Graefe's Arch Clin Exp Ophthalmol (1986) 224:147-151 Graefe's Archive for CliniCal and Expedmental

Ophthalmology © Springer-Verlag 1986

Nonrandom distribution of coated pits and vesicles in the connective tissue cells of the trabecular meshwork of rabbit

G. Diaz, S. Carta, and N. Orzalesi Institute of Human Anatomy and Eye Clinic, University of Cagliari, 1-09100 Cagliari, Italy

Abstract. Coated pits (CPs) and coated vesicles (CVs) are distributed nonrandomly along the surface of the connec- tive tissue cells of the rabbit trabecular meshwork. Morpho- metric and statistical analyses of the distances between con- secutive structures reveal a tendency to cluster, which is apparently higher among CVs than among CPs. The spatial relationship between CVs and CPs is also demonstrated by the analysis of their association (presence/absence) in each cell. The data suggest the hypothesis that each cluster is formed by structures in the same stage of maturation. In addition to the recent demonstration of the clustering of CPs in transformed cells in vitro (Pfeiffer et al. 1980), our findings indicate that: (a) clustering is characteristic of both CVs and CPs; (b) it also occurs in normal tissue cells; (c) it represents a further peculiarity, which confirms the specific nature of receptor-mediated endocytosis; (d) it reflects, at a higher level, the clustering of the receptor molecules, which are responsible for the high selectivity of the endocytic process.

Introduction

Coated pits (CPs) are distinct organelles actively involved in the uptake and transport of peptides, proteins, and lipo- proteins by means of receptor-mediated endocytosis (RME; Goldstein et al. 1979; Dickson et al. 1981). They represent small regions of the plasma membrane characterized by a bristled coat of cytoplasm that coincides with the presence of specific receptors on the outer cell surface. Despite some controversy on the mechanism of RME and the hypothesis of receptosomes (Pastan and Willingham 1983; Van Deurs et al. 1983; Willingham and Pastan 1984), it seems quite possible that the internalization of the receptor-bound sub- stances is accomplished through the invagination of these coated regions, followed by the detachment of pinocytic 'coated ' vesicles (CVs) which exhibit the original coating on their cytoplasmic aspect (Goldstein et al. 1979; Rees and Wallace 1980; Orzalesi et al. 1982).

In a previous investigation we studied CPs and CVs in the trabecular meshwork of man and rabbit (Diaz et al. 1982). The vesicles were of the large type (> 0.1 ~m in diam- eter) that is generally found in the peripheral cytoplasm (Wild 1980) and seemed to be distributed quite uniformly

Offprint requests to: G. Diaz

throughout the whole trabecular meshwork, including the aqueous plexus of rabbit and the Schlemm's canal of man; there were no differences in morphological appearance be- tween man and rabbit.

Recently it has been shown in J774.2 mouse macro- phages that CPs are distributed nonrandomly on the cell surface. The clustering of CPs is quite subtle in the un- treated macrophages, while it becomes readily evident after colchicine treatment when they accumulate over a microvil- lous protuberance that develops at one pole of the cell. CPs (or their molecular precursors) probably move by diffu- sion or by inclusion in membrane flow (Pfeiffer et al. 1980).

The aim of this paper was to investigate the distribution of CPs and CVs in the trabecular meshwork of rabbit in order to ascertain if the clustering observed in mouse mac- rophages would also apply to the trabecular tissue cells. This fact presents two points of interest: (1) an uneven distribution of coated structures (not only CPs, but also CVs) would represent a general feature of adsorptive pino- cytosis; (2) changes in topographical distribution and clus- tering could indicate critical mechanism by which the cells can modulate their responses to growth substances and hor- mones.

Materials and methods

The study was performed on the trabecular meshwork of the normal albino rabbits used as controls in the course of previous investigations on the effects of steroids on the anterior chamber angle (Fossarello et al. 1982). Specimens were fixed in 2% glutaraldehyde and 2% paraformaldehyde in 0.1 M cacodylate buffer, pH 7.4, for 2 h at room temper- ature; postfixed in 1% osmium tetroxide in the same buffer for 1 h; stained 'en block' in a saturated aqueous solution of uranyl acetate for 8-12 h; dehydrated in graded ethanol and toluene; embedded in Araldite M. Ultrathin sections were cut with an LKB Ultratome III, stained with bismuth subnitrate (Riva 1974), and examined with a Jeol 100 S electron microscope.

