Studies on host-virus interactions in the chick embryo-influenza virus system. X. An experimental analysis of the von Magnus phenomenon
Journal of Experimental Medicine (JEM)
Abstract
An analysis has been made of factors contributing to the von Magnus phenomenon; i.e., the emergence of increasing quantities of non-infectious hemagglutinins (NIHA) in successive passages in the allantois of chick embryos of undiluted allantoic fluids infected with influenza virus.
Using the PR8 strain, the von Magnus phenomenon was pronounced when the serial seeds were obtained under conditions which permitted extensive inactivation of infectious virus during individual passages. Correspondingly, it was reduced but not abolished when precautions were taken to avoid accumulation of inactivated virus in the inocula. Thus, inactivated virus may be taken as a contributing factor.
Preparations of infectious virus obtained under conditions largely excluding the presence of inactivated virus were capable of yielding some NIHA on passage as long as sufficient amounts were injected to permit each host cell to adsorb several infectious virus particles. However, the fact remains that more NIHA was found in the harvests when the inocula contained a large proportion of non-infectious virus material.
Following injection of various types of seeds NIHA appeared in the allantoic fluids as soon as liberation of virus became detectable. This time relationship and the rates of release of non-infectious virus components seemed to exclude that the NIHA obtained consisted entirely of infectious virus which had been inactivated during incubation in ovo. It was apparent rather that NIHA other than that due to heat-inactivated virus was released.
Correlations between the infectivities and hemagglutinating capacities of over 50 standard and undiluted passage seeds and the compositions of the harvests derived therefrom on passage without dilution indicated that the corresponding activities in the yields did not depend entirely upon the relative concentrations of infectious virus and non-infectious hemagglutinins in the inocula but that apparently different forms of NIHA were obtained in successive undiluted passages.
... Standard seeds (ST) were prepared by allantoic inoculation of about 104 IDs0 and the aUantoic fluids were collected after further incubation of the eggs at 36 to 37°C. for 24 hours; i.e., when the infectivity titers were still rising or just bad reached their peaks (7,8). Under these conditions the yields consisted almost entirely of infectious virus and revealed IDs0/HA ratios in the order of l0 e'6. ...
... Heated standard seeds (A ST) were derived from standard seeds by exposure to 37°C. in ~tro for several days according to the technic described (12). Undiluted passage seeds (UP) were produced essen-tia21y by the method of you Magnus (9) as recorded elsewhere (8). The incubation periods employed for the consecutive passages of undiluted infected a]]antoic fluids were restricted to 24 hours. ...
... Progeny.--It is evident from the above experiments that conditions yielding NIHA do not readily develop following inoculation of dilute inocula of virus, even though ultimately multiple infection of cells must occur. This may have several reasons: (a) the virus which will become available for additional adsorption onto individual cells during the incubation period will be fully infectious under the conditions under which it is produced, and multiple infection with fully infectious virus produces the yon Magnus effect only to a relatively small extent (8); and (b) there is no "burst" phenomenon which would lead to sudden overloading of remaining susceptible cells. On the contrary, a rather slow saturation is expected since all cells, whether infected initially by the seed or by the progenies of the first or subsequent cycles, will release virus at nearly constant rates for long periods of time (18). ...
Certain aspects of the formation of non-infectious hemagglutinins (NIHA) in the chick embryo infected with influenza virus have been analyzed.
It was shown by the use of combined in ovo-deembryonation technics that little or no NIHA is released following infection with small doses of standard virus during the most active and constant growth periods of the virus extending to about the 36th hour of incubation in spite of the fact that multiple infection of cells must have taken place in the latter half of that period. A slight decrease in the ID50/HA ratios of the yields obtained after the 36th hour, coinciding with the falling off of virus production and release may possibly be explained in terms of inactivation of completed virus or leakage of as yet incompleted virus from damaged cells.
