Angiogenesis on the optic tectum of albino Xenopus laevis tadpoles

Developmental Brain Research, 48 (1989) 197-213

197

Elsevier BRESD 50936

Angiogenesis on the optic tectum of albino Xenopus laevis tadpoles Carl M. Rovainen and Meena H. Kakarala Department of Cell Biology and Physiology, Washington University School of Medicine, St. Louis, MO 63110 (U.S.A.) (Accepted 14 February 1989)

Key words: Amphibian; Angiogenesis; Capillary; Endothelium; Eye removal; Hypoxia; Optic tectum; Regression; Sprouting; Temperature

Developing blood vessels were observed directly on the dorsal surface of the optic tectum of anesthetized, transparent albino

Xenopus ,taevis tadpoles, stages 41-54. Case histories of individual tadpoles indicated that pial capillaries developed by the classical mechanism of sprouting of endothelial cells from existing blood vessels. 'Deep sources' appeared on the tectal surface during development. These were sites of upwelling blood cells from capillaries within the nervous tissue of the rectum into vessels on the surface. Few 'deep sinks' were observed in the dorsal tectum of normal tadpoles. The earliest deep sources were probably formed by sprouts from the surface vessels through the basement membrane and into the nervous tissue; later ones may also have formed from internal sprouts back to the surface. Maps of deep sources and of surface vessels in case histories indicated that neural tissue and blood vessels in the caudal half of the tectum grew faster than in the rostral half. The medial venules on the dorsal tectum originated as ordinary-sized rostrocaudal capillaries. They enlarged in diameter as they drained the increasing flow of blood from the tectum into the choroid plexus over the 4th ventricle. Some capillaries disappeared or regressed during development. Our observations on the tectum were consistent with the classical sequence of loss of flow, narrowing, collapse of the lumen, and retraction of endothelial cells into adjacent vessels. Likely sites for regression were upstream from a deep source and at crosslinks between transverse vessels on the lateral tectum. Morphometric parameters for tectal angiogenesis were (a) surface density (ram-~) calculated as total length of surface vessels divided by the dorsally projected surface area, and (b) density of deep sources (ram-2) calculated as total number divided by surface area. From stages 41/42 to 50 average surface density approximately doubled, and average density of deep sources increased about 5-fold. Some of the factors which might be expected to alter brain angiogenesis include nervous activity, availability of Oz, and metabolic rate. Removal of one eye deprived the contralateral tectum of direct retinal inputs, while the ipsilateral side was a control in the same animal. Anterograde labeling of retinal axons with diI18 from the remaining eye confirmed projections only to the opposite side. No significant differences in densities of surface vessels or of deep sources were observed between the contralateral and ipsilateral sides of the tectum. Likewise, in tadpoles with both eyes removed, neither the density of surface vessels nor the density of deep sources was significantly different from normal tadpoles. Tadpoles reared at half atmospheric pressure did not have significantly increased densities of blood vessels compared to normal controls. Exposure to 1/3 atmosphere significantly increased the average densities both of surface vessels and of deep sources. Tadpoles gassed daily with 100% 0 2 had significantly lower densities of deep sources. The densities of surface vessels and of deep sources in the tectum increased significantly in tadpoles adapted to 30 °C. Blood flow in the tectum and the size of the heart were also increased compared to tadpoles of the same stage and length at room temperature. INTRODUCTION Angiogenesis is the process by which new capillaries and blood vessels are formed, especially by sprouting of endothelial cells from existing vessels. Considerable progress is being made towards determ i n i n g the cellular and molecular mechanisms which

control the growth of e n d o t h e l i u m and new blood vessels. The best characterized angiogenic molecules are basic and acidic fibroblast growth factors ( b F G F and a F G F ) 23, but a variety of other peptides and low molecular weight c o m p o u n d s also e n h a n c e angiogenesis directly and indirectlylm8. Capillary endothelium in vitro responds to angiogenic factors by

Correspondence: C.M. Rovainen, Department of Cell Biology and Physiology, Washington University School of Medicine, 660 South Euclid Avenue, Box 8101, St. Louis, MO 63110, U.S.A. 0165-3806/89/$03.50 © 1989 Elsevier Science Publishers B.V. (Biomedical Division)

198 release of urokinase and collagenase 24~5, increased cell division 2j and migration s2, and tubule formation 35. The brain and spinal cord are special cases for angiogenesis and induction of tissue-specific properties of endothelium. The initial blood vessels of the embryonic brain originate from a plexus of angioblasts and endothelium on the outer surface of the neural tube 8'15. Then endothelial sprouts from the surface vessels penetrate the basement membrane covering the brain and form capillary branches and loops within the nervous tissue 3"45. The retina is also an attractive model for angiogenesis, with an initial front of proliferating cells and vessels on the inner surface, regression of some capillary segments, and invasion of the deeper layers of tissue 9. Brain and retina are rich sources of angiogenic factors TM18,23,42, and cultured endothelium from the brain responds to them with increased incorporation of [3H]thymidine 44. Brain endothelium is specialized as the main interface of the blood-brain barrier, with restricted permeabilities 16'4°, high transmural electrical resistance 38, and distinctive histochemical properties 43"47. These properties appear during development 43"46 and are induced by central n e r v o u s tissue 47, probably astrocytes 1"28'49. The heads of living albino Xenopus laevis tadpoles are strikingly transparent (Fig. 1A). Blood vessels, cranial nerves, muscle fibers, individual cells in and below the epidermis, and the brain are easily resolved. The optic tectum was selected as a test region for angiogenesis because its dorsal surface forms a sheet under the epidermis, it is accessible for experimental injections and implants both in the third ventricle and in the pial spaces, and angiogenesis can be correlated with known neural development (see Discussion). The purpose of the present paper is to introduce this preparation and to chart the normal development of capillaries on the surface of the optic tectum by the classical case history approach used by Clark 6 in tadpole tail. MATERIALS AND METHODS Embryos of Xenopus laevis, albino strain, were obtained from Nasco, Fort Atkinson, WI. They were staged according to Nieuwk0Qp and Faber 37 and were provided with powdered frog food (Nasco) at

