Patterns of angiogenesis in neural transplant models: II. Fetal neocortical transplants

Allografts of CNS tissue possess a blood-brain barrier: III. Neuropathological, methodological, and immunological considerations

Development of a blood-brain barrier (BBB) within mammalian CNS grafts, placed either intracerebrally or peripherally, has been controversial. Published data from this laboratory have emphasized the presence or the absence of a BBB within solid mammalian tissue or cell suspension grafts is determined intrinsically by the graft and not by the surrounding host parenchyma (e.g., brain, kidney, testis, etc.). Nevertheless, correctly interpreting whether or not a BBB exists within brain grafts is manifested by methodologies employed to answer the question and by ensuing neuropathological and immunological consequences of intracerebral grafting. The present study addresses these issues and suggests misinterpretation for the absence of a BBB in brain grafts can be attributed to: (1) rupture of interendothelial tight junctional complexes in vessels of CNS grafts fixed by perfusion of the host; (2) damage to host vessels and BBB during the intracerebral grafting procedure; (3) graft placement in proximity to inherently permeable vessels (e.g., CNS sites lying outside the BBB) supplying the subarachnoid space/pial surface and circumventricular organs such as the median eminence, area postrema, and choroid plexus; and (4) graft rejection associated with antigen presenting cells and the host immune response. The latter is prevalent in xenogeneic grafts and exists in allogeneic grafts with donor-host mismatch in the major and/or minor histocompatibility complex. CNS grafts between non-immunosuppressed outbred donor and host rats of the same strain (e.g., Sprague Dawley or Wistar rats) can be rejected by the host; these grafts exhibit populations of immunohistochemically identifiable major histocompatibility complex class I+ and class EE+ cells (microglia, macrophages, etc.) and CD4+ T-helper and CD8+ T-cytotoxic lymphocytes. PC12 cell suspension grafts placed within the CNS of non-immunosuppressed Sprague Dawley rats are rejected similarly. Donor cells from solid CNS grafts placed intracerebrally and stained immunohistochemically for donor major histocompatibility complex (MHC) class I expression are identified within the host spleen and lymph nodes; these donor MHC expressing cells may initiate the host immune response subsequent to the cells entering the general circulation through host cerebral vessels damaged during graft placement. Rapid healing of damaged cerebral vessels is stimulated with exogenously applied basic fibroblast growth factor, which may prove helpful in reducing the potential entry of donor cells to the host circulation. These results have implication clinically for the intracerebral grafting of human fetal CNS cell suspensions.

Human fetal neocortical tissue grafted to rat brain cavities survives, leads to reciprocal nerve fiber growth, and accumulates host IgG

The human-to-rat xenograft approach offers possibilities to study aspects of primate cortex development and function without monkeys. Human fetal cortical tissue was grafted to prepared cortical cavities of immunosuppressed host rats. Fetal tissue fragments were collected after routine low-pressure vacuum aspiration abortions performed in the first trimester of gestation. Human derived neurons and human nerve fiber outgrowth were visualized by immunohistochemistry with antibodies against human neurofilament protein 70 kD (hNFP70). Ingrowth from rat host striatum or cortex into the grafts was analyzed by immunohistochemistry with antibodies against tyrosine hydroxylase. Astrocytes were evaluated by immunohistochemistry with antibodies against glial fibrillary acidic protein. The grafts grew into different sizes (1-10 mm in diameter) and contained large numbers of hNFP70-positive nerve fibers. All grafts gave rise to outgrowth of hNFP70-positive fibers into the host with partly a cortical layering; layers III and IV received a majority of the human fibers. In several cases, the graft-derived nerve fibers entered the host brain at restricted areas, while there was no crossing over of nerve fibers at the rest of the graft-host interface. Tyrosine hydroxylase-positive fibers were usually not abundant in the grafts. Interestingly, cases of massive ingrowth occurred from host striatum into the graft in a pattern suggesting "permissive sites" at the graft-host interface in the same way as outgrowth from graft to host was found. Additionally, tyrosine hydroxylase-immunoreactive fibers from host cortex were found to grow into the transplant. Glial fibrillary acidic protein immunoreactivity was increased at the interfaces between graft and host cortex or host striatum. Immunohistochemistry using antibodies against rat IgG indicated the presence of rat IgG within the grafts, and in bordering areas of host brain, possibly indicating a defective graft-host barrier. Taken together, these results show that human cortical tissue pieces grafted to cortical cavities of immunosuppressed rats survive grafting and develop, and that reciprocal nerve fiber growth between grafts and hosts occur. Human cortical neurons can grow into the rat host brain in a pattern which is partly determined by host cortical architecture.