Exposure of the entodermal cells of the allantois of eggs deembryonated shortiy after injection of saturation or near saturation inocula of standard seed to large quantities of infectious virus added to the media at various times after infection and not extending over more than 2 hours resulted in a decrease of the ID50/HA ratios of the progenies only during the first 2 or possibly 4 hours after the primary inoculation. Later addition did not influence the yields. As discussed, such sudden and heavy exposures of cells are not expected to occur during the infectious process induced by small inocula of standard seed.
The possible role of destruction of cell receptors in NIHA production has been analyzed in several ways. The addition of receptor-destroying enzyme (RDE) to the media of deembryonated eggs after near saturation inocula of standard seeds, if anything, increased the ID50/HA ratios of the progenies, and that only when added during the first few hours following infection, presumably by reducing the changes for high multiplicity of infection of cells. In contrast, ultraviolet-inactivated virus, which retains its enzymatic activity, lowered, if anything, the ID50/HA ratios of the progenies, when present in the media of deembryonated eggs from the 2nd to 4th or possibly 6th hour after infection. Excessive amounts of irradiated virus may still cause some degree of interference under these conditions. Later addition of irradiated viruses were without effect with respect to NIHA production or interference. In attempts to alter the cell receptors prior to infection by potassium periodate (KIO4), it was noted that the addition of glycerol led to the appearance and partial retention for at least 24 hours of substances in the allantoic fluids which were capable of inactivating considerable proportions of standard virus. These data indicate that destruction of external cell receptors plays little if any role in NIHA production.
The implications of these findings are discussed.
In den vorausgegangenen Kapiteln wurde über die Widerstandsfähigkeit und ihr Gegenteil, die Empfänglichkeit von Organismen gegenüber einer lebenden Noxe gesprochen. Unter Benutzung des gleichen Einteilungsprinzips kann aber auch über die Unempfindlichkeit von Lebewesen gegenüber nicht vermehrungsfähigen Noxen (Giften oder Toxinen), d. h. solchen Stoffen, die ihre Giftigkeit gegenüber anderen, für sie empfindlichen Organismen zeigen, zusammenfassend berichtet werden.
Die Krankheiten der Myxovirusgruppe umfassen solche des Menschen und der Tiere. Zum Teil sind es reine Menschen- oder Tierkrankheiten, zum Teil Tierkrankheiten, die auch auf den Menschen übertragbar sind (Zoo-Anthroponosen). Bei den Menschenkrankheiten handelt es sich vorwiegend um respiratorische Syndrome.
Indem die Herausgeber in dem mir zugeteilten Thema die beiden Bezeichnungen „Resistenz“ und „Immunität“ gleichgeschaltet und ungestaffelt nebeneinander stellten, brachten sie schon zum Ausdruck, daß sie das Wort Resistenz nicht in seiner weitesten Bedeutungsmöglichkeit innerhalb der Biologie verstanden wissen wollten. Es sollte also nicht etwa Widerstandsfähigkeit eines Organismus gegen eine Noxe ganz allgemein bedeuten; denn dann wäre ja Immunität nur ein eng umschriebener Unterabschnitt eines so weit gefaßten Resistenzbegriffes.
Studies have been reported concerning the relationships between virus materials found in the allantoic membranes and media of eggs deembryonated after injection of Standard (ST), heat-inactivated (37°C.) standard (ΔST), and undiluted passage (UP) seeds.
It was found that the membranes always contained relatively more non-infectious hemagglutinins (NIHA) than the media and, correspondingly, the ratios between infectious virus and hemagglutinin units (ID50/HA) in the tissues were up to 1.5 log10 units lower than in the liberated progeny. These differences were seen not only following inoculation of undiluted ST, ΔST, and UP seeds, the progenies of which always contain considerable proportions of NIHA, but also when dilute ST inocula were employed which lead to the liberation of only infectious virus. Essentially similar differences in the ID50/HA ratios were observed also in the allantoic membranes and fluids obtained from growth curve experiments in the intact chick embryo employing the various types of seeds.