stages 45 and beyond. Tadpoles were maintained individually in beakers with 100 ml or in small groups in 200-1000 ml growth medium: 60 mg Instant Ocean Salts (Aquarium Systems, Wickliffe, OH) per liter deionized water plus powdered food or 10 mM NaCI and 1 mM N a H C O 3 plus food). Tadpoles and late embryos were anesthetized by immersion in 0.3-0.5 mM benzocaine in 10 mM NaC1 plus 1 mM N a H C O 3. During microscopic observations the animals rested on 1-2 plies of gauze in a shallow layer of the benzocaine fluid in a Petri dish. The head was illuminated from below with a conventional Abbe condenser, and the brain was observed from above through a dry 40x objective with a working distance of 10 mm (Nikon). Outlines of blood vessels and their processes on the surface of the optic tectum were sketched by camera lucida (Nikon drawing tube) at a calibrated magnification of 20 ~m/cm. Figs. 2-5, and 10 are ink tracings from the original pencil sketches. For morphometry, rectangular outlines were drawn over each dorsal tracing of the tectum or half-tectum and the area was calculated (cm 2 corrected to the original mm2). Total length of capillaries and venules was measured with a metric map-reading wheel (cm corrected to the original ram). Surface density was calculated as the total length of blood vessels divided by surface area. Deep sources were recognized as upwellings of blood cells into the surface vessels from vessels at a lower focal plane in the nervous tissue of the rectum. Upwellings at the lateral and caudal edges of the rectum were surface capillaries and were not included. The density of deep sources was calculated from the number within the outline divided by the enclosed area. This is an index or at least a facet of internal capillaries and has the units of conventional capillary density (mm 2). Internal capillaries often could be resolved for some distance in the deeper neural tissue of the tectum. For simplicity, these were neither traced in their entirety nor included in the present diagrams. However, these could be included in a future 3-dimensional analysis of the developing tectal vasculature. In the present young tadpoles, the deep sources were the same size or smaller than surface capillaries and never approached the size or carried the substantial blood flow of the medial surface venules. Diameters of the optic tectum and

199 of the medial venules were estimated directly in anesthetized tadpoles with calibrated eyepiece reticles. Tectal diameter was taken as the maximum distance from the right to left pial-neural borders. The outer diameters of paired medial venules were taken over the caudal tectum or rarely the diameter of a single venule was taken when a pair was absent. Eyes were evulsed from tadpoles deeply anesthetized in 1 mM benzocaine. One eye was removed over a range of stages, 36-50, but usually 42-47. Both eyes were removed from tadpoles at stages 45-47. Tadpoles were transferred to 30 mM NaCI plus 1 mM NaHCO 3 for healing of the wound for one or a few days and then were maintained in growth medium until they were anesthetized and traced at 1-39 days after the operation. Similar numbers of tadpoles were traced first on the contralateral and then on the ipsilateral side and in the reverse order to control for possible decreasing circulation and declining visibility of capillaries under anesthesia. Retinotectal projections were labeled by diffusion of fluorescent diI18 in membranes 22 from the remaining eye 9, 17, 21, 28, 36, 44, and 51 days following removal of the first. Anesthetized tadpoles were fixed by immersion in 5% paraformaldehyde in 100 mM sodium phosphate buffer pH 7 at room temperature; the skin was peeled from the dorsal surface of the head to provide better fixation of the brain. The cornea and lens of the remaining eye were removed, and crystals of diI18 (Molecular Probes, Eugene, OR) were implanted in the eye cup. The tadpoles were incubated for 8-29 days in the dark in 2% paraformaldehyde in buffered saline. The skull and eye were dissected together as a sheet, the brain and optic nerve were viewed ventrally and dorsally by epifluorescent microscopy with rhodamine filters, and the patterns of fluorescent fibers in the contralateral and ipsilateral optic tracts and tectum were observed and photographed. For low oxygen, 10 or fewer early stage tadpoles were grown in a dosed 500 ml glass suction flask in 100 ml growth medium at room temperature. For the first 1-3 days the flask was evacuated to 2/3 atmosphere but then was maintained at 1/2 atmosphere (380 mm Hg) with daily replacement of the air and re-evacuation. Two groups of tadpoles at 1/2 atmosphere were further evacuated to 1/3 atmo-

sphere (253 mm Hg) and were traced 1-6 days later. These low pressures included the saturated vapor pressure of water (18.6 mm Hg at 21 °C). Within 4 days the hearts of the tadpoles in low oxygen were redder in appearance due to more erythrocytes; larger red hearts and blood vessels were especially conspicuous in tadpoles at 1/3 atmosphere. For high oxygen, similar groups of tadpoles in a dosed 500 ml flask were gassed daily with 100% 0 2 at ambient pressure. Samples of gas over the water were initially 90% or more 0 2. Tadpoles at high temperature were grown in open beakers in standard growth medium at 30 °C in an incubator or water bath. The hearts of tadpoles at 30 oC were larger by an average of 18% compared to tadpoles of the same length and stage at 21 °C; the enlargement of the heart was first noticeable after 2 days at 30 °C. Tadpoles at low temperature were kept in beakers of medium in a refrigerator at 12 °C. Tadpoles from all conditions of rearing were anesthetized by immersion in 0.3-0.5 mM benzocaine and were traced at room temperature while exposed to normal atmospheric oxygen. The blood flow was higher in the tectum of tadpoles adapted to 30 °C than in normals, even when both were observed under anesthesia at 20-21 °C. The average stages at the times of tracings and average treatment intervals are given in Table III. RESULTS

General organization of tectal blood vessels Capillaries and venules form a network over the dorsal surface of the optic tectum in Xenopus tadpoles. These vessels are immediately apparent in vivo by the movements of red blood cells. The general pattern of blood flow was upward through capillaries on the lateral surfaces of the tectum, across the dorsal surface towards the midline, and into paired medial venules, which drained caudally into vessels or sinuses in the choroid plexus over the 4th ventricle (see arrows in Figs. 2-5 and 10). During development additional deep capillaries formed. These were consistently sources of extra blood into the dorsal surface vessels and are defined as 'deep sources'. Conversely, 'deep sinks' were defined as capillaries which joined the surface vessels and which carded blood cells from the surface into the deeper layers of the tectum. Deep

200

201 sinks were only rarely observed on the dorsal surface of the tectum in normal tadpoles and accounted for far fewer than 1% of the total junctions of interior to dorsal surface vessels. Evidently, both surface and deep capillaries received blood from more ventral arteries, which were not resolved in the present experiments. Velocities of blood flow in individual capillaries were heterogeneous by qualitative observations. Some surface vessels lacked blood cells and visible flow, but occasional blood cells could pass through, and the outlines of the vessel walls could be traced to other vessels with visible circulation. Sometimes the empty vessels were narrow or tapered and perhaps collapsed; these may have represented an early stage in regression6 or new vessels formed by the joining of sprouts 7. Tapered or blunt-ended vascular branches in the tectum resembled capillary sprouts in the tail fin, but sequential tracings of growth were necessary to distinguish them from regressing branches.

