- Email: [email protected]
Distribution of sympathetic neuroeffector junctions in the juxtaglomerular region of the rabbit kidney
Journal of the Autonomic Nervous System, 40 (1992) 239-252l © 1992 Elsevier Science Publishers B.V. All rights reserved 0165-1838/92/$05.00
239
JANS 01319
Distribution of sympathetic neuroeffector junctions in the juxtaglomerular region of the rabbit kidney S u s a n E. L u f f
a,
Sandra G. Hengstberger
a, E l s p e t h M . M c L a c h l a n
b and W.P. Anderson
a
a Sir Thomas Ramsay Electron Microscopy Laboratory, Baker Medical Research Institute, Prahran, Victoria, Australia, and b Department of Physiology and Pharmacology, Unit,ersity of Queensland, Queensland, Australia (Received 11 March 1992) (Revision received and accepted 6 July 1992)
Key words: I n n e r v a t i o n ; K i d n e y cortex; U l t r a s t r u c t u r e
Abstract Two structurally distinct types of sympathetic axon (Type I and Type II) have recently been identified in the renal cortex of the rat and the rabbit. This study describes the distribution and density of the neuroeffector junctions made by these two types of axon on the different tissues from the juxtaglomerular region of the rabbit renal cortex, lmmunohistochemical studies showed that tyrosine hydroxylase-positive axons were located only in regions adjacent to the arteries and arterioles in the renal cortex. Ultrastructural studies of the juxtaglomerular region indicated that both types of axon formed junctions on vascular smooth muscle cells, epithelial cells of proximal tubules and renin-secreting granular epithelioid cells. The density of neuromuscular junctions (18 × 103/mm 2 of vessel surface) was more than twice as high on the afferent arteriole as on the efferent arteriole or proximal tubules immediately adjacent to the glomerular arterioles (both about 6 × 103/ram2). The junction density on granular epithelioid cells was much lower (about 2 × 103/mm 2) and were rarely observed on the distal tubule. Afferent arterioles preferentially received junctions from Type I axons at a relatively high density (14.2 × 103/mm 2) whereas junctions formed by Type II axons were less selectively distributed and occurred at lower densities on all other tissues (range, 1-6.3 × 103/mm2). Presynaptic membrane specialisations were identified only at junctions on arterioles and granular epithelioid cells and occurred more frequently at Type I than at Type II junctions. The data suggest that the predominant effect of the sympathetic innervation in the juxtaglomerular region of the renal cortex is on the afferent arteriole and that the two axon types within the kidney may have different functions.
Introduction W e have r e c e n t l y d e s c r i b e d two distinct structural types o f s y m p a t h e t i c p o s t g a n g l i o n i c axon (T ype I an d T y p e II). T h e s e axons f o r m j u n c t i o n s with t h e s m o o t h m u s c l e cells o f t h e i n t r a l o b u l a r a r t e r i e s an d t h e a f f e r e n t a n d e f f e r e n t juxtag l o m e r u l a r a r t e r i o l e s in b o t h rabbit a n d rat kidney [15]. B o t h types o f axon a r e n o r a d r e n e r g i c a nd have a n u m b e r o f m o r p h o l o g i c a l f e a t u r e s
Correspondence to: S.E. Luff, Baker Medical Research Institute, Commercial Road, Prahran, Vic. 3181, Australia.
that are distinct f r o m t h o se axons i n n e r v a t i n g vessels in o t h e r tissues [13,14,10]. T y p e I axons are typically l a r g e r t h a n T y p e II axons ( m e a n m i n i m u m intervaricosity d i a m e t e r 0.36 ~ m and 0.08 ~ m , respectively). T h e varicosities o f T y p e I axons are n o t w e l l - d e f i n e d , b e i n g only slight exp a n s i o n s of o n e side of the axon c o n t a i n i n g small (approx. 50 n m d i a m e t e r ) synaptic vesicles aggreg a t e d t o w ar d s t h e p r e j u n c t i o n a l m e m b r a n e . T h e r e m a i n i n g p o r t i o n o f t h e axon is filled with microt u b u l es an d m i c r o f i l a m e n t s that are c o n t i n u o u s into t h e a d j a c e n t i n t e r v a r i c o s e regions. T y p e II axons, on t h e o t h e r hand, are similar to t h o se
240
found on other arterial vessels in that they have well defined varicosities which lack clusters of microtubules running through them. However, they differ from other vasoconstrictor axons in that the minimum diameter of the intervaricosities is less (0.05 lzm cf. 0.1 ~ m ) and many of the varicosities are very small (minimum diameter 0.25 ~ m cf. 0.5 Fzm). The possible functional significance of these two types of sympathetic axon in the kidney was not addressed in our initial study. However, it has been proposed that the sympathetic nervous system directly influences a number of different renal mechanisms including selective control of afferent and efferent arteriolar constriction, glomerular filtration rate, renin secretion and tubular reabsorption [5,9]. We have therefore investigated the distribution of these axon types by means of a serial section ultrastructural analysis of the juxtaglomerular region. Our aim was to determine whether the junctions formed by Type I and II axons are differentially distributed between the arteriolar smooth muscle, tubular epithelial cells and renin-secreting granular epithelial cells, Initially we confirmed, using immunohistochemical methods, that the sympathetic innervation of the renal cortex was confined to regions adjacent to the distributing vessels, juxtaglomerular arterioles and the hilus of the glomerulus.
Materials and Methods
Immunohistochemistry Two rabbits were anaesthetised with pentobarbitone (Nembutal 60 m g / k g i.v.) and the kidneys perfused with 1% paraformaldehyde, 1% glutaraldehyde in 0.1 M phosphate buffer (pH 7.3) introduced via the dorsal aorta in a retrograde direction (pressure = 170 mmHg; flow rate = 150 m l / m i n ) see [15]. Transverse segments of kidney (approx. 5 mm thick and 1 cm square) were immersed in the same fixative and left overnight at 4°C. Vibratome sections (70 /xm) were cut from these slices, washed in phosphate buffered saline (PBS, 3 x 10 min) and placed in 50% ethanol for 20 min followed by 10% cyanoborohy-
dride in PBS for 31) min; this was followed by an additional wash in PBS (3 x 11) min). The sections were then incubated at room temperature for 24 h with a monoclonal antibody directed against rat tyrosine hydroxylase (TH), (1:100, Boehringer Mannheim Biochemica, Rohm Haas, PA) diluted in 10% normal horse serum in PBS. After washing again in PBS (3 x 10 min), sections were exposed for 1 h to a biotin conjugated secondary antibody to mouse IgG (Sigma Chemical Company, St. Louis, MO) diluted to 1 in 200, followed by avidin-biotin-horseradish peroxidase complex (1 : 1 0 0 0 - 1 : 1500, Vector Laboratories Inc, Burlingame, CA) for 1 h. After washing in PBS (3 x 10 min), the sections were transferred to (I.5 m g / m l diaminobenzidine for 1(t min, after which hydrogen peroxidc was added (final concentration, 0.075%). Sections were incubated in this mixture for a further 3 rain. The reaction product was intensified with 0.10% osmium tetroxide for 30 rain. Sections were dehydrated in ethanol, infiltrated in Durcupan resin (Polaron Equipment Ltd., Watford, UK), mounted in resin on glass slides and polymerised at 60°C.
Quantification of axon density in different cortical regions The relative densities of the TH-immunoreactive nerve plexus on the afferent and efferent arterioles of outer, mid and inner regions of the cortex were determined as follows. Four of the TH-immunoreacted sections (70 izm thick) were viewed under the light microscope. The cortex was divided into three regions (outer, mid and inner), each equivalent to approximately a third of the total thickness of the cortex. For each glomerulus in which both the afferent and efferent arterioles were visible, the density of the nerve plexus around the afferent arteriole was scored subjectively as being greater than, equal to, or less than that around the efferent arteriole.
Electron microscopy Kidneys from three additional rabbits were fixed by perfusion as described above but using 2% paraformaldehyde, 2% glutaraldehyde in 0.1 M sodium cacodylate buffer (pH 7.3). Pieces of tissue from the outer cortex (approx. 8 mm 3)
241 were prepared for electron microscopy by immersing them in the initial fixative for a further 2 h and then post-fixing in 1% osmium tetroxide in the same buffer for 4 h prior to dehydration and embedding in Epon [13]. Semi-thin sections were cut and viewed in the light microscope to select areas of tissue for analysis (see below). Serial thin sections (approx. 0.8 x 0.8 mm; 100 nm thick) were cut of the selected areas and mounted on Formvar-coated slot grids and subsequently examined in the electron microscope.
Tissue sampling Areas of tissue (approx. 800 /zm X 800 /xm) were selected for serial sectioning and analysis in which:
A /
#
(i) afferent and efferent arterioles were cut longitudinally through the mid region of the vessel and through the region where they entered the glomerulus. Series of sections were analysed from three afferent and two efferent arterioles. The length of vessel analysed ranged from 75-158 /zm, the thickness of tissue from 2.3-5.0/xm. The total surface area of the afferent and efferent arterioles analysed was 2400/zm 2 and 2692/zm 2 respectively; (ii) regions of macula densa were studied in two series of sections containing a total of four glomeruli. The thicknesses of tissue analysed in each series were 4.0 and 5 . 6 / z m respectively and the total surface area of macula densa analysed was 1068/xm2;
EFFERENT
- ~ " "
'
ARTERIOLE
~ -
~
x
/ I
I
GLOMERULUS
/ I
AFFERENT ARTERIOLE
/ /q
r,:
I
t
t
t I
lOOpm
/ /
I I
Fig. 1. A: Camera lucida drawing of a single glomerulus showing the arrangement of TH-immunoreactive sympathetic axons around a segment of an interlobular artery, an afferent arteriole, forming a ring around the hilus of the glomerulus (arrows) and along the efferent arteriole. B: Light micrograph of the same glomerulus.