In correlating the liberated virus materials in the media of deembryonated eggs to those in the membranes it was noted that in any given 2 hour interval during the phase of nearly constant production and release up to 10 times the quantity of infectious virus was shed as was present in the tissues at the onset of that period. In contrast, only about ¼ of the hemagglutinins were released during the same time. The viral (V) and soluble (S) complement-fixing antigens were found in the tissues but no detectable quantities were released during any 2 hour interval.
The NIHA in the membranes apparently is located within the cells since it could not be released by the action of RDE. Intracellular inhibitors of hemagglutination were readily inactivated following inoculation of undiluted ST, ΔST, or UP seeds but not when ultraviolet-inactivated virus was used. The inhibitor activity decreased in proportion to the hemagglutinins produced.
Transfer of infected deembryonated eggs to the cold room after production and liberation of progeny were well under way immediately halted further release but in the tissues the status quo was maintained and release was resumed on return to the 37°C. incubator. The addition of potassium cyanide to the medium of deembryonated eggs at 37°C. during the period of nearly constant production and release of virus material reduced immediately and to comparable extents the ID50 and HA titers in the tissues and liberation decreased in proportion. On removal of the cyanide 2 hours later, both titers in the tissues gradually returned to those of the untreated control eggs with a corresponding increase in liberation. The ID50/HA ratios were not affected by these manipulations. It is concluded that the NIHA in the membranes forms part of a dynamic process.
An attempt has been made in the discussion to integrate the present results with previous observations concerning the formation of incomplete forms of virus and their nature and role in the infectious process.
Multiplication of influenza viruses in tissue cultures of chick embryo kidney was investigated. The results were as follows:1.
Up to a certain amount of the inoculum there was a direct relationship between the amount of inoculum and of released hemagglutinin and infectious virus. A higher inoculum was not followed by an increased yield.
2.
An inoculation of full infectious virus yielded high I/H-relations only by infection of about 1% of the cells.
3.
Under optimal multiplication-conditions one cell produced 12,2 EID50-units and 400–500 HA-particles. At the point of maximal yield 1.79 × 104 cells produced one S-antigen- and 7.87×103 cells one V-antigen-unit.
4.
The amount of inoculated virus did not influence the length of the eclipse-phase of A/Singapore. With the strain B/Lee and D/Sendai, however, there exists a relation between the inoculum and the release of infectious virus. The release of free and cell-associated hemagglutinin and complement-fixing antigen depends on the quantity of the inoculum.
1. After inoculation of eggs with a large dose (10 8·6 ID 50 ) of influenza virus, chorio-allantoic membranes harvested from the 3rd to at least the 21st hr. (but not at the 1st or 2nd hr.) gave rise to large amounts of soluble antigen in 6 hr. in the chorio-allantoic membranes of sub-inoculated eggs (6 hr. soluble antigen production test). These amounts were larger than resulted from inoculation of equivalent numbers of ID 50 of standard allantoic fluid virus. By the conventional 72 hr. test, the infectivity of the chorio-allantoic membranes increased 1 hr. later, at the 4th hr. after inoculation.
2. Allantoic fluids and chorio-allantoic membranes harvested from passages of very large and of small amounts of influenza virus were tested by 6 hr. soluble antigen production. The relative amounts of soluble antigen formed by different seeds appeared to be determined by their ID 50 : HA ratios, irrespective of whether they were allantoic fluid or chorio-allantoic membrane preparations. Larger amounts of soluble antigen were formed in 6 hr. per ID 50 inoculated by preparations having lower ID 50 : HA ratios, i.e. containing larger amounts of ‘incomplete’ virus, and vice versa.
3. It is suggested that high infectivity in the 6 hr. soluble antigen test per ID 50 inoculated is due to a content of ‘incomplete’ virus, or of virus precursors: possibly these are in fact the same.
4. More particles of an incomplete virus preparation than of standard allantoic fluid virus must be inoculated to produce a given amount of soluble antigen in 6 hr. It is suggested that a process resembling multiplicity reactivation must occur in cells infected with more than one particle of ‘incomplete’ virus.
