with cytoplasmic penetration into Nucleopore filters. By JORMA ... diffusion rates through these filters do not support long range diffusion as a mechanism.
/ . Embryol. exp. Morph. Vol. 31, 3, pp. 667-682, 1974 Printed in Great Britain
667
Transfilter induction of kidney tubules: correlation with cytoplasmic penetration into Nucleopore filters By JORMA WARTIOVAARA 1 , STIG NORDLING, EERO LEHTONEN AND LAURI SAXEN From the Third Department of Pathology, the Electron Microscope Laboratory, and the Department of Zoology, University of Helsinki, Finland
SUMMARY The presence of cell contacts in transfilter mouse kidney tubule induction by spinal cord was investigated. The interacting tissues were separated by Nucleoporeflrt membrane filters of various pore sizes. Filters with a mean pore size of 0-2 /*m or more allowed tubule formation, whereas 013/tm pore filters prevented it. There was an inverse correlation between pore size and minimum time of transfilter culture required for induction to occur. The differences in the diffusion rates through these filters do not support long range diffusion as a mechanism for induction. Electron microscopy of the cultures showed abundant cytoplasmic penetration deep into filters with 0-2 /tm or larger pores. Processes from mesenchymal and spinal cord cells were closely apposed within the filter channels. No extracellular matrix or membrane vesicles were seen between the processes. In a few instances shallow penetration was seen in 01 /tm type filters, but no contacts were observed. The presence of close cell appositions in those filters which allow kidney tubulogenesis suggests that close cellular interactions, rather than long range diffusion of signal substances, is the most likely communicative mechanism in this transfilter induction. INTRODUCTION Morphogenetic tissue interactions involve signals transmitted between cells. Attempts to isolate signal substances which operate in normal development have been unsuccessful, and there is no conclusive evidence of their transmission from the inducing to the responding tissue. However, as co-operation between heterotypic cells leads to the determination and morphogenesis of one or both of the interactants, different mechanisms of interaction should be considered. Holtfreter (1955) has suggested that inductor molecules diffuse over long distances to the responding tissue. Other theories are based on the requirement of cell contacts (Weiss, 1947, 1958) or the morphogenetic action of extracellular matrix at the inductor/responding tissue interface (Grobstein, 1955; Bernfteld & Wessels, 1971). 1
Author's address: Third Department of Pathology, Haartmaninkatu 3, SF-00290 Helsinki 29, Finland.
668
J. WARTIOVAARA AND OTHERS
Much of the evidence favouring the diffusion hypothesis as well as the idea of transmissible extracellular materials comes from experiments in which nitrocellulose (Millipore) filters have been placed between the interacting tissues. These membrane filters are believed to prevent cytoplasmic contacts, although they allow passage of inductive signals (Grobstein, 1956; Grobstein & Dalton, 1957). However, in transfilter induction of kidney tubules, the passage of these signals is slow, which makes mechanisms based on long range diffusion unlikely in this system (Nordling, Miettinen, Wartiovaara & Saxen, 1971). Further experiments with interposition of filters were therefore designed in order to correlate pore size, porosity and rates of diffusion with passage of inductive messages, and to study the morphological relationship between the interacting tissues. Millipore filters did not seem well suited for such studies as they have a spongy structure with great variation in the size and distribution of pores (Lehtonen, Nordling & Wartiovaara, 1973). Nucleopore filters do not have these limitations as the pores are uniform in size and the channels straight (Cornell, 1969; Wartiovaara, Lehtonen, Nordling & Saxen, 1972). MATERIAL AND METHODS
Induction experiments Eleven-day-old hybrid mouse embryos CBA/T6T6 x A/Sn were used throughout. Metanephric rudiments were dissected from the embryos, and their mesenchymes removed by micro-manipulation and EDTA treatment. Six to eight mesenchymes were placed on a Nucleopore filter and a fragment of the thoracic region of the spinal cord was attached to the opposite side with 1 % agar. Explants were cultured for 18-72 h on a Trowell-type screen at the medium-gas interface at 37 °C in Eagle's minimum essential medium in Earle's balanced salt solution supplemented with 10% foetal calf serum (Microbiological Associates Inc., Bethesda, Md.) (Saxen & Saksela, 1971). For consistency, the inductor (spinal cord) was placed on the matt side of the filter, and the mesenchyme on the glossy side. The morphogenetic response, i.e. tubule formation, was examined from serial sections of paraffin-embedded explants. The Nucleopore® filters were obtained from General Electric Co. (Pleasanton, Calif.). According to the manufacturer they are made by exposing a polycarbonate tape to charged particles in a nuclear reactor followed by chemical etching of the sensitized tracks into uniform pores. The thin Nucleopore filters are transparent, and it is thus easy to locate the tissues on both sides of the filter and to see the gross development of the tissues. However, tissues are not easily attached to the filter and can slide away during culture, or be lost during fixation and dehydration. Finally, it was difficult to get good sections for light and electron microscopy.