To assure an unbiased sample, the same electron micro- graphs collected for the above study were reconsidered and selected only on the basis of magnification (minimum x 14,000) and good quality of images, both needed for a

correct morphological evaluation of CPs and CVs. In fact, as pointed out in previous articles (Diaz et al. 1982; Orzalesi et al. 1982), the identification of CVs is seriously condi-

148

Fig. 1. Micrograph showing the subdivision of the cell contour into tracts. For simplicity, only some of them are outlined. (x 15,000)

tioned by the possibility that some vesicles are pits still communicating with the cell surface through 'necks' out of the plane of section. In consideration of this fact, the definition of CVs was based on the absence of any morpho- logical connection between the vesicles and the plasma membrane.

Measurements were done'by means of a MOP-3 semiau- tomatic image analyzer (Carl Zeiss, Inc., FRG) on line with a CBM 8032 microcomputer (CBM, Inc., USA). We ana- lyzed 300 cell profiles whose contour exceeded 1 gm in length. For each profile we measured: (a) the contour length of the cell profile; (b) the length of CPs; (c) the number of CVs and CPs; (d) the distance between consecutive ~ zs coated structures along or near the cell surface (tracts CV- ~_ 30 CV, CP-CP and CV-CP); (e) in cases where the cell profile ,,.ca 2s

appeared interrupted by the photographic margin, we mea- 29 e~ lS sured the distance between one end of the cell contour and ~ z0

the nearest coated structure ('external' tracts) (Fig. 1). " s For (a), (b), (d) and (e) we estimated the means and

standard deviations. For (d) and (e) we also calculated the ss

median and its 95% confidence limits, the skewness coeffi- so

cient, the goodness-of-fit ~2 (test of normality), and t he ~ 4s significance of the difference between the two distributions ~ 4e (two-sample Kolmogorov-Smirnov test; see Kendall and ~ 3s Stuart 1973). ~ :m

o 2s The correlation between the above experimental data ~ 2o

and theoretical data resulting from an hypothesis of syn- " zs chronous development of coated structures was evaluated z0 by means of Kendall's nonparametric rank test. s

The association between CVs and CPs was tested by the X 2 test of independence from (c). ~)

The nonrandomness of spatial distribution of CVs and CPs in the trabecular meshwork was tested by the X 2 analy- sis o f ' quadrats'.

R e s u l t s

The length of CPs found in trabecular tissue cells of rabbit (0.35+0.03gm, mean_+SD) and their linear density (0.069 pits/gin of cell contour) are in close agreement with the data reported for mouse macrophages in culture (Pfeifer et al. 1980).

CVs and CPs appear uniformly distributed throughout the trabecular tissue, also at a quantitative evaluation of

' EXTERNRL" t le2)

$)

o 2 4 S S )O 12 LENGTtf LIw nlcnows)

C V - C P I,~e)

C P - C P i t s )

L o 2 4 s e 1o |2

L.E:HGTI"I tzN NZcSOHS)

Fig. 2. Frequency distribution of the length of the four tracts. The curves show the expected normal distributions. (The number of measurements is shown in parentheses)

149

30,

~ 2$,

20 ~o 15, ~:: 10, is.

S,

0

® B l g l E I g w B I I l l

8 16 24 32 40 40 LENGTH (H]CRO.S)

Fig. 3. Frequency distribution of the length of the cell contours. The curve shows the expected normal distribution

their dispersion in tissue ( ' q u a d r a t ' analysis, Z2=2.308, P,> 0.05).

