine into mice labels type B2 In summary, AAV4 targets two important cell types rather than type B1 cells. In this study, AAV4 transduced resident in the lateral ventricular wall of the brain, the astrocytes in the SVZ and RMS in addition to ependymal ependyma and astrocytes in the SVZ and RMS. Ependy- cells. The robust astroglial transduction did not correlate mal cell transduction by AAV4 is highly efficient and with transgene positive interneurons in the OB. Together independent of delivery route or animal age. On the with prior work our data suggest that AAV4 vectors may other hand, AAV4 vectors transduce astrocytes in the selectively target type B1 astrocytes when injected into SVZ and RMS only when delivered directly to the SVZ. the SVZ. Thus, AAV4 provides as a unique tool for These unique properties can be exploited for prolonged transduction of specific astroglial subtypes following secretion of proteins into the CSF for disease therapy, or directed delivery to the SVZ. for asking fundamental biological questions about the Astrocytes and ependymal cells have roles in adult SVZ and RMS astroctyes. neurogenesis distinct from their proposed roles as NPCs. For example, in Vax-1 knockout mice, the absence of astroctyes in the RMS is accompanied by failed neuro- Materials and methods blast migration and subsequent OB atrophy.25 AAV4 may provide a useful tool for further understanding the role Animals of this specific astroglial subpopulation, and the genes C57BL/6 mice were obtained from Jackson Laboratories they express, in migration. and from our own breeding colonies. Animal maintenance Ependymal cells secrete the bone morphogenic protein and experimental manipulations were approved by the (BMP) inhibitor noggin, blocking the gliogenic effects of University of Iowa Animal Care and Use Committee. BMPs and promoting a neuronal fate of neural stem cells.26 Also, EGF, BDNF and CNTF can promote the Viral vectors proliferation of NPCs or/and affect the migration or cell AAV4 vectors were prepared by triple transfection with fate.27–29 The efficient ependymal cell transduction packaging and helper plasmids.11 Reporter vectors used profile of AAV4 could be exploited to provide a reservoir in the study contained a nucleus-targeted Escherichia coli

