J. gen. Virol (I975), 28, 59-72
59
Printed in Great Britain
Adaptation Studies with Ross River Virus: Laboratory
Mice and Cell Cultures
By W. P. T A Y L O R *
AND I . D . M A R S H A L L
Department of Microbiology, John Curtin School of Medical Research,
Australian National University, Canberra, A.C.T. 26oH Australia
(Accepted 5 March I975)
SUMMARY
Ross River virus, an Australian group A arbovirus, was adapted by serial passage
to cell cultures and to day old mice. The results of titrations in mice of different ages
allowed the comparison of virulence between different stocks. Passage in cell
cultures depressed the virulence of virus while passage in mice raised the level of
virulence. Clones of original virus populations revealed heterogeneity with respect
to virulence but none of the 4I clones was as highly virulent as virus passed Io times
in mice. Clones selected in sequence during serial passage in mice indicated that
adaptation proceeded by the overgrowth of variants of increasingly higher
virulence, and that clones from relatively highly passaged strains were still heterogeneous in virulence.
INTRODUCTION
The recovery of viruses from either field or pathological samples frequently involves an
adaptation of the agent to a suitable laboratory host. At times several serial passages may be
needed to establish evidence of isolation or to raise pathogenicity for animals or cells to levels
considered necessary for laboratory experimentation. Bradish, Allner & Maber (1971) have
commented on the importance of alterations in virulence, protective properties and virus
population changes that may occur even within very few animal passages, but scant attention is usually given to the original characteristics of an isolate and its integrity in the
passaged variant finally studied. Recent reports from this laboratory (Mims et al. 1973;
Murphy et al. I973) indicate that important pathological events may be overlooked by using
fast growing, adapted virus stocks.
Adaptation to a novel environment, either in the laboratory or the field, implies a mechanism for variation, usually ascribed to random mutation or selection of variants (Fenner &
Cairns, 1959); with the arboviruses neither host induced variation (Hoskins, 1959) nor
genetic recombination (Burge & Pfefferkorn, 1966) has been demonstrated convincingly.
Constant exposure to a selection pressure ensures that the variant best suited to the system
predominates. Adaptation has been shown to proceed by the selection of a pre-existing
minority population (II'Enko et al. I968) or by the selection of de novo mutants arising
during the course of replication in the new host (Churchill, Chubb & Baxendale, 1969).
The availability of unmanipulated pools of Ross River virus (RRV), an Australian group
A arbovirus (Doherty et al. I963), and the alteration in pathogenicity observed during serial
passage in infant mouse brain (Gard, Marshall & Woodroofe, 1973) prompted a more
* Present address: Federal Department of Veterinary Research, Vom, Benue Plateau State, Nigeria.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 14 Jun 2017 21:10:15
60
W.P.
T A Y L O R A N D I. D . M A R S H A L L
detailed investigation of the processes underlying adaptation, the results of which are reported here and in an accompanying paper (Taylor & Marshall, I975).
METHODS
Viruses. Strains of RRV with the prefix NB were recovered from pools of Aedes vigilax
mosquitoes or Mansonia linealis (Gard et al. 1973) collected at Nelson Bay on the central
coast of New South Wales.
Strain NB5o92 was used as the original field mosquito pool (NB5o92/field), after one or
I2 passages in infant mouse brain (MB), (NB5o92/MB[ ; NB5o92/MBI2), as the fifth Vero
cell passage from NB5o92/MBI (NB5o92/MBI/Vero-5), or as virus maintained for five
alternating passages between infant mice and laboratory reared Aedes aegypti mosquitoes
(MOS) (NB5o92/MB 3 MOS2).
Strains NB6o24 and NB7t 13 were used as the original infected mosquito pools (NB6o24[
field; NB7II3/field) while NB6o M was also used after one infant mouse brain passage
(NB6o24/MBt).
The prototype RRV strain T48 (Doherty et al. 1963), was used after 12 passages in infant
mouse brain (T48/MBI2).