Case histories of angiogenesis Sequential tracings from individual tadpoles at 1-3 day intervals documented the growth of tectal blood vessels (Figs. 2-4). Capillaries with moving blood cells could be distinguished at stages 40-43, but the walls of empty vessels and branches were obscured by residual yolk granules. Visibility improved at stage 45 and onward, and tracings were more nearly complete. Some vessels, presumably collapsed, had less contrast and were sometimes missed (m in Fig. 2). However, despite these imperfections it was possible to identify individual capillaries, deep sources (d), sprouts (s), and regressions (r) from serial tracings during development. Case histories indicated that sprouting was the mechanism of angiogenesis on the surface of the optic tectum in Xenopus tadpoles. In Fig. 2C-E, sprout sl grew slowly to another surface vessel and fused to become a new capillary. New vessels could also form rapidly in 1-2 days or less, for instance in the region near d3 in Fig. 2B,C, in the caudal corners

in Fig. 2D,E, and sl-3 in Fig. 4A,B. A few sprouts also appeared and disappeared without forming vessels. The detailed contacts between sprouts or from sprouts to existing blood vessels were not resolved in the present study, but have been described in tadpole tail in vivo7. Individual capillaries on the tectum could also regress over periods of one to several days (r,rl-9, Figs. 2-4). Prior to regression such capillaries lacked flow and became smaller in diameter and less distinct (Fig. 3, r2, Fig. 4, s2-r2). Later, the vessel pinched off and retracted as tapered processes (Fig. 4, r5, r6). Surviving vessels were most often those carrying blood from the lateral edges to the medial venules. Regressing capillaries were commonly the crosslinks between these vessels (Fig. 3) or were segments upstream from a deep source (Fig. 4, s2-r2). In these cases the pressures at the two ends of the capillary may have been nearly equal, and the loss of perfusion may have initiated the process of regression (see Discussion). The numbers of deep sources in the tectum increased dramatically during development. One or two deep sources were often present in the rostral tectum at stages 42-46. In Fig. 2 the total numbers of visible deep sources increased to 4 in C, 8 in E, and 14 in F, exclusive of those at the lateral and caudal edges of the tectum. Deep sources are indicated by an opening with an emerging arrow in the figures, but only some of them are identified by individual labels, e.g., dl-3. Increased numbers of deep sources also are apparent in Figs. 4 and 5. One candidate for regression of a deep source is d3 in Fig. 4D,E.

Vascular maps indicate faster growth in the caudal tectum Deep sources were assumed to be embedded as anchors in the nervous tissue and were used as landmarks for its expansion. Deep sources were concentrated in the rostral half of the tectum. Distances were measured between identified deep

Fig. 1. Photographs of a living tadpole and of the medial venules on the dorsal side of the optic tectum. A: the head of an albino Xenopus laevis tadpole, stage 48. The view is dorsal with transmitted light, kindly photographed by Dr. Thomas Woolsey. The optic tectum is indicated by an arrow. In A - C of this figure rostral is to the left and caudal to the right, hut the convention in Figs. 2-5 and 10 is caudal to the top. B: medial venules in vivo of a stage 46 tadpole. The dark spots are overlying melanophores, which express some pigment in this strain. C: medial venule and a communicating capillary in a stage 50 tadpole. Blood cells flowed from rostral to caudal but are blurred in the time exposure.

202

E _.

..

,I,~

-....~------.

Fig. 2. Case history of vascular development on the dorsal tectum. Caudal is to the top, rostral to the bottom. Dashed lines indicate the walls of the underlying third ventricle and the lateral and caudal edges of the tectum. Arrows indicate direction of blood flow. d l - 3 : deep sources (locations of dl and d3 in E and F were estimated at the edge of the tracings), m: estimated locations of two empty unperfused capillaries which were missed during the tracings in C and D. r: capillary regression (r is a short connection in B, dashes in D, and stumps in C,E). sl-5: sprouts and new capillaries. The tracings were made on the following calendar dates. A: 9/18/88 (day 1), stage 42, length 6.6 mm. B: 9/19 (day 2), stage 45, length 7.9 mm. C: 9/20 (day 3), stage 45.5, length 8.7 mm. D: 9/22 (day 5), stage 45.5, length 8.8 mm. E: 9/24 (day 7), stage 46, length 9.6 mm, left side tilted slightly upward. F: 9/27 (day 10), stage 47, length 12.1 mm. The right side of the tadpole and of the tectum was rotated upward in F.

203

/z..,-

o

,.2

--.

/r

r/'.

I

E,,'....

I . " 7..,,x...

"

-."

Fig. 3. Case history of vascular development with multiple capillary regressions on the dorsal surface of the right optic tectum, dl-3: deep sources, rl-9: regressions, approximate former paths of vessels are indicated by thin dashed lines, sl-2: sprouts and new capillaries. Tracings were made on 6/10-20/88 with day 1 corresponding to the first tracing, not the age since fertilization. A: day 1, stage 46.5, length 11.6 mm. B: day 4, stage 46.5, length 12.4 mm. C: day 6,stage 46.5, length 12.7 (little growth). D: day 8, stage 47, length 14.5 mm, E: day 11, stage 48, length 16.8 mm.

sources in both the transverse and longitudinal directions and were c o m p a r e d to measurements of tectal diameter and distances from an identified deep source to the caudal edge of the tectum (Table I). T h e distances between deep sources expanded at the same rate as diameter. However, the expansion between the caudal tectum and a more rostral deep source was significantly greater. This indicated that the caudal tectum grew more rapidly than the rostral tectum. T h e growth of surface vessels in different regions was tested by the expansion of grid patterns. Orthogonal 50 # m grids were drawn over the surface vessels at day 1 of a case history and were projected onto the same surface vessels through subsequent tracings (Fig. 4). Intermediate tracings were used

but only the first and last are illustrated in Fig. 5. One result was that the grids did not expand uniformly or regularly. A n o t h e r result was that the sizes of grids in the lateral caudal quadrant increased more than those in the rostral tectum. That the grids and surface vessels at later stages bowed rostromedially (Figs. 4 and 5) also indicated a relative expansion from the caudal c o m e r s of the tectum. In addition, sprouting and elongation of individual capillaries appeared to be greater in the caudal tectum, and the appearance of deep sources there was delayed.

Vascular morphometry in normal tadpoles Four measurements were m a d e on the optic tectum and its blood vessels in normal tadpoles:

204

Fig. 4. Case history of vascular development on the right dorsal tectum of a third tadl~le, d l - l l : deep sources, rl-5: regressions (former positions in C-F are indicated by dashes), sl-7: sprouts and new capillaries. Note that capillary s2 is well peffused in B but then loses visible flow, narrows (s2-r2 in D and E), and regresses (r2) after 13 days in F. Likewise, its neighboring capillary r6 loses flow, narrows, and retracts as tapered processes. Also note the regression r3 between B and C. A candidate for regression of a deep source is (d3) in E anf F. The directions of blood flow change in the capillary above sl-s3 between C and D and in the new medial capillary s6 between D and E. Changes in local flow may have initiated the regression in vessels s5-r5 in F. The former position of r7 in F is shown by dashes, and a new vascular branch crosses it. Grid lines at 50 ~m intervals are superimposed on the surface vessels on day 1 in A and are extended to successive tracings with the surface vessels as landmarks. The distortion of the grid lines indicates that the surface vasculature does not expand uniformly on the teetum. Tracings were made on 6/21 to 7/5/88. A: day 1, stage 44, length 8.5 ram. B: day 2, stage 46, length 10 mm. C: day 4, stage 47, length 12.8 ram. D: day 7, stage 48, length 14.9 ram. E: day 10, stage 48, length 16 ram. F: day 15, stage 48, length uncertain.