Anatomical and functional consequences of induced rejection of intracranial retinal transplants

Retinae from embryonic rats transplanted over the midbrain of newborn host rats establish connections with visual centres of the host brain, which mediate a pupilloconstrictor response in the host eye when the transplant is stimulated by light. The changes in the size of the host pupil can be measured accurately with a pupillometry system. We have taken advantage of the additional observation that while grafts between rat strains, as between Long Evans and Sprague-Dawley strains, may survive indefinitely, they can be induced to reject by skin grafting from the strain providing the donor retinal tissue. Combining pupillometry with skin grafting provides a useful way of examining correlated anatomical and behavioural changes associated with graft rejection from its earliest stage to the point of overt destruction. Even within three days of skin grafting, the amplitude and speed of constriction as well as the response latency all showed significant enhancement from normal, and this was sustained for a further week or more. Response deterioration followed during the second week post-skin grafting, but the exact timing varied considerably among animals. Anatomical observations of the process of retinal rejection showed the first invasion of lymphocytes to occur between days 5 and 7 and total degeneration of the retinal transplant and its projections to occur by two to three weeks post-skin grafting. The lymphocytic infiltration was preceded by upregulation of microglia, which expressed both class I and II major histocompatibility antigens and by activation of astrocytes identified by their expression of glial fibrillary acidic protein. Within the target region of retinal transplant axons, major histocompatibility antigen expression and astrocytic responses preceded degeneration of transplant derived axons (demonstrated by the Fink-Heimer stain) and there was no evidence for any lymphocytic lymphocytic infiltration during transplant rejection. These observations show that the earliest stages of microglial activation are accompanied by an enhancement of response parameters, but that the functional failure finally occurs only at an advanced stage of graft destruction. The absence of lymphocytic infiltration into areas receiving terminals from axons of transplant origin, even though these contain significant numbers of reactive microglia, suggests that the terminal axonal processes are not a primary target for the immune response.

Functional development of fetal suprachiasmatic nucleus grafts in suprachiasmatic nucleus-lesioned rats

Recovery of circadian drinking rhythms in suprachiasmatic nucleus (SCN)-lesioned rats after fetal SCN grafting was related to the immunocytochemical appearance and fiber outgrowth of vasopressin (VP)-, vasoactive intestinal polypeptide (VIP)-, and somatostatin (SOM)-containing neurons in the implants. At 4 weeks postgrafting, the first recovered animal was found. After longer survival times, 38% of the animals showed recovery. Immunocytochemical evaluation indicated that full maturation of the SCN grafts was not reached until 4 weeks postgrafting. Grafted VP and VIP cells were always located together, whereas SOM cells were clustered nearby but separate. Neuropeptide Y fibers were observed with an increasing fiber density between 2 and 5 weeks posttransplantation and were clustered particularly at the level of the SOM cells. In all rhythm-recovered animals transplants of VP and VIP fibers had grown laterally into the hypothalamus. A few nonrecovered animals also showed ingrowth of such fibers, though more caudally to the lesioned SCN. Many of the nonrecovered rats showed similar stainings but without these efferent outgrowth to the host. We conclude that neither a humoral factor nor the presence of VP and VIP efferents in the host brain alone are enough for the restoration of circadian drinking rhythms.

Vascularization of fetal neocortical grafts implanted in brain infarcts in spontaneously hypertensive rats

The vascularization of neural grafts in ischemic brain was studied in spontaneously hypertensive rats grafted with a suspension of fetal neocortical tissue into the infarcted area five to six days after ligation of the middle cerebral artery. The brain vasculature was examined by scanning electron microscopy of corrosion vascular casts and the cortical microvasculature was stereologically quantified in light microscopy three months after the occlusion. Patent anastomoses were present between the middle cerebral artery distal to occlusion and the proximal part, as well as to the anterior and posterior cerebral arteries, in both grafted and non-grafted rats. A vascular plexus covering the infarct cavities and the grafts contained leptomeningeal vessels intermingled with a thin capillary network which is not normally found on the brain surface. The graft vessels were derived from this vascular plexus. The regular pattern of arterioles and venules penetrating from the cortical surface in normal neocortex was absent in the grafts but the capillary morphology was similar in both types of tissue. The grafts had a lower capillary density than normal tissue and lacked the laminar distribution of capillaries characteristic of normal neocortex. The results demonstrate the plasticity of the vascular system where remodeling of the vascular tree after an ischemic insult provides suitable conditions for the vascularization of neocortical grafts.

Immunocytochemical expression of the endothelial barrier antigen (EBA) during brain angiogenesis

The antibody to the endothelial barrier antigen (anti-EBA) is localized to the luminal plasma membrane of endothelia that have a blood-brain barrier (BBB) but not to other vessels, for instance those in the circumventricular organs, which lack barrier function. We have examined EBA expression in the rat in certain tissues and in brain microvessels in models of brain angiogenesis such as development, wound healing and neural transplantation. All brain microvessels including pial ones stained for anti-EBA whereas those of the dura, median eminence and choroid plexus did not. Vessels of the iris which are characterized by tight junctions and barrier function expressed EBA strongly. Embryonic day 18 brain did not stain at all for anti-EBA although vessels were readily localized with anti-laminin. Following stab wounds to mature brain, directly injured and adjacent microvessels lacked EBA expression for a period of approximately 2 weeks which is a similar time frame of BBB breakdown. Following this period, EBA expression gradually returned to a normal pattern by 3-4 weeks. Likewise, in intraparenchymal transplants of fetal neocortex EBA expression was not observed for 2 weeks and while at later times transplant vessels expressed EBA whereas some interface vessels associated with inflammatory cells did not. Permeable choroid plexus vessels vascularizing intraventricular transplants did not stain for anti-EBA at any time period and neither did vessels in adrenal medulla transplants. The present study shows that while EBA expression is a postnatal event unlike the development of a barrier to serum protein, its expression may be lost or delayed in injured vessels or ones associated with inflammatory cells or reactive astrocytes.