Procedures involving cell cultures. Plaque assays of virus infectivity were made in triplicate
using Vero cell monolayers inoculated with o-I ml samples from a tenfold dilution series
prepared in chilled borate-buffered gelatine saline, pH 7"2 (Fazekas de St. Groth, Graham &
Jack, I958). Plaques were counted after the addition of neutral red counterstain following
incubation for four days at 35"5 +- o'5 °C under tris buffered agar overlay; virus infectivities
are expressed as p.f.u./ml or p.f.u./g.
Clones were obtained from NB5o92]MBI, NB5o92/MBI2, NB6o24[field or from stocks
prepared during the serial passage of NB6o24[field. In each case virus was obtained by
aspirating the agar and underlying cell debris from well developed plaques on monolayers
of Vero cells. Homogeneity of the selected variant was indicated by one or two cycles of
replating, harvesting only from monolayers bearing single plaques. Final stocks of selected
clones of virus were grown in the same cell type held under fluid maintenance medium
until c.p.e, was apparent. Cell-associated virus was released by freezing and thawing and
the suspension clarified by centrifuging at low speed.
Tissue harvests. Mouse brains were aspirated into glass ampoules with a suction apparatus
without opening the cranium. Mouse carcases (MC), or parts thereof, were obtained by the
removal of the head, skin and viscera of exsanguinated mice. All stocks were prepared at
IO ~ w/v suspensions in normal saline.
Estimation of virus infectivity and virulence in experimental mice. An outbred multicoloured line of Walter and Eliza Hall Institute mice was used throughout. The age of experimental animals was an important feature of this work and group allocations were as
follows: 'day old' mice were used during the first 24 h of life, 'infant mice' were not older
than 48 h, while ' one week' old mice had completed seven but not eight full days of life.
Older age groups were obtained by holding mice for discrete increments of I week.
For titrations of virus infectivity, o.o2 ml samples of dilutions prepared as above were
inoculated i.c., i.p., or s.c. into mice of defined age, using eight titter-mates per dilution and
a 0.o2 ml inoculum. The incidence of clinical signs and mortality in infected litters was
recorded daily for 2 weeks.
Certain combinations of virus and age group of mice were characterized by almost benign or entirely subclinical infections. For this reason, the end point of all virus titrations
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 14 Jun 2017 21:10:15
Adaptation studies with Ross River virus
61
Table ~. Titration i.e. in day oM miee of virus of original mosquito pools NB5o92 and NB7I I 3
Original
virus
isolate
NB5o92
NB71 t3
Infectivity as log
(units/ml)
c
~
~
IDso
CDso*
4"2
4"1
3"7
4"1
%
morbidity
%
mortality
Survival
in
days
86
IOO
24
5
7 to I3 (Io mice)
Io (2 mice)
* CDso = dose for 50 % clinical response.
was determined by the challenge of mice surviving to the I4th day. Mice first used as infants
or day olds were inoculated i.p. with lO6.o day old mouse LDso of a mouse adapted stock
(T48/MBI2), and held for a further IO days for development of pathognomonic signs. The
absence of clinical signs was taken as evidence of prior infection. With RRV, however, there
is an age-dependant resistance to the development of clinical signs so that an alternative
system was required for mice first used when I week old or older. These animals were challenged as above, bled 48 h later, and individual samples screened for viraemia in a cellculture assay. It was established that susceptible mice between 3 and 8 weeks old had peak
viraemia of Io 3.4 to 1o6.3 p.f.u./ml at 48 h, and that the absence of detectable viraemia at this
time indicated the presence of antibodies to RRV detectable by haemagglutination inhibition
(HI).
For each virus stock examined in mice, infectivities were calculated as 50 % end points
(Reed & Muench, I938) to show the total infectivity (IDs0/ml or IDs0/g), infectivity in terms
of clinical signs (CDso/ml or CDs0/g) and, where appropriate, infectivity based on mortality
(LDso/ml or LDs0/g). Morbidity and mortality rates, obtained from the same samples, indicate the % of infected animals that developed either clinical signs or died. Average survival
times (AST) represent the arithmetic mean for all mice dying during the course of a titration
regardless of dose, since for most strains of RRV the death times are independent of dose.