The formation of incomplete virus by the influenza B/Lee strain on serial undiluted passages in the allantoic cavity has been studied in growth curve experiments with both the allantoic membranes and fluids. It is shown that large amounts of incomplete virus are produced by the B/Lee strain after 5 to 7 undiluted transfers.
Comparative studies on the A/PR8 and B/Lee strains show that maximal incompleteness is consistently attained in UP-2 to UP-4 with the PR8 strain and in UP-5 to UP-7 with the Lee strain indicating that the pattern of behavior is an inherent characteristic of the strains.
The Lee strain is found to be capable of eliciting at least the same amounts of incomplete virus as is the PR8 strain and this vitiates the sequence of the strains in the “incomplete virus gradient” suggested byFazekas de St. Groth andGraham, and makes its existence questionable.
After allantoic injection of chick embryos with a known amount of influenza virus, the process of adsorption of the agent onto host cells and infection of them can be interrupted at a given time by the administration of large quantities of heterologous virus inactivated by irradiation. A sudden great increase in the amount of free virus in the allantoic fluid occurring after 6 hours in the case of the PR8 strain, and 9 hours in that of the Lee strain, indicates that the untreated virus associated with the host cells has multiplied. The length of the period preliminary to this increase remains the same even though the concentration of the original inoculum is varied over a wide range. Since administration of the irradiated virus leaves no susceptible host cells, because of the interference phenomenon, and further adsorption of active virus is minimized or entirely prevented, practically the entire new increment of virus can be found in the allantoic fluid and assayed; for every ID(50) adsorbed about 50 ID(50) are released. Homologous irradiated virus, on the other hand, when injected after infection of the allantoic sac, reduces the yield of virus to a more or less considerable extent. Some inhibitory effect can still be observed when the homologous irradiated virus is given several hours after infection. This effect is linked to the virus particle and destroyed by prolonged irradiation.
Influenza A or influenza B virus rendered non-infective by ultraviolet radiation was found to be capable of producing interference
with the multiplication of active influenza viruses in the chick embryo. Certain temporal and quantitative relationships affecting
the interference phenomenon with this host-virus system were studied. An hypothesis of the mechanism of interference between
the influenza viruses is proposed and discussed.
The period and rate of liberation of influenza virus from entodermal cells of the allantois have been studied by deembryonating eggs within a few minutes after infection, exchanging the medium thereafter at hourly intervals and assaying the virus concentration in the harvests thus obtained (differential growth curves). If the inoculum was sufficiently large, presumably all available cells immediately became infected and only 1 infectious cycle was expected to occur. If the inoculum was small, so that only a fraction of the cells adsorbed virus, the infectious process was held to 1 cycle by continuous exposure of the remaining susceptible cells to RDE. In either case, the results obtained indicate that once cells have been infected they produce and liberate virus at nearly constant rates for periods of 30 hours or longer before the yields decrease rapidly. Evidence has been presented which strongly suggests that such prolonged periods of liberation are observed not only in deembryonated eggs but also in the intact chick embryo.
Attempts have been made in the discussion to reconcile these findings with previous estimates of the liberation period and to integrate them with histologic observations and electron micrographs of thin sections of infected allantoic membranes having a bearing on the mode of liberation.
