Attachment of Two Myxoviruses to Ciliated Epithelial Cells

Transcription

1 ,1. gen. Virol. (I97O), 9, Printed in Great Britain Attachment of Two Myxoviruses to Ciliated Epithelial Cells By R. R. DOURMASHKIN AND D. A. J. TYRRELL Clinical Research Centre Laboratories, National Institute for Medical Research, London, N.W. 7 (Accepted I I June I97o) SUMMARY It is thought that influenza and related viruses enter susceptible cells, such as those of tissue cultures and the chorioallantois, by being adsorbed to the surface and then taken in by an active process termed ' viropexis'. It has been suggested that this active process resembles phagocytosis (Fazekas de St Groth 1948). However, influenza viruses commonly invade the ciliated epithelium of the respiratory tract of the intact host, which is thought not to be actively phagocytic. Organ cultures of such epithelium are extremely susceptible to infection (Hoorn & Tyrrell, I969). It was therefore of interest to use these to investigate the mechanism of entry of influenza viruses into ciliated epithelial cells; it was uncertain whether the primary target cells would be the ciliated or the mucus-secreting ceils. As infection was so efficient, it seemed likely that the virus might 'exploit' in some way the sweeping action of the cilia and, rather than being moved on by their activity, might attach and then enter the cells directly or indirectly. Further studies on the entry of virus into non-ciliated cells appeared during this work and these will be discussed later. METHODS Tissue. Organ cultures of guinea-pig trachea were set up in 6o mm. Petri dishes using the techniques of Hoorn (Hoorn & Tyrrell, ~ 969), and were maintained at 33 overnight in Eagle's medium in 5 % CO2 in air. Only fragments showing definite ciliary activity next morning were used in experiments. Viruses. The influenza strain A 2/ENG 12/64 was used. Inactivated virus was used in the form of a pool stored at 4 with azide for months. It still had full haemagglutinin and neuraminidase activity and it was purified and concentrated first by adsorption and elution from human red cells, and then by centrifugation to a pellet which was subsequently fractionated by rate centrifugation in a sucrose density gradient. Active virus was obtained by harvesting eggs inoculated 2 days before, clarifying the fluid by low speed centrifugation and producing a pellet by centrifugation at 2o,ooo rev./min. for 3o min. The pellet was resuspended in Eagle's medium and used with minimum delay. The titres of haemagglutinin were usually I/IOO,OOO or greater. Active Sendai virus was prepared in a similar way. Adsorption of virus. Single pieces of tissue about 2 mm. square were transferred to a cup in a disposable plastic titration tray (Linbro) in the cold room and each received one drop of virus. Adsorption usually continued for I hr and then the cup was cut out and filled with fixative; it was floated first on a bath of warm water if periods of incubation were required for the experiment. 6 Vlr 9