Explorations of otic transplantation

Embryonic rat inner ears were transplanted to the anterior chamber of the eyes of adult rats. While considerable development was evident, the structures present were limited to the vestibular division. We hypothesized that this selective survival could be due to the rate of vascularization. To test the effects of graft vascularization we made transplants in which the internal structures were exposed by removing the apex and base of the developing cochlea. The transplants were rapidly vascularized by the iris. Many of the soft labyrinthine structures of the cochlea from 1-day-old donors showed considerable development, including the spiral limbus, basilar membrane, and organ of Corti. To test the possibility that the cochlea requires inductive or trophic support beyond Embryonic Day 15 (E15), we cotransplanted the embryonic inner ear with developing brain stem. In these transplants, we observed improved development of the cochlea, with spiral ganglion cells and an organ of Corti possessing hair cells, Deiter's cells, and pillar cells. To further address the effect of developing CNS tissue on the development of grafted inner ear, we transplanted E15 inner ears to either the cortex or the brain stem of neonatal rats. In these experiments we have seen evidence of both vestibular and cochlear sensory surfaces. In the cochlea, an organ of Corti-like structure can be seen. The possibility of neural connections with the host brain has yet to be investigated.

Long-term survival of mouse corpus callosum grafts in neonatal rat recipients, and the effect of host sensitization

Previous studies have suggested that the incidence of spontaneous rejection among immunogenetically mismatched neural transplants in neonatal recipients varies significantly depending on the cellular composition of the graft material. For example, neuron-rich grafts of embryonic mouse retina generally survive for extended periods without showing signs of rejection after implantation into neonatal rats, whereas cortical xenografts, which contain abundant glial and endothelial cells as well as neurons, typically undergo rejection 4-6 weeks after implantation. To determine whether the presence of donor glia is responsible for this high incidence of spontaneous rejection, we examined the fate of a non-neuronal graft material composed predominantly of xenogeneic glial cells (post-natal day 3, PD3, CD-1 mouse corpus callosum) implanted into the mesencephalon of PD1 Sprague-Dawley rats. The distribution and survival of donor astrocytes were assessed using a monoclonal antibody specific for a mouse astrocyte surface antigen, M2. Thirteen of 16 animals sacrificed within 2 months of implantation had detectable transplants. In these animals, M2-positive cells frequently migrated well away from body of the graft, clustering in large numbers in several characteristic regions of the host brain. Unlike cortical grafts of similar age, the vast majority (93%) of callosal transplants showed no histological signs of rejection or major histocompatibility complex antigen expression in and around the transplant-derived cells. As previously noted in the neonatal retinal transplant paradigm, however, well-integrated 1-month-old corpus callosum grafts could be induced to reject by appropriate sensitization of the host immune system, implying that the host was not immunologically tolerant to the foreign neural graft. With longer survival times in unsensitized hosts, a progressively smaller percentage of animals had detectable donor astrocytes (5 of 10 animals at 3 months postimplantation and 4 of 16 animals at 4 months); in those 9 animals with surviving grafts, only small numbers of M2-positive cells were seen within the graft bed and surrounding host brain. However, only 2 of the 26 "long-term" animals showed evidence of graft rejection. These results indicate that mouse astrocytes show characteristic patterns of migration into the host brain when implanted into neonatal rats; however, these xenogeneic cells have a limited duration of survival. The infrequency with which even subtle signs of spontaneous rejection were detected in animals that had received corpus callosum xenografts suggests that an immune-mediated process is unlikely to be responsible for the time-dependent elimination of the donor astrocytes.(ABSTRACT TRUNCATED AT 400 WORDS)

Employment of fibroblasts for gene transfer: applications for grafting into the central nervous system

Genetic modification of primary skin fibroblasts offers a new approach to the focal delivery of deficient transmitter-specific enzymes (e.g., TH) or trophic substances (e.g., NGF) to the damaged or diseased CNS. Although fibroblasts are unable to provide anatomical corrections to defective neural connectivity, they can serve as biological pumps for the enzymes and growth factors in vivo. The capability of genetically engineered cells to ameliorate disease phenotypes in animal models of CNS disorders may ultimately results in the restoration of function. At this time, primary skin fibroblasts appear to be a convenient cellular population for the application of gene transfer and intracerebral grafting for the animal model of Parkinson's disease. It is now important for future investigations to provide data concerning the long-term stable expression of the transgene product (e.g., TH) following intracerebral implantation, as well as determining optimal conditions for the survival of primary cells grafted into the nervous system.

Repair of the blood-brain barrier following implantation of polymer capsules

Past studies of polymer-encapsulated cell lines implanted in the brain indicated their usefulness for transmitter replacement therapy in animal models. Such grafts may have potentially important clinical applications, but their placement into neural parenchyma may cause a traumatic injury resulting in a leaky blood-brain barrier around the implant. This study investigated whether or not injury repair and reformation of the barrier takes place near a polymer capsule implanted in the brain of Sprague-Dawley rats. The two methods used for detection of a leaky barrier were immunocytochemical localization of extravasated serum albumin and circulating Evans blue that binds to serum albumin. Immunocytochemical staining for glial filament protein provided a measure for evaluating injury associated gliosis. Polymer capsules implanted for 10, 16 and 18 days were surrounded by microvessels that leaked detectable quantities of serum albumin into interstitial spaces and, by secondary uptake, into some nearby neurons and reactive astrocytes. Reactive astroglia were observed within the outer regions of the capsule wall and in the near vicinity of the implant after these early survival times. In contrast, at post-implantation times of 46 and 54 days, serum albumin was no longer detected in the neural parenchyma near the macrocapsules and only few reactive astrocytes remained. These findings show that polymer capsules implanted within the cerebrum permit (a) reformation of the blood-brain barrier and (b) occurrence of repair processes that lead to minimal deposition of reactive astroglia near the implanted polymer capsule.