At the cellular level, the dis tr ibut ion of CVs and CPs was analyzed by compar ing the distance between consecu- tive coated structures (tracts C V C V , C P - C P and CV-CP) with the length of r andom segments of the cell contour (' external ' tracts). In fact, the dis tr ibut ion of the ' external ' tracts is not perfectly homogeneous, since it tends to reflect in par t the dis tr ibut ion of the cell contour length (Figs. 2 and 3). This allowed the dis tr ibut ion of the ' ex te rna l ' tracts to be used as reference distr ibution to test the randomness of CV-CV, C P - C P and C V - C P tracts. F o r all, the Kolmo- gorov-Smirnov test rejected the null hypothesis of homoge- neity (Table 1), thus proving a tendency of CVs and CPs to be associated with each other in clusters (Fig. 4).

Clustering is part icular ly evident for CVs (56% of tracts C V - C V shorter than 1 gm). A weaker associat ion is found

Table 1. Statistical analysis of the length of the four tracts (in microns)

CV-CV CP-CP CV-CP External

Number of measurements 45 79 78 182 Mean 2.0 3.1 3.5 4.5 Standard deviation 2.6 3.8 4.1 3.9 Median 0.9 1.8 1.9 3.4 95% Confidence lower 0.6 1.1 1.6 2.8

limit of median 95% Confidence upper 1.4 2.5 3.0 4.2

limit of median Skwness coefficient (gl) 1.9 1.7 1.2 0.8 Z z Goodness of fit test 59.0 91.4 47.6 84.8

for normality Degrees of freedom 5 8 8 10 Significance P < 0.001 0 . 0 0 1 0.001 0.001

Kolmogorov-Smirnov D" 0.402 0.226 0.190 - test vs External

Significance P < 0.001 0.01 0.05 -

a Largest absolute difference between cumulative standardized fre- quencies

among CPs and between CVs and CPs (38% and 30% of tracts C P - C P and C V C V , respectively, less than I pm). However, even in these cases, clustering is still present when we consider that only 15% o f ' e x t e r n a l ' tracts are less than 1 gm long (Fig. 2).

The test of independence, appl ied to a 2 x 2 frequency table according to the presence of CVs and/or CPs or nei-

Fig. 4. Trabecular tissue cells showing clusters of coated structures (as ter isks) along the cell surface. ( x 15,000)

150

Table 2. Classification of the cell profiles according to the presence or absence of CVs and CPs

CPs CVs

Present Absent Total

Present 55 76 131 Absent 41 128 169 Total 96 204 300

Zz=9.86, P<0.005

VE$ I CLE$

" 2,," • 2 \

ERRL¥ P IT$ MflTURE P IT$

Fig. 5. Ideographic picture showing the relationship between 3D clusters of synchronous structures (circles: early pits, mature pits and vesicles) and 2D tracts visible in section (dashed lines: 1, CV- CV; 2, CV-CP; 3, CP-CP)

also maintained inside the cell in the CVs which have sepa- rated from the cell surface. Thus, clustering o f the coated structures could be considered as a general feature of recep- tor-mediated endocytosis (RME).

The different degrees of association that result from the analysis o f tracts can be explained by assuming the existence of clusters of structures in the same stage of maturat ion - for example, clusters formed by early pits, mature pits and vesicles, as illustrated in Fig. 5. Apar t from arbitrary sizes (number of structures per cluster, distance between clusters, etc.), the frequency and length of the tracts derived from the model correspond to the length and frequency of the tracts experimentally observed. The nonparametric correlation is shown in Table 3. The presence of homoge- neous clusters of coated structures implies that cyclic (time- related) and localized (space-related) waves of R M E are produced in the cell.

Clustering of CVs and CPs represents a further peculiar- ity of coated structures which confirms their specific nature. The phenomenon refects, at a higher level, the clustering of specific receptor molecules which, on the outer CPs' sur- face, are involved in the process of adsorptive pinocytosis (Anderson etal. 1976; Willingham etal. 1979; Dickson et al. 1981 ; Van Deurs et al. 1982).

No information is currently available on the nature and role of the molecules, which are internalized by means of R M E in the connective tissue cells of the trabecular mesh- work. Trabecular cells may be submitted to various experi- mental procedures (changes in intraocular pressure, steroid treatment, plasmoid aqueous formation, etc.) in vivo as well in culture. Therefore, further investigations on clustering of coated structres should throw some light on the biologi- cal significance of this phenomenon. It is also possible that clustering may represent an alternative cue for the evalua- tion of the cell viability with respect to metabolic efficiency and structural integrity.

