Pineal Gland Transplants into the Cerebral Hemisphere of Newborn Rats: A Study of the Blood Brain Barrier and Innervation

Pineal glands from neonatal (0-1 day) Long-Evans black-hooded rats were transplanted into the cerebral hemispheres of litter mates for periods of 1 to 5.5 months. Grafts exhibited differentiated pinealocytes that were intensely immunoreactive for serotonin. Transplant vasculature was permeable to endogenous IgG, comprised fenestrated endothelia with wide pericapillary spaces typical of in situ glands, and had a volume density intermediate to that of surrounding cortex and in situ pineals. Along the periphery, transplant capillaries tended to have continuous endothelia similar to those of host cortex. This peripheral zone was impermeable to endogenous IgG and appeared to increase in size in older grafts. The presence of noradrenergic-like fibers within the perivascular compartment suggested that transplants were innervated by peripheral sympathetic neurons from the superior cervical ganglia. In animals which had been superior cervical ganglionectomized, noradrenergic-like fibers were absent or degenerating. Neural regulation of transplant metabolic activity was suggested by the increased frequency of pinealocyte synaptic ribbons in denervated grafts. These findings are consistent with the hypothesis that factors from both graft and host influence vasculature physiology and differentiation in neural transplants. Furthermore, grafts appeared to receive appropriate neural input from the peripheral sympathetic system.


INTRODUCTION
The pineal gland, a neuroendocrine gland of diencephalic origin, exhibits several characteristics that make it a useful model for transplantation studies concerning mechanisms of the blood brain barrier (BBB) and innervation. First, the pineal gland of rats is normally vascularized by fenestrated, permeable blood vessels with wide pericapillary spaces/11,22/. Second, it receives a well defined sympathetic innervation from the superior cervical ganglion (SCG) which is necessary for the normal circadian production of the pineal hormone, melatonin /26/. The sympathetic innervation is identified ultrastructurally by the presence of characteristic dense-core vesicles within the sympathetic terminals /21/. Third, the pineal gland contains high concentrations of the indoleamine, serotonin (5HT), and is the principal source of circulating melatonin. These indoles, which exhibit robust circadian rhythms and are easily measured by both biochemical and histochemical methods, provide important markers of the functional integrity of the transplant. Finally In the present report, intracerebral grafts of the neonatal pineal gland into newborn hosts demonstrated a permeable vasculature to circulating macromolecules. These findings corresponded to ultrastructural observations of fenestrated endothelial cells comparable to the normal in situ gland. The innervation of grafts by apparent sympathetic fibers prompted further studies to determine whether this noradrenergic input influenced grafted pinealocyte function.

MATERIALS AND METHODS
Long-Evans, black-hooded rats were used. The pineal gland of 0-1 day-old rats was dissected under hypothermic anesthesia into sterile Ringer's solution and gently aspirated into a glass pipet glued to the tip of a 50/m Hamilton syringe. Simultaneously, small neocortical lesions were made by aspiration into litter mates also anesthetized by hypothermia. The recipient lesions were made 1-2 mm from the midline just rostral to the coronal suture and grafts were immediately placed into the lesion cavity by slowly ejecting them from the syringe canula with the aid of a micromanipulator. The grafts were held in place by a flap of the cartilaginous skull and the pups were warmed by an incandescent lamp before returning them to their mothers. Donor rats were killed by decapitation while under hypothermic anesthesia. The results reported in this study are from 28 rats receiving such transplants.
After 1 to 5.5 months, animals were overdosed with sodium pentobarbital and perfused with physiological saline followed by a 4% paraformaldehyde/0.1% glutaraldehyde solution. In eight rats the SCG were surgically removed (SCGX) bilaterally under sodium pentobarbital anesthesia (45 mg/kg) 5-7 days prior to sacrifice.
Tissue blocks containing the transplant were vibratome sectioned at 50-100 m and examined either with Nissl or immunohistochemically for serotonin and IgG as previously described /32/. Transplant blood vessel permeability to endogenous immunoglobulin (IgG) was analyzed by incubating tissue in biotinylated rabbit anti-rat IgG (1:200). Serotonin positive pinealocytes were demonstrated using rabbit anti-serotonin antiserum (INCstar, Stillwater, MN at 1:100-1:1500) followed by incubation with biotinylated goat anti-rabbit IgG (1:200). Avidin-Biotin histochemistry was performed according to the protocol in the Vectastain ABC kit (Vector Labs, Burlington, CA) using the chromogen diaminobenzidine. Immunoreacted sections were also 115 processed for routine electron microscopy. Sections were post-fixed in osmium tetroxide, dehydrated in ethanol followed by propylene oxide and flat embedded on plastic slides with epoxy resin. After the blocks were cured, the transplants were photographed for purposes of orientation and the transplants cut out and remounted on plastic stubs for sectioning. Toluidine blue 1/am sections from these blocks were photographed for purposes of orientation in the electron microscope and were used to estimate volume density of blood vessels (V%v) with a 90-point grid on a transparent overlay and applying standard stereological point-counting methodologies/35/. Thin sections (70-80 nm) were examined with an Hitachi H-600 TEM operated at 75 kV. Counts of pinealocyte synaptic ribbons were obtained by systematically scanning three grid spaces (300 mesh; Pelco, Tustin, CA) at a magnification of 6,000 on the microscope screen. Each of the grids was coded so that the individual performing the counts was unfamiliar with the experimental group. Differences between the means were statistically analyzed by the Student's ttest (2-tailed).

