Differentiation of the RN33B Cell Line into Forebrain Projection Neurons after Transplantation into the Neonatal Rat Brain

Experimental Neurology 175, 370 –387 (2002) doi:10.1006/exnr.2002.7888

Differentiation of the RN33B Cell Line into Forebrain Projection Neurons after Transplantation into the Neonatal Rat Brain Cecilia Lundberg, Ulrica Englund, Didier Trono,* Anders Bjo¨rklund, and Klas Wictorin 1 Wallenberg Neuroscience Center, Department of Physiological Sciences, BMC A11, Lund University, S-221 84 Lund, Sweden; and *University of Geneva Medical School, Geneva, Switzerland

INTRODUCTION The rat neural cell line RN33B has a remarkable ability to undergo region-specific neuronal differentiation after transplantation into the CNS. To further study its neurogenic properties in vivo, we used a recombinant lentiviral vector to genetically label the cells with the Green Fluorescent Protein (GFP) gene before implantation into the striatum/cortex, hippocampus, or mesencephalon of newborn rats. Three weeks after implantation, about 1–2% of the GFP-expressing cells had developed morphologies typical of neurons, astrocytes, or oligodendrocytes, the rest remained as either immature or undifferentiated nestinpositive cells. At 15–17 weeks postgrafting, the immature cells had disappeared in most graft recipients and only cells with neuronal or glial morphologies remained in similar numbers as at 3 weeks. The GFP distributed throughout the expressing cells, revealing fine morphological details, including dendrites with spines and extensive axonal projections. In all forebrain regions, the grafted cells differentiated into neurons with morphologies characteristic for each site, including large numbers of pyramidal-like cells in the cortex and the hippocampus, giving rise to dense projections to normal cortical target regions and to the contralateral hippocampus, respectively. In lower numbers, it was also possible to identify GFP-positive granulelike cells in the hippocampus, as well as densely spiny neurons in the striatum. In the mesencephalon by contrast, cells with astrocytic features predominated. The ability of the grafted RN33B cells to undergo region-specific differentiation into highly specialized types of forebrain projection neurons and establish connections with appropriate targets suggests that cues present in the microenvironment of the neonatal rat brain can effectively guide the development of immature progenitors, also in the absence of ongoing neurogenesis. © 2002 Elsevier Science (USA) Key Words: lentiviral vector; GFP; reporter gene; neuron; glial cell.

1 To whom correspondence and reprint requests should be addressed. Fax: 46-46-2220561. E-mail: [email protected].

0014-4886/02 $35.00 © 2002 Elsevier Science (USA) All rights reserved.

Trophic factor-expanded or immortalized neural precursors are currently used in experiments with transplantation into the CNS both to develop potential therapeutic approaches and to address basic developmental issues (13, 34, 36). Their differentiation following implantation is known to depend on both intrinsic properties of the grafted cells and on host environmental conditions, such as ongoing neurogenesis. Thus, for instance, trophic factor-stimulated adult rat hippocampus-derived progenitors, grafted into adult recipients, have been found to differentiate into neurons only after implantation into the rostral migratory pathway/olfactory bulb or into the hippocampal granular cell layer (33). Similar influences from ongoing host neurogenesis have been described also for the immortalized rodent neural cell lines HiB5 (24) and C17-2 (29) and for growth-factor-expanded human neural progenitors (12). In all these cases, however, the grafted cells have been observed to differentiate primarily into interneurons with only local axonal projections. A different, but still region-specific, differentiation pattern has been documented for the rat medullary raphe-derived cell line RN33B. These cells can differentiate into neurons with morphologies typical of cortical, hippocampal, and striatal projection neurons following implantation into the corresponding structures of either neonatal or adult recipients (18, 21, 26 –28). This intriguing feature of the RN33B cell line makes it particularly interesting as a tool to explore mechanisms of region-specific neuronal differentiation, long-distance axonal outgrowth, and pathway reconstruction in the CNS. For grafted RN33B cells, the use of the lacZ reporter gene has allowed for a detailed morphological analysis (see, e.g., 26), but has provided no information on axonal projections into the host. Evidence that the RN33B cells implanted into the intact adult striatum can project, at least into the host globus pallidus, has come from retrograde tracing studies (18). The reporter gene Green Fluorescent Protein (GFP) is known to distribute its gene product throughout the expressing cells and has in several CNS applications allowed for a

370

RN33B CELLS DIFFERENTIATE INTO PROJECTION NEURONS

detailed analysis of axons (6, 16, 19, 20). Recently, we found that different neural cell lines, including the RN33B cells, can be efficiently transduced in vitro with a lentiviral vector carrying GFP and that the transduced cells continue to express the transgene at high levels also after grafting to the brain (10). In the current experiment, we grafted the GFP-expressing RN33B cells into the striatum/cortex, hippocampus, and mesencephalon of newborn rats to explore their developmental potential as assessed at 3 and 15–17 weeks after grafting. We were thus able to demonstrate the multipotent and region-specific differentiation of this cell line after transplantation, including the formation of forebrain projection neurons with longdistance axonal projections and astrocyte- as well as oligodendrocytelike cells. MATERIALS AND METHODS

Cell Culture and Transduction The cell line RN33B, derived from the rat embryonic day (E) 12.5 medullary raphe region, immortalized using the temperature-sensitive large T-antigen, and genetically labelled with the reporter gene lacZ, was cultured as previously described (21, 37). In brief, the cells were grown in polyornithine-coated flasks (Nunc, Denmark), at the permissive temperature (33°C), in DMEM/F12 medium (Cat. No. 32500; Gibco, Life Technologies, Grand Island, NY) with 1% glutamine and 0.2% NaHCO 3 and supplemented with 10% fetal bovine serum. The cells were passaged at near confluency. The recombinant lentiviral vector was produced as described previously (39). Briefly, the transfer plasmid pHR⬘CMV-GFP-W (38) was cotransfected with pMD.G and pCMV⌬R8.91 into 293T cells. Supernatants were collected on days 2 and 3 after transfection and concentrated at 116,000g by ultracentrifugation. The titer of recombinant viral vectors was determined by infecting 293T cells with serial dilutions of the concentrated supernatant. The titer (TU, transducing units per milliliter) was calculated from the percentage of GFP expressing cells at 48 h postinfection. Dilutions resulting in 15% or less expressing cells were used for the calculations. For the RN33B cells, in this report, a multiple of infection (m.o.i.) of 5 was used (based on the TU/ml on 293T cells). The viral vector was added to the culture medium and left for 48 h before the medium was changed. For further details, see Ref. 10, where minor parts of the current data are presented. Following the transduction of the RN33B cells, about 85% of the cells expressed the GFP protein at levels detectable in the epifluorescence microscope. A slow but gradual decline in the proportion of GFPexpressing cells, after the transduction, was observed. Therefore, selected cultures at passage four after the