242
(iii) renin secreting granular epithelioid cells were identified in series of sections of six different afferent arterioles (three from (i) above). The length of these series ranged from 2.3-5.5 tzm. The total surface area of renin-containing cells analysed was 3063 ixmZ; and (iv) proximal and distal tubules adjacent to the afferent and efferent arterioles identified in (i) were studied plus one additional series of sections of tubules adjacent to another afferent arteriole (tissue thickness = 3.4 ~m). The total surface area of proximal and distal tubules analysed was 5340 izm 2 and 1650/xm 2 respectively. A ?ialysis
In each series, the whole section area was scanned (approx. 6400 lzm 2) but the analysis concentrated on the specific regions in question. The location of all axons identified in each series of sections was noted on low power electron micrographs, and each axon was followed through sequential sections. In some cases, each profile of a particular axon was photographed and the analysis conducted from the photographs so as to obtain measurements of contact area and terminal size. In other cases, the axon terminals were simply followed through the series of sections in the microscope and representative photographs taken of all junctions. The axons were identified as either Type I or Type 1I and the numbers of junctions of each type counted. Junction density was calculated by dividing the number of junctions by the surface area of tissue scanned in each analysis (i.e. number of junctions per mm 2 of tissue surface). Surface areas of arterioles, renin-secreting ceils, proximal and distal tubule cells were measured from low power electron micrographs.
D a t a are throughout.
presented
as
m e a n + S.E.M.
Statistical test The A'2 test was used to determine the significance of differences in the distribution of junctions between tissues. Differences were tested against those predicted, assuming an even distribution of junctional density.
Results lmmunohistochemistry TH-positive axons in the rabbit renal cortex were associated with the arterial vessels around which they formed a branching and anastomosing network. The axons entered the kidney at the hilus and were present around the arcuate arteries, interlobular and intralobular arteries and the juxtaglomerular arterioles. Axons around the afferent arteriole extended around the hilus of the glomerulus in two distinct tracts to reach the efferent arteriole. The axons did not appear to penetrate the lacis (Fig. 1). The plexus of axons around the efferent arteriole extended to the vasa recta region (in the case of the inner cortical glomeruli), where there were also networks of TH-positive axons. Some of the axons accompanying the efferent arterioles of outer cortical glomeruli extended to the outer surface of the kidney (see also [22]). The afferent arterioles (approximately 20-25 txm diameter) bore a denser axonal plexus than did the efferent arterioles (12-15 p~m diameter, Fig. 1). This was true irrespective of the location of the glomerulus in the cortex and was confirmed in an analysis of four 70-/~m thick sections
Fig. 2. A: Electron micrograph showing the typical features of a Type I axon; the varicose region consists of a slight swelling of the axon containing mitochondria and small synaptic vesicles which are confined to a region facing the smooth muscle cell, Several microtubules run continuously through between intervaricosity regions of the axon. The axon is partially surrounded by a Schwann cell. The axon does not form a junction in this section. B: Electron micrograph of a junction formed by a Type I axon on an afferent arteriole smooth muscle cell showing a presynaptic membrane specialisation (arrows) associated with a cluster of synaptic vesicles. C: Electron micrograph of a junction formed by a Type I1 axon showing a varicosity filled with synaptic vesicles and no visible microtubules. The axon has a narrow intervaricosity containing microtubules. This junction with a smooth muscle cell of an efferent arteriole has two regions of presynaptic membrane specialisation (arrows). s = synaptic vesicles; sch = Schwann cell; sm = smooth muscle cell; v = varicosity; mt = microtubules. Calibration bar = 1 p m throughout.
244
in which a comparison was made of the relative densities of the TH-positive plexus around the afferent and efferent arterioles of glomeruli in the outer, mid and inner regions of the renal cortex. The innervation density on the two arterioles was compared subjectively in material as illustrated in Fig. 1. The plexus was denser on the afferent than the efferent arteriole in 2 4 / 2 6 cases in the outer cortex and in all of 21 and nine cases in the mid and inner cortical regions. In the two exceptions the density of the plexus was similar, i.e. there was no instance of the neural plexus being greater on the efferent compared to the afferent arteriole. In addition, there was no obvious gradient in innervation density along the length of either vessel but some particularly large varicosities or clusters of varicosities lay near the junction of the arterioles with the glomerulus (Fig. 1). Occasionally, TH-positive axons branched away from the plexus around either the afferent or efferent arterioles into the tubule tissue (Fig. 1). However, they extended for very short distances ranging from 6.8-28.4/xm (mean = 15.8 _+ 1.7/.~m, n = 16) before they petered out. Thus it was apparent from the pattern of TH-immunoreactivity that the catecholaminergic innervation is very restricted in the outer cortex, so that the possibility of sympathetic actions beyond the regions near the vasculature can be excluded.
Ultrastructural features of neuroeffector junctions Axons occurred either singly or in small bundles ( 1 - 6 axons). The axon bundles were always surrounded by Schwann cell, whereas a proportion of the single axons had no Schwann cell or other support cell wrapped around them (see [15]). The distribution of the axons around the afferent and efferent arterioles are compared in Table If. Axons of both Type I and Type II were identified according to the criteria described in detail previously [15]. The distinctive features of these two types of axons are illustrated in Fig. 2. Both Type I and Type II axons formed neuroeffector junctions (Fig. 2B, C). These were present on most of the different tissues in this region (see below). A junction was defined as an axon
swelling (varicosity) containing vesicles and mitochondria, lying closely apposed (approx. 50 nm) to the basal lamina of an effector cell. The junctional cleft always contained a single layer of basal lamina. At all junctions, small (approx. 50 nm diameter) synaptic vesicles, some with electron dense granules, were clustered towards the presynaptic m e m b r a n e (see [12]). Large (approx. 100 nm diameter) dense-cored vesicles were also present in both types of axon terminal. Within the region of the arterioles, most varicosities (78%) formed junctions with either smooth muscle cells, renin-secreting granular epithelioid cells, proximal tubule cells or the parietal layer of the Bowman's capsule. Most of the varicosities that did not form junctions were small and contained few synaptic vesicles (see [13]). Presynaptic m e m b r a n e specialisations in the form of an electron dense thickening of a small region of the presynapfic m e m b r a n e (Fig. 2B, C) (see [13]) were identified in some junctions of both types. These structures were always confirmed to be present in two or .more successive sections and were associated with a small cluster of vesicles. Occasionally, there was more than one presynaptic membrane specialization at a junctional region (Fig. 2C). Only a proportion (16%) of junctions bore these specialisations which occurred more frequently at Type I than Type II junctions (see later). In the Type II axons they were more common in the larger varicosities (i.e. > 0.5/xm in diameter).
Distribution and density of neuroeffector junctions The distribution of neuroeffector junctions on different tissues is summarized in Table I.
TABLE I
Distribution of Type I and H junctions on different tissues Tissue
Surface area
Numbers of junctions
scanned p~mz Total Type I Type II Afferent arteriole Efferent arteriole Proximal tubule Granular epithelioid cells
2400 2692 5 340
42 19 28
28 2 9
14 17 19
3063
l0
7
3
245
(i) Arterioles. Junctions were readily identified on both the afferent and efferent arterioles and were often, but not exclusively, located over the cleft between two smooth muscle ceils, sometimes forming junctions with both ceils (Fig. 3). Some muscle cells received more than one neuromuscular junction from either the same or different axons (see Fig. 3). In one afferent arteriole, along which a single axon could be followed for tens of micrometers, each smooth muscle cell was found to be innervated by at least one junction [15]. However, there was no specific distribution of junctions along equivalent lengths (up to 180 ~m) of either arteriole in relation to the glomerulus (Fig. 4). The numerical data describing the innervation of afferent and efferent arterioles are sum-
marised in Tables II and III. About 70% of varicosities associated with each of the afferent and efferent arterioles, formed junctions (Table H). Mean junctional density was about three times greater for the afferent compared to the efferent arteriole (Table III). Further, the density of Type I junctions on the afferent arteriole was higher (Table II) than on the efferent arteriole (X 2 test P < 0.025). In contrast, the Type I junctions occurred at very low densities on efferent arterioles which were predominantly innervated by Type II axons (see Fig. 4). However, the density of Type II junctions was similar on both vessels (Table II). The contact areas of some junctions on the smooth muscle cells of the afferent arteriole were slightly larger than those on the efferent arteriole (see [15], Table 2) although on average they were
Fig. 3. A: E l e c t r o n m i c r o g r a p h of two Type II axons with j u n c t i o n s l o c a t e d b e t w e e n two s m o o t h m u s c l e cells. B: E n l a r g e m e n t of box in A. sm = s m o o t h m u s c l e cell; v = varicosity. C a l i b r a t i o n bars = 1 # m .
]!
246 A
43_
2-
'
II
O4111 0
i
r
i
i
~
J
i
i
B
i
4-
1. 0
i
I i
~
J
i
i
i
i
i
I
~l
q..
o
4-
E
Z•-z 3"
3 iHinil'
2-
2-
1-
0
i
0
i
i
40
i
i
80
Distance from
i
i
120
i
i
i
160
glomerulus(gm)
0
1111
1) ' 4~) ' 80 ' 120 ' 1(~0 ' Distance from glomerulus (Ixm)
Fig. 4. Distribution of junctions with distance from the glomerulus along three afferent (A) and two efferent arterioles (B). Hatched columns = Type 1 axon junctions; open columns = Type 1I axon junctions. Arrow heads = length of vessel analysed.
not significantly different (Table III). However, presynaptic membrane specialisations occurred twice as frequently in junctions on the afferent compared to the efferent arteriole and the percentage was higher in Type I than in Type II axon junctions (Table II). (ii) Granular epithelioid cells. There were only one or two granular epithelioid cells containing renin granules per afferent arteriole in these series of sections. These cells were always located at the junction of the arteriole with the glomerulus and were never encountered in efferent arterioles. The junctions on these cells, like those on the smooth muscle cells, were restricted to the adventitial surface.
Six axons (three Type I and three Type II) formed a total of ten junctions on granular epithelioid cells (e.g. Fig. 5D, see Table I). Although there was considerable variation between the different series analysed, the mean junction density (see Table III) was significantly lower than the density on arteriolar smooth muscle cells (X 2 test P < 0.025). The junction density was higher for Type I than Type II axons (see Fig. 6, Table I). The areas of junctional contact were similar to that in other tissues (see Table 111). Presynaptic m e m b r a n e specialisations were found in two of the junctions, one formed by a Type I axon and the other by a Type II axon.