RESULTS
Transplants were located within the cortical lesion cavity near the cortical surface ( Fig. 1), bordering the lateral ventricle ( Fig. 2) or within the lateral ventricle ( Fig. 3). Pineal tissue was easily distinguished from surrounding cortex with Nissl stain (Figs. lc, 2a) and revealed positive immunoreactivity for both IgG (Figs. la,b,2b) and serotonin (Fig. 3). The IgG immunoreactivity sometimes had a trabecular appearance corresponding to the vasculature of the gland (Fig. la). In some transplants, especially older ones, the IgG immunoreactivity was greatest in the central part of the transplant and declined in intensity toward the periphery of the graft (Fig. lb,2b).  .017-.057 The fine structure of transplant pinealocytes was indistinguishable from that of in situ cells (Fig. 4a).
Coplasmic organelles were well differentiated and included the synaptic ribbon which is characteristic of this cell type (Fig. 4b).  5A,7). The nerve fibers tended to be more common in grafts located within the cortex as compared to those lying within the lateral ventricle. Superior cervical ganglionectomy resulted in the complete degeneration of noradrenergic-like perivascular terminals in six out of eight in situ glands analyzed. Counts of synaptic ribbons in these six animals were significantly elevated above unoperated controls ( Table 2). Although all eight animals which had undergone SCGX experienced bilateral ptosis, in two of these animals noradrenergic terminals were infrequently observed within the gland, suggesting that denervation was not complete. The counts of synaptic ribbons in these two animals did not differ from unoperated controls and were therefore not included in the statistical analysis presented in Table 2.  rt---7) *Indicates significant differences at p < 0.05. Noradrenergic-like terminals were never observed within transplants of animals that had been sympathectomized (as judged by ptosis and lack of terminals within the fl situ gland). In one case an apparent degenerating terminal was found within the transplant of an animal that had undergone SCGX (Fig. 7c). Noradrenergic-like terminals were also not detectable in one graft located within the ventricle from an animal that had not been sympathectomized. Terminals were observed in the remaining transplants from innervated controls and from animals that had been partially sympathectomized. The frequency of nerve endings within these transplants was highly variable. When grafts were grouped according to the presence/absence of noradrenergic terminals, the frequency of synaptic ribbons was significantly greater in transplants lacking noradrenergic-like innervation ( Table 2). Counts of synaptic ribbons in grafts with sympathetic terminals were not significantly different from in situ control glands.