371

transduction were sorted for GFP autofluorescence using FACS (fluorescence activated cell sorting; Vantage Turbo Upgrade, Becton–Dickinson, Franklin Lakes, NJ). Cells were suspended as single cells in HBSS (without calcium and magnesium, Gibco) with 1% bovine serum albumin (Sigma) and sorted to obtain a near 100% GFP-positive cell population. The fraction to be sorted was defined as cells with a fluorescence between 300 and 10,000 arbitrary fluorescence units, which amounted to 60% of the whole cell population. Antibodies The primary antibodies used were as follows: GFP rabbit polyclonal (1:1000; Clontech, Palo Alto, CA), GFP chicken polyclonal (1:5000; Chemicon, Temecula, CA), ␤-galactosidase rabbit polyclonal (␤-gal, 1:500; 5 Prime–3 Prime Inc., Boulder, CO), neuronal nuclei mouse monoclonal (NeuN, 1:100; Chemicon), glial fibrillary acidic protein rabbit polyclonal (GFAP, 1:500; DAKO, Denmark), SV40 large T-antigen mouse monoclonal (T-ag, 1:50; Calbiochem, La Jolla, CA), nestin rabbit polyclonal (1:500; Dr. R. D. G. McKay, NIH, Bethesda, MD), vimentin rabbit polyclonal (1:25; DAKO), NG2 rabbit polyclonal (1:500; Dr. W. B. Stallcup, Burnham Institute, La Jolla, CA), and perlecan rabbit polyclonal (1:1000; Dr. R. Fa¨ ssler, Lund University, Lund, Sweden). Secondary antibodies were either biotinylated and raised against rabbit or chicken (1: 200, Vector, Burlingame) or Cy2- or Cy3-conjugated rabbit, mouse, or chicken antibodies (1:400; Jackson, West Grove, PA). Immunocytochemistry For immunocytochemistry, to determine the proportion of GFP- and ␤-gal-positive cells prior to implantation, the cells were plated overnight at a low density in four-well plates and fixed in 4% paraformaldehyde [10 min, room temperature (RT)]. After rinses with potassium phosphate buffered saline (KPBS), the cultures were preincubated with 5% normal serum, of the same species as the secondary antibody, whereafter the cells were incubated with the primary antibodies recognizing either GFP or ␤-gal (overnight, ⫹4°C), and then with the appropriate secondary antibodies, followed by avidin– biotin–peroxidase complex (ABC, Vectastain Elite, Vector), and with 3,3-diaminobenzidine (DAB, 25 mg/ml; Sigma, St. Louis, MO) as chromogen. Cell Preparation and Transplantation On the day of transplantation, the cell culture medium was replaced by Hank’s balanced salt solution (HBSS, Gibco) without calcium and magnesium (1 min) and then by 0.1% trypsin, dissolved in the same buffer (4 min, 37°C), whereafter serum-containing medium was added. The detached cells were transferred into an

372

LUNDBERG ET AL.

TABLE 1 Transplantation Coordinates Group

cx/str

hpc

mes

A (unsorted)

A:⫹0.5 L:2.2 V:⫺3.0; ⫺2.5 and A:⫹0.9 L:1.9 V:⫺3.0; ⫺2.5 (4x0.25ul) A:⫹0.5 L:2.2 V:⫺3.0; ⫺2.5 (2x0.5ul)

A:⫺1.2 L:1.5V:⫺1.8; ⫺1.5 (2x0.5ul)

A:⫺3.4 L:1.3V:⫺3.0; ⫺2.5 (2x0.5ul)

A:⫺1.2 L:1.5V:⫺1.8; ⫺1.5 (2x0.5ul)

A:⫺3.4 L:1.3V:⫺3.4; ⫺3.0 and A:⫺2.9 L:1.3V:⫺3.2; ⫺3.0 (4x0.25ul)

B (FACS)

Abbreviations: cx, cortex; hpc, hippocampus; mes, mesencephalon; str, striatum. A (anterior) and L (lateral) in relation to bregma; V (ventral) from the dura.

Eppendorff tube and centrifuged (5 min, 600 rpm), resuspended, and counted in a hemocytometer and finally prepared into a suspension of 100,000 cells/␮l in HBSS (with calcium and magnesium). Neonatal rats (postnatal days 1 and 2) were used as recipients (Sprague–Dawley, B&K Universal, Stockholm, Sweden). All animal-related procedures were conducted in accordance with local ethical guidelines and approved animal care protocols. For surgery, the rat pups were temporarily removed from the mother and anesthetized using hypothermia and placed in a neonatal device mounted onto a Kopf stereotaxic frame (8), and the injections were made from glass capillaries attached to a 2-␮l Hamilton syringe (Hamilton Bonaduz AG, Bonaduz, Switzerland). The transduced RN33B cells were implanted into the striatum/cortex, hippocampus, and mesencephalon. In total, 1 ␮l of suspension (i.e., 100,000 cells) was injected into the striatum/cortex (i.e., with coordinates aimed at striatum, but with cells subsequently distributed in both of these structures; see also below), hippocampus (hpc), or mesencephalonmes, as described in Table 1. The coordinates were slightly adjusted between the two groups to avoid leakage from the first needle tract when injecting in a second tract (for cx/str) and to avoid damaging a large venous sinus (for mes). The injections were each made over 1 min, and the glass capillary left in place for 2 extra min before retraction. Two rounds of transplantation were performed: group A, with unsorted cells from passage three after transduction, and group B, with cells from passage five after three rounds of FACS. Twenty-three rats were grafted into one or two regions each (see Table 2), with the the total numbers of grafts/region as follows: striatum/cortex, 17 (4 in A and 13 in B); hippocampus, 13 (7 in A and 6 in B); and mesencephalon, 9 (2 in A and 7 in B). To determine whether GFP potentially released from any dying grafted cells could be taken up and detected in host cells, separate control animals were included. Thus, five additional P1 rats received regular grafts of

cells implanted into striatum/cortex and two additional rats were injected with the same cell suspension, but after killing the cells through five repeated freeze (liquid nitrogen)/thaw cycles. Immunohistochemistry At 3 or 15–17 weeks after transplantation (see Table 2), the graft recipients were anesthetized by intraperitoneal injections of an overdose of pentobarbital and perfused transcardially, with phosphate-buffered saline (PBS) (at room temperature, 1 min), followed by 200 –250 ml 4% paraformaldehyde (ice-cold) dissolved in phosphate buffer (PB, 0.1 M). The brains were removed, postfixed in the same fixative for 2– 4 h, and then incubated overnight in 20 –25% sucrose in PB. The brains were then sectioned in eight series in the coronal or sagittal (three rats only) plane, on a freezing microtome, at 30- or 40-␮m section thicknesses. The mesencephalic graft placement was analyzed only at the 3-week survival time. TABLE 2 Summary of All Transplants Transplantation region Groups (n ⫽ total no. of grafts)

cx/str

hpc

mes

3 weeks survival A B Total

2 6 8

4 3 7

2 7 9

15–17 weeks survival A B Total

2 7a 9

3 3 6

— — 0

Note. Group A, unsorted cells; Group B, FACS-sorted cells. Abbreviations: cx, cortex; hpc, hippocampus; mes, mesencephalon; str, striatum. a Two specimens were sagitally sectioned and not quantified.