Intracerebral grafting of cultured autologous skin fibroblasts into the rat striatum: an assessment of graft size and ultrastructure

To identify a suitable donor cell population for gene therapy applications to the central nervous system, primary fibroblasts isolated from skin biopsies and maintained in culture are employed as autologous cells for intracerebral grafting within the adult rat striatum. Results from the present investigation reveal that cultured primary skin fibroblasts cease to proliferate once they reach confluence; these cells are thus contact inhibited in vitro. Following implantation within the striatum, the volume of the primary fibroblast grafts, stained immunohistochemically for fibronectin, does not differ significantly at 3 and 8 weeks. The graft size is dependent on the density of the cell suspension, but not dependent on either the number of passages the cells are taken through in culture prior to grafting or on the postoperative survival period. Ultrastructural evidence reveals that at 8 weeks the grafts are composed primarily of collagen and fibroblasts with rough endoplasmic reticulum and vesicles. Reactive astrocytic processes and phagocytic cells are also present in the grafts. The grafts are extensively vascularized with capillaries composed of nonfenestrated endothelium; intercellular junctions are evident at sites of apposition between endothelial cells. It is concluded that primary skin fibroblasts are able to survive for at least 8 weeks following intracerebral implantation and continue to synthesize collagen and fibronectin in vivo. Also, the grafts maintain a constant volume between 3 and 8 weeks, thereby indicating that primary skin fibroblasts do not produce tumors. Finally, dynamic host-to-graft interactions--including phagocytic migration, astrocytic hypertrophy and infiltration within the grafts, and angiogenesis--are features that constitute the structural integration of primary skin fibroblasts grafted within the adult rat central nervous system.

Allografts of CNS tissue possess a blood-brain barrier. II. Angiogenesis in solid tissue and cell suspension grafts

Angiogenesis and patency of blood vessels were analyzed qualitatively in solid CNS and peripheral tissue syngeneic, allogeneic, and xenogeneic grafts and in individual cell suspension grafts of astrocytes, fibroblasts, PC12, and three additional tumor cell lines placed intracerebrally in adult host mice. Postgrafting survival times were 1 day through 4 weeks. The patency of graft vessels was determined in sections from immersion-fixed tissues incubated to reveal the endogenous peroxidase activity of host red cells trapped within the lumen of blood vessels. Additionally, horseradish peroxidase (HRP) was administered intravenously to live hosts; HRP labels host brain and graft vessels on the luminal surface and reveals the presence or absence of a blood-brain barrier (BBB) within the grafts. The origins of blood vessels supplying solid tissue xenografts were identified immunohistochemically with primary antibodies against host (athymic AKR mice) and donor (fetal Lewis rats) major histocompatibility complex (MHC) class I. Blood vessels supplying solid CNS grafts at 1-7 days post-transplantation were identified ultrastructurally and possessed interendothelial tight junctional complexes; however, they were not perfused with either host blood or blood-borne HRP prior to 8 days. Graft vessels at 10 days were outlined consistently by peroxidase-positive red cells in immersion-fixed material and labeled with blood-borne HRP. These vessels provided a BBB to the circulating HRP and exhibited interendothelial tight junctions. Evidence of angiogenesis within solid anterior pituitary grafts and the variety of cell suspension grafts was obtained prior to 3 days post-transplantation in immersion-fixed preparations; the vessels, with the notable exception of those supplying astrocyte cell suspensions, failed to present a BBB to blood-borne peroxidase. Endothelia in the solid pituitary allografts and the PC12 cell grafts were highly fenestrated and exhibited open interendothelial junctions; those in the tumor and fibroblast cell grafts, for the most part, appeared nonfenestrated, and many possessed open interendothelial junctional complexes. Immunostaining for host and donor MHC class I revealed that donor blood vessels predominate over host vessels in CNS xenografts and supply pituitary xenografts exclusively; in both preparations, donor vessels were not identified within the host CNS. Because cell suspension grafts were derived from endothelia-free preparations grown in culture, blood vessels supplying these grafts were necessarily of host CNS origin and manifested a morphological transformation from a BBB to a non-BBB endothelium. The data suggest that angiogenesis in solid CNS grafts placed into the adult host CNS, compared to similarly placed solid peripheral tissue/cell suspension grafts, is not rapid.(ABSTRACT TRUNCATED AT 400 WORDS)

Permeability to blood-borne protein and 3HGABA in CNS tissue grafts. I. Intraventricular grafts