247 T A B L E I1
Ultrastuctural details o f the innervation o f afferent and efferent arterioles by Type I and H axons Innervation Axons % of Type | axons (of total) % of Type II axons (of total)
Afferent arteriole
Efferent arteriole
62 29
38 71
% of axons that occured singly: Total (% all axons) Type I (% all single axons) Type II (% all single axons)
30 56 44
17 14 86
% axons without support cells
17
3
Mean varicosity volume, ~tm 3 +_S.E.M. Junctions % total varicosities that form junctions * Mean junction density: × 1 0 3 / m m 2 _+S.E.M. Type I Type II
0.52 +_0.25 n= 8
70
14.2 +_4.4 5.8+_ 1.1
0.36_+ 0.13 n= 8
69
0.7, 0.8 4.1, 8.2
% junctions with presynaptic m e m b r a n e specialisations: Total 21 10 Type l 25 0 Type lI 14 12 * Data obtained from 3 afferent and two efferent arterioles (see Materials and Methods).
(iii) Proximal tubules. The innervation of the proximal tubules was determined in the same series of sections as for the afferent and efferent arteriole. Junctions were only observed on the tubule surface facing the arterioles. In the six series, five Type I and nine Type II axons formed a total of 28 junctions (nine Type I and 19 Type II) on proximal tubule epithelial cells (e.g. Fig.
5A). The density of junctions varied considerably between the different series of sections analysed. In one series, we found only Type I junctions; in two others, only Type II junctions. In another, there were approximately equal numbers of both types of junction, whereas in two others no junctions were observed. Overall, the mean density of junctions formed by both axon types was lower than on arteriolar smooth muscle cells (X 2 test P < 0.025) and was similar on tubules close to both afferent and efferent arterioles (7 and 5 x 103/ram 2 respectively). However, the mean density of Type II junctions was higher than Type I junctions and was similar to that on the efferent arteriole (Table I, Fig. 6). The areas of junctional contact were similar to that in other tissues (see Table III); however, presynaptic membrane specialisations were not observed at any of these junctions. (iv) Bowman's capsule. Of the five glomeruli that were analysed for the innervation of the
T A B L E llI
Junction density and contact area on the different jurtaglomerular tissues Tissue
* Junction density Mean + S.E.M.
Range
M e a n + S.E.M.
Range
Afferent arteriole
20.0 + 4.7 n=3 n=2 2.25 4- 2.24 n=6 4.7 +_ 2.8 n=6
10.6-25.3
0.48 _+ 0.32 n=4 0.38 + 0.05 n=6 0.42 _+ 0.13 n=9 0.43 + 0.02 n=15
0.02-1.19
Efferent arteriole Granular epithelioid cells Proximal tubules
Junction contact area ~ m 2
4.48, 8.8 0-10.6 0-16.4
* Junction density expressed as n u m b e r of junctions X 1 0 3 / m m 2 tissue surface analysed.
0.25-0.57 0.13-1.04 0.06-1.08
248
Fig. 5. J u n c t i o n s on effectors o t h e r than smooth muscle. A: Type II junction on a proximal t u b u l e e p i t h e l i a l cell. B: Type I junction on a distal tubule e p i t h e l i a l cell. C: Type 11 j u n c t i o n on m a c u l a d e n s a cell. D: Type II junction on g r a n u l a r e p i t h e l i o i d cell. E: Type I1 junction on a cell forming B o w m a n ' s capsule. A r r o w s = p r e s y n a p t i c m e m b r a n e specialisations; v = varicosity; p t - proximal tubule e p i t h e l i a l cell; dt = distal t u b u l e cell; r = renin granule; bc = B o w m a n ' s capsule cell; md = m a c u l a d e n s a cell; sch = Schwann cell. C a l i b r a t i o n bar = 1 /xm applies t h r o u g h o u t .
249 20 16 12 0 t~
8
•m•
4 o
:~
20
'~ 16
"E 12
"~ ~
Type I axons
o 2o
Type II axons
0 arterlolar ~proximel renln smooth tubule secreting muscle n=6 cells n=5 n=6 Fig. 6. Densities of neuroeffector junctions on different tissues in the outer renal cortex. Note that smooth muscle cells include afferent and efferent arterioles (see Table II for details). * Values for the proximal tubule cells are for those tubules located adjacent to afferent and efferent arterioles. Renin-secreting cells (granular epithelioid cells) are cells of the afferent arterioles, n = number of series analysed. Error bars = S.E.M.
arterioles, three Type II axons were found to form a total of five junctions on the epithelial cells that form the parietal layer of the Bowman's capsule (e.g. Fig. 5E). A presynaptic m e m b r a n e specialisation was identified at one of these junctions. (v) Distal tubules and macula densa. No junctions were found on the epithelial cells of the distal tubules in any of the detailed analyses. However, two junctions, adjacent to each other
on the same tubule cell, were found close to the macula densa region in a series of sections cut specifically to study this region in greater detail (see below). Both axons were Type I and were in the same axon bundle (see Fig. 5B). T h e r e were no presynaptic m e m b r a n e specialisations at either of these junctions. The axons located near the macula densa occur as a bundle that rings the lacis (see Fig. 1), so that it was not possible to locate these axons in sections cut longitudinally through the arterioles. Maculae densae were therefore studied associated with four additional glomeruli (see Materials and Methods). The plane of section through the macula densa was selected at random. Seven axons (two Type I, five Type II) were found lying within 8 / x m of the specialised distal tubular cells at the outer edge of three of the four maculae densae. Two axons, both Type II, formed junctions on the basal lamina overlying the macula densa cells (e.g. Fig. 5C). It should be noted that macula densa cells have a very convoluted plasm a l e m m a and the basal lamina often lay some distance away from the cell membrane. Hence, the separation between the m e m b r a n e s of the axon and the macula densa cell at both junctions was significantly greater than at the other junctions described above (see Fig. 5C). There were no presynaptic m e m b r a n e specialisations at either of these junctions. (vi) Mesangial cells. No axons were observed close to the extraglomerular mesangial cells in any of the series of sections studied. This was consistent with the finding that no TH-immunoreactive axons extended into this region (see above). (vii) Axons contacting more than one cell type. One axon was found to form junctions sequentially with an afferent arteriole smooth muscle cell, a renin-secreting granular epithelioid cell and the surface of Bowman's capsule. No other instance was found in which an axon innervated more than one type of tissue. Serial sections were cut of an additional 4.8 /zm thickness of tissue following on from a series of sections of an afferent arteriole in which we had identified three axons that formed junctions with proximal tubule ceils (i.e. total thickness of tissue analysed was 9.8
25O >m). None of these axons formed junctions with arteriolar smooth muscle cells.
Discussion This paper reports the distribution of specialised neuroeffector junctions in the outer renal cortex of the rabbit formed by the two types of sympathetic axon recently identified by us in both the rat and rabbit kidney [15]. A neuroeffector junction was identified as a close apposition between a varicosity and an effector cell with the intervening junctional cleft containing a single layer of basal lamina and the synaptic vesicles clustered towards the presynaptic membrane. This structural arrangement was identical to that of neuromuscular junctions in other vascular beds (see [12]). As well as describing the location of junctions on different cell types, we have quantified and compared the densities of junctions made by each type of axon on the different tissues in the region of the glomerulus. This ultrastructural description of the innervation of the juxtagiomerular region in the rabbit considerably expands the reports of others in different species [1,7,8,20]. The vast majority of axons in this region are sympathetic postganglionic noradrenergic axons as demonstrated by the presence of electron dense cores in small-diameter synaptic vesicles, and by their ability to take up 6-hydroxydopamine [15]. Immunohistochemistry of the kidney tissue demonstrated that TH-immunoreactive axons reached the juxtaglomerular region via the arcuate, interlobular and intralobular arteries, and that the axon plexus was denser around the afferent than the efferent arterioles. Although we have concentrated on the ultrastructural studies on the glomeruli in the outer cortex, TH-immunoreactivity associated with deeper glomeruli was distributed in a similar pattern, suggesting that these ultrastructural findings apply throughout the renal cortex. This contrasts with a previous autoradiography study which concluded that there is a decreasing gradient of innervation density of the afferent compared to the efferent arteriole in glomeruli from the capsule to the
medulla in the rat [3]. A more recent report [4] supports our findings of a uniform innervation pattern. Consistent with our other studies [12,13], the majority (78%) of perivascular noradrenergic varicosities formed specialised neuroeffector junctions.
Relati~.,e densities of neuroeffector junctions on different tissues The greatest density of junctions was round on the vascular smooth muscle cells of the afferent arteriole (Table III). Afferent arterioles bore junctions at a density of 20 x 103/mm 3, whereas junctional density on efferent arterioles was less than half this value and was even lower on the other tissues (Table III). These values for density have been determined by reconstructing all junctions within each series and so are not equivalent to the p a r a m e t e r we used previously to define the relative density of junctions on a range of arterial vessels [14]. This latter parameter reflected the frequency with which junctions were found in studies of random sections. It is possible to estimate the density of junctions on the 50 # m diameter submucous arterioles of the guinea pig small intestine by allowing for the average area of individual junctions (1.26 /.tin 3, see [12]). The frequency of junctions in random sections was 109 x 103/mm 2 [14], so that junctional density can be estimated to be about 9 x 1 0 3 / m m 2. This estimate is confirmed by the real junctional density on these arterioles (10 x 103/mm 2) when it is determined in the same way as has been done here for the juxtaglomerular tissues (S.E. Luff, unpublished observations). As the submucous arterioles were the most densely innervated of the arterial vessels previously studied, it seems likely that the afferent arterioles are amongst the most densely innervated vessels in the body. It is, however, notable that mean junctional area on the juxtaglomerular arterioles was smaller ( < 0.5/.~m 3 Table III) than on the submucous arterioles. The neurovascular junctions on the juxtaglomerular arterioles were distributed evenly along the vessels studied (Fig. 4). This contrasts with the distribution of junctions on guinea pig
251
submucous arterioles of equivalent diameter to the juxtaglomerular arterioles (15-25 #m), where the junctions on the former occur clustered at the vessel branch points [13]. The majority of varicosities formed junctions and most smooth muscle cells appeared to be directly innervated as intervaricosity lengths ( < 1.5/zm) were comparable to the diameter of the smooth muscle cells (i.e. 2 / z m ) (see [15]). This result differs from that reported in the rat [2] in which only 20% of the agranular cells of both the afferent and efferent arterioles appeared to be innervated. Neuroeffector junctions were found on effector cells of most other tissue types lying adjacent to the vascular smooth muscle cells but at densities significantly lower than those on the vasculature. The renin-secreting granular epithelioid ceils formed part of the afferent arteriole close to the glomerulus and were absent from the efferent arteriole in the rabbit (cf. rat, [22]). The mean junction density on renin-secreting cells was much lower (Table III) than that found on adjacent vascular smooth muscle cells. Junctions were observed on some proximal tubule epithelial ceils lying adjacent to the afferent and efferent arterioles. The density of junctions was low but similar on tubules close to both afferent and efferent arterioles (7 and 5 x 103/mm 2 respectively). However, as tubule tissue located away from the vasculature was not innervated it must be concluded that the innervation of this tissue is very sparse ([14], see also [11,19]). The density of junctions on the distal tubules and associated macula densa cells was even lower than that found on the proximal tubules. Junctions were identified only in series which concentrated on the macula densa, and only close to its outer edge, consistent with the arrangement of TH-immunoreactive axons. These observations indicate that the density of junctions on the distal tubules in the outer cortex is extremely low. Junctions were also found on the epithelial cells forming Bowman's capsule (see also [8]). One of these was formed by an axon which also innervated the afferent arteriole. It is difficult to envisage what function could be served by these junctions.