The usefulness of the deembryonation technic has been analyzed as a tool in the study of various problems in the growth cycle of influenza virus in the entodermal cells of the allantoic of chick embryos. Various improvements in the deembryonation technic have been described. The method readily permits repeated sampling of the medium at various stages after infection (cumulative growth curves) or frequent exchanges of the medium (differential growth curve). However, the yield of infectious virus or of hemagglutinins is less than that observed in the intact chick embryo. The difference observed is greater than can be accounted for by the reduction in the available host cells and is assumed, therefore, to be due in part to interruption of blood and nutrient supply to the cells. This handicap can be overcome by the combined in ovo-deembryonation technic, in which deembryonation is performed at any desired time after infection of the intact chick embryo, and the medium is collected and analyzed after 1 to 3 hours of further incubation. The value of the technic is demonstrated by the fact that liberation of virus from infected cells can be detected earlier than in the intact egg. Furthermore, it continues at a nearly constant rate for many hours, thus proving to be erroneous previous inference which had been based upon in ovo experiments. The technic also permits readily the addition and subsequent removal of substances that might interfere with viral propagation. As an example a study was made of the effect of the receptor-destroying enzyme of V. cholerae (RDE) when added to the medium of eggs infected prior to deembryonation. By carefully grading the dose of virus and using an appropriate amount of RDE, one-step growth curves were obtained indicating that those cells not directly invaded by the seed virus were subsequently protected against infection by action of the enzyme. The smaller the amount of virus the less RDE was required in order to note a protective effect. With a decrease in the period of exposure to RDE regeneration of cell receptors became increasingly more apparent in that correspondingly greater amounts of virus were produced and liberated late in the incubation periods. These results confirmed and extended those reported by Stone. More extensive applications of these technics will be reported in subsequent papers of this series.
Various strains of influenza virus produce a cytopathogenic effect in cultures of HeLa cells. The virus could not be passed in series. Virus partially or even completely inactivated with respect to infectivity by exposure to 37°C. or ultraviolet light retained some of its cytopathogenic effect.
No evidence has been obtained of an increase in infectious virus in HeLa cultures, but an increase in hemagglutinins and in both viral and soluble complement-fixing antigens became detectable during incubation. These virus materials apparently were not released from these cells prior to their destruction.
These results suggested that HeLa cells are capable of supporting an incomplete reproductive cycle of influenza virus. The fact that radioactive phosphorus was readily incorporated into the hemagglutinin supplies strong evidence for this interpretation.
Procedures which make possible the enumeration of both infective and hemagglutinating influenza A virus particles have been developed and used in a quantitative investigation on the reproduction of the agent. Infective particles were found to be highly unstable and their half-life was only 147 minutes in allantoic fluid at 35°C. both in vitro and in vivo. The instability of infective particles provides an explanation for the rapid accumulation of non-infective particles which retained the hemagglutinating property. The number of non-infective (N) particles was determined from the difference between the number of hemagglutinating (H) particles and the number of infective (I) particles as indicated by the relation: [N] = [H]– [1]. When the half-life of infective particles was taken into account, both infective and hemagglutinating particles were found to disappear from the allantoic fluid; i.e., were adsorbed by the allantoic membrane, at the same logarithmic rate after inoculation. Inoculation of any number of particles up to 3 x 107 was followed by a constant and progressive decrease in the proportion of unadsorbed particles from 0 to 4 hours. Approximately 20 per cent of particles were unadsorbed at 2 hours and about 5 per cent at 4 hours. Inoculation of 3 x 108 or more particles led to a larger proportion of unadsorbed particles at 4 hours. The maximum number of particles adsorbed was computed to be about 1.6 x 109. The concentration of both infective and hemagglutinating particles increased rapidly in the allantoic fluid after 4 hours when any number of infective particles up to 3 x 107 was inoculated. With such inocula, the rate of increase during the logarithmic period was constant and the time to double the concentration of infective or hemagglutinating particles was 46 minutes. With larger inocula, i.e. 3 x 108 particles, the concentrations of infective and hemagglutinating particles did not increase until after 8 hours and the rate of increase was much slower. The time to double the concentration of either then became 92 minutes. The number of infective particles was approximately equal to the number of hemagglutinating particles during the logarithmic increase period when any number of infective particles up to 3 x 106 was inoculated and no more than 106 non-infective particles were included in the inoculum. This finding was taken to indicate that all or almost all particles produced and released under these conditions were infective. That such particles became inactivated rapidly and led to the accumulation of an increasing number of non-infective particles after the logarithmic period can be explained by the short half-life of infective particles. The number of infective particles was no larger than one-tenth the number of hemagglutinating particles during the logarithmic increase period after 3 x 107 or more infective particles had been inoculated or when smaller inocula were used which also contained 3 x 107 or more non-infective particles. Non-infective particles prepared in vitro at 35° or 22°C. were as effective as those which accumulated in vivo in diminishing the proportion of infective particles in the yield. The extent of the reduction in the proportion of infective particles was directly related to the number of non-infective particles included in the inoculum. The yield of hemagglutinating particles was diminished when the inoculum contained 3 x 107 or more non-infective particles. The rate of increase was reduced so that the time to double the concentration became 92 minutes when the inoculum contained 3 x 108 non-infective particles. It appears from these findings that the single condition which will lead to the emergence of non-infective particles during the logarithmic period is a high initial particle-cell ratio. Because non-infective particles are equally as effective as infective particles in producing this result, it seems probable that the appearance of non-infective but hemagglutinating particles is not a necessary accompaniment of the reproductive process.