diameters of the tectum and of the medial venules were m e a s u r e d directly with reticles, and densities of surface vessels and of deep sources were calculated from tracings. The diameter of the optic tectum from right to left pial surfaces was taken as an index of its growth (Fig. 6, Table II). However, as described in the previous section the caudal tectum appeared to elongate more rapidly than the rostral parts. The medial venules were the channels for draining the progressively increasing blood flow from the tectum to the choroid plexus. During development

the qualitatively increased blood flow was correlated with increased diameters of these vessels (Fig. 7, Table II). O u t e r diameters were the easiest to resolve during routine measurements. F r o m stages 46 to 53 the walls of the medial venules were simple endothelial sheets 1 ktm or less in thickness except at sites with endothelial nuclei, which were 3 - 5 / ~ m in width. The inner diameters were thus 1-2 /~m smaller than the outer ones. A light zone just inside the l u m e n in older tadpoles was not included in wall thickness. This was a 3 - 5 / ~ m layer of plasma from

205

I

I

,

i;

/

~

~e

~,~",,~-~---~'

Fig. 5. Expansion of surface grids based on case histories of patterns of vessels on the dorsal tectums of four tadpoles. Intermediate stages were traced and plotted as in Fig. 4, but only the first (upper drawings) and last days (bottom) are illustrated here. Day 1 grid intervals are 50/~m. A: day 1, stage 44.5; day 8, stage 47. B: day 1, stage 42.5; day 7, stage 48. C: day 1, stage 46; day 14, stage 49. D: day 1, stage 44.5; day 15, stage 49. Note that the grids do not expand uniformly. Growth is most pronounced in the caudal lateral tectum. Grids and vessels are generally bowed rostromedially.

which red blood cells were excluded at high flow rates. Capillary density is conventionally expressed as total length (mm) divided by tissue volume (mm3),

1.5 -t :.

|

TABLE I

Normalized expansion after 4-9 days of tectal diameter, distances between identified deep sources, and distance from an identified deep source to the caudal edge of the tectum

2.

~1.0 _o o

Is

•

n.s., not significantly different from expansion of tectal diameter.

•

t L|--

$

•

":'o•.a• 3.

:.

.-,...~....~:~..•

0.5

Measurement Tectal diameter mediolateral Distance between deep sourcesmediolateral Distance between deep sources rostrocaudal Distance from deep source to caudal edge

Mean expansion

S.D.

1.236

0.108 21

1.236

0.180

17

n.s.

1.214

0.127

18

n.s.

1.607

0.223 20

n

stage

P < 0.1%)

Fig. 6. Growth of the optic tectum in albino Xenopus tadpoles. Diameters were measured from the right to left sides at the inner pial borders in vivo in anesthetized tadpoles of various developmental stages of Nieuwkoop and Faber a7. For mean values, see Table II.

206 T A B L E II

Average values in normal tadpoles by stages Means _+ S.E.M. (n).

Stage

Surface density (ram-I)

Deep density (mm --2)

Diameter of tectum (l~m)

Outer diameter of medial venules (12m)

41 42 43 44 45 46 47 48 49 50 51 52

9 . 3 + 1 . 3 (6) 11.1 + 0.9 (9) 11•3+ 1.3 (5) 12•9 _+ 0.7 (9) 15.4 + 1.0 (12) 15.9 + 0.5 (43) 16.3 + 0.5 (24) 15•8 + 0.4 (23) 17•8 + 0.5 (20) 17.0 + 0.7(15)

2 . 3 + 2 . 3 (6) 6.3 + 1.8 (9) 7.1 _+4•4 (5) 9.6 _+ 2.3 (9) 15.8 + 4.1 (12) 17.3 + 1.4 (43) 22.5 _+ 2.0 (25) 29.6 + 3.0 (23) 39.2 -+ 3.4 (20) 34•4 + 3.2(16)

412+11(16) 433 _+ 14 (11) 4 6 2 + 1 1 (9) 492 _+ 16 (13) 509 -+ 11 (21) 590 + 9 (48) 632 + 9 (48) 700 + 23 (39) 789 -+ 15 (28) 793 + 46(32) 1197_+ 63 (5) 1325 + 44 (4)

12.7 + 1.4 (6) 15.7_+ 1.2 (6) 15.2 _+ 1.5 (6) 16•8 + 0.6 (16) 19•2 + 0.9 (13) 19.5 + 0.6 (22) 24.7 + 0.7 (21) 24.1 + 1.1 (13) 28.4 + 1.1 (15) 33.6_+ 2.6 (9) 32.9 + 2.9 (7)

11.3+1.2

(3)

conventional u n i t s ( m m -2) and was calculated as the number of capillaries emerging at the dorsal surface divided by surface area. Both the density of surface capillaries (Fig. 8, Table II) and the density of deep

but in the case of the tectum it was more appropriate to calculate surface density as the summed length of surface vessels, including the medial venules (mm), divided by area (mm2). Density of deep sources had 60

5(1

40

g

~>20 :!-

-g

o.

E

40

41

4'2

4'3

4'4

4'5

4'6

4'7'

48

stage

,49

50

51

5'2

53

54

Fig. 7. Outer diameters of the medial venules in anesthetized tadpoles of different developmental stages• The walls of the vessels were 1 ,um or less in thickness. The diameter plotted at stage 40 is that of a sprout• For mean values, see Table II.

207 80 -r

7O

25

E E

eo

..'9

.[

| 20

:

•

•

| is

~

.:

: :

~ 5o

:,~,."

• ~. ~. ~.

I

E

:..

$ •

~ 40

..

". ?

~ ao :.

:

•

.

.• o

•

," .':

•.-.'" : . . . : . .

: , "

o

lO

>,20

I

:.

' •

t

•

t it"•:''" .. ...::

: ¢

"

"

t

,

"

"

i

t

m

g~

•. •

'o

t

i

0 41

t

42

i

i

43

i

i

44

i

i

i

45

i

46

i

i

47

i

i

48

i

i

49

i

i

50

i

i

51

stage

Fig. 8. Densities of blood vessels on the surface of the tectum. The trend was an approximately 2-fold increase during this period of development. For mean values, see Table II.

sources (Fig. 9, Table II) increased during normal development. These taken together with the increase in diameter and length of the tectum demonstrate statistically the dramatic growth of capil-

I

t

41

l

t

t

42

.

i

i

43

i

Y " . .

i

44

l

"• .

i

; • ' 7 ::.

I

I

I

t

45 46 41 stage

t

t

48

t

49

i

,50 51

Fig. 9. Densities of observed deep forces increased about 5-fold during this period of development. For mean values, see Table II.

laries during development. All of these parameters are plotted as functions of the developmental stages of Nieuwkoop and Faber 37. This has the effect of expanding time or

i

~al

Fig. 10, Comparison of vascular patterns on the fight (control) and left (deprived) dorsal surfaces of the optic tectum 20 days following the removal of the right eye. The tadpole was stage 44 and 8 mm in length at the time of the operation and grew to stage 47.5 and 15 mm at the time of tracing. The overall diameter of the rectum (700/~m), densities of surface vessels (18.4 and 15.5 ram-l), and numbers of deep sources (10 and 13) were normal.