The higher density of junctions on the afferent compared to the efferent arteriole is compatible with the physiological effects of renal nerve stimulation [21]. Conversely, the relatively sparse (and patchy) innervation of the proximal tubular cells seems at variance with physiological studies, mainly performed in the rat, in which marked actions of nerves on tubular function have been reported (e.g. see [5,17]). Apart from the direct effects of sympathetic nerves on the renin-secreting cells of the afferent arteriole [5], it has been proposed that the sympathetic innervation also indirectly modulates renin secretion via a direct neural effect on the macula densa [6,18]. However, the functional significance of a highly localised structural arrangement of the innervation of the macula densa tissue is unknown. Similarly, this structural study does not support claims that macula densa ceils of the distal tubule and the glomerular mesangium are richly innervated [16]. In the rabbit we have not been able to find any direct innervation of the mesangium despite extensive searching. Only one instance of a single axon forming junctions with different effectors was observed. The axon concerned formed junctions sequentially with a vascular smooth muscle cell, a reninsecreting cell and with Bowman's capsule. This result contrasts with reports in the rat that axons frequently form contacts with more than one tissue [2].
Distribution of Type I and H junctions The two structural types of axon in the outer renal cortex [15] both formed neuroeffector junctions with the various effector cells, but they were distributed differentially between them. Type I axons contacted afferent arterioles and reninsecreting cells more often than other tissues. The greatest density of Type I axon junctions was on the afferent arteriole, whereas the density on efferent arterioles, renin-secreting cells and juxtaglomerular proximal tubules cells was low (see Table III, Fig. 5). Type II axons, on the other hand, were distributed at similar densities over afferent and efferent arterioles and proximal tubules. The rare junctions formed on the distal
252
tubule cells near the macula densa were made by both types of axon. The low density of Type I axon junctions on the efferent arteriole and other non-vascular effectors indicates that the majority of these axons must terminate at the glomerulus although some extend for some distance along the efferent arteriole (Fig. 4). The difference in the distribution of junctions formed by these two types of axon between tissues suggests different functional roles for the Type I and II axons. Presynaptic membrane specialisations were present at about 16% of junctions; this is a similar proportion to that observed for other arteries [13]. Most presynaptic specialisations were detected at junctions on arterioles. However, specialisations were present at a much higher proportion of Type I than Type II junctions (see Table II). If such structures reflect release sites (as at other junctions and synapses), these findings suggest that the afferent arterioles receive a more effective innervation: (i) because of the high density of Type I junctions on this vessel; and (ii) because of the higher likelihood that these contain specialised release sites. The structural data also strongly support the view that the major functional effect of the innervation is on the vasculature. Previous studies have demonstrated junctional contacts on several effector types in the rat kidney [2,3,8]. Barajas and MiJller [2] also compared innervation density of the different tissues in the juxtaglomerular region. However, it is difficult to compare their data directly with ours as they used a different criterion for estimating innervation density (i.e. the proportion of cells receiving direct innervation). The definition of a 'neuroeffector contact' in both earlier studies included varicosity appositions with effector cells with membrane separations of 300 nm and containing two distinct layers of basal lamina. Because we found the majority of varicosities formed specialised junctions, there is no reason at present to believe that varicosities merely lying 'close' to an effector cell are involved in neurotransmitter release. They concluded, as we do, that the innervation density of the tubule cells was much lower than that of the arterioles, but they found that a higher proportion of granular epithelioid cells received
junctions than did smooth muscle cells. The data are not currently available to compare innervation density between the two species. In conclusion, this study has identified two important aspects of the juxtaglomerular innervation in the cortical region in the rabbit: (i) Type I and Type II noradrenergic axons are differentially distributed between effector ceils, implying distinct functional roles; and (ii) presynaptic membrane specialisations occur more often at Type I junctions on afferent arterioles than at junctions (of either type) on all other tissues. The findings that the Type I axons have a higher proportion of varicosities that form junctions, and have a higher proportion of junctions with presynaptic membrane specialisations than Type II axons, suggest that Type I axons are likely to be more effective in controlling effector function. The distribution of junctions indicates that Type I axons are predominantly vasoconstrictor as they are either absent or present at very low densities on other tissues whereas Type II axons may have a more gene.ralised effect as they are similarly distributed on different tissues. The way in which the sparse innervation found on non-vascular effectors directly influences the function of these tissues remains to be clarified.
Acknowledgements We wish to thank The National Heart Foundation of Australia and the National Health and Medical Research Council for their financial support of this project, and Simone Young for her valuable technical assistance.
References 1 Barajas, L., Innervation of the renal cortex, Fed. Proc., 37 (1978) 1192-1208. 2 Barajas, L. and Miiller, J., The innervation of the juxtaglomerular apparatus and surrounding tubules: a quantitative analysis by serial section electron microscopy, J. Ultrastruct. Res., 3 (1973) 107-132. 3 Barajas, L. and Powers, K.V., Innervation of the thick ascending limb of Henly, Am. J. Physiol., 255 (1988) F340 F348.
253 4 Barajas, L. and Powers; K.V., Monoaminergic innervation of the rat kidney: a quantitative study, Am. J. Physiol., 259 (1990) F503-F511. 5 DiBona, G.F., The functions of renal nerves. Rev. Physiol. Biochem. Pharmacol., 94 (1982) 75-181. 6 DiBona, G.F., Neural control of renal function: cardiovascular implications, Hypertension, 13 (1989) 539-548. 7 Ferguson, M., Ryan, G.B. and Bell, C., The innervation of the renal cortex in the dog: an ultrastructural study. Cell Tissue Res., 253 (1988) 539-546. 8 Gorgas, K., Struktur und innervation des juxtaglomerular apparatus der rat. Adv. Anat. Embryol. Cell Biol., 54 (1978) 5-84. 9 Gottschalk, C.W., Moss, N.G. and Colondres, R.E., Neural control of renal function in health and disease. In D.W. Seldin and G. Griebisch (Eds.), The Kidney, Physiology and Pathophysiology, (Vol. 1), Raven Press, New York, 1985, pp. 581-611. 10 Klemm, M., Luff, S.E. and Van Helden, D.F., Sympathetic varicosities form specialised junctions on mesenteric veins of the guinea pig, Proc. Aust. Physiol. and Pharmacol. Soc., 22 (1991) 147P. 11 Kriz, W. and Kaissling, B., Structural organisation of the mammalian kidney. In D.W. Seldin and G. Griebisch (Eds.), The Kidney: Physiology and Pathophysiology, (Vol. 1), Raven Press, New York, 1985, pp. 265-306. 12 Luff, S.E., McLachlan, E.M. and Hirst, G.D.S., An ultrastructural analysis of the sympathetic neuromuscular junctions on arterioles of the submucosa of the guinea pig ileum, J. Comp. Neurol., 257 (1987) 578-594. 13 Luff, S.E. and McLachlan, E.M., The form of sympathetic postganglionic axons at clustered neuromuscular junctions near branch points of arterioles in the submucosa of the guinea pig ileum, J. Neurocytol., 17 (1988) 451-463.
14 Luff, S.E. and McLachlan, E.M., Frequency of neuromuscular junctions on arteries of different dimensions in the rabbit, guinea pig and rat, Blood Vessels, 26 (1989) 95-106. 15 Luff, S.E., Hengstberger, S.G., McLachlan E.M. and Anderson, W.P., Two types of sympathetic axon innervating the juxtaglomerular arterioles of the rabbit and the rat kidney differ structurally from those supplying other vessels, J. Neurocytol., 20 (1991) 781-795. 16 Maddox, D.A. and Brenner, B.M., GIomerular ultrafiltration. In B.M. Brenner and F.C. Rector Jr. (Eds.), The Kidney, (Vol. 1) W.B. Saunders, Philadelphia. 17 Moss, N.G., Renal function and renal afferent and efferent nerve activity, Am. J. Physiol., 243 (1982) F425-433. 18 Osbourn, J.L., Thames, D. and DiBona, G.F., The role of macula densa in renal nerve modulation of renin secretion, Am. J. Physiol., 242 (1982) R367-R371. 19 Reinecke, M. and Forssmann, W.G., Neuropeptide (neuropeptide Y, neurotensin, vasoactive intestinal polypeptide, substance P, calcitonin gene-related peptide, somatostatin) immunohistochemistry and ultrastructure of renal nerves, Histochemistry, 89 (1988) 1-9. 20 Simpson, F.O. and Devine, C.E., The fine structure of autonomic neuromuscular contacts in the arterioles of the sheep renal cortex. J. Anat., 100 (1966) 127-137. 21 Tucker, B.J., Peterson, O.W., Munger, K.A., Bird, J.E., Mitchell, M., Pelayo, J.C. and Blantz, R.C., Glomerular hemodynamic alterations during renal nerve stimulation in rats on high- and low-salt diets, Am. J. Physiol., 258 (1990) F133-FI43. 22 Taugner, R. and Hackenthal, E., The Juxtaglomerular Apparatus: Structure and Function. Springer-Verlag, Berlin, 1989.