The mouse virulence of each RRV stock was assessed using these parameters, although it is
realised that conclusions may only be valid for this particular model system.
RESULTS
The responses of mice to unadapted NB strains of R R V
The severity of RRV infection in laboratory mice depends on the origin and treatment of
the strain used, and the age of the indicator host. Between extremes of subclinical and
rapidly lethal infections, a distinctive intermediate syndrome is seen (Mims et al. I973),
characterized by various degrees of hind limb paralysis from which the infected mice frequently recover. Murphy et al. 0973) have described the necrosis of striate muscle which
underlies this clinical sign. Two field isolates in processed mosquito pools, NB5o92/field and
NB7I I3/field, were titrated by i.e. inoculation in day old mice to give results that typify this
mild syndrome (Table I).
Such unadapted field strains produced clinical signs in the majority of infected day old
mice (86 to ioo ~) after an incubation period of 5 to 6 days, but the mortality rate (5 to
24 ~o) was too low to allow an LDs0 end point. Deaths occurred during the second week of
infection amongst the most severely paralysed mice. Recovery was frequently complete by
day 14 although in some mice residual signs persisted for up to a further IO days, accompanied at times by severe wasting of muscle.
5
viR 28
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 14 Jun 2017 21:10:15
62
W. P. TAYLOR
AND
I. D . M A R S H A L L
Table 2. Comparative titrations of unadapted and mouse brain adapted stocks of NB5o9 2 in
infant mice inoculated by different routes
Virus
NB5o92/MB 3 MOS2
i.c.
Infectivity
morbidity
% mortality
AST (days) _s.e.
5*
94
4
12.0+ I-O
NB5o92/MBI2
i.p.
s.c.
i.c.
i.p.
s.c.
5"4*
98
2t
5'4*
94
6
8"3t
1oo
1oo
8"or
Ioo
Ioo
8"2t
Ioo
Ioo
11'o_+0"9
I2.7_+o'9
5"6-+0"2
6.0___0-2
6"0-+0'2
* log (IDs0/ml).
t log (LDs0/ml).
Comparative titrations by different routes of inoculation
Alternating passage in infant mice and mosquitoes does not alter the virulence of field
strains of RRV (Taylor & Marshall, 1975), but passage in infant mice produce an increase in
virulence (vide infra). To compare the influence of route of inoculation, a mouse brain
adapted stock NB5o92/MB~/, and a stock maintained by alternate mouse and mosquito
passage, NB5o92/MB3 MOS2, were used. Parallel titrations in infant mice were made with
each stock inoculated i.c., i.p., or s.c. (Table z).
The infectivities for NB5o92/MB3 MOSz were similar for all inoculation routes. The
morbidity of over 93 ~ reflected a small number of subclinical infections but mortality rates
were never above 2i ~ . Most deaths occurred in the group inoculated i.p., although other
experiments with RRV strains of different virulence have not shown that infection by the i.p.
route is more lethal than that by the i.c. or s.c. routes. On the basis of the response to i.c.
infection virus stock NB5o92[MB3 MOS2 is similar to the field pools NB5o9z or NB7I 13
(Table I).
On the other hand, mice inoculated by any route with the mouse-adapted NB5o92]MBI2
strain invariably died within 5 to 7 days, but, again, differences in infectivity were not
evident.
When 3-week old mice were infected i.e. or i.p. with the same virus stock of strain NB5o92/
MB3 MOS2, infectivities were io 3'5 and io 3"7 IDs0/ml respectively, although all infections
were benign. It was concluded therefore, that route of inoculation did not influence significantly the outcome of RRV infections in laboratory mice.
Alterations in virulence with passage in cell cultures
Before cell culture passage, the virulence of clone 62 was typical of a group of clones
isolated from NB6o24/field (vide infra) each of which produced frank clinical signs in all
infected day old mice and sufficient mortality (20 to 60 ~ ) to allow the calculation of an
LDso end point. This level of virulence was slightly higher than that of isolates NB5o92/field
and NB7I I3/field (see Table I). In contrast, clone 64 from the same material was clearly
more virulent and invariably killed all the day old mice it infected.