208 TABLE III Tests of different conditions on vascular densities of the developing tectum

n.s.. not significantly different. Treatment

Stage at observation mean +_S.D.

Interval (days) until observation mean -+ S.D.

Surface density mean + S. E.M. (n) (mm- i)

Density of deep sources mean + S. E.M. (n) (mm-2)

Removal of one eye Contralateral tectum Ipsilateral tectum

48.7 -+ 2.0 48.7 + 2.3

16.8 _+ 1(I.4 17.5 _+ 10.7

17.42 + 0.44 (36) n.s. 17.37 + 0.57 (28) n.s.

34.3 + 3.5 (34) n.s. 34.7 _+5.2 (28) n.s.

Removal of both eyes

49.4+2.6

21.(I_ + 13.6

17.80_+5.83 (9)n.s.

29.4+5.8

Low O, 1/2 atm l/3 atm

47.5 + 1.3 48.0+ 1.1

11.5 4- 7.7 2.3 _+2.0 (additional days)

16.99 + 0.43 (33) n.s. 18.90 + 0.58 (6) P < 0. 1%*

35.0 + 3.4 (33) n.s. 66.5 + 10.3 (6) P < 0.1%*

High O z (gassed daily with 100% 02)

49.1 + 0.9

14.6 _+6.8

16.38 + 0.48 (10) n.s.

24.0+4.3 (10) P < 1%*

30 °C

47.8 _+ 1.2

7.6 __+6.3 (2-29 days)

17.74 + 0.37 (68) P < 0.2%*

42.8 + 2.5 (68) P<0.1%*

12 °C

46.0 + 0.9

11.0 __+6.1

14.35 + 0.39 (13) n.s.

16.6 + 4.3 (13) n.s.

(9)n.s.

* Significant differences are by comparison to the nearest stages in normal tadpoles, Table II.

chronological age at early stages and compressing it at later ones. U n d e r the growth conditions in our l a b o r a t o r y tadpoles changed from stage 41 to 44 in one day, 44-45 in one day, 45-47 two days per stage, and 4 7 - 5 0 a p p r o x i m a t e l y 5 days p e r stage. In terms of growth, the average tadpole u n d e r our l a b o r a t o r y conditions was 6 m m in length at stage 41, 10 m m at stage 46, 15 m m at stage 48, 20 m m at stage 51, 30 m m at stage 53, and 40 m m at stage 55. Eye removal

D e n e r v a t i o n or visual deprivation might be expected to reduce the vascularization of the optic tectum. The retinal projection to the tectum is c o m p l e t e l y crossed prior to m e t a m o r p h o s i s in normal S e n o p u s 19"2°'26. Thus, removal of one eye should deprive the contralateral optic tectum of its direct retinal innervation, while the tectum ipsilateral to the o p e r a t i o n should be normally innervated by the intact eye and serve as a control in the same tadpole. In m e t a m o r p h i c tadpoles, stage 56, removal of one eye elicits a new projection from the intact eye to its ipsilateral tectum 19. H o w e v e r , this ipsilateral p r o j e c t i o n has not been observed after eye

removal in y o u n g e r tadpoles 26. We tested retinotectal projections in our series of o n e - e y e d tadpoles by the spread of the fluorescent p r o b e diI from the remaining eye. Conspicuous labeling was present across the optic chiasma and in the contralaterai optic rectum without any noticeable ipsilateral projections in 13 tadpoles, stages 48-53, at 9-51 day intervals after r e m o v a l of the first eye. Fig. 10 is a pair of tracings of b l o o d vessels on the dorsal surface of the optic tectum in a t a d p o l e 20 days after r e m o v a l of the right eye. T h e vasculature is essentially similar on the two sides. Quantitatively, the density of surface vessels on the contralaterai e x p e r i m e n t a l tectum was 15.5, and on the ipsilateral control side 18.4. T h e n u m b e r s of d e e p sources were 13 on the contralateral side and 10 on the ipsilateral side. Both groups of numbers are in the n o r m a l range for this stage. Densities of surface vessels and of d e e p sources in all of the o n e - e y e d t a d p o l e s were c o m p a r e d statistically on the contralateral and ipsilateral sides of the tectum as functions b o t h of intervals after eye removal and of d e v e l o p m e n t a l stages. Densities on the two sides were not significantly different at any

209 stage, 46-54, or at any interval, 1-39 days. The overall means and standard errors, which were not significantly different, are listed at the top of Table III. The diameters of the medial venules and blood flow were measured in another group of 9 stage 47 tadpoles 20 days after the removal of one eye. Mean outer diameter of medial venules on the side contralateral to eye removal was 19.7 + 1.2 #m (S.E.M.) and on the control ipsilateral side 19.2 + 1.3/~m. Maximum velocities of red blood cells in the medial venules were estimated by the flying spot technique 5 and were 369 + 57/~m/s and 443 + 90 pm/s (means + S.E.M.), respectively, on the sides contralateral and ipsilateral to eye removal. Blood flows through the medial venules were calculated on the basis of inner diameters and laminar flow and were 52 + 14 and 61 + 24 pl/s (means + S.E.M.) contralateral and ipsilateral to eye removal. None of these differences is statistically significant. Removal of both eyes did not produce any significant change in densities of surface vessels or of deep sources (Table III) compared to normal tadpoles at the same stage (Table II).

Low and high oxygen Capillary densities in tissues should match metabolic rates and diffusion distances for oxygen. Low partial pressure of 02 might provide an additional stimulus for angiogenesis and produce greater capillary density, while high 0 2 might decrease angiogenesis and capillary density. Tadpoles reared at one half atmospheric pressure were compared to normal controls both by stages, 46-50, and as a whole at an average stage of 47.5 (Table III), and no significant differences in densities of surface vessels or deep sources were observed. However, severe hypoxia at one third atmospheric pressure produced significantly higher densities of both surface vessels and of deep sources (Table III); these particular tadpoles also had high vascular densities when they were still at 1/2 atm. Conversely, tadpoles gassed with 100% 0 2 had significantly lower densities of deep sources (Table III).