THE term `incomplete' virus was used by von Magnus1 for designating products appearing on serial intra-allantoic passage of large inocula of influenza virus, and distinguished from `standard' virus obtained after inoculation of diluted seed material by the apparent lack of infectivity with retained high hæmagglutinating capacity. In sedimentation studies Gard, von Magnus et al. 2,3 found the incomplete virus to be more inhomogeneous than standard virus and to have a considerably lower sedimentation constant. Birch-Andersen and Svedmyr4, examining several preparations in the electron microscope, failed to demonstrate any significant differences in the mean particle sizes, although the incomplete virus tended to show a somewhat wider variation in size and morphology.
Excerpt
The trend of virus research in recent years has been to shift emphasis from the isolated viral particle itself to the virus-cell complex as an integrated and unique biological unit. The interplay of the two component parts of the complex expresses itself on one hand in the usurpation or modification of cellular functions under the influence of the virus, and on the other, in the manner in which the cell's supply of various limiting factors seems to determine how much and what kind of virus is going to be produced. Or else, the two co-exist in a perfectly balanced internal equilibrium, as in lysogenic systems. Under the influence of various factors, this equilibrium is subject to upsets which may operate in favor of either cell or virus. Recognition of the biological unity of the virus-cell complex is in tune with the realization that the extra-cellular “resting” forms are relatively unrewarding...
Particle counts of influenza virus preparations were made by two independent techniques ; there was good agreement in the counts found by the two methods. Counts made on four strains of influenza virus and one strain of 'incomplete' virus showed that at the agglutination end-point there was about one virus particle per red cell. Parallel infectivity measurements of these strains made under optimal conditions showed that about ten virus particles corresponded to one ID 50. Many studies of the quantitative aspects of influenza virus growth have involved assumptions about the number of virus particles which constitute one infective dose. It has seemed worth while to test these assumptions by counting the number of virus particles seen by electron microscopy which correspond to one egg infective dose. Two independent techniques were used for making the particle counts, the first based on the spray technique used by Luria, Williams & Backus (1951) in which counts are made of the virus particles mixed with polystyrene latex particles of known concentration, and the second based on the method of Dawson & Elford (1949) in which the virus particles are adsorbed on red cell ghosts. In the spray technique the virus particles were counted before and after absorption with red blood cells and the difference gives the number of particles which should be specifically adsorbed to the red cell ghosts. This figure was found to correspond extremely well with the estimated number of virus particles on the surface of the red cell ghosts. In this way an estimate of the number of virus particles in different preparations of influenza virus was obtained, and this figure correlated with simultaneous measurements of the agglutination titre and infectivity of the preparation. This paper describes the results obtained. Materials
Summary1. The mean number of allantoic cells per square centimeter of the allantoic membrane in 10-day embryonated chicken eggs has been estimated by direct counting with the phase contrast microscope. 2. In the chronic (CA) portion of the allantoic membrane there were 1.66 × 105, and in the amnion-yolk sac (AYA) portion 3.00 × 105 allantoic cells per cm2. 3. The total area of the allantoic membrane was measured directly. The area of the CA portion was 57.9 cm2, that of the AYA portion was 29.1 cm2, and that of the entire membrane was 87.0 cm2. 4. The number of cells lining the allantoic cavity of the 10-day chick embryo was estimated to be 1.8 × 107.