RN33B CELLS DIFFERENTIATE INTO PROJECTION NEURONS

Throughout the study, GFP was analyzed using immunohistochemistry, rather than by direct fluorescence, in order to increase the sensitivity of the analysis. For single immunohistochemistry to detect either GFP or ␤-gal, the sections were first quenched with 3% H 2O 2 and 10% methanol in KPBS to remove endogenous peroxidase and thereafter preincubated in 5% of normal serum and 0.25% Triton X-100 in KPBS (1 h at RT). Incubations with the primary antibody were then followed by appropriate secondary antibodies, ABC and the DAB reaction (see above). For double immunostaining (GFP and ␤-gal, NeuN, GFAP, T-ag, nestin, NG2, vimentin, or perlecan), the two primary antibodies were applied in parallel followed by Cy2-/Cy3conjugated secondary antibodies. The sections were mounted onto chrome-alum coated slides and dehydrated and coverslipped with DPX mountant (BDH, UK) (DAB-reacted sections) or directly coverslipped with PVA/DABCO (for Cy2-/Cy3-stained sections (23). The material was analyzed using light-, fluorescence-, and confocal laser scanning (MRC 1024UV; Bio-Rad, UK) microscopy (also using the Openlab software). Throughout most of the study, the rabbit-anti-GFP antibody (Clontech; no longer available from the manufacturer) was used. For immunofluorescent doublestaining, a chicken-anti-GFP antibody (Chemicon) was instead utilized (as in Figs. 2 and 7C and 7F). Also this chicken antibody resulted in similar GFP immunostaining, with labeling of cells with the various morphologies and cellular details described below, although, using our standard procedures, with less intensity and fewer processes revealed.

373

technique. The total numbers of neuron- and astrocytelike cells, respectively, were thus estimated by counting all such cells in every eighth section throughout the various structures followed by corrections with the Abercrombie formula, P ⫽ 1/f ⫻ A ⫻ M/(D ⫹ M) (where P is the total cell number, A is the raw count, D is the average cell diameter, M is the section thickness, and f is the frequency of selected sections) (1). Mean cell diameters were calculated for astrocytes and for neurons from each region separately, measuring approximately 30 – 60 cells per structure, with a calibrated ocular grid. Representative examples of cells fulfilling the criteria of neuron- and astrocytelike cells are shown, e.g., in Figs. 3 and 5–7 (neurons, in most cases with large cell bodies and multiple, branching dendrites, often with spines; astrocytes, with smaller cell bodies, rich in branching fibers). The cell numbers were compared using an ANOVA followed by Fisher’s PLSD post hoc test. Data presented as mean (range) in the text. For quantification, the animals from groups A and B (see Table 2) were pooled for each region. The slight changes in transplantation coordinates resulted in no visible differences in graft placements and cellular distribution. Moreover, both the in vitro and in vivo characteristics of the RN33B cell line remained unchanged through repeated passaging and FACS sorting, in agreement with previous studies (18). For total cell numbers, the cortex and striatum were assessed together, whereas the counting of differentiated cells was performed for each of these two structures separately. RESULTS

Quantification The total numbers of GFP-positive cells in the different regions of the 3-week specimens were counted using a stereological method, with unbiased and systematic sampling. Cells were thus counted in the striatum and cortex together (n ⫽ 8). hippocampus (and overlying cortex) (n ⫽ 6), and mesencephalon (n ⫽ 8), with only the specimens cut coronally included. The numbers of cells were estimated using the optical dissector (14) and by subsequently applying the fractionator formula N ⫽ SQ ⫻ F1 ⫻ F2 ⫻ F3, where N is the total number of GFP expressing cells, SQ is the sum of cells counted; F1 is the fraction of sections used (i.e., 8 in the present study), F2 is the fraction of tissue depth used to collect data (1.33), and F3 is the fraction of tissue area used to collect data. The GRID software (Interactivision, Denmark) was used to generate unbiased counting frames. To separately quantify the differentiated GFP-positive cells with neuron- and astrocytelike morphologies, in both the short- and long-term animals, we decided to apply the Abercrombie method. The low total numbers of cells practically ruled out the use of the stereological

Survival and Overall Distribution of Grafted, GFP-Positive Cells The transduced RN33B cells propagated well in vitro, and although the proportion of GFP-expressing cells was slowly but gradually reduced over time, the GFP-positive fraction remained high through repeated passages. At the time of implantation, about 85% of the cells in group A (i.e., at passage three after transduction) and 75% of the cells in group B (i.e., at passage five after the last FACS) were GFP-positive in culture. Since the vast majority of the cells were also ␤-galimmunoreactive, most cells thus expressed both reporter genes. All of the 23 grafted, short- and long-term specimens, of the nonsorted and FACS-sorted cells (groups A and B), as well as the five rats which received live cells in the control group, contained GFP-immunoreactive cells. The groups A and B animals were analyzed together, since very similar transplantation parameters were used. GFP-positive cells were present throughout all the implanted structures (Fig. 1). Since the cells grafted into the striatum were distributed in both striatum and cortex (Figs. 1A and 1B), that injec-

374

LUNDBERG ET AL.

FIG. 1. Distribution of GFP-expressing RN33B cells in the various grafted regions and of their axonal projections at 3 weeks after transplantation. (A) Microphotograph of cells implanted into the striatum, present both throughout the striatum (str) and within large areas of the adjacent cortex (cx). (B) Schematic coronal drawings of the GFP-positive neurons present in the cortex and striatum (small dots in the

RN33B CELLS DIFFERENTIATE INTO PROJECTION NEURONS

tion site is referred to as striatum and cortex, and the total number of cells was calculated for these two regions together. The stereological cell counts thus revealed on average 146,202 GFP-positive cells in the striatum/cortex, 66,623 in the hippocampus, and 112,898 in the mesencephalon (100,000 cells injected in each case) (Table 2). The animals receiving the freezethawed suspension of killed cells contained no GFPpositive cells (data not shown). At 3 weeks after grafting, the vast majority of the GFP-expressing cells had immature or undifferentiated morphologies in all regions, often forming long chains of cells associated with smaller blood vessels (Fig. 2A). As revealed by GFP and T-ag double staining in two specimens, most of these undifferentiated cells were T-ag-positive, whereas the cells with mature neuronal or glial morphologies were not (data not shown). In sections double stained for GFP and perlecan, a marker of the vascular basal lamina (7), it was clear that the grafted GFP-positive cells were present on the parenchymal side of the endothelial sheet (data not shown). The majority of these grafted, perivascular cells expressed the neural precursor markers nestin (15, 11) (Figs. 2C–2E) and vimentin (11) (Figs. 2F–2H), and a few examples of cells expressing the glial progenitor marker NG-2 (32) were also detected (Figs. 2I–2K). In two of the grafted rats, large numbers of small rounded cells, with few and short processes, were present also within or adjacent to the ependyma and subventricular zone (Fig. 2B). These cells had presumably integrated into the regions near to the ventricles, after leakage of cells from the injections aimed at the striatum or hippocampus. These cells had immaturelooking morphologies and expressed nestin (Fig. 2L– 2N) and occasionally vimentin (data not shown). The cells in the ependymal/subependymal regions were intermingled with host GFAP-positive cells, but with no clear evidence using confocal microscopy of GFP/GFAP double-labeled cells (data not shown). Interestingly, cells integrated on the septal side, close to the ventricle, had a different appearance, with more loosely arranged cells and more mature, including neuronal, morphologies (Fig. 2B). GFP-positive cells had also migrated into frontal cortical areas. Sections from more rostral regions, such as the rostral migratory stream (rms) and olfactory bulb (ob), were not analyzed. In one of the few specimens sectioned in the