In the present study, solid grafts of fetal CNS tissue from the rat neocortex, cerebellum, or ventral mesencephalon were placed into the lateral, III or IV ventricles of young adult hosts. Survival periods ranged from 2 days to 20 months. To study the permeability to protein and potential changes in the blood-brain barrier (BBB), macromolecules such as HRP, HRP-human serum albumin, and HRP-human IgG were administered intravascularly and circulated for periods between 3 minutes and 1 hour. Younger grafts were completely filled with the protein, even at 2 days, when the graft vasculature already contained host macrophages, whereas all older grafts showed variability in permeation with protein ingress initiating at the graft-host interface and subsequently diffusing through the extracellular spaces. Permeation was from several sources: permeable vessels of the circumventricular organs and the choroid plexus which grew into the grafts, the perivascular spaces surrounding these vessels, or from the normally impermeable vessels of the pia mater, which, because of their engulfment by the graft and subsequent angiogenesis, may have been rendered permanently leaky. Invading vessels were often "cuffed" by lymphocytic cells. Many grafts were only partially filled by the glycoprotein conjugates; ventral mesencephalic grafts allowed the least diffusion even when vascularized by choroidal vessels. Fenestrated vessels were not directly observed even though petechial leaks were evident and vessels indigenous to the CNS grafts retained BBB properties. To determine endogenous protein exudation, noninjected animals were immunocytochemically examined for rat serum albumin (RSA). The distribution of RSA mimicked that of the injected proteins at interface regions, although in most instances the entire graft was filled by a light, diffuse labeling suggesting a steady-state protein leakage over the life of the graft. When HRP was delivered intraventricularly, the intraventricular grafts were nearly filled with reaction product by 20 minutes. The depth of penetration in the grafts from the CSF interface was generally threefold greater than in normal brain. The increase in permeation suggests that solutes may flow through these grafts (out of or into the CSF) at an increased rate. Lastly the neurotransmitter tritiated gamma-aminobutyric acid (3HGABA) which does not cross the BBB was vascularly administered to hosts bearing neocortical grafts. These experiments not only confirmed the permeability in these grafts but showed that the blood-borne amino acid could be directly sequestered by grafted neurons or glia.(ABSTRACT TRUNCATED AT 400 WORDS)

Oxygen transport in intraspinal fetal grafts: graft-host relations

In the present experiments, we determined tissue oxygen tension (PtO2) levels within the injured spinal cords of adult rats following transplantation of fetal spinal cord tissue. Partial resection cavities were made at L1-L2 levels after which whole pieces of 14-day donor tissue were placed into the cavities. Analysis of recordings obtained from graft tissues at 1 and 2 months after transplantation revealed low PtO2 values in many cases. Even the more extensively developed transplants at 3 months continued to show mean PtO2 levels lower than those taken from normal, mature spinal cord tissue. Measurements from host tissue, adjacent to a lesion in which no graft was introduced, showed normal or elevated PtO2 levels. In contrast, where extensive host and graft integration had occurred, the PtO2 levels of adjacent host resembled those obtained within the transplants. On the other hand, in cases of poor host-graft integration, characterized by either cellular or fibrotic graft-host interfaces or large cysts, the PtO2 tensions exceeded normal levels. Therefore, the present results show that when fetal grafts are placed acutely into an aspiration cavity within the adult spinal cord, the transplants quickly establish an oxygen microenvironment resembling that found during normal fetal development. Oxygen transport is therefore a regulated variable in the graft neuropil as it is in the normally developing spinal cord. Furthermore, in the presence of closely approximated fetal transplants, adjacent host tissue assumes tissue oxygen levels that mimic those in the graft. This "inductive" effect gradually diminishes as development proceeds and may be the hallmark of successful graft-host integration.

Cerebrovascular responsiveness to hypercapnia in intracerebral tissue transplants

Cerebral blood flow was measured using [14C]iodoantipyrine quantitative autoradiography in rats which had previously undergone unilateral ibotenate-induced nucleus basalis lesion followed by intracortical implantation of foetal basal forebrain cell suspensions. Transplants had no effect upon host cortical blood flow, although within the transplant itself, blood flow was significantly lower than the contralateral site. Both the transplant and host cortex exhibited a similar degree of hyperaemia in response to hypercapnia.

Brain angiogenesis: variations in vascular basement membrane glycoprotein immunoreactivity

Changes in the distribution and quantity of laminin and fibronectin within the basement membranes of developing or regenerating CNS blood vessels were investigated using two immunocytochemical techniques. Three models of angiogenesis were studied: normal pre- and postnatal development, wound healing, and vascularization of fetal neocortical transplants placed in the adult rat brain. Although all brain vessels were stained in enzymatically pretreated immunoreacted paraffin sections, those associated with wound and transplant sites were the most intensely reactive with both antisera during the first postoperative week. When 40-microns vibratome sections of normal adult brains were immunoprocessed, only the meninges and vessels of the circumventricular organs were stained. The remainder of the brain vasculature was immunoreactive only if sections were enzymatically treated prior to immunoprocessing. In contrast, the nascent vasculature in developing brain and the regenerating vessels at wound and transplant sites were reactive to both antisera without enzymatic pretreatment of the sections. This immunoreactivity decreased by 11 days postnatal in normal animals and 4 weeks postoperative in experimental animals, coinciding with the period of astrocytic contact and complete vascular basement membrane formation in both cases. The variations in staining pattern and intensity may be reflections of differences in the quantity of laminin and fibronectin within the basement membranes of proliferating and/or non-blood-brain barrier vasculature. However, the results of the different experimental protocols suggest that immature vascular basement membranes may have a molecular configuration that does not require an enzymatic unmasking step to react with the antisera. Alternatively, the looseness of the surrounding neuropil inherent in developing and injured CNS could allow the antisera greater access to basement membrane antigens.