Table 3. Nonparametric correlation between experimental data of Table 1 and theoretical data derived from Fig. 5

Experimental Theoretical

Ranked Maximum CV-CP/CP-CP CV-CP/CP-CP frequency Middle -

Minimum CV-CV CV-CV

Ranked Maximum C V C P CV-CP mean length Middle CP-CP CP-CP

Minimum CV-CV CV-CV

~= 1, P<0.005 by Kendall rank correlation test

ther in each profile ~ 2 = 9 . 8 6 , P<0 .005) provided further evidence for a spatial relationship between CVs and CPs (Table 2).

Discussion

Clustering of CPs that was previously shown in mouse mac- rophages in vitro (Pfeiffer et al. 1980) is now demonstrated in normal cells of living tissue, such as the trabecular cells of the rabbit angle.

Moreover, our findings provide evidence that clustering is not confined to CPs found on the cell surface but is

References

Anderson RGW, Goldstein JL, Brown MS (1976) Localization of low density lipoprotein receptors on plasma membrane of normal human fibroblasts and their absence in cells from a familiar hypercholesterolemia homozygote. Proc Natl Acad Sci USA 73:2434-2438

Diaz G, Orzalesi N, Fossarello M, Carta S, Del Fiacco G (1982) Coated pits and coated vesicles in the endothelial cells of trabec- ular meshwork. Exp Eye Res 35:99-106

Dickson RB, Nicolas JC, Willingham MC, Pastan I (1981) Interna- lization of g2 macroglobulin in receptosomes. Studies with monovalent electron microscopic markers. Exp Cell Res 132:488~493

Fossarello M, Carta S, Del Fiacco G, Diaz G, Orzalesi N (1982) Quantitative ultrastructural study of the anterior chamber angle of the rabbit with corticosteroid-induced ocular hypertension. Ophthalmic Res 14:40--45

Goldstein JL, Anderson RGW, Brown MS (1979) Coated pits, coated vesicles and receptor-mediated endocytosis. Nature 279 : 679-685

Kendall MG, Stuart A (1973) The advanced theory of statisitcs, vol II. Griffin, London

Orzalesi N, Fossarello M, Carta S, Del Fiacco G, Diaz G (1982) Identification and distribution of coated vesicles in the retinal pigment epithelium of man and rabbit. Invest Ophthalmol Vis Sci 23 : 689-696

Pastan I, Willingham MC (1983) Receptor-mediated endocytosis, coated pits, receptosomes and Golgi. Trends Biochem Sci 8:250-253

151

Pfeiffer JR, Oliver JM, Berlin RD (1980) Topographical distribu- tion of coated pits. Nature 286: 727-729

Rees A, Wallace K (1980) Coated vesicles and receptor biology. In: Ockleford CDI, Whyte A (eds) Coated vesicles. Cambridge University Press, Cambridge, pp 219-242

Riva A (1974) A simple and rapid staining method for enhancing the contrast of tissues previusly treated with uranyl acetate. J Microsc (Paris) 19:105-108

Van Deurs B, Nilausen K, Faergeman O, Meinertz H (1982) Coated pits and pinocytosis of cationized ferritin in human skin fibroblasts. Eur J Cell Biol 27 : 270-278

Van Deurs B, Peterson OW, Bundgaard M (1983) Do coated pino- cytic vesicles exist? Trends Biochem Sci 8:400-401

Wild AE (1980) Coated vesicles: a morphologically distinct sub- class of endocytic vesicles. In: Ockleford CDI, Whyte A (eds) Coated vesicles. Cambridge University Press, Cambridge, pp 1-24

Willingham MC, Pastan I (1984) Do coated vesicles exist? Trends Biochem Sci 9 : 93-94

Willingham MC, Maxfield FR, Pastan I (1979) ~2 macroglobulin binding to the plasma membrane of cultured fibroblasts. Dif- fuse binding followed by clustering in coated regions. J Cell Biol 82:614-625

Received June 29, 1984 / Accepted June 21, 1985