These clones were passaged eight times in monolayers of Vero cells and the virulence
examined by i.c. inoculation of day old mice at passages I, 4 and 8 (Table 3).
Both clones were less virulent for mice after passage in Vero ceils. Clone 62 showed a
steadily declining mortality rate, from 5o ~ with the original stock, to 8 ~ after 8 passages
and at this latter level some infections were now subclinical. Essentially similar results were
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 14 Jun 2017 21:10:15
Adaptation studies with Ross River virus
63
Table 3. Comparative titrations of clones 62 and 64 from NB6o24[fieM during passage in Vero
cell monolayers, using day oM mice inoculated i.c.
Designation and
passage level
62 clone stock
62/Vero-I
62]Vero-4
~
Infectivity as log (unitslml)
'
~
]Ds0
CDso
LDso
62]Veto-8
8.2
8"5
7"7
7"0
64 clone stock
64]Vero-I
64/Vero-4
64/Vero-8
7"9
8"2
8'8
6"4
8.z
8'5
7"7
~
morbidity
~
mortality
AST
+ s.c.
6.2
5"4
No LDs0
No LDso
No LDs0
Ioo
IOO
lOO
96
5o
48
30
8
11.4 +
_ I"7
lO'7+-o"7
11-6+o"9
13'2 + o'8
7'9
8"2
8'8
5'8
7"9
8-2
8.8
4"9
IOO
lOO
lOO
8o
IOO
I oo
IOO
52
7'2 +_0"3
7"6 _+o-2
7'8 _+o'2
8'5 +_o'5
Table 4. Comparative i.p. titrations in mice of different ages of samples derived from RRV
strains NB5o92]MBI2 and T48]MB]2 during passage in monolayers of Veto cells
Designation and
passage history
of virus strain
Age of
mice
NB5o92/MBI 2]Vero-2
NB5o92/MBI2/Vero-8
I week
I week
T48/MBI2/Vero-I
I week
2 weeks
3 weeks
I week
2 weeks
3 weeks
T48]MBI2/Vero-8
Infectivity as log (units]ml)
~
~
~
IDso
CDso
LDso
4'7
3"9
8"4
98.2
98"2
7'5
7'6
6"4
%
morbidity
~
mortality
AST+ s.c.
(days)
4"7
3'9
4"7
3"4
Ioo
IOO
IOO
67
lO'4+o'6
14"o+o'3
8'4
8.0
7'9
7'5
6-9
No CDso
8"4
6"7
7'3
7'3
No LDso
No LDso
IOO
92
83
IOO
8o
33
IOO
46
71
94
I7
4
5-5+o'3
I3"2-+o"6
II'5+o'5
8"8+_0'4
i2.2_+o.6
9"o
obtained with two independently passaged (uncloned) stocks of virus, NB6o24/MBx and
NB5o92]MBI, neither of which differed in initial virulence from clone 62.
After 8 passages, clone 64 killed only 52 ~ of mice and 2o ~ of infections were subclinical.
Similar manipulations were made with two other stocks of virulent virus previously
adapted by passage in mouse brain, (NB5o9z/MBIz and T48/MBI2 ). Of these, T48/MBI2
was more virulent since it caused 7o ~ of deaths in 3 week old mice, whereas NB5o92/MBI2
was unable to produce clinical infections, let alone deaths, in mice of this age. Titrations in
I-week old mice (Table 4) indicated that NB5o92/MBI2 declined in virulence between Veto
passages 2 and 8, since the survival time increased (P < o-oot) and the mortality rate declined. For virus strain T48/MB 12, although mortality was little affected in I-week old mice,
the longer survival was again a feature of passage (P < o.ooI ; Table 4). In 2- and 3-week old
mice there was a significant reduction in morbidity and mortality rates. Thus virus strain
T48/MBI2]Vero-I was significantly less virulent after 7 further passages in Vero cells.
The similar passage of virus strains in monolayers of BHK cells yielded different results.