Gene Therapy AAV4 targets ependyma and astrocytes in the SVZ and RMS G Liu et al 1507 b-galactosidase or an eGFP gene with expression driven References off the Rous sarcoma virus long terminal repeat (RSV) promoter. The expression cassettes were flanked by 1 Gage FH. Mammalian neural stem cells. Science 2000; 287: AAV2 ITR sequences. The viral titers were determined 1433–1438. by quantitative PCR and ranged from B5 Â 1011 to 2 Taupin P, Gage FH. and neural stem cells of 1 Â1012 pt/ml. the central in mammals. J Neurosci Res 2002; 69: 745–749. Injections 3 Alvarez-Buylla A, Garcia-Verdugo JM, Tramontin AD. A unified Newborn C57BL/6 mice (n ¼ 8) at postnatal day 0 or day hypothesis on the lineage of neural stem cells. Nat Rev Neurosci 1 were immobilized with hypothermia. AAV4 vectors 2001; 2: 287–293. (1.0 ml/animal) were injected to the right hemisphere 4 Doetsch F, Garcia-Verdugo JM, Alvarez-Buylla A. Cellular targeting the lateral ventricle. Adult mice (4–8 weeks old) composition and three-dimensional organization of the subven- tricular germinal zone in the adult mammalian brain. J Neurosci were anesthetized with ketamine/xylazine (ketamine 1997; 17: 5046–5061. 100–125 mg/kg, xylazine 10–12.5 mg/kg). AAV4 vectors 5 Falk A et al. Gene delivery to adult neural stem cells. Exp Cell Res were injected into the right lateral ventricle (n ¼ 6, 10 ml/ 2002; 279: 34–39. animal) or the SVZ (5 ml/animal; n ¼ 6) using a stereo- 6 Yoon SO et al. Adenovirus-mediated gene delivery into neuronal tactic apparatus. Injections were performed using a 30- precursors of the adult mouse brain. Proc Natl Acad Sci USA gauge Hamilton syringe driven by a microinjector (Micro 1996; 93: 11974–11979. 1; World Precision Instruments, Sarasota, FL, USA) at 7 Consiglio A et al. Robust in vivo gene transfer into adult 0.5 ml per minute. The coordinates for intraventricular mammalian neural stem cells by lentiviral vectors. Proc Natl injections were 0.4 mm rostral to the bregma, 1.0 mm Acad Sci USA 2004; 101: 14835–14840. lateral to midline, and at 2 mm depth to the pia surface. 8 Hughes SM, Moussavi-Harami F, Sauter SL, Davidson BL. Viral- The coordinates for SVZ injections were 1.0 mm rostral, mediated gene transfer to mouse primary neural progenitor 1.7 mm lateral and 2.5 mm deep. cells. Mol Therapy 2002; 5: 16–24. 9 Wu P, Ye Y, Svendsen CN. Transduction of human neural Histochemistry and immunofluorescent staining progenitor cells using recombinant adeno-associated viral Animals were transcardially perfused with PBS followed vectors. Gene Therapy 2002; 9: 245–255. 10 Passini MA, Lee EB, Heuer GG, Wolfe JH. Distribution of a by 4% paraformaldehyde in 0.1 M phosphate buffer, pH 7.4, at 4 weeks postinjection. The brains were postfixed in lysosomal enzyme in the adult brain by axonal transport and by cells of the rostral migratory stream. J Neurosci 2002; 22: the same fixative at 41C overnight, cryoprotected in 30% 6437–6446. sucrose for 24 h, and were frozen in OCT compound. 11 Davidson BL et al. Recombinant adeno-associated type 2, 4 and 5 Serial sagittal sections were cut at 16 mm thickness on vectors: transduction of variant cell types and regions in the a cryostat. One out of every five sections was examined. mammalian CNS. Proc Natl Acad Sci USA 2000; 97: 3428–3432. b-Galactosidase-positive cells were identified by 5- 12 Peretto P, Merighi A, Fasolo A, Bonfanti L. The subependymal bromo-4-chloro-3-indolyl b-D-galactoside (X-Gal) stain- layer in rodents: a site of structural plasticity and cell migration ing or immunohistochemistry (IHC). IHC was performed in the adult mammalian brain. Brain Res Bull 1999; 49: 221–243. with the following primary antibodies: Alexa 488 or 13 Peretto P, Merighi A, Fasolo A, Bonfanti L. Glial tubes in Alexa 568 conjugated rabbit IgG specific for E. coli the rostral migratory stream of the adult rat. Brain Res Bull 1997; b-galactosidase antibodies at 1:150 (Biodesign Interna- 42: 9–21. tional, Kennebunkport, ME, USA); biotinylated mouse 14 Luskin MB. Restricted proliferation and migration of postnatally monoclonal anti-NeuN at 1:100 (Chemicon, Temecular, generated neurons derived from the forebrain subventricular CA, USA); and a Cy3-conjugated mouse monoclonal zone. 1993; 11: 173–189. anti-glial fibrillary acidic protein (GFAP) at 1:2000 15 Lois C, Alvarez-Buylla A. Long-distance neuronal migration in (Sigma, St Louis, MO, USA). Alexa 350 or Alexa 488 the adult mammalian brain. Science 1994; 264: 1145–1148. conjugated streptavidin was used at 1:500 (Molecular 16 Tramontin AD, Garcia-Verdugo JM, Lim DA, Alvarez-Buylla A. Probes, Eugene, Oregon, USA). Briefly, sections were Postnatal development of radial glia and the ventricular zone blocked for 2 h at room temperature in PBS with 3% BSA, (VZ): a continuum of the neural compartment. Cereb 0.1% Triton X-100, 5% normal goat serum and 5% normal Cortex 2003; 13: 580–587. et al. mouse serum, then incubated overnight with primary 17 Spassky N Adult ependymal cells are postmitotic and are derived from radial glial cells during embryogenesis. J Neurosci antibodies diluted in PBS with 3% BSA and 0.1% Triton 1 2005; 25: 10–18. X-100 at 4 C. If needed, sections were washed and 18 Merkle FT, Tramontin AD, Garcia-Verdugo JM, Alvarez-Buylla A. incubated with secondary antibodies for 2 h at room Radial glia give rise to adult neural stem cells in the subven- temperature. Labeled sections were viewed by a laser- tricular zone. Proc Natl Acad Sci USA 2004; 101: 17528–17532. scanning confocal microscope (Zeiss LSM510). 19 Anthony TE, Klein C, Fishell G, Heintz N. Radial glia serve as neuronal progenitors in all regions of the . Neuron 2004; 41: 881–890. Acknowledgements 20 Johansson CB et al. Identification of a in the adult mammalian central nervous system. Cell 1999; 96: 25–34. We thank members of the Davidson Laboratory for 21 Doetsch F et al. Subventricular zone astrocytes are neural stem critical discussion, Stephanie M Hughes for assisting cells in the adult mammalian brain. Cell 1999; 97: 703–716. with early studies, Colleen S Stein for reviewing the 22 Laywell ED et al. Identification of a multipotent astrocytic stem manuscript and Christine McLennan for manuscript cell in the immature and adult mouse brain. Proc Natl Acad Sci preparation. This work was supported by the NIH (HD USA 2000; 97: 13883–13888. 33531) and the University of Iowa Graduate Program in 23 Chiasson BJ, Tropepe V, Morshead CM, van der Kooy D. Adult Neuroscience (GL). mammalian forebrain ependymal and subependymal cells de-

Gene Therapy AAV4 targets ependyma and astrocytes in the SVZ and RMS G Liu et al 1508 monstrate proliferative potential, but only subependymal cells 27 Craig CG et al. In vivo growth factor expansion of endogenous have neural stem cell characteristics. JNeurosci1999; 19: 4462–4471. subependymal neural precursor cell populations in the adult 24 Capela A, Temple S. LeX/ssea-1 is expressed by adult mouse mouse brain. J Neurosci 1996; 16: 2649–2658. CNS stem cells, identifying them as nonependymal. Neuron 28 Benraiss A et al. Adenoviral brain-derived neurotrophic factor 2002; 35: 865–875. induces both neostriatal and olfactory neuronal recruitment 25 Soria JM et al. Defective postnatal neurogenesis and disorganiza- from endogenous progenitor cells in the adult forebrain. tion of the rostral migratory stream in absence of the Vax1 J Neurosci 2001; 21: 6718–6731. homeobox gene. J Neurosci 2004; 24: 11171–11181. 29 Emsley JG, Hagg T. Endogenous and exogenous ciliary 26 Lim DA et al. Noggin antagonizes BMP signaling to create a neurotrophic factor enhances forebrain neurogenesis in adult niche for adult neurogenesis. Neuron 2000; 28: 713–726. mice. Exp Neurol 2003; 183: 298–310.

Gene Therapy