2 78 R. R. DOURMASHKIN AND D. A. J. TYRRELL Fixation and staining. To obtain good fixation of both membranes and intracellular structures, two schedules were followed. In the first (sequential fixation) the organ cultures were flooded with freshly prepared 3 Yo glutaraldehyde in o- I M-phosphate buffer solution. After 5 rain. they were washed in o.i M-phosphate buffer, and then fixed in I ~o osmic acid in phosphate buffer (Millonig, 1962 ). They were then washed, dehydrated in alcohol and embedded in Araldite in the usual way, taking care to ensure adequate infiltration by slow rotation of the tissues suspended in resin. This procedure gave good fixation of the virus particles and cilia, but did not reveal much cellular detail. The second schedule was the double fixation method of Hirsch & Fedorko (t968), using glutaraldehyde and osmic acid simultaneously. This method, although showing cellular detail of the tracheal epithelium, resulted in swelling of the ciliary membranes. However, the continuity of the membranes remained intact whereas they appeared broken in techniques involving the use of bufferedosmic acid alone. Flat embeddings were used, the epithelium being oriented under a dissecting microscope. Thin sections were cut and stained with uranyl acetate. RESULTS Virus particles were readily identified in sections after exposure to influenza A 2. After 5 rain., most seemed to be entangled in the mucus overlying the cilia. After I hr particles were floating free, a few were attached to microvilli, but most were seen lined up along the membranes of cilia, and apparently attached to them (Fig. I). Calculations based on haemagglutinin titrations suggested that only a small fraction of the virus added was attached to the epithelium. The number of attached particles of either infectious or noninfectious virus was reduced rapidly to less than 3o % by warming at 37 for 3 rain. (Table I). The effect was seen whether large or moderate numbers of particles were attached. Tissue was also washed in cold saline after the adsorption step and then incubated as in the preparation of specimens for electron microscopy. About lo units of haemagglutinin was released into the warmed medium, but a similar amount was also released from washed cultures held in medium at 4. The mechanism by which virus particles attached to cold cells was carefully studied. It seemed that both influenza and Sendal virus often stuck to the membrane by quite a small area of contact, but occasionally the virus particle or cell profile was altered so that the two membranes were apposed, sometimes closely, over a larger area. Sometimes the approximated areas seemed rather thickened or blurred, but this was never seen in areas where the plane of section ran perpendicular to that of the cell membrane, as shown by the fact that the triple-layered unit membrane was clearly resolved at both sides of the point to which the virus was attached (Figs. 2, 3). The membranes of cilia were sometimes interrupted or possibly torn, but it was not possible to relate this to the adsorption of virus and, after virus had disappeared from warmed cultures, the membranes were indistinguishable from those of uninfected cultures (Fig. 4). In warmed cultures, and mainly in those warmed for 3 rain., profiles were seen which showed unequivocally (Fig. 5) that the membranes of Sendai virus and the cell fused and the virus nucleocapsid entered the cilium, became partly uncoiled and apparently spread down between the membrane and the filaments (Fig. 6). A similar process can apparently occur when Sendai virus adsorbs to microvilli. No such appearances were seen in cultures exposed to influenza virus. Instead, in cultures

3 Attachment of" two myxoviruses to ciliated epithelial cells 79 Fig. I. Electron micrograph of an organ culture of guinea-pig trachea, incubated with influenza virus for I hr at 4. Large numbers of virus particles are adsorbed to the cilia. ' Double fixation' with glutaraldehyde and osmium tetro ide (Methods). Table t. Numbers of influenza virus particles attached to cilia Number of particles r Times at two Counts/#m. of Counts as percentage of those temperatures ciliary membrane after I hr of adsorption at 4 A ~ z 4 37 Expt... I 2 3 I z 3 60 o 0"56 0"58 3"20 I I00 60 I -- 0"44 o' I3 6o 3 o"13 o'i5 o' "24 o'

4 8O i!iiill i~; ~f!i ~ i~ R. R. DOURMASHKIN AND D. A. J. TYRRELL Fig. 2. Organ culture incubated with influenza virus for I hr at 4. Note appearance of virus particles, marked by arrows, overlapping the membranes of cilia. Double fixation.

5 Attachment of two myxoviruses to ciliated epithelial cells 81 Fig. 3. Influenza virus adherent to membranes of cilia, from an organ culture incubated with virus for I hr at 4. Sequential fixation. Fig. 4- Organ culture incubated with influenza virus for ~ hr at 4, then 3 min. at 37. Sequential fixation. Particles adhered to cilia as above; in warmed cultures they were less numerous.

6 82 R.R. DOURMASHKIN AND D. A. J. TYRRELL containing active virus, intact virus particles appeared to be present within the cilia (Fig. 2). Serial sections indicated that the appearance of virus particles within a cilium, or crossing the ciliary membrane, always represented a portion of a virus particle, the remainder being completely outside the cilium. Thus, electron micrographs of virus appearing to cross the Fig. 5. Organ culture incubated with Sendai virus for I hr at 4, then 3 rain. at 37. Many particles adhere to the cilia; fusion of virus membrane to membranes of cilia is seen with virus nucleocapsid distributed along the microtubules of lhe cilia.

7 Attachment of two rnyxoviruses to ciliated epithelial cells 83 Fig. 6. Influenza virus particles adsorbed to microvilli and to the surface of an epithelial cell. Note fibres projecting from microvilli. (a) Organ culture incubated for I hr at 4. Double fixation. (b) Organ culture incubated I hr at 4, then I min. at 37. Sequential fixation. Fig. 7. Portion of a mucus-secreting cell, from an organ culture incubated with virus for I hr at 4, then ~ rain. at 37. Note that the appearance of the microvilli differs from that of other epithelial cells. Sequential fixation.