Blood-brain barrier permeability is not altered by allograft or xenograft fetal neural cell suspension grafts

Alterations in blood-brain barrier (BBB) function after brain grafting seem dependent on the donor phenotype and possibly on the grafting technique. Intracerebral blood grafts of nonneural tissue permanently disrupt the host BBB, while fetal neural block grafts probably do not. Cell suspensions, an alternative technique in brain grafting, disrupt the extracellular matrix of the graft. Fetal cell suspension allografts appear to form a functional BBB. We confirm and extend this finding to include fetal neural xenografts. Allograft and xenograft fetal neural cell suspensions were intracerebrally injected, and the BBB was examined using intravenous horseradish peroxidase (HRP). Neither graft type showed disruption of the BBB at the graft site from 2 weeks to more than 6 months after grafting. Vascular supply was prominent at all time points. Xenograft survival was improved with cyclosporine, yet cyclosporine did not affect BBB permeability. Cyclosporine did not interfere with repair of the BBB after simple brain trauma was induced by a control injection of saline. We conclude that fetal allograft and xenograft neural cell suspensions rapidly form and maintain a BBB impermeable to intravenous HRP.

Normal cerebrovascular regulatory mechanisms are present in intracerebral neuronal transplants

Local cerebral blood flow and local cerebral glucose utilization were measured using quantitative autoradiography in parallel groups of rats (n = 5-7) which 12-15 weeks previously had undergone limited unilateral ibotenate-induced lesion of the nucleus basalis magnocellularis, followed by implantation into ipsilateral neocortex of primordial basal forebrain cell suspensions. Surviving transplants were visualized by acetylcholinesterase histochemistry. Neither lesion alone nor the presence of a transplant produced significant side-to-side differences in either blood flow or glucose use in any of the 20 brain areas measured. Glucose use within the transplant was independent of the site of implantation. When sited in neocortex, glucose use in the transplant (66 +/- 4 mumol/100 g per min) was significantly lower than in the corresponding contralateral site (113 +/- 3 mumol/100 g per min), whereas when sited in subcortical white matter, glucose use (53 +/- 3 mumol/100 g per min) was significantly higher than in the contralateral side (29 +/- 4 mumol/100 g per min). In the host brain as a whole, the ratio of blood flow to glucose use ipsilateral to the transplant (m = 1.27, r = 0.88) was not significantly different from that of the contralateral side (m = 1.30, r = 0.94). This relationship was also observed within the transplanted tissue itself despite the fact that alkaline phosphatase histochemistry revealed a relative hypervascularization associated with the implantation site.(ABSTRACT TRUNCATED AT 250 WORDS)

Patterns of immune rejection of mouse neocortex transplanted into neonatal rat brain, and effects of host immunosuppression

We studied the histological and immunological characteristics of graft rejection in the rodent central nervous system (CNS) using embryonic mouse neocortex transplanted into the CNS of neonatal rats. Grafts from animals aged 8-145 days (n = 210) were examined using standard histological techniques for demonstrating cell morphology and fiber projections. Immunohistochemical techniques were used to identify graft projections into the host CNS. The incidence of graft rejection was 18% for animals between 18 and 30 days of age, but increased abruptly to 73% for animals older than 30 days. No graft rejection was seen in animals younger than 18 days. In a smaller group of xenograft recipient rats sacrificed at specific time points before and after one month of age, detailed immunohistochemical studies were performed to correlate the histological appearance of the graft with the level of major histocompatibility complex (MHC) class I and II immunoreactivity, and microglial, astrocytic and lymphocytic staining within the graft and host brain. Evidence of mild rejection as manifested by the appearance of scattered lymphocytes within the graft coincided with the development of Class I and II immunoreactivity within the graft and at the graft-host interface, which was demonstrated in some animals as early as 24 days. At 29 days of age, rejecting grafts showed diffuse MHC expression within the graft and at the graft-host interface; in contrast, unrejected grafts failed to show MHC immunoreactivity. Thirty-four day-old grafts often showed severe rejection with perivascular lymphocytic cuffing within the graft and in host parenchyma remote from the graft associated with increased MHC immunoreactivity within the host brain. In grafts older than 34 days there was frequently a violent rejection reaction with disruption of the cytoarchitecture of the graft and surrounding host tissues, and widespread MHC antigen expression. Immunosuppression with cyclosporin A was effective in avoiding rejection. The high incidence of rejection with neocortical xenografts is in striking contrast to the much lower incidence seen with retinal xenografts. This suggests that there are immunological features unique to neocortex which incite host recognition and rejection.

Short-term post-grafting morphological alterations of glia from an adult brain transplant

Fragments of corpus callosum from adult rabbit have been implanted into the brain of newborn mice. Previous studies had shown that under such conditions transplant-derived astroglial cells differentiate in the host and survive for at least 2 months. The present study was devised to clarify the fate of the differentiated astrocytes present in the adult transplant by using combined ultrastructural and immunohistochemical approaches. These mature cells are shown to degenerate and die within 2 days after the implantation. Therefore, we suggest that stem cells present in adult tissue would account for the surviving population of transplant-derived glial cells.