Over eight passages there was no indication of the attenuation of clone 62, NB5o92/MBI
or NB6o24/MBI. However there was a decline in the virulence of c/one 64, with the morbidity in one day old mice falling from IOO ~ to 75 ~ and mortality from IOO ~ to 53 ~o after
8 passages in BHK cells.
Thus the passage of RRV strains in cultures of some cells may select variants of reduced
virulence, although this is not a regular effect.
5-2
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 14 Jun 2017 21:10:15
64
w.P.
.~
12
-6.~ 10
.~
8
~._g
6
~
<
4
2
I
I
TAYLOR AND I. D. MARSHALL
I
I
I
I
I
÷
÷
÷
i
_
100
90
80
1
Mortality rate in
I Day-old mice /
_
~ l-week-old
miceI
7o
El 2-week-oldmic
60
~ 50
40
~] 3-week-oldmi q
Morbidity r a t ? [
~
,
20
10
1
2
3
6
/
9
10
(MB1) (MBI/MC2) (MB1/MC7)
(MB1/MC9)
(MB1/MC1) (MB1/MC5)
(MBI/MC8)
Passage level
Fig. I. The adaptation to mice of RRV strain NB6o24 by serial passage in day old mice. The first
passage (MBI) was from the brain of a day old mouse infected i.c. with a terminal dilution of the
field mosquito pool NB6oz4. Subsequent passages were with high doses inoculated s.c. of virus
prepared from a mouse carcase (MC) from the previous passage. The histogram shows increasing
morbidity and mortality rates in mice of given ages infected with virus strains at increasing passage
levels. The upper graph is of the mean survival times (+s.e.) of day old mice infected with virus
strains at given passage levels.
Alterations in virulence with passage in day oM mice
A sample of the mosquito pool NB6o24[field was titrated by i.c. inoculation in day old
mice, and by clinical signs the terminal dilution appeared to be at lO -3 . The brains of all
mice inoculated with the lO -3 and IO-4 dilutions were collected individually at day 7 and
tested for virus. Three of 7 brains inoculated with the I o -z dilution contained virus, and none
of 8 inoculated with Io -4. Virulence assays indicated that the 3 mouse brain stocks collected
from the terminal dilution were similar to each other and to the majority o f R R V strain
NB6o24 clones derived from plaques on Vero cells. As the 3 strains appeared to behave
similarly during two further passages in mice, the adaptation series was continued with only
one line.
After the initial selection by terminal dilution, all transfers were made with the highest
dose practicable in order to enhance transmission of variants of higher virulence. The first
passage was by mouse brain (MB) harvest after i.c. inoculation, but all subsequent transfers
were by s.c. inoculation to be more comparable to alternate mosquito-mouse passage
experiments reported elsewhere (Taylor & Marshall, I975) and, as it was now appreciated
that R R V is myotropic (Murphy et al. I973), virus for further passage was recovered from
mouse carcase (MC). In effect, transfers were usually at a i : io dilution of IO ~o carcase
suspension, which in early transfers represented a dose of about io ~ IDs0/mouse, and up to
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 14 Jun 2017 21:10:15
Adaptation studies with Ross River virus
65
Table 5. Titration of clones of RRV strain NB5o92[MBI by i.c. inoculation
of infant mice
Infectivityas log (units/ml)
Clone
no.
~--- . . . .
IDso
i
3
5
I5
I6
17
19
8.I
8.2
8-6
7"6
7"3
7'1
6'4
~
CDso
%
morbidity
%
mortality
81
ioo
96
IOO
90
98
I00
2
6I
5
Io
2
o
I0
6.8
8-2*
8"3
7"6
7"0
7"O
6"4
* Clone 3 LDs0 = 5"5.
lOs IDs0/mouse in later transfers. Virulence assays at each passage were made by i.c. titration
in day old mice and i.p. titration in older mice.