239
JANS 01319
Distribution of sympathetic neuroeffector junctions in the juxtaglomerular region of the rabbit kidney S u s a n E. L u f f
a,
Sandra G. Hengstberger
a, E l s p e t h M . M c L a c h l a n
b and W.P. Anderson
a
a Sir Thomas Ramsay Electron Microscopy Laboratory, Baker Medical Research Institute, Prahran, Victoria, Australia, and b Department of Physiology and Pharmacology, Unit,ersity of Queensland, Queensland, Australia (Received 11 March 1992) (Revision received and accepted 6 July 1992)
Key words: I n n e r v a t i o n ; K i d n e y cortex; U l t r a s t r u c t u r e
Abstract Two structurally distinct types of sympathetic axon (Type I and Type II) have recently been identified in the renal cortex of the rat and the rabbit. This study describes the distribution and density of the neuroeffector junctions made by these two types of axon on the different tissues from the juxtaglomerular region of the rabbit renal cortex, lmmunohistochemical studies showed that tyrosine hydroxylase-positive axons were located only in regions adjacent to the arteries and arterioles in the renal cortex. Ultrastructural studies of the juxtaglomerular region indicated that both types of axon formed junctions on vascular smooth muscle cells, epithelial cells of proximal tubules and renin-secreting granular epithelioid cells. The density of neuromuscular junctions (18 × 103/mm 2 of vessel surface) was more than twice as high on the afferent arteriole as on the efferent arteriole or proximal tubules immediately adjacent to the glomerular arterioles (both about 6 × 103/ram2). The junction density on granular epithelioid cells was much lower (about 2 × 103/mm 2) and were rarely observed on the distal tubule. Afferent arterioles preferentially received junctions from Type I axons at a relatively high density (14.2 × 103/mm 2) whereas junctions formed by Type II axons were less selectively distributed and occurred at lower densities on all other tissues (range, 1-6.3 × 103/mm2). Presynaptic membrane specialisations were identified only at junctions on arterioles and granular epithelioid cells and occurred more frequently at Type I than at Type II junctions. The data suggest that the predominant effect of the sympathetic innervation in the juxtaglomerular region of the renal cortex is on the afferent arteriole and that the two axon types within the kidney may have different functions.
Introduction W e have r e c e n t l y d e s c r i b e d two distinct structural types o f s y m p a t h e t i c p o s t g a n g l i o n i c axon (T ype I an d T y p e II). T h e s e axons f o r m j u n c t i o n s with t h e s m o o t h m u s c l e cells o f t h e i n t r a l o b u l a r a r t e r i e s an d t h e a f f e r e n t a n d e f f e r e n t juxtag l o m e r u l a r a r t e r i o l e s in b o t h rabbit a n d rat kidney [15]. B o t h types o f axon a r e n o r a d r e n e r g i c a nd have a n u m b e r o f m o r p h o l o g i c a l f e a t u r e s
Correspondence to: S.E. Luff, Baker Medical Research Institute, Commercial Road, Prahran, Vic. 3181, Australia.
that are distinct f r o m t h o se axons i n n e r v a t i n g vessels in o t h e r tissues [13,14,10]. T y p e I axons are typically l a r g e r t h a n T y p e II axons ( m e a n m i n i m u m intervaricosity d i a m e t e r 0.36 ~ m and 0.08 ~ m , respectively). T h e varicosities o f T y p e I axons are n o t w e l l - d e f i n e d , b e i n g only slight exp a n s i o n s of o n e side of the axon c o n t a i n i n g small (approx. 50 n m d i a m e t e r ) synaptic vesicles aggreg a t e d t o w ar d s t h e p r e j u n c t i o n a l m e m b r a n e . T h e r e m a i n i n g p o r t i o n o f t h e axon is filled with microt u b u l es an d m i c r o f i l a m e n t s that are c o n t i n u o u s into t h e a d j a c e n t i n t e r v a r i c o s e regions. T y p e II axons, on t h e o t h e r hand, are similar to t h o se
240
found on other arterial vessels in that they have well defined varicosities which lack clusters of microtubules running through them. However, they differ from other vasoconstrictor axons in that the minimum diameter of the intervaricosities is less (0.05 lzm cf. 0.1 ~ m ) and many of the varicosities are very small (minimum diameter 0.25 ~ m cf. 0.5 Fzm). The possible functional significance of these two types of sympathetic axon in the kidney was not addressed in our initial study. However, it has been proposed that the sympathetic nervous system directly influences a number of different renal mechanisms including selective control of afferent and efferent arteriolar constriction, glomerular filtration rate, renin secretion and tubular reabsorption [5,9]. We have therefore investigated the distribution of these axon types by means of a serial section ultrastructural analysis of the juxtaglomerular region. Our aim was to determine whether the junctions formed by Type I and II axons are differentially distributed between the arteriolar smooth muscle, tubular epithelial cells and renin-secreting granular epithelial cells, Initially we confirmed, using immunohistochemical methods, that the sympathetic innervation of the renal cortex was confined to regions adjacent to the distributing vessels, juxtaglomerular arterioles and the hilus of the glomerulus.
Materials and Methods
Immunohistochemistry Two rabbits were anaesthetised with pentobarbitone (Nembutal 60 m g / k g i.v.) and the kidneys perfused with 1% paraformaldehyde, 1% glutaraldehyde in 0.1 M phosphate buffer (pH 7.3) introduced via the dorsal aorta in a retrograde direction (pressure = 170 mmHg; flow rate = 150 m l / m i n ) see [15]. Transverse segments of kidney (approx. 5 mm thick and 1 cm square) were immersed in the same fixative and left overnight at 4°C. Vibratome sections (70 /xm) were cut from these slices, washed in phosphate buffered saline (PBS, 3 x 10 min) and placed in 50% ethanol for 20 min followed by 10% cyanoborohy-
dride in PBS for 31) min; this was followed by an additional wash in PBS (3 x 11) min). The sections were then incubated at room temperature for 24 h with a monoclonal antibody directed against rat tyrosine hydroxylase (TH), (1:100, Boehringer Mannheim Biochemica, Rohm Haas, PA) diluted in 10% normal horse serum in PBS. After washing again in PBS (3 x 10 min), sections were exposed for 1 h to a biotin conjugated secondary antibody to mouse IgG (Sigma Chemical Company, St. Louis, MO) diluted to 1 in 200, followed by avidin-biotin-horseradish peroxidase complex (1 : 1 0 0 0 - 1 : 1500, Vector Laboratories Inc, Burlingame, CA) for 1 h. After washing in PBS (3 x 10 min), the sections were transferred to (I.5 m g / m l diaminobenzidine for 1(t min, after which hydrogen peroxidc was added (final concentration, 0.075%). Sections were incubated in this mixture for a further 3 rain. The reaction product was intensified with 0.10% osmium tetroxide for 30 rain. Sections were dehydrated in ethanol, infiltrated in Durcupan resin (Polaron Equipment Ltd., Watford, UK), mounted in resin on glass slides and polymerised at 60°C.
Quantification of axon density in different cortical regions The relative densities of the TH-immunoreactive nerve plexus on the afferent and efferent arterioles of outer, mid and inner regions of the cortex were determined as follows. Four of the TH-immunoreacted sections (70 izm thick) were viewed under the light microscope. The cortex was divided into three regions (outer, mid and inner), each equivalent to approximately a third of the total thickness of the cortex. For each glomerulus in which both the afferent and efferent arterioles were visible, the density of the nerve plexus around the afferent arteriole was scored subjectively as being greater than, equal to, or less than that around the efferent arteriole.
Electron microscopy Kidneys from three additional rabbits were fixed by perfusion as described above but using 2% paraformaldehyde, 2% glutaraldehyde in 0.1 M sodium cacodylate buffer (pH 7.3). Pieces of tissue from the outer cortex (approx. 8 mm 3)
241 were prepared for electron microscopy by immersing them in the initial fixative for a further 2 h and then post-fixing in 1% osmium tetroxide in the same buffer for 4 h prior to dehydration and embedding in Epon [13]. Semi-thin sections were cut and viewed in the light microscope to select areas of tissue for analysis (see below). Serial thin sections (approx. 0.8 x 0.8 mm; 100 nm thick) were cut of the selected areas and mounted on Formvar-coated slot grids and subsequently examined in the electron microscope.
Tissue sampling Areas of tissue (approx. 800 /zm X 800 /xm) were selected for serial sectioning and analysis in which:
A /
#
(i) afferent and efferent arterioles were cut longitudinally through the mid region of the vessel and through the region where they entered the glomerulus. Series of sections were analysed from three afferent and two efferent arterioles. The length of vessel analysed ranged from 75-158 /zm, the thickness of tissue from 2.3-5.0/xm. The total surface area of the afferent and efferent arterioles analysed was 2400/zm 2 and 2692/zm 2 respectively; (ii) regions of macula densa were studied in two series of sections containing a total of four glomeruli. The thicknesses of tissue analysed in each series were 4.0 and 5 . 6 / z m respectively and the total surface area of macula densa analysed was 1068/xm2;
EFFERENT
- ~ " "
'
ARTERIOLE
~ -
~
x
/ I
I
GLOMERULUS
/ I
AFFERENT ARTERIOLE
/ /q
r,:
I
t
t
t I
lOOpm
/ /
I I
Fig. 1. A: Camera lucida drawing of a single glomerulus showing the arrangement of TH-immunoreactive sympathetic axons around a segment of an interlobular artery, an afferent arteriole, forming a ring around the hilus of the glomerulus (arrows) and along the efferent arteriole. B: Light micrograph of the same glomerulus.