375

sagittal plane, a few cells were detected in the rms, but not in the ob. Neuronal and Glial Differentiation at Three Weeks after Transplantation Cells with mature morphologies of neurons or astrocytes accounted for about 1–2% of the total number of GFP-positive cells (Table 3), and their morphologies differed markedly between the various implanted regions. In the cortex (n ⫽ 8), large numbers of GFP-positive cells were found not only close to the needle tract, dorsal to the injection site in the striatum, but also medially and laterally in the lower layers of the cortex (Fig. 1A). A large proportion of the differentiated cells closely resembled cortical pyramidal neurons and were mostly situated in layers V–VI, possessing large somas and rich trees of branching and spine-bearing apical and basal dendrites (Figs. 3A–3C). The apical dendrites had the normal vertical orientation, extending toward the pial surface, perpendicular to the cortical layers. In addition, the cortex contained smaller, multipolar neuronlike cells, presumably representing interneurons. There were also numerous GFP-positive cells with characteristic astrocytic morphologies (Fig. 3D), as well as many examples of oligodendroglial-like cells, situated within or adjacent to the corpus callosum (Fig. 3E). The Abercrombie-adjusted cell counts for the cortex (Table 3) revealed on average 1837 (range 543–2838) neuronlike and 394 (49 – 697) astrocytelike cells in the cortical gray matter. The striatum (n ⫽ 8) contained fewer GFP-expressing and differentiated cells (Table 3), i.e., on average 46 (range 5–92) neuronlike (Figs. 3F and 3G) and 154 (65–291) astrocytelike cells, as well as occasional oligodendroglial-like cells, associated with the internal capsule bundles (Fig. 3H). While some of the neuronlike cells were densely spiny, most had only a low density of spines. As illustrated in Figs. 1B and 4, the GFP-staining visualized extensive axonal projections emanating from the cells situated in the cortex and striatum. Apart from a dense fiber network within the ipsilateral cortex, projections could also be followed through the corpus callosum into the contralateral cortex. At more caudal levels, numerous GFP-positive fibers were present within the internal capsule fiber bundles run-

most rostral image) and giving rise to extensive projections within the cortex, through the corpus callosum (cc) and internal capsule (ic), and into the thalamus (thal) and the mesencephalon. (C) Following transplantation into the hippocampus (hpc), GFP-positive cells occurred mostly within lateral portions, including parts of CA3. (D) Also the neuron-like cells of the hippocampus (small dots) provided a dense network of efferent projections, both within the ipsilateral (see also in C) and the contralateral hippocampus. (E) Following transplantation into the mesencephalon, large numbers of cells were present above all in dorsal and lateral regions, including the superior colliculus (sc), and the medial geniculate complex (MG). Abbreviations: cc, corpus callosum; cp, cerebral peduncle; DG, dentate gyrus; fi, fimbria; ml, medial lemniscus; sn, substantia nigra; v, ventricular system. (A, C, and E) Scale bar: 500 ␮m.

376

LUNDBERG ET AL.

377

RN33B CELLS DIFFERENTIATE INTO PROJECTION NEURONS

TABLE 3 Quantifications of Numbers of GFP-Positive Cells after Transplantation 3 weeks Region cx/str

Total numbers 146,202 ⫾ 18,413 (8)

cx str

hpc

66,623 ⫾ 24,078 (6)

mes

112,898 ⫾ 18,906 (8)

15–17 weeks Differentiated cells

Differentiated cells

N: 1837 ⫾ 248 (8) a A: 394 ⫾ 75 (8) N: 46 ⫾ 10 (8) A: 154 ⫾ 25 (8) N: 534 ⫾ 105 (6) A: 161 ⫾ 52 (6) N: 75 ⫾ 16 (8) A: 886 ⫾ 198 (8) a

N: 1871 ⫾ 581 (7) a A: 187 ⫾ 169 (7) N: 111 ⫾ 46 (7) A: 61 ⫾ 55 (7) N: 499 ⫾ 219 (6) A: 41 ⫾ 48 (6) — —

Note. Values given as mean ⫾ S.E.M. (and numbers of analyzed transplants). a Significant difference in numbers of neuron- and astrocytelike cells, P ⬍ 0.05 (ANOVA followed by Fisher’s PLSD post hoc test). Abbreviations: cx, cortex; hpc, hippocampus; mes, mesencephalon; str, striatum. —, not analyzed.

ning through the striatum (Figs. 3F and 4A). These projections could be traced into the thalamus (Fig. 4B), where they formed dense fiber networks with arborizations and terminal-like swellings, above all in the ventrolateral and ventrobasal nuclei (Figs. 4B– 4D). The projections then continued further caudally along the cerebral peduncle into the mesencephalon, where some of these fibers were found to terminate within the substantia nigra (Figs. 4E and 4F). In the hippocampus (n ⫽ 7), large numbers of GFPexpressing cells with neuronal morphologies were present in the pyramidal layers of CA1–CA3 (Figs. 1C, 1D, and 5A). The majority of these cells resembled normal pyramidal cells and possessed multiple branching dendrites, with the normal orientation perpendicular to the cell layer (Fig. 5D). Extensive and similarly distributed GFP-positive projections were found throughout the ipsilateral and contralateral rostral hippocampus, with high fiber densities above all in the stratum oriens, but also in the stratum radiatum and stratum lucidum (Figs. 1C, 1D, and 5B). The fibers were seen to cross the midline in the ventral hippocampal commissure (Fig. 5C) and were present also in the fimbria and dorsal septum. On average, the grafted hippocampi contained 534 (range 166 – 813) cells with neuronal morphologies and lower numbers (161) of astrocytelike cells (17–319) (Table 3). Although most cells were located in the lateral hippocampal regions, more medial parts also contained cells (see, e.g., Fig. 1C). Occasional neurons had

become integrated into the granule layer of the dentate gyrus; these cells presented the characteristic morphological features of dentate granule cells with branched apical dendrites extending into the overlying molecular layer (Figs. 5E and 5F). Medially situated cells were found also contralateral to the implant site, presumably as a result of migration from cells leaking into the ventricles. Similar to the injections aimed at the striatum, also the hippocampal injections resulted in large numbers of cells distributed in the adjacent neocortex. In the mesencephalon (n ⫽ 9), most cells were distributed dorsally and laterally, above all within the superior colliculus and the medial geniculate nucleus (Fig. 1E). In five specimens, GFP-positive cells were found also in more ventral areas, including the substantia nigra. Among the differentiated cells, on average 886 (410 –1619) presented typical astroglial features (Figs. 6A and 6B; Table 1) and less than 10%, i.e., on average 75 cells (53–185), had neuronal morphologies. They were mostly found more laterally, in structures like the medial geniculate complex (Fig. 6B). Only occasional cells with a neuronal morphology were found within the substantia nigra. Overall, ␤-gal immunostaining revealed similar cell types as observed with GFP labeling, although the two reporter genes were mostly detected in separate cell populations (Figs. 7A–7E). GFP staining clearly revealed more of the precise morphologies of the maturelooking grafted cells, including further details of the