Temperature High temperature decreases the solubility of 0 2 and would be expected to raise metabolic rates. Tadpoles adapted to 30 °C for 2 or more days had

larger hearts and increased blood flow in surface and deep capillaries of the tectum. The diameters of the tectum and of the medial venules were similar in tadpoles of the same stages at 30 °C and at room temperature. The vascular parameter which increased most significantly at 30 °C was the density of deep sources. At each stage 46-50 the mean densities of deep sources were 50-70% higher in 30 °C tadpoles than in the corresponding stages of tadpoles at room temperature. The overall mean for density of deep sources at the average stage of 47.8 (Table III) was significantly higher than in normal tadpoles at stage 47 or 48 (Table II). Deep sinks on the dorsal surface of the rectum were also more common, 13-533 deep sources, in the groups of tadpoles adapted to 30 °C. Similar numbers of sprouts were observed in 30 °C tadpoles compared to those at room temperature. Many of the surface sprouts at 30 °C became connected to deep sources. Regressions also occurred at 30 °C, especially at cross-links between transverse surface vessels. Tadpoles did not grow well at 12 °C and remained near developmental stage 46; their vascular densities were not significantly different from stage 46 tadpoles at room temperature (Tables II and III). DISCUSSION The present results have depended on the visibility of blood vessels on the surface of the optic tectum in living Xenopus tadpoles. The resolution of morphological details in the tectum was less than in the tail fins of the same animals. This was due in part to the additional thickness of underlying tissues of the head. Metamorphic tadpoles beyond stage 52 were less favorable for in vivo observations because their dorsal meningeal tissue became thicker and because the vascular networks on the brain became too complex for easy tracings. The earliest perfused capillaries on the optic tectum were recognized by blood cells moving inside them at stages 40-43, but yolk granules interfered with the resolution of unperfused vessels and branches. Resolution and blood flow increased significantly at stages 45 and 46. For the earlier stages of angiogenesis in Xenopus special vital or histochemical markers for endothelial cells and angioblasts should be developed. One approach may be specific antibodies, as developed

210 for quail endothelium 8'31'39"41. Another possible approach is intracellular injection of cytoplasmic markers into blastomeres 13'29'36 and following descendent endothelial cells when they are suitably labeled. Blood pressure contributed in two ways to the visualization of capillaries on the optic tectum. First, moving blood cells were more conspicuous than stationary structures. Second, the contrast of the walls of capillaries was better when they were distended, even if no blood cells were flowing through them. If blood flow to the tectum ceased, some capillaries were no longer visible, perhaps because they had collapsed. This raised the possibility that cessation of flow in capillaries might render some of them invisible. Case histories provided the opportunity to check for missed capillaries in sequences of tracings. In Fig. 2C,D dotted lines labeled with ' m ' indicate positions of vessels which were probably missed, because they were present in preceding and subsequent tracings. In this tadpole the missed capillaries resulted in a small underestimate of density of surface vessels. Diminished circulation was a more serious problem for deep sources. These were most often recognized by the movement of blood cells into surface vessels, and some were probably missed in the absence of such flow. Part of the scatter in the density of deep sources in Fig. 9 probably is due to different observed fractions of visible to total deep sources due to variations in total and regional blood flow through the tectum. Thus, the middle and upper regions of the scatter plot in Fig. 9 are probably the more realistic estimation of densities of deep sources.

In models for angiogenesis the first step is stimulation of endothelium. The identification of angiogenic factors and the mechanisms of action have been reviewed recently 11'1s'23. Mechanical factors such as shear stress and vasodilation may also promote angiogenesis in vivo 27. For brain angiogenesis in particular, the initial blood vessels are formed over the outer surface of the neural tube 8' 15.45. As the brain grows, angiogenic factors may be released by the nervous tissue to stimulate endothelium of the surface vessels. In addition, there should be some message for the direction of sprouts to and through the basement membrane on the brain surface. One of the early responses of activated

endothelium is the release of urokinase, which activates tissue plasminogens, and of collagenases TM 35 These proteases help open gaps in basement membranes. Then endothelial cells migrate through the gap as an initial sprout and proliferate as an additional source of cells 2. Ultimately, the growing sprout contacts another sprout or pre-existing blood vessel, adheres, and forms a loop or connection. The details of the initial contact and statistics of sprout to sprout or sprout to vessel connections were not determined in the present study. Finally, a lumen forms within the cord of endothelial cells or even within single cells St. In Xenopus tadpoles the internal carotid arteries and their branches are ventral to the brain, and the venous drainage from the rectum is dorsal, establishing a general ventral to dorsal pressure gradient. Thus, blood flow in internal capillaries is also ventral to dorsal and produces the upwelling of blood cells from the deep sources. We suggest that deep sources are formed from sprouts which penetrate the rectum from the surface vessels. However, once some internal vessels have formed, these may also form sprouts, which may grow out to rejoin vessels and sprouts at the surface. Besides sprouting, endothelial growth is also necessary for the expanding luminal area of blood vessels during development. Tectal blood vessels in Xenopus grow both in length and in diameter, the latter dramatically in the medial venules (Fig. 7). Classically, Thoma 5° proposed that blood vessels grew in diameter in response to increased rates of blood flow. In the tadpole tail Clark 6 proposed that increased blood flow enhanced both enlargement of existing vessels and sprouting of new ones. An increase in medial venule diameter might have been expected in tadpoles at 30 °C with increased blood flow. However, under conditions of laminar flow and constant pressure gradient, the diameter would only have needed to increase in proportion to the fourth root of the perfusion rate, and the small change might not have been detected. In contrast, the increases in blood flow and in diameter of the medial venules between early and late stages were much more substantial and conform to Thoma's law. For the future, the flow rates need to be measured directly, and proliferation of endothelium should also be tested. Hypoxia at high altitude stimulates cerebral blood

211 flow, induces polycythemia, and increases the volume of capillaries and blood in the mammalian brain. It is not yet clear whether endothelial sprouting is stimulated, and if it is, whether the direct stimuli are low oxygen, metabolic stress, and/or the mechanical effects of increased flow and hematocrit. Capillary diameters compared to those in control cerebral cortices were increased after acclimatization to high altitude in young aa and neonatal rats 3 and in dogs 32. Mean capillary densities were also increased in rats at high altitude 3"14and in human infants which died after prolonged arterial hypoxia 14. Total fractional volumes of cerebral capillaries were increased in puppies born at high altitude 4. In the provocative experiments by Miller and Hale a4 polycythemia was induced in rats by high altitude, by stimulation of erythopoietin release by cobalt, and by infusion of rat erythrocytes. The reported densities of blood vessels in the cerebral cortex of controls were several fold lower than in other studies TM and may have been based on the most prominently marked vessels; however, increases in vessels of 21-29% were significant in all three types of polycythemia. Miller and Hale 34 suggested that vascularization could have been stimulated by increased blood viscosity, blood volume, or blood flow to the brain. In Xenopus tadpoles hypoxia at 1/3 atmosphere but not at 1/2 atmosphere significantly increased the densities of blood vessels on the optic tectum. Whether the stimulus was low oxygen or polycythemia remains to be tested. Regression of capillaries was well documented by Clark 6 in tadpole tails and was preceded by cessation of flow. Blood flow is a plausible mechanism for validation of capillaries during angiogenesis. If blood flow were rapid the blood vessels would be retained and might form additional sprouts, whereas unperfused capillaries would be physiologically ineffective and would regress. Perhaps fluid flow over the endothelium is itself a signal for maintenance or additional growth. High shear stress stimulates proliferation of aortic endothelium in culture 12, but flow effects should be tested with cultured capillary endothelium, especially as configured for sprouting into loose collagen 35. Tissues outside blood vessels may provide the stimulatory and inhibitory signals both for blood flow and for sprouting and regression. Thus, the cessation of blood flow may occur because