242
(iii) renin secreting granular epithelioid cells were identified in series of sections of six different afferent arterioles (three from (i) above). The length of these series ranged from 2.3-5.5 tzm. The total surface area of renin-containing cells analysed was 3063 ixmZ; and (iv) proximal and distal tubules adjacent to the afferent and efferent arterioles identified in (i) were studied plus one additional series of sections of tubules adjacent to another afferent arteriole (tissue thickness = 3.4 ~m). The total surface area of proximal and distal tubules analysed was 5340 izm 2 and 1650/xm 2 respectively. A ?ialysis
In each series, the whole section area was scanned (approx. 6400 lzm 2) but the analysis concentrated on the specific regions in question. The location of all axons identified in each series of sections was noted on low power electron micrographs, and each axon was followed through sequential sections. In some cases, each profile of a particular axon was photographed and the analysis conducted from the photographs so as to obtain measurements of contact area and terminal size. In other cases, the axon terminals were simply followed through the series of sections in the microscope and representative photographs taken of all junctions. The axons were identified as either Type I or Type 1I and the numbers of junctions of each type counted. Junction density was calculated by dividing the number of junctions by the surface area of tissue scanned in each analysis (i.e. number of junctions per mm 2 of tissue surface). Surface areas of arterioles, renin-secreting ceils, proximal and distal tubule cells were measured from low power electron micrographs.
D a t a are throughout.
presented
as
m e a n + S.E.M.
Statistical test The A'2 test was used to determine the significance of differences in the distribution of junctions between tissues. Differences were tested against those predicted, assuming an even distribution of junctional density.
Results lmmunohistochemistry TH-positive axons in the rabbit renal cortex were associated with the arterial vessels around which they formed a branching and anastomosing network. The axons entered the kidney at the hilus and were present around the arcuate arteries, interlobular and intralobular arteries and the juxtaglomerular arterioles. Axons around the afferent arteriole extended around the hilus of the glomerulus in two distinct tracts to reach the efferent arteriole. The axons did not appear to penetrate the lacis (Fig. 1). The plexus of axons around the efferent arteriole extended to the vasa recta region (in the case of the inner cortical glomeruli), where there were also networks of TH-positive axons. Some of the axons accompanying the efferent arterioles of outer cortical glomeruli extended to the outer surface of the kidney (see also [22]). The afferent arterioles (approximately 20-25 txm diameter) bore a denser axonal plexus than did the efferent arterioles (12-15 p~m diameter, Fig. 1). This was true irrespective of the location of the glomerulus in the cortex and was confirmed in an analysis of four 70-/~m thick sections
Fig. 2. A: Electron micrograph showing the typical features of a Type I axon; the varicose region consists of a slight swelling of the axon containing mitochondria and small synaptic vesicles which are confined to a region facing the smooth muscle cell, Several microtubules run continuously through between intervaricosity regions of the axon. The axon is partially surrounded by a Schwann cell. The axon does not form a junction in this section. B: Electron micrograph of a junction formed by a Type I axon on an afferent arteriole smooth muscle cell showing a presynaptic membrane specialisation (arrows) associated with a cluster of synaptic vesicles. C: Electron micrograph of a junction formed by a Type I1 axon showing a varicosity filled with synaptic vesicles and no visible microtubules. The axon has a narrow intervaricosity containing microtubules. This junction with a smooth muscle cell of an efferent arteriole has two regions of presynaptic membrane specialisation (arrows). s = synaptic vesicles; sch = Schwann cell; sm = smooth muscle cell; v = varicosity; mt = microtubules. Calibration bar = 1 p m throughout.
244
in which a comparison was made of the relative densities of the TH-positive plexus around the afferent and efferent arterioles of glomeruli in the outer, mid and inner regions of the renal cortex. The innervation density on the two arterioles was compared subjectively in material as illustrated in Fig. 1. The plexus was denser on the afferent than the efferent arteriole in 2 4 / 2 6 cases in the outer cortex and in all of 21 and nine cases in the mid and inner cortical regions. In the two exceptions the density of the plexus was similar, i.e. there was no instance of the neural plexus being greater on the efferent compared to the afferent arteriole. In addition, there was no obvious gradient in innervation density along the length of either vessel but some particularly large varicosities or clusters of varicosities lay near the junction of the arterioles with the glomerulus (Fig. 1). Occasionally, TH-positive axons branched away from the plexus around either the afferent or efferent arterioles into the tubule tissue (Fig. 1). However, they extended for very short distances ranging from 6.8-28.4/xm (mean = 15.8 _+ 1.7/.~m, n = 16) before they petered out. Thus it was apparent from the pattern of TH-immunoreactivity that the catecholaminergic innervation is very restricted in the outer cortex, so that the possibility of sympathetic actions beyond the regions near the vasculature can be excluded.
Ultrastructural features of neuroeffector junctions Axons occurred either singly or in small bundles ( 1 - 6 axons). The axon bundles were always surrounded by Schwann cell, whereas a proportion of the single axons had no Schwann cell or other support cell wrapped around them (see [15]). The distribution of the axons around the afferent and efferent arterioles are compared in Table If. Axons of both Type I and Type II were identified according to the criteria described in detail previously [15]. The distinctive features of these two types of axons are illustrated in Fig. 2. Both Type I and Type II axons formed neuroeffector junctions (Fig. 2B, C). These were present on most of the different tissues in this region (see below). A junction was defined as an axon
swelling (varicosity) containing vesicles and mitochondria, lying closely apposed (approx. 50 nm) to the basal lamina of an effector cell. The junctional cleft always contained a single layer of basal lamina. At all junctions, small (approx. 50 nm diameter) synaptic vesicles, some with electron dense granules, were clustered towards the presynaptic m e m b r a n e (see [12]). Large (approx. 100 nm diameter) dense-cored vesicles were also present in both types of axon terminal. Within the region of the arterioles, most varicosities (78%) formed junctions with either smooth muscle cells, renin-secreting granular epithelioid cells, proximal tubule cells or the parietal layer of the Bowman's capsule. Most of the varicosities that did not form junctions were small and contained few synaptic vesicles (see [13]). Presynaptic m e m b r a n e specialisations in the form of an electron dense thickening of a small region of the presynapfic m e m b r a n e (Fig. 2B, C) (see [13]) were identified in some junctions of both types. These structures were always confirmed to be present in two or .more successive sections and were associated with a small cluster of vesicles. Occasionally, there was more than one presynaptic membrane specialization at a junctional region (Fig. 2C). Only a proportion (16%) of junctions bore these specialisations which occurred more frequently at Type I than Type II junctions (see later). In the Type II axons they were more common in the larger varicosities (i.e. > 0.5/xm in diameter).
Distribution and density of neuroeffector junctions The distribution of neuroeffector junctions on different tissues is summarized in Table I.
TABLE I
Distribution of Type I and H junctions on different tissues Tissue
Surface area
Numbers of junctions
scanned p~mz Total Type I Type II Afferent arteriole Efferent arteriole Proximal tubule Granular epithelioid cells
2400 2692 5 340
42 19 28
28 2 9
14 17 19
3063
l0
7
3
245
(i) Arterioles. Junctions were readily identified on both the afferent and efferent arterioles and were often, but not exclusively, located over the cleft between two smooth muscle ceils, sometimes forming junctions with both ceils (Fig. 3). Some muscle cells received more than one neuromuscular junction from either the same or different axons (see Fig. 3). In one afferent arteriole, along which a single axon could be followed for tens of micrometers, each smooth muscle cell was found to be innervated by at least one junction [15]. However, there was no specific distribution of junctions along equivalent lengths (up to 180 ~m) of either arteriole in relation to the glomerulus (Fig. 4). The numerical data describing the innervation of afferent and efferent arterioles are sum-
marised in Tables II and III. About 70% of varicosities associated with each of the afferent and efferent arterioles, formed junctions (Table H). Mean junctional density was about three times greater for the afferent compared to the efferent arteriole (Table III). Further, the density of Type I junctions on the afferent arteriole was higher (Table II) than on the efferent arteriole (X 2 test P < 0.025). In contrast, the Type I junctions occurred at very low densities on efferent arterioles which were predominantly innervated by Type II axons (see Fig. 4). However, the density of Type II junctions was similar on both vessels (Table II). The contact areas of some junctions on the smooth muscle cells of the afferent arteriole were slightly larger than those on the efferent arteriole (see [15], Table 2) although on average they were
Fig. 3. A: E l e c t r o n m i c r o g r a p h of two Type II axons with j u n c t i o n s l o c a t e d b e t w e e n two s m o o t h m u s c l e cells. B: E n l a r g e m e n t of box in A. sm = s m o o t h m u s c l e cell; v = varicosity. C a l i b r a t i o n bars = 1 # m .
]!
246 A
43_
2-
'
II
O4111 0
i
r
i
i
~
J
i
i
B
i
4-
1. 0
i
I i
~
J
i
i
i
i
i
I
~l
q..
o
4-
E
Z•-z 3"
3 iHinil'
2-
2-
1-
0
i
0
i
i
40
i
i
80
Distance from
i
i
120
i
i
i
160
glomerulus(gm)
0
1111
1) ' 4~) ' 80 ' 120 ' 1(~0 ' Distance from glomerulus (Ixm)
Fig. 4. Distribution of junctions with distance from the glomerulus along three afferent (A) and two efferent arterioles (B). Hatched columns = Type 1 axon junctions; open columns = Type 1I axon junctions. Arrow heads = length of vessel analysed.
not significantly different (Table III). However, presynaptic membrane specialisations occurred twice as frequently in junctions on the afferent compared to the efferent arteriole and the percentage was higher in Type I than in Type II axon junctions (Table II). (ii) Granular epithelioid cells. There were only one or two granular epithelioid cells containing renin granules per afferent arteriole in these series of sections. These cells were always located at the junction of the arteriole with the glomerulus and were never encountered in efferent arterioles. The junctions on these cells, like those on the smooth muscle cells, were restricted to the adventitial surface.
Six axons (three Type I and three Type II) formed a total of ten junctions on granular epithelioid cells (e.g. Fig. 5D, see Table I). Although there was considerable variation between the different series analysed, the mean junction density (see Table III) was significantly lower than the density on arteriolar smooth muscle cells (X 2 test P < 0.025). The junction density was higher for Type I than Type II axons (see Fig. 6, Table I). The areas of junctional contact were similar to that in other tissues (see Table 111). Presynaptic m e m b r a n e specialisations were found in two of the junctions, one formed by a Type I axon and the other by a Type II axon.