FIG. 2. Characterization of GFP-positive cells associated with small blood vessels and of cells located in the periventricular regions at 3 weeks after transplantation. (A) Microphotograph of GFP-positive cells (arrowheads) outlining host capillaries. (B) GFP-positive cells distributed periventricularly, laterally within the ependymal and subependymal zones, adjacent to the striatum (str), and medially in the septal region (sept). (C–E) Confocal microscopy images of perivascular GFP-positive cells (C) expressing the neural progenitor marker nestin (D) as visualized in the merged image (E) (examples indicated by arrowheads). (F–K) In addition, the cells associated with the capillaries, also expressed the early progenitor marker vimentin (F–H) and occasionally the early glial marker NG-2 (I–K) as indicated by arrowheads in the two merged images (H and K, respectively). (L–N) The GFP-positive cells present in the the ependymal and subependymal zones (L) were found to express nestin (M) as exemplified in the merged image (N, arrowheads). Arrows in K and N indicate single-labeled GFP-positive cells. Scale bars: (A and B) 50 ␮m, (C–K) 20 and ␮m, (L–N) 15 ␮m.

378

LUNDBERG ET AL.

RN33B CELLS DIFFERENTIATE INTO PROJECTION NEURONS

dendritic trees and more extensive axonal projections (Figs. 7A and 7B). Thus, for instance in the thalamus, both GFP- and ␤-gal-labeled fibers were detectable, but the GFP-immunoreactive projections occurred at a much higher density and possessed terminal arborizations and swellings (Figs. 7F–7H). Astrocytelike cells were observed also with ␤-gal-staining, as well as a few cells resembling oligodendrocytes, although not nearly as many as revealed in the GFP analysis (data not shown). GFP and NeuN double staining showed that many of the GFP-positive cells with typical neuronal morphologies, for instance, in the hippocampus and the cortex, expressed this neuron-specific protein (Figs. 7I–7K). 15- to 17-Week-Old Transplants In these long-term grafts, the cells with immature or undifferentiated features had largely disappeared. Thus, in 9 of 11 animals only GFP-positive cells with neuronal or glial morphologies were found. The other two animals possessed both differentiated GFP-positive cells as well as regions rich in smaller and immature-looking cells, often surrounding blood vessels, as observed at 3 weeks. These immature cells (but not the differentiated cells) were seen to remain T-ag-positive to a large extent (data not shown). The numbers of cells with neuronal phenotypes were similar to those found at 3 weeks in the cortex, striatum, and hippocampus (Table 3). Cells with glial morphologies were generally fewer, with a marked variation from animal to animal (e.g., from 0 to 1201 cells in the cortex) and with 8 of 20 transplantation sites containing no or only very few astrocytelike cells. GFP-positive oligodendrocytelike cells detected in none of the long-term specimens. The GFP-immunoreactive neurons appeared very similar to those found in the short-term specimens. Thus, the cortex (n ⫽ 9) contained above all numerous pyramidal-like cells, with extensive dendritic trees (Figs. 8A and 8C). Bundles of axons projected through the internal capsule, into the thalamus, where elaborate ramifications and terminal swellings were found (Figs. 8D and 8E). In one of the rats with grafted cells in the cortex, analyzed in sagittal sections, GFP-labeled fibers could be detected as far caudally as in the pons (data not shown). The striatum (n ⫽ 9) contained more sparsely distributed GFP-positive neurons (Fig.

379

8B). The grafted hippocampi (n ⫽ 6) contained large numbers of GFP-positive pyramidal neurons situated in the CA1–CA3 regions, mostly in the lateral portions (Figs. 8F and G), and extensive projections were found throughout both the ipsilateral and contralateral hippocampus (data not shown). The overall density and distribution of the GFP-positive projections in these long-term specimens appeared similar to that of the short-term animals. Mesencephalic grafts were not analyzed at this time point. DISCUSSION

Here, we demonstrate further the site-specific neuronal differentiation of the RN33B cell line, including the formation of forebrain projection neurons with extensive axonal projections, after grafting into the newborn rat brain. The genetic labeling with a GFP-carrying lentiviral vector thus revealed the entire cytoplasm of the expressing neurons and in addition visualized numerous astrocyte- as well as oligodendrocytelike cells. The regional specificity was also reflected in that more cells with neuronal as compared to astrocytic morphologies were found in the engrafted cortex and hippocampus, whereas the astrocytelike cells predominated in the mesencephalon. At 3 weeks after grafting, the vast majority of the GFP-expressing cells were either immature or undifferentiated. However, in most of the 15- to 17-week-old recipients only mature neuron- or glial-like cells were still detectable at similar numbers as after 3 weeks. GFP Expression in Grafted RN33B Cells The detection of around 150,000 GFP-positive cells in the striatum and cortex at 3 weeks, after implantation of 100,000 cells, most likely reflects an ongoing cell division initially after implantation, in agreement with a previous report using [ 3H]thymidine and the reporter gene lacZ to analyze the grafted cells (18). Many of the implanted cells probably downregulate the transduced reporter gene (18), as also suggested in the current study by the frequent occurrence of the two reporter genes GFP and lacZ in morphologically similar but separate cell populations. The cell counts at 3 weeks indicated that only 1–2% of the GFP-expressing cells

FIG. 3. GFP-positive cells within the neocortex (A–E) and striatum (F–H), at 3 weeks after transplantation. (A) While most cells had an immature or undifferentiated appearance, a subpopulation displayed distinct neuronlike morphologies with large cell bodies, and with multiple branching dendrites. Pyramidal-like cells were mostly located in the lower cortical layers, as approximately denoted with roman numerals. The small asterisk indicates an area with numerous GFP-positive cells surrounding small blood vessels, as characterized further in Fig. 2. (B) Higher magnification of the boxed area in A, with two pyramidal-like cells indicated by arrows. (C) Detailed image of a neuron situated in layer II and indicated in A, with arrows pointing to spine-bearing dendrites. In addition, many GFP-expressing astrocytelike cells were present in the cortex (D), as well as in cells resembling oligodendrocytes (E), situated in white matter tracts. (F) Also in the striatum, the majority of the cells displayed immature or undifferentiated morphologies, but there were also low numbers of cells with typical neuronal (arrow) or astrocytelike (arrowhead) features. (B) Higher magnification of the neuron indicated by an arrow in A. (C) Oligodendrocytelike cell situated within a bundle of the internal capsule. Scale bars: (A) 200␮m, (B and F) 100 ␮m, (C–E and G and H) 25 ␮m.

380

LUNDBERG ET AL.