Cell contacts and induction
669
Diffusion experiments The diffusion through Nucleopore niters was determined for 3H-thymidine (New England Nuclear Co., Boston, Mass.), bovine serum albumin (BSA) (Bovine Albumin Powder, Fraction V, Armour Pharmaceutical Co., Eastbourne, U.K.), and 3H-labelled polio virus (courtesy of Dr A. Yaheri, Department of Virology, University of Helsinki) made radioactive by cultivation in the presence of 3H-phenylalanine, and purified by two isopycnic bandings in caesium chloride. The procedure for determining the rate constants of diffusion was identical to that previously described (Nordling et ah 1971) and involved a diffusion chamber with two sides separated by the filter to be investigated. The test substance was added to one side and the concentration of the substances was determined spectrophotometrically (using Lowry's method for BSA) or by liquid scintillation spectrometry, at various intervals. The thickness of the filter was measured with a micrometer in which the displacement of an inductive sensor was electronically recorded. The reproducibility of the measurements was better than 1 /*m. At least 10 different measurements were performed on each filter type. Determination of pore size and morphology Filters were examined with a scanning electron microscope. Small strips of Nucleopore filter were attached with silver-containing glue to specimen supports and covered with a thin layer of carbon and gold in a Balzer's MicroBA3 evaporator and examined with a JSM-U3 scanning electron microscope. The diameters of the pores were measured from scanning electron micrographs. The borders of the pores were not sharp, making exact measurements difficult. The measurements are therefore accurate only to about 10%. The porosity (percentage of filter occupied by pores) was also measured from these micrographs. About 100 pores were measured. Penetration studies For transmission electron microscopy explants were covered with a layer of 1 % agar and fixed for 60 min at 4 °C with 1-25 % glutaraldehyde in 0-1 M cacodylate buffer, pH 7-2. After thorough rinsing in 0-2 M sucrose in 0*1 M cacodylate buffer the tissues were postfixed in cold 1 % OsO4 in 0-1 M cacodylate buffer, pH 7-2, for 60 min. After dehydration in ethanol, specimens were embedded in Epon 812. Thin sections, cut with diamond knives (IVIC, Caracas) were stained with 3 % uranyl acetate in 50% ethanol and with 0-13 % lead citrate in 0-01 N NaOH. Micrographs were taken with a Philips EM 300 electron microscope. Thick (1 to 2 //m) Epon sections for light microscopy were stained with toluidine blue.