8 8 4 R.R. DOURMASHKIN AND D. A. J. TYRRELL ciliary membrane probably represent a superimposition of virus and cell membrane within one section. Virus particles were seen to adsorb to microvilli in the tracheal organ cultures, in the same way as to cilia (Fig. 6a, b). It may be noted that the filamentous material emanating from the microvilli, best seen in sequentially fixed cultures, might play a part in virus entrapment. Fig. 8. Influenza virus adsorbed to human red cells. (a) The cell membrane was sectioned tangentially. Virus particles overlap the edge of tl'te membrane and appear to be fused with it. (b) Cell membrane sectioned to be visible as a triple layer. There is no evidence of fusion of menabranes of virus and erythrocyte.

9 Attachment of two myxoviruses to ciliated epithelial cells 85 No virus was seen adsorbed on the microvilli of mucus-secreting cells (Fig. 7). Another approach to the examination of the relationship of virus to cell membrane was to study the appearance of influenza virus particles adsorbed to red blood cells. Virus appeared to be fused to cell membranes only when the membranes were sectioned tangentially (Fig. 8 a). In every instance where the double layer of the cell membrane was clearly seen, the virus particles adhered only by means of their peripheral projections (Fig. 8b). Inactivated and active virus seemed to behave in exactly the same way. r Times at two temperatures Table 2. Uptake of influenza virus by Tetrahymena pyriformis 4 37 hr o I hr 1 hr lo rain. I hr 4o rain. I hr o I hr I hr i hr i hr o o o o o Mixture used Virus haemagglutinin titre of supernatant fluid Virus and Tetrahymena I [, Virus only Virus and Tetrahymena 1/96 1/8 Virus and Tetrahymena I/, 6 Virus and Tetrahymena I/2 Virus and Tetrahymena disrupted by blending i/i28 Virus and Tetrahymena previously incubated at 33 in cholera filtrate 1/32 Virus and Tetrahymena previously incubated in broth 1/16 Virus only I/256 Tetrahymena were sedimented at low speed and mixed with diluted allantoic fluid containing influenza virus A 2/Eng/344/68. After the indicated procedure and another low speed centrifugation the supernatant fluids were titrated for haemagglutinating activity. To study in more detail the relationship between viruses and cilia, some preliminary experiments were made with a protozoon from which the cilia can be recovered in a pure state. It was shown first that influenza virus attached rapidly to the ciliate Tetrahymena pyriformis. An axenic culture was centrifuged down at low speed and resuspended in distilled water at approximately IO %. The suspension was mixed with an equal volume of allantoic fluid diluted by approximately one-third in distilled water. The results are shown in Table 2. The haemagglutinin reappeared in the medium after a time and could be recovered by disruption of the organism, but not by treatment with cholera filtrate. On electron microscopy it became clear that the virus was nowhere attached to cilia, but was ingested into the protozoon's food vacuole, from which presumably it was later regurgitated (Fig. 9). DISCUSSION It is of great interest that representative viruses which invade respiratory epithelium can stick firmly to the cilia; these observations tend to confirm our original hypothesis. It is desirable now to determine whether viruses of other groups which also enter the same sort of epithelium, attach in the same way. The mechanism of the entry of Sendai virus nucleocapsid into the cilia is clear and Morgan & Howe 0968) have shown that the virus enters chorioallantoic cells in the same way. We are not convinced from our micrographs that influenza virus does so in the same way, and only one micrograph of Morgan & Rose (I968) actually shows the moment

10 86 R. R. DOURMASHKIN AND D. A. J. TYRRELL Fig. 9a. Tetrahymena pyriform& incubated with influenza virus; this section shows the cilia of the buccal overture and virus particles engulfed in the food vacuole (V). Virus particles do not adhere to cilia. Fig. 9b. Insert: A higher magnification of the area marked in Fig. 9a.