FIG. 4. GFP-immunoreactive projections extending from the cells situated in the cortex and striatum, at 3 weeks after grafting in darkfield and brightfield images. (A–D) Fibers running through the caudal parts of the striatum (str) and globus pallidus (gp), with high fiber densities in the bundles of the internal capsule (ic) (A) and in the thalamus (thal) (B–D). (E and F) GFP-positive projections detected further caudally, e.g., in the substantia nigra (sn) and in the cerebral peduncle (cp). Inset in F shows the fiber indicated with an arrow, at higher magnification, with terminal-like swellings. Scale bars: (A, B, E, and F) 200 ␮m and (C and D) 100 ␮m.

RN33B CELLS DIFFERENTIATE INTO PROJECTION NEURONS

381

FIG. 5. GFP-labeled cells in the hippocampal formation at 3 weeks postgrafting in darkfield and brightfield images. (A and D) Apart from cells with undifferentiated or immature characteristics, there were large numbers of GFP-positive cells, with morphologies of pyramidal neurons (D), mostly situated laterally in the CA3-region within the pyramidal cell layer (pcl). (B) GFP-positive projections extending into the contralateral hippocampus, with highest fiber densities in the stratum oriens (so) and stratum radiatum (rad) and with fibers detected also in the fimbria (fi). (C) GFP-stained axons crossing the midline in the ventral hippocampal commissure (vhc). (E and F) Small numbers of GFP-expressing neurons, as well as cells with astrocytic features, were also located in the dentate gyrus (DG), also on the side contralateral to the implanted hippocampus (as in E). Arrows indicate three examples of neurons located within the granule cell layer (GL). Arrow 1 points to the same cell also in the low magnification darkfield inset of (F), which gives an overview, including the area illustrated in E. Abbreviation: cc, corpus callosum. Scale bars: (A–C and F) 200 ␮m (D and E) 50 ␮m.

had differentiated into cells with neuronal or glial morphologies. The numbers of these differentiated cells remained at similar levels also at the longer survival times, in line with results of previous studies (27). The observations in the T-ag/GFP double-stained sections suggest that the immortalizing gene is down-regulated in the differentiating cells, while at least some of the

undifferentiated cells appear to continue to express the T-ag. The undifferentiated or immature cells were no longer detectable in 9 of 11 long-term animals. Although this could be due to transgene down-regulation, it more likely reflects a loss of this cell population over time, as suggested from previous lacZ- and [ 3H]thymidine-based investigations (18, 27). In vitro, undifferen-

382

LUNDBERG ET AL.

FIG. 6. GFP-immunoreactive cells in the mesencephalon at 3 weeks after transplantation. (A) Among cells with more differentiated morphologies, astrocytelike cells (arrows) predominated and were distributed above all in the more dorsal regions, including the superior colliculus (sc). (B) Examples of cells with neuronal or astrocytelike (arrows) morphologies in the medial geniculate complex (MG). Scale bars: 50 ␮m.

tiated RN33B cells have previously been found to express nestin and vimentin (37). At 3 weeks after transplantation, double staining with nestin and vimentin revealed that most of the cells associated with blood vessels were indeed immature cells, still expressing these early neural progenitor markers. Also, the cells situated in the ependymal or subependymal regions had similar immature features. Thus, it appears that only a small proportion of the implanted cells had received, or were able to respond to, the appropriate signals necessary for a further differentiation into more mature neural phenotypes. The transplantation of cells killed through freezing and thawing resulted in no uptake of GFP by host cells, parallelling previous observations on lacZ-expressing cells (27). Indeed, GFP is known from cell lineage studies to remain in an expressing cell and its progeny and not to be transferred fortuitously from cell to cell (20). Moreover, the lentiviral vector is well suited for this type of grafting experiment, with its high transduction efficiency, genomic integration and by being replication-defective (4, 10, 39). The repeated passaging and FACS sorting of the cells, prior to implantation, further assured the specificity of the GFP labeling.

Region-Specific Differentiation into Neurons and Glial Cells The host environment is well known to provide a strong influence on the development of in vitro-expanded precursor cells following transplantation. For many cell types, the differentiation into neurons has been shown to depend on or be strongly influenced by ongoing neurogenesis (12, 33) or a lesion-induced reappearance of appropriate signals (28, 30). The hippocampus-derived and large T-antigen-immortalized cell line HiB5 (24) has been found to differentiate significantly into neurons (up to 20 –25%) when grafted into the neonatal dentate gyrus and cerebellum (24), but only to a small extent in the intact adult striatum (17) and not at all in the adult neocortex (5). The differentiating subpopulation of the RN33B cells are known to develop into cells with morphologies typical of projection neurons, for instance, in the neonatal neocortex and hippocampus (26). This was observed also in the present study and could reflect that still developing although no longer neurogenic regions can provide factors stimulating neuronal differentiation for this particular cell line. For the RN33B cells, however,

FIG. 7. Comparison of ␤-gal immunostaining with GFP/NeuN double labeling. (A and B) Light microscopical images from the cortex of a long-term specimen, with immunohistochemical detection of GFP (A) and ␤-gal (B). Note the more detailed staining of neurites with GFP staining. (C–E) Fluorescence double labelling of GFP- (C) and ␤-gal- (D) immunoreactive cells in the cortex, and a merged image (E). This illustrates that the reporter gene products were mostly detected in two separate cell populations. (F–H) Similarly double-stained sections, from the thalamus, with more projection fibers immunoreactive for GFP (F), as compared to ␤-gal (G), as also observed in the merged image (H). (I–K) Section through the hippocampus, double stained to detect GFP (I) and the neuron-specific protein NeuN (J), and merged image (K), with three examples of double-stained GFP- and NeuN-positive cells. Scale bars: (A and B) 100 ␮m and (C–K) 50 ␮m.

RN33B CELLS DIFFERENTIATE INTO PROJECTION NEURONS

383

384

LUNDBERG ET AL.

RN33B CELLS DIFFERENTIATE INTO PROJECTION NEURONS

similar differentiation patterns have been observed in the cortex and hippocampus also in adult recipients, suggesting that these stimulatory and regionally specific signals persist into adulthood (27). Indeed, we have seen similar site-specific neuronal differentiation from GFP-labeled RN33B cells implanted into cortex or hippocampus in a few adult rats, including fibers projecting into the contralateral hippocampus (CL, UE, AB, KW, unpublished observations). Thus, the present results demonstrate two remarkable features of the RN33B cell line; first, the propensity to develop into several types of forebrain projection neurons, in particular cortical and hippocampal projection neurons; and second, the ability to integrate into already established cell layers. This integration takes place, for instance, in the hippocampal pyramidal and dentate granule layers, as well as in the deep layers of the cerebral cortex, where the cells adopt the characteristic morphological features of the neurons present at each of these locations. Moreover, the cells assume the proper orientation and become appropriately aligned with the host neurons. While the first property probably reflects an intrinsic feature, or commitment, of the cells themselves, the second property is likely to depend on cues present in the local host microenvironment. Overall, the differentiation pattern of any expanded stem or precursor cell type following transplantation clearly depends both on intrinsic properties related to, for instance, the regional origin, developmental status/degree of commitment and immortilization/cell culture methods, and on influences from the host environment. A greater neuronal as compared to astrocytic differentiation was observed in the cortex and hippocampus, but with an astrocytic predominance in the mesencephalon. This again shows the responsiveness of these cells to regionally specific instructive molecules or signals. Although the RN33B cells were first reported to differentiate only into neurons in vivo (27), later ␤-galand [ 3H]thymidine-based studies have revealed their development also into astrocytes (18). The current GFP labeling, in addition, visualized numerous oligodendrocytelike cells (cf. Ref. 35). This further illustrates the interesting multipotent properties of this cell line, with its ability to differentiate into distinct neuronal phenotypes, combined with a broader capacity to develop into all major cell types of the CNS.