nearby tissues are already well supplied with 0 2 and metabolites, and regression of capillaries may occur as a result of diminished release of angiogenic factors. The neural development of the optic tectum has been studied intensively in amphibians 1°'3°. In Xenopus tadpoles the first axons from retinal ganglion cells arrive at the tectum near stage 38/39, and the first electrical responses to bright visual stimuli can be recorded in the tectum at stage 4025. In the present study, movements of blood cells in tectal capillaries were observed as early as stages 40-42, approximately coincident with the onset of visual function. The tectum continues to grow by the proliferation of neurons, but the pattern is not uniform. It is generally agreed that the tectum grows by adding neurons predominantly caudally and medially, as indicated by higher incorporation of [3H]thymidine in these regions in Xenopus and Rana tadpoles 1°'3°'4s. This differential growth was confirmed in the present study by using surface blood vessels and deep sources as landmarks. Individual blood vessels became longer as the tectum expanded, as classically stated as a rule for vascular growth by Thoma 5° and demonstrated by Clark 6 in the tadpole tail. The rate of lengthening was more rapid in the caudal than rostral tectum, and as areas expanded more rapidly caudally, sprouting of new capillaries was also more apparent. Deep sources were most numerous in the older and thicker rostral and lateral regions of the tectum but were more rarely observed caudally and medially. Retinotectal connections shift during the development of the tadpole to preserve the retinotectal map in the face of generation of ganglion cells in the periphery of the retina and generation of neurons in the caudal and medial tectum 1°'2°. Deep sources presumably remain embedded among the same neurons and glia in the tectum and might be used as additional markers during the neuroanatomical changes. The surface vessels may or may not be attached securely to the pia, and some slippage and non-uniform lengthening may occur during growth of the tectum. This may have contributed to the irregular expansion of grids in Figs. 4 and 5. Removal of one eye in Rana embryos is followed by a reduction in mitosis and retardation of growth in the contralateral optic tectum, especially in older

212

tadpoles near metamorphic climax 3°. Similar effects

Neurobiology at Washington University under the

have been observed in other species, including stage

direction of Dr. Gerald Fischbach and from Bio-

63 X e n o p u s 3°. The X e n o p u s tadpoles in the present

medical Research Support G r a n t SO7RR05389-26

study were younger, stages 45-54, and of the albino

through Washington University and by US PHS

strain. None of the m e a s u r e m e n t s on blood vessels

Grants NS09367 and HL41075. Very special thanks

of the contralateral tectum was significantly different

are due to Dr. Judah F o l k m a n and his laboratory in

from those on the side ipsilateral to eye removal or

the D e p a r t m e n t of Surgery, the Children's Hospital, Boston, MA, for sabbatical training of C. R o v a i n e n

on normal controls. A n extended examination of the interrelations of angiogenesis and neural develop-

and for the initial discussions of the amphibian

m e n t is a n o t h e r of the opportunities for future

tadpole as a preparation for angiogenesis. We are

investigation of this system.

also grateful to Drs. Dale Purves, Thomas Woolsey,

ACKNOWLEDGEMENTS

and Philip D. Stahl for their e n c o u r a g e m e n t and support of this project, to Dr. Paul Grobstein for

This work was supported by start-up grants from the M c D o n n e l l C e n t e r for Cellular and Molecular

suggesting the albino strain of X e n o p u s , and to Ms. Mary Ellen Spence for care of tadpoles and measurements of capillaries.

REFERENCES 1 Arthur, EE., Shivers, R.R. and Bowman, P.D., Astrocytemediated induction of tight junctions in brain capillary endothelium: an efficient in vitro model, Dev. Brain Res., 36 (1987) 155-159. 2 Ausprunk, D.H. and Folkman, J., Migration and proliferation of endothelial cells in preformed and newly formed blood vessels during tumor angiogenesis, Microvasc. Res., 14 (1977) 53-65. 3 B~ir, T., The vascular system of the cerebral cortex, Adv. Anat. Embryol. Cell Biol., 59 (1980) 1-60. 4 Becker, E.L., Cooper, R.G. and Hataway, G.D., Capillary vascularization in puppies born at a simulated altitude of 20,000 feet, J. Appl. Physiol., 8 (1955) 166-168. 5 Brhnemark, P.-I. and Jonsson, I., Determination of the velocity of corpuscles in blood capillaries. A flying spot device, Biorheology, 1 (1963) 143-146. 6 Clark, E.R., Studies on the growth of blood-vessels in the tail of the frog larva - - by observation and experiment on the living animal, Am. J. Anat., 23 (1918) 37-88. 7 Clark, E.R. and Clark, E.L., Microscopic observations on the growth of blood capillaries in the living mammal, Am. J. Anat., 64 (1939) 251-301. 8 Coffin, J.D. and Poole, T.J., Embryonic vascular development: immunohistochemical identification of the origin and subsequent morphogenesis of the major vessel primordia in quail embryos, Development, 102 (1988) 735-748. 9 Connolly, S.E., Hores, T.A., Smith, L.E.H. and D'Amore, P.A., Characterization of vascular development in the mouse retina, Microvasc. Res., 36 (1988) 275-290. 10 Constantine-Paton, M., A neural pattern unfolding: properties of retinotectal differentiation in frog tadpoles. In E.D. Pollack and H.D. Bibb (Eds.), Developmental Neurobiology of the Frog, Liss, New York, 1988, pp. 231-253. 11 D'Amore, P.A. and Braunhut, S.J., Stimulatory and inhibitory factors in vascular growth control. In U.S. Ryan (Ed.), Endothelial Cells, Vol. 2, CRC Press, Boca Raton, FL, 1988, pp. 13-36. 12 Davies, P.F., Remuzzi, A., Gordon, E.J., Dewey, C.F., Jr.

