A Rapid Replacement of Vimentin-Containing Radial Glia by Glial Fibrillary Acidic Protein-Containing Astrocytes in Transplanted Telencephalon

The present study follows the early events in the development of astroglia in rat embryonic (El5) tissue grafted into the cortex of adult rats. Astroglial elements (radial glia and astrocytes) were studied by glial fibrillary acidic protein (GFAP) and vimentin immunohistochemistry on post-transplantation (PT) days 7, 11, 14, 17, and 21. At PT7, GFAP-immunopositive elements were only scarce fibers in the transplants. At PTll, a dense network of long, GFAP-immunopositive fibers enmeshed the entire transplant, and astrocytes were already recognized. The fibers also showed vimentin immunoreactivity. By PT14, astrocytes became the predominant GFAP-labeled elements, although a few long fibers persisted. When compared with in situ development, the grafts showed earlier GFAP-immunoreactivity and earlier appearance of astrocytes, as well as a more rapid transition from the immature to the mature form of the glial system.


INTRODUCTION
Neural transplantation is not only a promising clinical method/8,29/but also a powerful tool for studying neurohistogenesis. As clinical results can be expected only in cases of long-surviving transplants, most data (reviewed in /14,25/) have been published on transplants that had survived at least for one month or longer. Early post-operative events, however, are crucial for the survival of the transplant, and during this period the transplant can used as a model system for certain neurohistogenetic processes. Experimental evidence has suggested that after a successful transplantation, cell maturation follows a schedule that is similar to that found in situ/9,20,25/. The data from such studies correlate with other observations concerning in vitro development/1,43/.
The development and fate of transplanted glia are of interest because grafted immature astroglia can diminish post-lesional neurodegenerative processes /13,32,38,39/ and support axonal regrowth, whereas transplanted oligodendroglial cells are capable of remyelinization/13/. Although several groups have demonstrated the presence of astrocytes in transplants /9,10-12,16,23,37/, the first days of astroglial development after transplantation have not yet been adequately investigated.
The decisive stage of astroglial development in situ is the transition from the immature and vimentin-containing radial glia to the mature and glial fibrillary acidic protein (GFAP)-containing astrocytes (reviewed in /24,41/). The radial glia consists of straight fibers stretching between the ventricular and pial surfaces of the central nervous tissue. This system has an important role in guiding the migrating neuroblasts in the developing brain (reviewed in /35/). After the migration period, however, radial glia in mammals transforms into astrocytes and disappears/28,436,42/. GFAP/21/is the characteristic cytoskeletal protein of mature astrocytes (reviewed in /6/), whereas vimentin, originally described in fibroblasts, is characteristic for immature glia in mammals /18/ and almost completely disappears during glial maturation /17,30,34/. It should be noted that this description is only correct for mammals. Concerning other vertebrates, the reader is referred to other reports /2,26,33,44/, among others. In primates, the transition described above occurs during the fetal period, with the replacement of vimentin by GFAP preceding the transition of the radial glial system into the definitive astrocytes /3,15,28,30,36/. In rodents and carnivora, however, which litter rather immature offspring, the aforementioned processes are concurrent in early postnatal life /7,17,22, 31,34,40,42), thus in such animals only a few or no radial fibers contain GFAP.
Central nervous tissue has transplantation capability only during the early stage of development. In the present study, we followed astroglial development in fetal rat cortical transplants during the post-operative days 7 and 21. The grafts were transplanted on the 15th embryonic day (El5), which is the lower limit of the time window/25/37/during which the cortical tissue is in an optimal stage for transplantation. Such a schedule allowed the immature tissue to develop in the alien environment for as long as possible. At El5, the rat brain consists of radial glia and glioand neuroblasts; vimentin is the glial cytoskeletal protein/18,24,25,37/. Rat brains developing in situ were removed at different postnatal days corresponding to the equivalent post-transplantation days and fixed and processed in the same way as the transplants. The in situ developmental period involved the peak of transition from radial glia into astrocytes. Our results suggest that the transition of radial glia into astrocytes occurs sooner and more rapidly in the transplanted cortex that it does in situ.

METHODS
To obtain dated embryos, we prepared vaginal smears from female albino CFY (Carworth Farm Y, G6d6116, Hungary) rats that had mated with male rats overnight. The day of sperm-positivity was considered as E0. At El5, the embryos were removed from the mother in deep ketamine-xylazine narcosis (20 and 80 mg/kg body weight, respectively, and samples of their neopallium were immediately dissected out. Adult recipient rats (CFY strain) were anesthetized as described above. A piece (-3 x 3 mm) of the left parietal bone was removed from each recipient, the dura was pierced, and a pit was formed in the cortical tissue (under gentle suction using a water-jet pump). The embryonic tissue sample was placed in the pit using a pair of microsurgical forceps, the piece of skull bone that had been removed was then repositioned, and the scalp wound was accurately sutured. The recipient rats received antibiotics both locally and systemically.
Brains were removed and postfixed overnight in the same fixative used for perfusion. The transplants and the surrounding host tissue were dissected out and embedded in agarose. Serial sections were cut using, a Vibratome and washed overnight in phosphate buffer. Parallel sections were processed for either GFAP or vimentin immunohistochemistry.
Before each reagent change, the sections were rinsed with phosphate buffer at pH 7 To form a series of in situ-developing control tissues, we removed the brains of developing rats at postnatal days P0, P4, P7, P10, and P14 and from the donor rats at El5, corresponding to the respective post-transplant days PT7, PT11, PT14, PT17, and PT21 (considering that the gestation period in rats is 22 days). The control brains were fixed, cut into serial coronal sections, and processed in the same way as the transplants.