Spinal cord repair: is tissue oxygenation an important variable?

We have demonstrated the reliability and feasibility of making PtO2 recordings from graft and host tissue in the injured spinal cord. The data suggest that the oxygen microenvironment of developing graft and host spinal tissue is clearly different from that found in normal spinal tissue or in transplants that have not survived or integrated well. These same constraints seem to apply to cavitation in developing grafts and poorly developed graft/host interfaces. The similarity between these findings and those from previous studies in other fetal vertebrates suggests that oxygen tensions in the spinal cord probably reflect the developmental status of the regenerating tissue. Our future studies will seek to define the relationship between anatomical development of transplant tissue and these functional (PtO2, microvascular development and tissue metabolism) indicators of graft development. These investigations should also provide a background for those later studies which seek to establish the mechanisms by which these relationships come about i.e. oxygen consumption of host/transplant tissue, blood flow to transplants, studies of glycolytic metabolism (2-DG autoradiography), etc. In this way, we can begin to understand the role of tissue metabolism in graft-mediated repair.

Mosaic reconstruction of blood vessels in mouse neocortical tissue transplanted into the third ventricle of rat brain

Newborn mouse neocortical tissue was transplanted into the third ventricle of rat brain and the reconstruction and the origin of the blood vessels were investigated by using a monoclonal antibody against mouse endothelial surface antigen-1 (MESA-1). It was clearly demonstrated that some of the blood vessels in the graft originated in the donor mouse neocortical tissue. An India ink perfusion experiment revealed that the blood was supplied to the MESA-1-positive blood vessels. Furthermore, electron microscopic immunohistochemical studies demonstrated the existence of a mosaic reconstruction of blood vessels which consisted of mouse- and rat-derived vascular endothelial cells. It was concluded that the blood vessels originating in the donor tissue and those originating in the host tissue inoculate with each other in the grafted tissue.

Allografts of CNS tissue possess a blood-brain barrier. I. Grafts of medial preoptic area in hypogonadal mice

This study represents the first part of a three-part investigation of blood vessels supplying CNS tissue transplanted within the brains of adult mammalian hosts. The results emphasize that blood vessels in solid CNS grafts contribute a blood-brain barrier to that of the host. Neurosecretory cells in basal forebrain grafts placed intraventricularly on the dorsal surface of the host median eminence, a neurosecretory site containing fenestrated blood vessels, do not stimulate similar blood vessels to inhabit the transplanted tissue. Solid grafts of the medial preoptic area containing neurons that synthesize and secrete gonadotropic hormone-releasing hormone (GnRH) were obtained from AKR mice and placed into the third cerebral ventricle of hypogonadal (HPG) mice genetically incapable of synthesizing GnRH. GnRH neurons in the allografts were confirmed immunohistochemically. Blood vessels supplying the host median eminence and the allograft at 10 days to 3 months post-transplantation were analyzed with peroxidase cytochemistry applied in three ways: to HPG mice injected systemically with native horseradish peroxidase; to HPG mice infused into the aorta with peroxidase subsequent to perfusion fixation; and to HPG mice brains fixed by immersion and incubated for endogenous peroxidase activity in red cells retained within blood vessels. The median eminence of the HPG mouse was innervated by GnRH neurons residing within the graft, and blood vessels traversing the median eminence-allograft interface were seen rarely. The allografts contained no fenestrated endothelia, and no extravasations of blood-borne HRP were related directly to leaky blood vessels supplying the grafted tissue. Endothelial cells throughout the CNS grafts were similar morphologically to blood-brain barrier endothelia; they were nonfenestrated, exhibited interendothelial tight junctional complexes and an endomembrane system of organelles, and they endocytosed blood-borne HRP that eventually was sequestered within dense body lysosomes. The results support the belief that blood vessels supplying CNS tissue transplanted to a host brain manifest endothelial characteristics identical to those of the tissue in normal life and to those of the host CNS.

Vascularization of fetal cell suspension grafts in the excitotoxically lesioned adult rat thalamus

Several studies have considered the establishment of vascularization in intracerebral solid transplants of neural tissue. The widely supported interpretation of the results is that the vascular network of the solid grafts is already present before implantation into the host brain. The situation is different when dissociated fetal tissue is transplanted as a cell suspension because in these conditions the fetal vascular network is disrupted. The present study has, therefore, been undertaken to follow the angiogenesis in a transplant of dissociated fetal cells implanted into the excitotoxically neuron-depleted thalamus. The vascular network is compared to that observed in the intact and in the lesioned thalamus both in terms of morphology of the capillaries and of the function of the blood-brain barrier (BBB). In the transplant, capillaries, stained by Indian ink, are very few in number and have very fine calibers during the first 20 days after grafting. Some structures can be identified as immature blood vessels at the electron microscopic level. The blood vessels are progressively more numerous in the graft and they demonstrate mature ultrastructural features 2 months after grafting. Last, there is no leakage of the BBB for peroxidase. The vascularization seems to follow a pattern of maturation comparable to that described during development in the literature. In contrast, in the lesioned area, there is a reactive angiogenesis: 10 days after the excitotoxic injection (shortest time studied), there are many wide caliber vessels with expanded perivascular spaces engorged with mesodermal cells. A microvascularization also develops transiently during the first two months. Capillaries are abnormal from the functional point of view, since there is a leakage of the BBB to macromolecules. The use of an experimental model in which transplant had to grow in a lesioned area permits to determine two types of vascularization: an apparently normal developmental timetable, normal morphological and functional characteristics, in the transplant; a reactive angiogenesis, in the lesioned area.