670
J. WARTIOVAARA AND OTHERS Table 1. Effect of pore size on tubule induction
(/tm)
Number of positive explants/total
Proportion of positive cases
01 0-2 0-5 3
3/30 27/31 25/25 15/15
01 0-9 10 10
Filter type
RESULTS
Transmission of the inductive stimulus Limiting effect of the Nucleopore filters. One hundred and one successful explants were examined after 72 h of culture (explants in which tissues could not be seen on both sides of the filter were discarded). All explants on filters of 3 and 0-5 /tm type showed tubule formation, 27 out of 31 explants on 0-2 //m type filters were positive (Table 1). Only three out of 30 explants on 0-1 /tm type filters showed a weak positive response. These results are 'qualitative', as explants with only one tubule or pretubular condensate were recorded as positive. Quantitative differences existed; on 0-2 /tm type filters only a few small tubules were usually seen, in contrast to an abundance of well-differentiated tubules in explants on 3 /im type filters. The survival of the tissues was usually good. Pore size and the minimum induction time. Minimum induction time meant the minimum time of transfilter contact required for the subsequent autonomous differentiation of the mesenchyme. After various periods of culture with spinal cord on the opposite side of the filter, the mesenchymes were removed and placed on Millipore 'TA' filters and subcultured without inductor (spinal cord) for 72 h. Ninety-eight successful explants were examined (Fig. 1). The minimum induction time was longer on small than on large pore size filters, and the time difference between filters of 0-5 and 0-2 //m type was about 12 h. Structural features of the filter Transmission electron microscopy of Nucleopore filters showed straight channels with even walls, but the orientation of these passages was not often parallel (Fig. 2). Scanning electron microscopy revealed that the distribution of pores was uneven (Figs. 3, 4, 5 and 6). Some channels had common or overlapping orifices, but they seemed to separate deeper in the filter. Some pore openings were elliptical because of the oblique course of the channels. Nucleopore filters have a glossy and a matt side. In scanning electron micrographs of the glossy side there were only pore openings (Fig. 6). On the matt side there were also irregularly shaped shallow impressions or lacunae (Fig. 7).
671
Cell contacts and induction
Time of contact (h) Fig. 1. Transmission of induction indicated as percentage of mesenchymes with kidney tubules as a function of pore size of Nucleopore filters (3, 0-5 and 0-2 /tm) and of time of transfilter contact with spinal cord tissue. After transfilter cultivation mesenchymes were cultivated alone for 72 h.
The pore size in filters of the same type varied (Fig. 8). The mean pore diameters measured from the scanning electron micrographs deviated from the nominal ones (Table 2). The porosity measured from the scanning electron micrographs was low and the pore density decreased with increasing pore size (Table 2). The thickness of the filters varied, but within the same batch there was only slight variation (Table 2). Transmission of molecules by diffusion In a diffusion chamber, where a substance dissolved in one chamber diffuses through a filter with a cross-sectional area of S into another chamber with an equal volume V, the following equation is valid (Nordling et al. 1971): In i l = / ) — / = —-/ m K) Ac Vh Hh where c0 is the concentration of the substance in chamber A when / = 0, Ac = cA — Cji (the concentrations of the substance at time t in chambers A and B respectively), h is the thickness of the filter and H the depth of the chamber. The permeability of Nucleopore filters was tested with different substances whose rate constants of diffusion In co/Ac/ have been tabulated in Table 3. For 3H-thymidine and BSA the rate constants of 0-1 [im. type filters are lower than the rate constants of the other three types of filters. The three larger pore size filters have about the same rate constants for thymidine. For polio virus the rate constant increases considerably with increasing pore size. 42
EMB
31
672
J. WARTIOVAARA AND OTHERS
FIGURES
2-7
Fig. 2. Transmission electron micrograph of thin section of a 0-2 /tm type Nucleopore filter. The channels have a straight but non-parallel course, x 15000. Figs. 3-6. Scanning electron micrographs of the glossy side of 0 1 , 0-2, 0-5 and 3 /*m type Nucleopore filters. The distribution of pores on the filter surfaces is uneven. Some confluent orifices are present. Magnifications x 15000. Fig. 7. Scanning electron micrograph of the matt side of Nucleopore filter with 3 /tm pores. Unlike the glossy side (Fig. 6) the surface of the matt side has shallow impressions, x 15000.
673
Cell contacts and induction 01 0-6 -
0-5 0-6 -
0-4 -
0-4r—
0-2 -
0-2-
—I 01
0-6 -
0-6 -
0-4 -
0-4-
0-2 -
0-2-
0-3
0-5
1
1-5
Pore size (/mi)
Fig. 8. Distribution of the pore sizes in different Nucleopore filters ( 0 1 , 0-2, 0-5, 3 /