Mortality in day old mice increased rapidly from the initial 39 ~ and all died when
inoculated with NB6o24/MBI/MC2 and subsequent passage strains (Fig. I). AST decreased
from i I days for NB6o24/MBI until it reached 4 to 5 days for MBI/MC8. NB6o24/MBI
failed to produce morbidity or mortality in 1-week old mice, but morbidity was IOO 9/oofor
MB I/MC2, and all mice of this age died when inoculated with MB I/MC6 and higher passage
strains. The AST was again shortened by MBI/MC8, the initial death time of 12 to I4 days
decreasing to 7 to 8 days. Similarly, the proportion of 2-week old mice that recovered fell
between passage levels 6 and 9, and when inoculated with MBI/MC9 and subsequent
passage strains all mice showed clinical signs and lOO ~ mortality. In 3-week old mice
mortality fluctuated between 40 ~ and 60 ~ for all passage levels after MBI/MC8, but in 4week old mice morbidity did not reach more than about 5o ~ and mortality was less than
20 ~. Virulence for laboratory mice assessed by these observations of the morbidity,
mortality, AST and infectivity in mice of different age groups did not increase significantly
beyond passage strain NB6o24/MBI/MC9.
The evidence of rising virulence associated with passage in mice was open to interpretation in two ways. Either the starting material (NB6o24/field) contained one or more
highly virulent sub-populations, the proportion of which rose during passage, or, during the
course of serial transmission, one or more virulence mutations occurred again followed by
enrichment. Bearing in mind that the passage series described was established by a terminal
dilution we favoured the latter interpretation. Nevertheless further evidence was sought by
examining the virulence of field material in more detail, by inoculating mice with known
mixtures of virulent and avirulent viruses and finally by experiments with mouse passaged
virus.
Characterization of clones from fieM isolations of virus
A total of 3o clones were selected from a range of dilutions of RRV strain NB5o92]MBI,
and of these I5 were examined by i.c. and s.c. inoculation of lO4 p.f.u, of the clone
stock into groups of 6o to 7o infant mice. By this technique I4 clones were reasonably
homogeneous, having a mean i.c. mortality rate of 24 ~. One, clone 3, was apparently more
virulent, with an i.c. mortality of 77 ~- There was no relationship between the pathogenicity
of any clone and the route of inoculation.
To elucidate the relationship between mortality and virus dose, 7 clones were titrated by
i.c. inoculation in infant mice at dilutions down to Io -7 (Table 5). Morbidity rates were
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 14 Jun 2017 21:10:15
66
W. P.
TAYLOR
AND
I. D.
MARSHALL
Table 6. Titration of clones of RR V strain NB6oz4/field by i.c. inoculation of infant mice
Clone
no.
62
63
64
65
66
72
73
77
78
85
88
Infectivity as log (units/rnl)
~
,
~
lDs0
CDs0
7'I
7'3
7"I
7"I
6"8
8.I
8"4
6-8
7'I
~ 8.2
6'7
~
morbidity
mortality
IOO
96
IOO
IO0
IOO
IOO
93
ioo
Ioo
ioo
IOO
II
54
IOO
0
28
48
4o
30
I9
63
46
7"1
7"I
7'I*
7"1
6-8
8-I~
7'7
6-8
7't
~ 8.2
6"7
* Clone 64 LDs0 = 7"I.
~ Clone 72 LDs0 = 4"3.
always high but, with the exception of clone 3, mortality rates were low (mean 6 ~ in total
of 3o5 infected mice). In this experiment clone 3 had a mortality rate of 62 ~ and an LDs0
of Io~'S/ml. Aside from clone 3, pooled data showed that for virus dilutions from Io° to
Io -7 the number of deaths per number of mice infected at each dilution were 3/48, 6/48,
1/48, 4[48, o/48, ~/4t, x/i8 and 2/6. As the probability of a fatal outcome did not vary with
virus dose, LDs0 estimates could rarely be made with this type of virus.
In a final test, each of the 3o NB5o9z/MB I clones was inoculated i.p. into x-week old mice at
a fixed high dose. Transient clinical signs were produced by 9 clones but only clone 3 gave
macroscopic evidence of muscle necrosis at post-mortem.
A further I I clones were selected at random from a sample of NB6o24[field which had
been diluted zoo-fold and plated in its entirety. Titrations performed by i.c. inoculation of
infant mice differentiated one clone of relatively high virulence, clone 64, from the remaining
'background' group (Table 6). Confirmation of exalted virulence was obtained in x-week
mice, in which clone 64 gave the highest morbidity and was unique in its ability to kill
(I7"5 ~ mortality).