11 Attachment of two myxoviruses to ciliated epithelial cells 87 of penetration; this is unconvincing. There are enormous differences in the electron micrographic appearances of red cells incubated with influenza and Sendai virus. Not only are cells lysed by exposure to Sendai virus, but nucleocapsid is released; we have confirmed that this does not occur in influenza virus treatment of red cells. It has been shown that at least some influenza virus remains accessible to neutralizing antibody for many minutes after it attaches to chorioallantoic membrane in a way which cannot be reversed by RDE (Ishida & Ackermann, ~956). After mixing for a few minutes with the normal cytoplasmic particles which bud from these cells, the virus is fully infectious but seems to be disrupted when examined by negative contrast (Hoyle, Horne& Waterson, t962). We have seen nothing corresponding to this in virus particles which have interacted with cilia. It is therefore likely that Sendai and influenza viruses react with cells by different mechanisms; in our view there is no evidence that the nucleocapsid which enters the cell from the Sendai virus particle is that which initiates infection. The entry of influenza virus into cells has not been visualized for certain. During this work, a controversy developed between Morgan and his colleagues, who demonstrated the entry of myxoviruses into cells by membrane fusion, and Dales and his colleagues who added to their earlier evidence that viruses such as herpes simplex and vesicular stomatitis enter cells by engulfment (Dales & Silverberg, ~969; Simpson, Hauser & Dales, I969; Morgan, Rose & Mednis, ~968). Viruses of different groups are constructed and multiply in such different ways that it is to be expected that they also enter cells in different ways. It is important to know which mechanism of entry relates to each virus, and whether the virus can then initiate infection; one type of virus may enter different cells in different ways. Thus, influenza viruses have been seen inside apparent phagocytic vacuoles where they may be digested; in order to replicate, however, they must penetrate the membrane and enter the cytoplasm and this process has not been clearly seen. The disagreement between the views of the groups of Morgan and Dales may be slight, since only herpes simplex virus has been shown to enter by both membrane fusion and by engulfment, and the experimental systems used differed appreciably. We wish to thank Dr G. Schild for supplying inactivated virus and Dr D. Lee for Tetrahymena pyriformis. We also thank Mrs R. Buckland, Mrs Joan Richmond, Misses Helen Woodward and Angela Stone, and Mr S. J. A. Tyrrell for technical assistance. REFERENCES DALES, S. & SmVERBERG, H. (I969)- Viropexis of herpes simplex virus by HeLa cells. Virology 37, 475- FAZEKAS DE ST GROTH, S. (1948). Viropexis, the mechanism of influenza virus infection. Nature, London x62, 294. HmSCH, V. G. & FEDORKO, M. E. (I968). Ultrastructure of human leukocytes after simultaneous fixation with glutaraldehyde and osmium tetroxide and 'postfixation' in uranyl acetate. Journal of Cell Biology 38, 615. HOORN, B. & TYRRELL, D. A. J. (I969). Organ cultures in virology. Progress in Medical Virology H, 4o8. HOYLE, L., r:orne, R. W. & WATERSON, A. P. (I962). The structure and composition of the myxoviruses. Ill. The interaction of influenza virus particles with cytoplasmic particles derived from chorioallantoic membrane cells. Virology x7, 533. ISH:DA, N. & ACKERMANN, W. W. (I956). Growth characteristics of influenza virus. Properties of the initial cell-virus complex. Journal of Experimental Medicine xo4, 5oI. M~LLONIG, G. (I962). Proceedings of the 5th International Conference of Electron Microscopy, vol. II, p. 8. New York: Academic Press. MORGAN, C. & HOWE, C. (I968). Structure and development of viruses as observed in the electron microscope. IX. Entry of parainfluenza I (Sendal) virus. Journal of Virology 2, : I22.

12 88 R.R. DOURMASHKIN AND D. A. J. TYRRELL MORGAN, C. & ROSE, H. M. (1968). Structure and development of viruses as observed in the electron microscope. VIII. Entry of influenza virus. Journal of Virology 2, 925. MORGAN, C., ROSE, H. M. & MEONIS, B. (I968). Electron microscopy of herpes simplex virus. I. Entry. Journal of Virology 2, 507. SIMPSON, R. W., HAUSER, R. E. & DALES, S. (i969). Viropexis of vesicular stomatitis virus by L cells. Virology 37, 285. (Received 8 April I970)