Integrity of the blood-brain barrier in retinal xenografts is correlated with the immunological status of the host

The purpose of this study was to determine the immunological correlates of blood-brain barrier breakdown in retinal xenografts in rats by utilizing skin grafting to initiate a timed immune response to the transplanted neural tissue. Embryonic day 13-14 CD-1 mouse retinae were grafted into the brainstem parenchyma of neonatal Sprague-Dawley rats. In one group of animals a 100 mm2 CD-1 skin graft was placed on the flank 21 days after the initial neural transplant in order to provoke an immune response to the neural graft. Control animals received no skin graft. Animals were injected with horseradish peroxidase (HRP) in the femoral vein 2-8 days after skin grafting. Brains were processed for Nissl, HRP-tetramethylbenzidine, and anti-M-6, -lymphocyte, -macrophage, and -astrocyte antibodies. Experimental and control animals injected 2-4 days after skin grafting showed no leakage of reaction product in the grafted tissue. A small percentage (one of eight) of 5-day animals showed isolated, patchy leakage, but no evidence of rejection of the neural graft. At 6 days all of the grafts showed evidence of leakage, and 71% of these grafts showed infiltration of lymphocytes. By 7-8 days extensive leakage of HRP and widespread infiltration of lymphocytes and macrophages were clearly evident. The present study demonstrates that blood-brain barrier breakdown is correlated closely with the sequence of immunological rejection of the graft. While these results confirm that a barrier exists in healthy neural transplants, they suggest that immunological factors should be considered in cases in which grafts are not protected by an intact barrier.

Neovascularization of transplanted central nervous tissue suspensions: an immunohistochemical study with laminin

Neovascularization of the dissociated central nervous tissue transplanted into the lateral ventricle of the rat was studied using laminin immunohistochemistry. A very high immunoreactive response to laminin was demonstrated in the presumably newly formed vessels within the transplants and the graft-host borders. The growing tips and fine spike-like sprouts called 'streamers' were also highly stained with laminin immunoreactivity. In contrast, laminin immunoreactivity was negligible in the vessels in the host brain. Therefore, these results indicate that laminin may be utilized as a marker for the newly formed vessels in neural transplantations.

The fine structure of vascular-astroglial relations in transplanted fetal neocortex

The vascular development within allografts of rat fetal neocortex was examined ultrastructurally with particular attention to astroglial-endothelial relationships. Grafts placed in the fourth ventricle exhibited a progressive astrogliosis around the host pial or choroidal vessels incorporated within the transplant which was evident by 1 month postoperative. Immunostaining with antisera to laminin showed intense reactivity around such neovessels at the light microscopic level. Transplants located intraparenchymally within the host parietal cortex also developed reactive astroglial "cuffs" around their marginal vessels by 1 week postoperative, although the degree and location of this reaction varied considerably with time. The origin of the reactive astroglia could not be directly determined from this study, but it is possible that they were stimulated by the collagen and fibroblasts present around vascularizing host pial and choroidal vessels in intraventricular grafts and by meningeal elements that entered the wound created for the intraparenchymal grafts. The marked astroglial reactivity within the grafts raises issues concerning their metabolic activity and their intimate relationship with brain endothelium. The close proximity of reactive astroglia to the graft vasculature would not appear to enhance the blood-brain barrier capabilities of transplant neovasculature, especially in intraventricular transplants, as might be suggested by many in vitro studies.

The astroglial response to autonomic tissue grafts

Autonomic (superior cervical) ganglia were grafted either into the IV ventricle where minimal trauma occurred or directly into the cerebral cortex which was necessarily traumatic. Previous studies have shown that host astroglia may migrate into autonomic tissue grafts. The purpose of the present study was to compare and contrast the astroglial response in allo- and autografts. By monitoring the host response in the two model sites using glial fibrillary acidic protein (GFAP) immunostaining in 1 micron plastic sections we sought to determine the role of injury stimulus in astroglial migration. In addition, these models could be used to investigate any potential differences in glial reactivity produced by allo- or autograft antigenic stimulation. In both ventricular and parenchymal locations, astroglia migrated progressively into allografts. Migration, which could have taken place along anastomotic vascular connections, began after one week and was continual, eventually replacing graft neural tissue. Astrocytic processes appeared enlarged and highly immunoreactive only as they entered the allografts or were in close association with the choroid plexus; adjacent host astrocytes were unaltered. Glial migration was greatly reduced in ventricular autografts but in the parenchymal site was nearly comparable to that of allografts. It was suggested that certain immunological factors may be involved in glial reactivity or migration considering the observed differences in the non-traumatic model whereas tissue damage stimulus played a major role in migration in both allo- and autografts. In no instances were typical astrocytic end-feet found on the autonomic graft vessels. The host astrocytic response to grafted autonomic tissue occurred significantly later (5-7 days) than the host endothelial response. This observation indicates that the graft vessels were original, intrinsic ones and the astrocytic invasion played no role in influencing endothelium with regards to brain-barrier properties.