247 T A B L E I1
Ultrastuctural details o f the innervation o f afferent and efferent arterioles by Type I and H axons Innervation Axons % of Type | axons (of total) % of Type II axons (of total)
Afferent arteriole
Efferent arteriole
62 29
38 71
% of axons that occured singly: Total (% all axons) Type I (% all single axons) Type II (% all single axons)
30 56 44
17 14 86
% axons without support cells
17
3
Mean varicosity volume, ~tm 3 +_S.E.M. Junctions % total varicosities that form junctions * Mean junction density: × 1 0 3 / m m 2 _+S.E.M. Type I Type II
0.52 +_0.25 n= 8
70
14.2 +_4.4 5.8+_ 1.1
0.36_+ 0.13 n= 8
69
0.7, 0.8 4.1, 8.2
% junctions with presynaptic m e m b r a n e specialisations: Total 21 10 Type l 25 0 Type lI 14 12 * Data obtained from 3 afferent and two efferent arterioles (see Materials and Methods).
(iii) Proximal tubules. The innervation of the proximal tubules was determined in the same series of sections as for the afferent and efferent arteriole. Junctions were only observed on the tubule surface facing the arterioles. In the six series, five Type I and nine Type II axons formed a total of 28 junctions (nine Type I and 19 Type II) on proximal tubule epithelial cells (e.g. Fig.
5A). The density of junctions varied considerably between the different series of sections analysed. In one series, we found only Type I junctions; in two others, only Type II junctions. In another, there were approximately equal numbers of both types of junction, whereas in two others no junctions were observed. Overall, the mean density of junctions formed by both axon types was lower than on arteriolar smooth muscle cells (X 2 test P < 0.025) and was similar on tubules close to both afferent and efferent arterioles (7 and 5 x 103/ram 2 respectively). However, the mean density of Type II junctions was higher than Type I junctions and was similar to that on the efferent arteriole (Table I, Fig. 6). The areas of junctional contact were similar to that in other tissues (see Table III); however, presynaptic membrane specialisations were not observed at any of these junctions. (iv) Bowman's capsule. Of the five glomeruli that were analysed for the innervation of the
T A B L E llI
Junction density and contact area on the different jurtaglomerular tissues Tissue
* Junction density Mean + S.E.M.
Range
M e a n + S.E.M.
Range
Afferent arteriole
20.0 + 4.7 n=3 n=2 2.25 4- 2.24 n=6 4.7 +_ 2.8 n=6
10.6-25.3
0.48 _+ 0.32 n=4 0.38 + 0.05 n=6 0.42 _+ 0.13 n=9 0.43 + 0.02 n=15
0.02-1.19
Efferent arteriole Granular epithelioid cells Proximal tubules
Junction contact area ~ m 2
4.48, 8.8 0-10.6 0-16.4
* Junction density expressed as n u m b e r of junctions X 1 0 3 / m m 2 tissue surface analysed.
0.25-0.57 0.13-1.04 0.06-1.08
248
Fig. 5. J u n c t i o n s on effectors o t h e r than smooth muscle. A: Type II junction on a proximal t u b u l e e p i t h e l i a l cell. B: Type I junction on a distal tubule e p i t h e l i a l cell. C: Type 11 j u n c t i o n on m a c u l a d e n s a cell. D: Type II junction on g r a n u l a r e p i t h e l i o i d cell. E: Type I1 junction on a cell forming B o w m a n ' s capsule. A r r o w s = p r e s y n a p t i c m e m b r a n e specialisations; v = varicosity; p t - proximal tubule e p i t h e l i a l cell; dt = distal t u b u l e cell; r = renin granule; bc = B o w m a n ' s capsule cell; md = m a c u l a d e n s a cell; sch = Schwann cell. C a l i b r a t i o n bar = 1 /xm applies t h r o u g h o u t .
249 20 16 12 0 t~
8
•m•
4 o
:~
20
'~ 16
"E 12
"~ ~
Type I axons
o 2o
Type II axons
0 arterlolar ~proximel renln smooth tubule secreting muscle n=6 cells n=5 n=6 Fig. 6. Densities of neuroeffector junctions on different tissues in the outer renal cortex. Note that smooth muscle cells include afferent and efferent arterioles (see Table II for details). * Values for the proximal tubule cells are for those tubules located adjacent to afferent and efferent arterioles. Renin-secreting cells (granular epithelioid cells) are cells of the afferent arterioles, n = number of series analysed. Error bars = S.E.M.
arterioles, three Type II axons were found to form a total of five junctions on the epithelial cells that form the parietal layer of the Bowman's capsule (e.g. Fig. 5E). A presynaptic m e m b r a n e specialisation was identified at one of these junctions. (v) Distal tubules and macula densa. No junctions were found on the epithelial cells of the distal tubules in any of the detailed analyses. However, two junctions, adjacent to each other
on the same tubule cell, were found close to the macula densa region in a series of sections cut specifically to study this region in greater detail (see below). Both axons were Type I and were in the same axon bundle (see Fig. 5B). T h e r e were no presynaptic m e m b r a n e specialisations at either of these junctions. The axons located near the macula densa occur as a bundle that rings the lacis (see Fig. 1), so that it was not possible to locate these axons in sections cut longitudinally through the arterioles. Maculae densae were therefore studied associated with four additional glomeruli (see Materials and Methods). The plane of section through the macula densa was selected at random. Seven axons (two Type I, five Type II) were found lying within 8 / x m of the specialised distal tubular cells at the outer edge of three of the four maculae densae. Two axons, both Type II, formed junctions on the basal lamina overlying the macula densa cells (e.g. Fig. 5C). It should be noted that macula densa cells have a very convoluted plasm a l e m m a and the basal lamina often lay some distance away from the cell membrane. Hence, the separation between the m e m b r a n e s of the axon and the macula densa cell at both junctions was significantly greater than at the other junctions described above (see Fig. 5C). There were no presynaptic m e m b r a n e specialisations at either of these junctions. (vi) Mesangial cells. No axons were observed close to the extraglomerular mesangial cells in any of the series of sections studied. This was consistent with the finding that no TH-immunoreactive axons extended into this region (see above). (vii) Axons contacting more than one cell type. One axon was found to form junctions sequentially with an afferent arteriole smooth muscle cell, a renin-secreting granular epithelioid cell and the surface of Bowman's capsule. No other instance was found in which an axon innervated more than one type of tissue. Serial sections were cut of an additional 4.8 /zm thickness of tissue following on from a series of sections of an afferent arteriole in which we had identified three axons that formed junctions with proximal tubule ceils (i.e. total thickness of tissue analysed was 9.8
25O >m). None of these axons formed junctions with arteriolar smooth muscle cells.
Discussion This paper reports the distribution of specialised neuroeffector junctions in the outer renal cortex of the rabbit formed by the two types of sympathetic axon recently identified by us in both the rat and rabbit kidney [15]. A neuroeffector junction was identified as a close apposition between a varicosity and an effector cell with the intervening junctional cleft containing a single layer of basal lamina and the synaptic vesicles clustered towards the presynaptic membrane. This structural arrangement was identical to that of neuromuscular junctions in other vascular beds (see [12]). As well as describing the location of junctions on different cell types, we have quantified and compared the densities of junctions made by each type of axon on the different tissues in the region of the glomerulus. This ultrastructural description of the innervation of the juxtagiomerular region in the rabbit considerably expands the reports of others in different species [1,7,8,20]. The vast majority of axons in this region are sympathetic postganglionic noradrenergic axons as demonstrated by the presence of electron dense cores in small-diameter synaptic vesicles, and by their ability to take up 6-hydroxydopamine [15]. Immunohistochemistry of the kidney tissue demonstrated that TH-immunoreactive axons reached the juxtaglomerular region via the arcuate, interlobular and intralobular arteries, and that the axon plexus was denser around the afferent than the efferent arterioles. Although we have concentrated on the ultrastructural studies on the glomeruli in the outer cortex, TH-immunoreactivity associated with deeper glomeruli was distributed in a similar pattern, suggesting that these ultrastructural findings apply throughout the renal cortex. This contrasts with a previous autoradiography study which concluded that there is a decreasing gradient of innervation density of the afferent compared to the efferent arteriole in glomeruli from the capsule to the
medulla in the rat [3]. A more recent report [4] supports our findings of a uniform innervation pattern. Consistent with our other studies [12,13], the majority (78%) of perivascular noradrenergic varicosities formed specialised neuroeffector junctions.
Relati~.,e densities of neuroeffector junctions on different tissues The greatest density of junctions was round on the vascular smooth muscle cells of the afferent arteriole (Table III). Afferent arterioles bore junctions at a density of 20 x 103/mm 3, whereas junctional density on efferent arterioles was less than half this value and was even lower on the other tissues (Table III). These values for density have been determined by reconstructing all junctions within each series and so are not equivalent to the p a r a m e t e r we used previously to define the relative density of junctions on a range of arterial vessels [14]. This latter parameter reflected the frequency with which junctions were found in studies of random sections. It is possible to estimate the density of junctions on the 50 # m diameter submucous arterioles of the guinea pig small intestine by allowing for the average area of individual junctions (1.26 /.tin 3, see [12]). The frequency of junctions in random sections was 109 x 103/mm 2 [14], so that junctional density can be estimated to be about 9 x 1 0 3 / m m 2. This estimate is confirmed by the real junctional density on these arterioles (10 x 103/mm 2) when it is determined in the same way as has been done here for the juxtaglomerular tissues (S.E. Luff, unpublished observations). As the submucous arterioles were the most densely innervated of the arterial vessels previously studied, it seems likely that the afferent arterioles are amongst the most densely innervated vessels in the body. It is, however, notable that mean junctional area on the juxtaglomerular arterioles was smaller ( < 0.5/.~m 3 Table III) than on the submucous arterioles. The neurovascular junctions on the juxtaglomerular arterioles were distributed evenly along the vessels studied (Fig. 4). This contrasts with the distribution of junctions on guinea pig
251
submucous arterioles of equivalent diameter to the juxtaglomerular arterioles (15-25 #m), where the junctions on the former occur clustered at the vessel branch points [13]. The majority of varicosities formed junctions and most smooth muscle cells appeared to be directly innervated as intervaricosity lengths ( < 1.5/zm) were comparable to the diameter of the smooth muscle cells (i.e. 2 / z m ) (see [15]). This result differs from that reported in the rat [2] in which only 20% of the agranular cells of both the afferent and efferent arterioles appeared to be innervated. Neuroeffector junctions were found on effector cells of most other tissue types lying adjacent to the vascular smooth muscle cells but at densities significantly lower than those on the vasculature. The renin-secreting granular epithelioid ceils formed part of the afferent arteriole close to the glomerulus and were absent from the efferent arteriole in the rabbit (cf. rat, [22]). The mean junction density on renin-secreting cells was much lower (Table III) than that found on adjacent vascular smooth muscle cells. Junctions were observed on some proximal tubule epithelial ceils lying adjacent to the afferent and efferent arterioles. The density of junctions was low but similar on tubules close to both afferent and efferent arterioles (7 and 5 x 103/mm 2 respectively). However, as tubule tissue located away from the vasculature was not innervated it must be concluded that the innervation of this tissue is very sparse ([14], see also [11,19]). The density of junctions on the distal tubules and associated macula densa cells was even lower than that found on the proximal tubules. Junctions were identified only in series which concentrated on the macula densa, and only close to its outer edge, consistent with the arrangement of TH-immunoreactive axons. These observations indicate that the density of junctions on the distal tubules in the outer cortex is extremely low. Junctions were also found on the epithelial cells forming Bowman's capsule (see also [8]). One of these was formed by an axon which also innervated the afferent arteriole. It is difficult to envisage what function could be served by these junctions.