In situ cortical astrogliogenesis
In the El5 donor brain, only vimentinimmunopositive radial fibers were seen; no radial fiber stained positively for GFAP. The same glial pattern was predominant in the in-situ developing cortex from P0-P7. Astrocytes were not found in either the vimentin-or GFAP-immunopositive sections. It should be noted that in other parts of the brain (corpus callosum, internal capsule, and anterior commissure), GFAP-immunopositive astrocytes were already recognizable during the first postnatal week, indicating that the GFAP immunostaining technique was effective.
We detected the first scarce GFAPimmunopositive elements, both radial fibers and astrocytes, at P 10 in our material from the cortex developing in situ. Until P 14, the number of GFAPpositive astrocytes increased, whereas the number of vimentin-positive radial fibers gradually decreased. The GFAP-immunopositive radial fibers remained scarce, and vimentin-positive astrocytes appeared only exceptionally. As such phenomena have been described before /17,31,34,40/, Fig. shows only one characteristic view of an in situ cortex stained for GFAP immunoreactivity at P10, for comparison with the glial structure of the transplants. VOLUME 6, NO. 1,1997 Astrogliogenesis in cortical transplants In transplants processed at PT7, GFAPimmunoreactive elements were represented by only a few fibers, which were seen predominantly at the vessels in the periphery of the transplant (Fig. 2). PTll transplants were characterized by GFAPimmunopositive long fibers that densely penetrated the transplants (Fig. 3). As astrocytes were also recognizable, the gratt actually showed a mixed population of GFAP-immunopositive elements. Perivascular glia were also seen. The frequency of GFAP-positive elements was characteristically higher in the transplants than in the host cortex. Around the graft, a reactive gliosis developed that seems to have been formed by the host glia. Figure 4 shows that immunostaining for vimentin labeled only a mesh of long fibers, which showed neither a parallel arrangement resembling the original radial glia nor an orientation toward the host/graft border. But nevertheless, we assume that the fibers corresponded to radial glia. The irregular arrangement may be due to a deformation of the telencephalic wall during transplantation and an irregular regrowth of the fibers. In the host cortex, only reactive astrocytes were vimentin-positive around the transplant.
In P 14 transplants, mature GFAP-immunoreactive astrocytes had replaced the fibers. In some transplants (Fig. 5), however, a few long fibers persisted. Although the astrocytes were similar to those usually observed in the adult rat brain in situ, their distribution was denser. Vimentin immunostaining was also seen, but only in a subpopulation of astrocytes, and the staining was less intense than that in the GFAP-labeled elements. Transplants older than 14 days showed no significant difference from the 14-day transplants (not shown). DISCUSSION In the transplant experiments described here, transformation of vimentin-immunopositive radial glia into GFAP-positive astrocytes differed from the in situ development in the following details: 1. The transformation in the transplants occurred earlier than that in situ. Our observations on the in situ development are congruent with published data /15,32,38/, suggesting that the transformation occurs mainly between the second and third postnatal weeks. In our transplants, however, such transformation occurred during the second postoperative week (corresponding to the first in-situ postnatal week). Indeed, the PT 11 transplant (corresponding to P4 in situ) contained not only more astrocytes but also far more GFAP than the in situ P 10 cortex. 2. In transplants, the duration of transformation was shorter (several days) than that in situ (two weeks). Such observations agree with the results of Bignami and Dahl/7/, which showed that a lesion in the neonatal brain evokes the appearance of GFAP without forming a permanent glial scar, as seen after stabbing in adults (reviewed in/5,6/). 3. The presence of many GFAP-positive long fibers suggests that in transplants, the vimentin-GFAP transition starts sooner than the transformation of radial glia into astrocytes. Unlike the in-situ process, the vimentinimmunopositivity in transplants does not decrease proportionally with the increase in GFAP immunopositivity. The abundance of astroglia in the transplants relative to the number in the in-situ cortex has been noted by several authors /9-11,16,23/. This phenomenon could be due to a greater survival ratio VOLUME 6, NO. 1, 1997 in astroglia than in neurons, or it may correspond to a reactive gliosis after the lesion, by which the grafted tissue was separated from the donor brain /10,11,16/. That immature astroglia do not form a permanent glial scar has been suggested; in fact, the astroglia diminish the effect of lesions in the adult brain/4,7,32,38,40/. Our observation that the reactive gliosis around the transplant appeared to be generated by the host rather than by the grafted glia (for similar observations) agrees with this opinion. The presence of vimentin immunopositivity in reactive gliosis (shown in Figs. 4A and 4B) and in the astrocytes of transplants has been previously reported/19,27,37/. In the latter case, the presence of vimentin can be either a sign of a reactive characteristic or a feature that has persisted since the immature stage of astrocytes in the transplant.
In conclusion, aPter transplantation, astroglia undergo an accelerated histogenesis that is likely to be provoked by the effects of lesions in both the donor and the host tissues