DISCUSSION
The present study revealed that transplants of the pineal gland into the cerebral hemisphere of newborn rats survived and developed anatomical features indistinguishable from those of in situ glands. By one month transplant pinealocytes were highly differentiated ultrastructurally, the capillaries exhibited fenestrated endothelia with wide pericapillary spaces, and the grafts were innervated by sympathetic fibers. Transplants also appeared capable of indoleamine biosynthesis as evidenced by the intense immuno- Fig. 6: A. Capillary located along the periphery of 1-month.old graft located in the cortex. The endothelium is continuous with a narrow pericapillary space. Pinealocytes completely surround this capillary. Bar indicates 2 #m. B. Non-fenestrated capillary along the border of a graft in the lateral ventricle. Note that the pericapillary space is wider (asterisk) where it borders pinealocytes (P) compared to where it borders ependyma (E). Bar indicates 2 #m. reactivity to serotonin antisera. Survival and differentiation of pineal tissue from adult rats and hamsters into the fourth and lateral ventricles of adult hosts has been previously reported/2,10,18/.
Previous work involving the transplantation of fetal neocortex into the cerebral hemisphere of newborn rats demonstrated a transplant vasculature that was impermeable to exogenous or endogenous macromolecules/32/. In the present report, similarly placed newborn pineal grafts demonstrated a permeable BBB. These contrasting findings suggest that transplant vascular differentiation and physiology is determined by the graft and not the host tissue. Because the neonatal grafts and hosts used in these studies were undergoingendothelial differentiation and proliferation at the time of transplant, it is conceivable that graft vascularization comprised both native and host vessels. Concerning pineal grafts, the high volume density of capillaries, which in some grafts was identical to that of in situ glands, further suggests the proliferation of native transplant vessels. In this regard it is noteworthy that pineal tissue contains substantial amounts of "endothelial cell stimulating angiogenesis factor (ESAF)" /33/. The possibility that host vessels also supply the graft is supported by the occurrence of capillaries with continuous endothelia along the periphery of transplants. These peripherally located vessels showed a morphology typical of CNS capillaries suggesting that host vessels penetrated the graft to anastomose with existing capillaries /15,17/. An ingrowth of renal vessels into neocortical tissue grafted in the subcapsular space of the kidney further supports the possibility that host vessels are able to penetrate transplants of neural tissue/28/. Furthermore, nonneuronal host cells have been shown to invade fetal tectal tissue grafted to the midbrain of newborn rats 12o/. IgG immunoreactivity within the parenchyma of the grafts as well as in the host circumventricular organs and the median eminence /32/ correspond to the presence of a permeable BBB that is morphologically characterized by fenestrated endothelial cells. The absence of IgG immunoreactivity along the margins of some of our grafts indicates an impermeability that corresponded to the observed non-fenestrated endothelium in these regions. Appearing larger in some of our older surviving animals, the peripheral impermeable zone may reflect an age-related increasing invasion of host vessels into the transplants. This possibility could be tested by examining transplants beyond the 5.5 month survival period used in the present experiment.
Instead of representing an ingrowth of host vessel,% the impermeable graft margins may alternatively Transplants exhibited a pattern of innervation by noradrenergic-like neurons within the perivascular spaces which is typical of that observed h situ. These observations prompted experiments to determine whether the noradrenergic fibers were of CNS origin (locus coeruleus) or of peripheral origin (SCG).
Normally, there is a widespread ascending noradrenergic innervation of the brain from the locus coeruleus /12,27/ in addition to peripheral sympathetic fibers from the SCG which follow the vasculature of the CNS. After specific brain lesions (medial septum, globus pallidus), a proliferation and invasion of peripheral noradrenergic fibers from the SCG was observed/6,7/. Consequently, our placement of grafts into neocortical lesion cavities may have induced an ingrowth of peripheral fibers as a result of the lesion. In support of this possibility, noradrenergic terminals were either absent or in degenerative stages in the transplants of animals that sustained complete SCGX.
This evidence for SCG innervation, which is also supported by the observed increase in pinealocyte synaptic ribbons after SCGX, is qualified by one animal that was not surgically sympathectomized and yet whose pineal transplant also lacked evidence of noradrenergic innervation. Accordingly, the absence of terminals in the surgical group is not proof that they were removed by peripheral sympathectomy.
The pineal gland normally exhibits a 24-hour rhythm in melatonin production that is dependent on its SCG innervation/26/. If, as our denervation experiments suggest, the transplants are innervated by sympathetic fibers from the SCG, it would be important to determine whether this connectivity establishes the rhythmic production of melatonin over the 24-hour light:dark cycle. Cyclical production of melatonin has been described in pineals transplanted into the anterior chamber of the eye after reinnervation by sympathetic fibers which normally innervate the iris /1,19,25/.
Our data on the frequency of pinealocyte synaptic ribbons suggest that the sympathetic innervation to transplants is functional. The significantly greater number of ribbons found in transplants without noradrenergic terminals paralleled the rise observed in denervated in situ glands. Pinealocyte synaptic ribbons normally exhibit quantitative changes related to alterations in the melatonin production by the gland /el 23/. Several studies have further suggested that these organelles play a role in B-adrenergic mechanisms controlling the rate of melatonin biosynthesis /13,14,24,34/. Because synaptic ribbon frequency in transplants appears dependent upon sympathetic innervation, this organelle could be an important marker to further test the hypothesis that sympathetic innervation of the grafts conveys appropriate information capable of synchronizing graft metabolic activities to the light:dark cycle.