The higher density of junctions on the afferent compared to the efferent arteriole is compatible with the physiological effects of renal nerve stimulation [21]. Conversely, the relatively sparse (and patchy) innervation of the proximal tubular cells seems at variance with physiological studies, mainly performed in the rat, in which marked actions of nerves on tubular function have been reported (e.g. see [5,17]). Apart from the direct effects of sympathetic nerves on the renin-secreting cells of the afferent arteriole [5], it has been proposed that the sympathetic innervation also indirectly modulates renin secretion via a direct neural effect on the macula densa [6,18]. However, the functional significance of a highly localised structural arrangement of the innervation of the macula densa tissue is unknown. Similarly, this structural study does not support claims that macula densa ceils of the distal tubule and the glomerular mesangium are richly innervated [16]. In the rabbit we have not been able to find any direct innervation of the mesangium despite extensive searching. Only one instance of a single axon forming junctions with different effectors was observed. The axon concerned formed junctions sequentially with a vascular smooth muscle cell, a reninsecreting cell and with Bowman's capsule. This result contrasts with reports in the rat that axons frequently form contacts with more than one tissue [2].
Distribution of Type I and H junctions The two structural types of axon in the outer renal cortex [15] both formed neuroeffector junctions with the various effector cells, but they were distributed differentially between them. Type I axons contacted afferent arterioles and reninsecreting cells more often than other tissues. The greatest density of Type I axon junctions was on the afferent arteriole, whereas the density on efferent arterioles, renin-secreting cells and juxtaglomerular proximal tubules cells was low (see Table III, Fig. 5). Type II axons, on the other hand, were distributed at similar densities over afferent and efferent arterioles and proximal tubules. The rare junctions formed on the distal
252
tubule cells near the macula densa were made by both types of axon. The low density of Type I axon junctions on the efferent arteriole and other non-vascular effectors indicates that the majority of these axons must terminate at the glomerulus although some extend for some distance along the efferent arteriole (Fig. 4). The difference in the distribution of junctions formed by these two types of axon between tissues suggests different functional roles for the Type I and II axons. Presynaptic membrane specialisations were present at about 16% of junctions; this is a similar proportion to that observed for other arteries [13]. Most presynaptic specialisations were detected at junctions on arterioles. However, specialisations were present at a much higher proportion of Type I than Type II junctions (see Table II). If such structures reflect release sites (as at other junctions and synapses), these findings suggest that the afferent arterioles receive a more effective innervation: (i) because of the high density of Type I junctions on this vessel; and (ii) because of the higher likelihood that these contain specialised release sites. The structural data also strongly support the view that the major functional effect of the innervation is on the vasculature. Previous studies have demonstrated junctional contacts on several effector types in the rat kidney [2,3,8]. Barajas and MiJller [2] also compared innervation density of the different tissues in the juxtaglomerular region. However, it is difficult to compare their data directly with ours as they used a different criterion for estimating innervation density (i.e. the proportion of cells receiving direct innervation). The definition of a 'neuroeffector contact' in both earlier studies included varicosity appositions with effector cells with membrane separations of 300 nm and containing two distinct layers of basal lamina. Because we found the majority of varicosities formed specialised junctions, there is no reason at present to believe that varicosities merely lying 'close' to an effector cell are involved in neurotransmitter release. They concluded, as we do, that the innervation density of the tubule cells was much lower than that of the arterioles, but they found that a higher proportion of granular epithelioid cells received
junctions than did smooth muscle cells. The data are not currently available to compare innervation density between the two species. In conclusion, this study has identified two important aspects of the juxtaglomerular innervation in the cortical region in the rabbit: (i) Type I and Type II noradrenergic axons are differentially distributed between effector ceils, implying distinct functional roles; and (ii) presynaptic membrane specialisations occur more often at Type I junctions on afferent arterioles than at junctions (of either type) on all other tissues. The findings that the Type I axons have a higher proportion of varicosities that form junctions, and have a higher proportion of junctions with presynaptic membrane specialisations than Type II axons, suggest that Type I axons are likely to be more effective in controlling effector function. The distribution of junctions indicates that Type I axons are predominantly vasoconstrictor as they are either absent or present at very low densities on other tissues whereas Type II axons may have a more gene.ralised effect as they are similarly distributed on different tissues. The way in which the sparse innervation found on non-vascular effectors directly influences the function of these tissues remains to be clarified.
Acknowledgements We wish to thank The National Heart Foundation of Australia and the National Health and Medical Research Council for their financial support of this project, and Simone Young for her valuable technical assistance.
References 1 Barajas, L., Innervation of the renal cortex, Fed. Proc., 37 (1978) 1192-1208. 2 Barajas, L. and Miiller, J., The innervation of the juxtaglomerular apparatus and surrounding tubules: a quantitative analysis by serial section electron microscopy, J. Ultrastruct. Res., 3 (1973) 107-132. 3 Barajas, L. and Powers, K.V., Innervation of the thick ascending limb of Henly, Am. J. Physiol., 255 (1988) F340 F348.
253 4 Barajas, L. and Powers; K.V., Monoaminergic innervation of the rat kidney: a quantitative study, Am. J. Physiol., 259 (1990) F503-F511. 5 DiBona, G.F., The functions of renal nerves. Rev. Physiol. Biochem. Pharmacol., 94 (1982) 75-181. 6 DiBona, G.F., Neural control of renal function: cardiovascular implications, Hypertension, 13 (1989) 539-548. 7 Ferguson, M., Ryan, G.B. and Bell, C., The innervation of the renal cortex in the dog: an ultrastructural study. Cell Tissue Res., 253 (1988) 539-546. 8 Gorgas, K., Struktur und innervation des juxtaglomerular apparatus der rat. Adv. Anat. Embryol. Cell Biol., 54 (1978) 5-84. 9 Gottschalk, C.W., Moss, N.G. and Colondres, R.E., Neural control of renal function in health and disease. In D.W. Seldin and G. Griebisch (Eds.), The Kidney, Physiology and Pathophysiology, (Vol. 1), Raven Press, New York, 1985, pp. 581-611. 10 Klemm, M., Luff, S.E. and Van Helden, D.F., Sympathetic varicosities form specialised junctions on mesenteric veins of the guinea pig, Proc. Aust. Physiol. and Pharmacol. Soc., 22 (1991) 147P. 11 Kriz, W. and Kaissling, B., Structural organisation of the mammalian kidney. In D.W. Seldin and G. Griebisch (Eds.), The Kidney: Physiology and Pathophysiology, (Vol. 1), Raven Press, New York, 1985, pp. 265-306. 12 Luff, S.E., McLachlan, E.M. and Hirst, G.D.S., An ultrastructural analysis of the sympathetic neuromuscular junctions on arterioles of the submucosa of the guinea pig ileum, J. Comp. Neurol., 257 (1987) 578-594. 13 Luff, S.E. and McLachlan, E.M., The form of sympathetic postganglionic axons at clustered neuromuscular junctions near branch points of arterioles in the submucosa of the guinea pig ileum, J. Neurocytol., 17 (1988) 451-463.
14 Luff, S.E. and McLachlan, E.M., Frequency of neuromuscular junctions on arteries of different dimensions in the rabbit, guinea pig and rat, Blood Vessels, 26 (1989) 95-106. 15 Luff, S.E., Hengstberger, S.G., McLachlan E.M. and Anderson, W.P., Two types of sympathetic axon innervating the juxtaglomerular arterioles of the rabbit and the rat kidney differ structurally from those supplying other vessels, J. Neurocytol., 20 (1991) 781-795. 16 Maddox, D.A. and Brenner, B.M., GIomerular ultrafiltration. In B.M. Brenner and F.C. Rector Jr. (Eds.), The Kidney, (Vol. 1) W.B. Saunders, Philadelphia. 17 Moss, N.G., Renal function and renal afferent and efferent nerve activity, Am. J. Physiol., 243 (1982) F425-433. 18 Osbourn, J.L., Thames, D. and DiBona, G.F., The role of macula densa in renal nerve modulation of renin secretion, Am. J. Physiol., 242 (1982) R367-R371. 19 Reinecke, M. and Forssmann, W.G., Neuropeptide (neuropeptide Y, neurotensin, vasoactive intestinal polypeptide, substance P, calcitonin gene-related peptide, somatostatin) immunohistochemistry and ultrastructure of renal nerves, Histochemistry, 89 (1988) 1-9. 20 Simpson, F.O. and Devine, C.E., The fine structure of autonomic neuromuscular contacts in the arterioles of the sheep renal cortex. J. Anat., 100 (1966) 127-137. 21 Tucker, B.J., Peterson, O.W., Munger, K.A., Bird, J.E., Mitchell, M., Pelayo, J.C. and Blantz, R.C., Glomerular hemodynamic alterations during renal nerve stimulation in rats on high- and low-salt diets, Am. J. Physiol., 258 (1990) F133-FI43. 22 Taugner, R. and Hackenthal, E., The Juxtaglomerular Apparatus: Structure and Function. Springer-Verlag, Berlin, 1989.