385

Anatomical Integration of the Implanted Cells For grafts of primary neural cells, it is well established that reciprocal anatomical graft– host integration can occur after implantation into the developing, but also adult, CNS (3). That also in vitro-expanded neural stem or progenitor cells establish contacts with the recipient CNS is known, for instance, from studies on the immortalized cell lines C17-2 (29) and RN33B (26), in both cases with electronmicroscopical evidence of synapse formation between host axons and reporter gene-expressing precursor cells. More recently, functional synaptic contacts onto grafted, trophic factorexpanded cortical stem cells were electrophysiologically documented (2). With regard to efferents, HiB5derived cerebellar and dentate granule cells have been found to extend appropriate local axons (24), and for growth factor-expanded human precursors, projections into the recipient rat brains have recently been demonstrated using species-specific antisera (12, 22, 25). For RN33B cells, retrograde tracing from the adult host globus pallidus has revealed efferent connections from intrastriatal grafts (18), whereas the ␤-gal-based analysis has provided no information on axonal projections (26, 27). In agreement with previous experiments utilizing the GFP reporter gene (2, 6, 9, 19), the transduction of RN33B cells with the lenti-GFP vector allowed the visualization of the entire cytoplasm of the expressing cells, including dendrites and long axonal projections. Using this approach, we were able to establish that the RN33B cells implanted into the neonatal cortex and hippocampus not only differentiate morphologically into pyramidal neurons, but in addition extend long-distance projections into appropriate target sites. The ability to form regionally specific projections has previously been shown, e.g., for grafts of primary cortical tissue implanted into the cortex of neonatal recipients (see, e.g., Ref. 31). The present results demonstrate that also the immature cells of the cell line RN33B can develop specifically to adopt the particular projection patterns of neurons normally present at that implantation site. The terminal arborizations and swellings were suggestive of an actual formation of direct contacts onto host elements, although the occurrence of functional synaptic connections remains to be established by electron microscopy and/or electrophysiological recordings.

FIG. 8. Nine of 11 of the long-term (15- to 17-week-old) specimens contained only GFP-expressing cells with more differentiated morphologies, typical of neurons or glial cells. (A) In the cortex (cx), large numbers of pyramidal-like neurons were present, with rich dendritic trees and networks of fibers. (B) Examples of GFP-expressing neurons within the striatum (str), some with densely spiny dendritic trees (arrow), but most with only sparsely spiny or aspiny dendrites (arrowheads). (C) Higher magnification of a pyramidal-like cell within the cortex, with GFP-positive spines (arrows). (D) Darkfield microphotograph of dense GFP-positive projections extending into the thalamus. (E) Higher magnification of axonal branches with terminal-like swellings within the thalamus. (F) GFP-expressing pyramidal-like neurons situated in the CA1–CA3 regions of the hippocampus. (G) Higher magnification of GFP-expressing cells in the pyramidal cell layer of CA1. Scale bars: (A, B, F, and G) 100 ␮m, (C) 20 ␮m, and (D and E) 50 ␮m.

386

LUNDBERG ET AL. 8.

Conclusions The present study provides further evidence of the remarkable ability of the RN33B cell line to differentiate into distinct and specific neuronal as well as glial cell types after transplantation into the neonatal rat brain. Most importantly, the genetic labeling of the cells with GFP allowed for a very detailed morphological analysis, visualizing the newly formed forebrain projection neurons, including also their long-distance axonal projections. The functional effects of most CNS cell replacement strategies are thought to largely depend on the establishment of graft– host connections (3). The current findings demonstrate that at least certain types of in vitro-expanded and transplanted precursor cells are capable of responding to signals from the host CNS and integrate anatomically to an extent where they also establish seemingly appropriate anatomical connections. The results are therefore of relevance for the ongoing efforts to introduce stem or progenitor cells as an alternative to primary embryonic tissue for transplantation into the CNS. However, for any future clinical application, much work still remains with development of appropriate protocols for cell derivation and propagation. ACKNOWLEDGMENTS We thank Dr. S. R. Whittemore (University of Louisville School of Medicine, Louisville, KY) for his generous gift of the RN33B parent cell line. We are also grateful to Anna-Karin Olde´ n, Birgit Haraldsson, Ulla Jarl, and Bengt Mattsson for technical assistance. The project was supported by grants from the Swedish MRC (12536, 11817, 13479), the Foundations of Anna-Stina and John Mattson, Segerfalk, Crafoord, and AFA (Arbetsmarknadens Fo¨ rsa¨ kringsaktiebolag). C.L. is a Wenner-Gren Foundation postdoctoral fellow.

9.

10.

11.

12.

13. 14.

15.

16.

17.

18.

REFERENCES 19. 1. 2.

3. 4.

5.

6.

7.

Abercrombie, M. 1946. Estimation of nuclear population from microtome sections. Anat. Rec. 94: 239 –247. Auerbach, J. M., M. V. Eiden, and R. D. Mckay, 2000. Transplanted CNS stem cells form functional synapses in vivo. Eur. J. Neurosci. 12: 1696 –1704. Bjo¨ rklund, A. and O. Lindvall, 2000. Cell replacement therapies for central nervous system disorders. Nat. Neurosci. 3: 537–544. Blomer, U., L. Naldini, T. Kafri, D. Trono, I. M. Verma, and F. H. Gage, 1997. Highly efficient and sustained gene transfer in adult neurons with a lentivirus vector. J. Virol. 71: 6641– 6649. Catapano, L. A., V. L. Sheen, B. R. Leavitt, and J. D. Macklis, 1999. Differentiation of transplanted neural precursors varies regionally in adults striatum. NeuroReport 10: 3971–3977. Chamberlin, N. L., B. Du, S. de Lacalle, and C. B. Saper, 1998. Recombinant adeno-associated virus vector: Use for transgene expression and anterograde tract tracing in the CNS. Brain Res. 793: 169 –175. Costell, M., E. Gustafsson, A. Aszodi, M. Morgelin, W. Bloch, E. Hunziker, K. Addicks, R. Timpl,and R. Fassler, 1999. Perlecan maintains the integrity of cartilage and some basement membranes. J. Cell. Biol. 147: 1109 –1122.

20.

21.

22.

23.

24.