and Gimbrone, M.A., Jr., Turbulent fluid shear stress induces vascular endothelial cell turnover in vitro, Proc. Natl. Acad. Sci. U.S.A., 83 (1986) 2114-2117. 13 Dale, L. and Slack, J.M.W., Fate map for the 32-cell stage of Xenopus laevis, Development, 99 (1987) 527-551. 14 Diemer, K., Capillarisation and oxygen supply of the brain. In D.-W. Liibbers, U.C. Luft, G. Thews and E. Witzleb (Eds.), Oxygen Transport in Blood and Tissue, Thieme, Stuttgart, 1968, pp. 118-123. 15 Evans, H.M., On the development of the aortae, cardinal and umbilical veins, and the other blood vessels of vertebrate embryos from capillaries, Anat. Rec., 3 (1909) 498-518. 16 Fenstermacher, J.D. and Rapoport, S.I., Blood-brain barrier. In Handbook of Physiology, Section 2, The Cardiovascular System, Vol. 4, American Physiological Society, Bethesda, MD, 1984, pp. 969-1000. 17 Folkman, J., Haudenschild, C.C. and Zetter, B.R., Longterm culture of capillary endothelial cells, Proc. Natl. Acad. Sci. U.S.A., 76 (1979) 5217-5221. 18 Folkman, J. and Klagsbrun, M., Angiogenic factors, Science, 235 (1987) 442-447. 19 Fraser, S.E., Late LEO: a new system for the study of neuroplasticity in Xenopus. In R.D. Freeman (Ed.), Developmental Neurobiology of Vision, Plenum, New York, 1978, pp. 319-329. 20 Gaze, R.M., Keating, M.J. and Chung, S.H., The evolution of the retinotectal map during development in Xenopus, Proc. R. Soc. Lond. B, 185 (1974) 301-330. 21 Glaser B.M., D'Amore, P.A., Michels, R.G., Patz, A. and Fenselau, A., Demonstration of vasoproliferative activity from mammalian retina, J. Cell Biol., 84 (1980) 298-304. 22 Godement, P., Vanselow, J., Thanos, S. and Bonhoeffer, E, A study in developing visual systems with a new method of staining neurones and their processes in fixed tissue, Development, 101 (1987) 697-713. 23 Gospodarowicz, D., Neufeld, G. and Schweigerer, L., Fibroblast growth factor: structural and biological properties, J. Cell Physiol., Suppl.5 (1987) 15-26. 24 Gross, J.L. Moscatelli, D. and Rifkin, D.B., Increased capillary endothelial cell protease activity in response to

213 angiogenic stimuli in vitro, Proc. Natl. Acad. Sci. U.S.A., 80 (1983) 2623-2627. 25 Holt, C.E. and Harris, W.A., Order in the initial retinotectal map in Xenopus: a new technique for labelling growing nerve fibres, Nature (Lond,), 301 (1983) 150-152. 26 Hoskins, S.G. and Grobstein, P., Development of the ipsilateral retinothalamic projection in the frog Xenopus laevis. I. Retinal distribution of ipsilateraily projecting cells in normal and experimentally manipulated frogs, J. Neurosci., 5 (1985) 911-919. 27 Hudlick~i, O. and Tyler, K.R., Angiogenesis. The Growth of the Vascular System, Academic Press, London, 1986, pp. 1-221. 28 Janzer, R.C. and Raft, M.C., Astrocytes induce bloodbrain barrier properties in endothelial cells, Nature (Lond.), 325 (1987) 253-257. 29 Kimmel, C.B. and Warga, R.W., Tissue-specific cell lineages originate in the gastrula of the zebrafish, Science, 231 (1986) 365-368. 30 Kollros, J.J., Toward an understanding of tectal development in frogs. In E.D. Pollack and H.D. Bibb (Eds.), Developmental Neurobiology of the Frog, Liss, New York, 1988, pp. 207-229. 31 Labastie, M.-C., Poole, T.J., P6ault, B.M. and LeDouarin, N.M., MB 1, a quail leukocyte-endothelium antigen: partial characterization of the cell surface and secreted forms in cultured endothelial cells, Proc. Natl. Acad. Sci. U.S.A., 83 (1986) 9016-9020. 32 Merker, H. and Opitz, E., Die Gef/iss der Pia mater h6henangepasster Kaninchen, Pflagers Arch., 251 (1949) 117-122. 33 Mercker, H. and Schneider, M., Uber Capillarver/inderungen des Gehirns bei H6henanpassung, Pfliigers Arch., 251 (1949) 49-55. 34 Miller, A.T., Jr. and Hale, D.M., Increased vascularity of brain, heart, and skeletal muscle of polycythemic rats, Am. J. Physiol., 219 (1970) 702-704. 35 Montesano, R., Vassalli, J.D., Baird, A., Guillemin, R. and Orci, L., Basic fibroblast growth factor induces angiogenesis in vitro, Proc. Natl. Acad. Sci. U.S.A., 83 (1986) 7297-7301. 36 Moody, S.A., Fates of the blastomeres of the 32-ceil-stage Xenopus embryo, Dev. Biol., 122 (1987) 300-319. 37 Nieuwkoop, P.D. and Faber, J., Normal Table of Xenopus laevis (Daudin), North Holland, Amsterdam, 1956. 38 Olesen, S.-P. and Crone, C., Substances that rapidly augment ionic conductance of endothelium in cerebral venules, Acta Physiol. Scand., 127 (1986) 233-241.

39 Pardanaud, L., Altman, C., Kitos, P., Dieterlen-Lievre, E and Buck, C.A., Vasculogenesis in the early quail blastodisc as studied with a monoclonal antibody recognizing endothelial cells, Development, 100 (1987) 339-349. 40 Pardridge, W.M., Recent advances in blood-brain barrier transport, Annu. Rev. Pharmacol. Toxicol., 28 (1988) 25-39. 41 Peault, B.M., Thiery, J.P. and LeDouarin, N.M., Surface marker for hemopoietic and,endothelial cell lineages in quail that is defined by a monoclonal antibody, Proc. Natl. Acad. Sci. U.S.A., 80 (1983) 2976-2980. 42 Risau, W., Developing brain produces an angiogenesis factor, Proc. Natl. Acad. Sci. U.S.A., 83 (1986) 3855-3859. 43 Risau, W., Hallman, R. and Albrecht, U., Differentiationdependent expression of proteins in brain endothelium during development of the blood-brain barrier, Dev. Biol., 117 (1986) 537-545. 44 Robertson, P.L., Du Bois, M., Bowman, P.D. and Goldstein, G.W., Angiogenesis in developing rat brain: an in vivo and in vitro study, Dev. Brain Res., 23 (1985) 219-223. 45 Roncali, L., Ribatti, D. and Ambrosi, G., Vasculogenesis in the chick embryo optic tectum, Acta Anat., 122 (1985) 229-234. 46 Stewart, P.A. and Hayakawa, E.M., Interendothelial junctional changes underlie the developmental 'tightening' of the blood-brain barrier, Dev. Brain Res., 32 (1987) • 271-281. 47 Stewart, EA. and Wiley, M.J., Developing nervous tissue induces formation of blood-brain barrier characteristics in invading endothelial cells: a study using quail-chick transplantation chimeras, Dev. Biol., 84 (1981) 183-192. 48 Straznicky, K. and Gaze, R.M., The development of the tectum in Xenopus laevis: an autoradiographic study, J. Embryol. Exp. Morphol., 28 (1972) 87-115. 49 Tao-Cheng, J.H., Nagy, Z. and Brightman, M.W., Tight junctions of brain endothelium in vitro are enhanced by astroglia, J. Neurosci., 7 (1987) 3293-3299. 50 Thoma, R., IJber die Histomechanik des Gaf~isssystems und die Pathogenese der Angiosklerose, Virchows Arch. Pathol. Anat., 204 (1911) 1-74. 51 Wolff, J.R. and B/ir, T., 'Seamless' endothelia in brain capillaries during development of the rat's cerebral cortex, Brain Res., 41 (1972) 17-24. 52 Zetter, B.R., Migration of capillary endothelial cells is stimulated by tumour-derived factors, Nature (Lond.), 285 (1980) 41-43.