Cunningham, M. G. and R. D. Mckay, 1993. A hypothermic miniaturized stereotaxic instrument for surgery in newborn rats. J. Neurosci. Methods 47: 105–114. Davies, S. J., D. R. Goucher, C. Doller, and J. Silver, 1999. Robust regeneration of adult sensory axons in degenerating white matter of the adult rat spinal cord. J. Neurosci. 19: 5810 –5822. Englund, U., C. Ericson, C. Rosenblad, R. J. Mandel, D. Trono, K. Wictorin, and C. Lundberg, 2000. The use of a recombinant lentiviral vector for ex vivo gene transfer into the rat CNS. NeuroReport 18: 3973–3977. Frederiksen, K. and R. D. Mckay, 1988. Proliferation and differentiation of rat neuroepithelial precursor cells in vivo. J. Neurosci. 8: 1144 –1151. Fricker, R. A., M. K. Carpenter, C. Winkler, C. Greco, M. A. Gates, and A. Bjo¨ rklund, 1999. Site-specific migration and neuronal differentiation of human neural progenitor cells after transplantation in the adult rat brain. J. Neurosci. 19: 5990 – 6005. Gage, F. H. 2000. Mammalian neural stem cells. Science 287: 1433–1438. Gundersen, H. J., P. Bagger, T. F. Bendtsen, S. M. Evans, L. Korbo, N. Marcussen, A. Moller, K. Nielsen, J. R. Nyengaard, B. Pakkenberg, and et al. 1988. The new stereological tools: Disector, fractionator, nucleator and point sampled intercepts and their use in pathological research and diagnosis. Apmis 96: 857– 881. Hockfield, S., and R. D. Mckay, 1985. Identification of major cell classes in the developing mammalian nervous system. J. Neurosci. 5: 3310 –3328. Ikawa, M., S. Yamada, T. Nakanishi, and M. Okabe, 1999. Green Fluorescent Protein (GFP) as a vital marker in mammals. Curr. Top. Dev. Biol. 44: 1–20. Lundberg, C., A. Martinez-Serrano, E. Cattaneo, R. D. Mckay, and A. Bjo¨ rklund, 1997. Survival, integration, and differentiation of neural stem cell lines after transplantation to the adult rat striatum. Exp. Neurol. 145: 342–360. Lundberg, C., C. Winkler, S. R. Whittemore, and A. Bjo¨ rklund, 1996. Conditionally immortalized neural progenitor cells grafted to the striatum exhibit site-specific neuronal differentiation and establish connections with the host globus pallidus. Neurobiol. Dis. 3: 33–50. Naldini, L. 1998. Lentiviruses as gene transfer agents for delivery to non-dividing cells. Curr. Opin. Biotechnol. 9: 457– 463. Okada, A., R. Lansford, J. M. Weimann, S. E. Fraser, and S. K. Mcconnell, 1999. Imaging cells in the developing nervous system with retrovirus expressing modified green fluorescent protein. Exp. Neurol. 156: 394 – 406. Onifer, S. M., S. R. Whittemore, and V. R. Holets, 1993. Variable morphological differentiation of a raphe-derived neuronal cell line following transplantation into the adult rat CNS. Exp. Neurol. 122: 130 –142. Ostenfeld, T., M. A. Caldwell, K. R. Prowse, M. H. Linskens, E. Jauniaux, and C. N. Svendsen, 2000. Human neural precursor cells express low levels of telomerase in vitro and show diminishing cell proliferation with extensive axonal outgrowth following transplantation. Exp. Neurol. 164: 215–226. Peterson, D. A. 1999. The use of fluorescent probes in cell counting procedures. In Quantitative Methods in Neuroscience—A Neuroanatomical Approach (A. M. Janson, S. M. Evans, and A. Mo¨ ller, Eds.). Oxford Univ. Press, Oxford. Renfranz, P. J., M. G. Cunningham, and R. D. Mckay, 1991. Region-specific differentiation of the hippocampal stem cell line HiB5 upon implantation into the developing mammalian brain. Cell 66: 713–729.

RN33B CELLS DIFFERENTIATE INTO PROJECTION NEURONS 25.

26.

27.

28.

29.

30.

31.

Rosser, A. E., P. Tyers, and S. B. Dunnett, 2000. The morphological development of neurons derived from EGF- and FGF-2-driven human CNS precursors depends on their site of integration in the neonatal rat brain. Eur. J. Neurosci. 12: 2405–2413. Shihabuddin, L. S., J. P. Brunschwig, V. R. Holets, M. B. Bunge, and S. R. Whittemore, 1996. Induction of mature neuronal properties in immortalized neuronal precursor cells following grafting into the neonatal CNS. J. Neurocytol. 25: 101–111. Shihabuddin, L. S., J. A. Hertz, V. R. Holets, and S. R. Whittemore, 1995. The adult CNS retains the potential to direct region-specific differentiation of a transplanted neuronal precursor cell line. J. Neurosci. 15: 6666 – 6678. Shihabuddin, L. S., V. R. Holets, and S. R. Whittemore, 1996. Selective hippocampal lesions differentially affect the phenotypic fate of transplanted neuronal precursor cells. Exp. Neurol. 139: 61–72. Snyder, E. Y., D. L. Deitcher, C. Walsh, S. Arnold-Aldea, E. A. Hartwieg, and C. L. Cepko, 1992. Multipotent neural cell lines can engraft and participate in development of mouse cerebellum. Cell 68: 33–51. Snyder, E. Y., C. Yoon, J. D. Flax, and J. D. Macklis, 1997. Multipotent neural precursors can differentiate toward replacement of neurons undergoing targeted apoptotic degeneration in adult mouse neocortex. Proc. Natl. Acad. Sci. USA 94: 11663– 11668. Sorensen, J. C., Castro, A. J., B. Klausen, and J. Zimmer, 1992. Projections from fetal neocortical transplants placed in the

32.

33.

34.

35.

36.

37.

38.

39.

387

frontal neocortex of newborn rats: A Phaseolus vulgaris-leucoagglutinin tracing study. Exp. Brain Res. 92: 299 –309. Stallcup, W. B. and L. Beasley, 1987. Bipotential glial precursor cells of the optic nerve express the NG2 proteoglycan. J. Neurosci. 7: 2737–2744. Suhonen, J. O., D. A. Peterson, J. Ray, and F. H. Gage, 1996. Differentiation of adult hippocampus-derived progenitors into olfactory neurons in vivo. Nature 383: 624 – 627. Svendsen, C. N. and A. G. Smith, 1999. New prospects for human stem-cell therapy in the nervous system. Trends Neurosci. 22: 357–364. Szuchet, S. 1995. The morphology and ultrastructure of oligodendrocytes and their functional implications. In Neuroglia (H. Kettenman and B. R. Ransom, Eds.), pp. 23– 43. Oxford Univ. Press, New York. Whittemore, S. R. and E. Y. Snyder, 1996. Physiological relevance and functional potential of central nervous system-derived cell lines. Mol. Neurobiol. 12: 13–38. Whittemore, S. R. and L. A. White, 1993. Target regulation of neuronal differentiation in a temperature-sensitive cell line derived from medullary raphe. Brain Res. 615: 27– 40. Zufferey, R., J. E. Donello, D. Trono, and T. J. Hope, 1999. Woodchuck hepatitis virus posttranscriptional regulatory element enhances expression of transgenes delivered by retroviral vectors. J. Virol. 73: 2886 –2892. Zufferey, R., D. Nagy, R. J. Mandel, L. Naldini, and D. Trono, 1997. Multiply attenuated lentiviral vector achieves efficient gene delivery in vivo. Nat. Biotechnol. 15: 871– 875.