A simultaneous comparison by titration of clones derived from NB5o9z/MBr and
NB6oz4/field in more accurately aged (day old) mice allowed the following deductions:
background NB6oz4 clones were slightly more virulent than their counterparts from
NB5o92]MBI, with mean mortalities of 6o and 27 ~ respectively, and also, because of a
dose response, regularly gave LDs0 end points in day old mice. Clone 3 (NB5o92) in these
mice was indistinguishable from NB6oz4 background clones, while clone 64 remained the
most virulent virus found in unadapted material and regularly caused Ioo ~o mortality in
day old mice.
These experiments, although indicating sub-population heterogeneity, failed to identify a
virus in unadapted field material equivalent in virulence to that present in highly mouseadapted stocks.
Inoculation of mice and cell cultures with mixtures of virulent and avirulent viruses
Following the demonstration that field virus contained sub-populations of strains of
different virulence for mice, experiments were carried out to determine the interaction
of pre-determined strains when inoculated as mixtures into mice or cell cultures. A small
plaquing clone (S0 which killed all day old mice with an AST of about 5 days, and 85 ~ of
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 14 Jun 2017 21:10:15
Adaptation studies with Ross River virus
67
Table 7- Mortality rates and average survival times for day old mice inoculated i.e. with a
virulent R R V strain (SI/NB5o92[MBI2), an avirulent R R V strain (NB5o92/MB~/Vero-5),
and with mixtures of the two in given ratios
Dose* ratio of virulent
to avirulent RRV
in total inoculum
%
mortality
AST + s.e.
(days)
i o ~: o
ioo
~o 4 : o
I0 ~ :0
IO : 0
~oo
I00
IO0
4"9+0"2
5-I +o.2
5"I ~ O ' I
5-5__+0.3
lO 7 : IO 6"6
106 : I06'6
I o s : IO ~'~
IO 4 : I o n ' °
i 0 a • i06.6
I02 : IO 6'6
I o : I o ~'s
IO0
I00
IOO
IOO
I00
I00
67
4"8±0"I
5"6_+0'2
6-3_+O'2
6 " 3 - O-2
9-2 _ 0-2
8"6+0"3
Io-6+o.8
0 : I O °'~
25
I2"O±O" 7
• Day old mice
ID~o.
~-week mice in about I2 days, had been obtained from NB5o92/MBI2. This clone was
paired with the fifth Vero passage stock from NB5o9z/MBI, a virus capable of killing only
25 % of day old mice in an AST of i2 days, and completely avirulent in I-week mice.
Serial tenfold dilutions of SI/NB5o9z/MBI 2 were mixed with equal volumes of undiluted
NB5o92[MBi[Vero-5 and each pool inoculated i.e. into randomized day old mice in groups
of ~6; control mice received dilutions of S x/NB5o92/MB Iz only, or undiluted NB5o92/MBI]
Vero-5 (Table 7). With each mixture, the syndrome produced was influenced by each virus.
In contrast to the 25 % mortality associated with the avirulent strain, mice inoculated with
mixtures invariably died until the ratio of virulent to avirulent virus fell below 1:4oooo,
and even at 1:4ooooo the mortality rate was significantly elevated. Yet in each group the
presence of the avirulent virus contributec to the outcome by prolonging the survival time,
as may be seen by comparing with controls receiving similar doses of SI/NB5o9z[MBI2
only (Table 7).
Since SI/NB5o92/MBI2 is lower in virulence than the virus strain selected after Io serial
mouse passages of NB6o24 (Fig. I), but is still readily detected even in low doses in the
presence of an avirulent partner, it is most unlikely that virus of a highly virulent character
exists undetected in unadapted field strains of RRV.
When blood from mice inoculated with the mixtures containing IO6 and lO4 virulent to
Io 6n avirulent virus was plated on Vero monolayers there were more small than large
plaques (virulent S I]NB5o9z/MB I2 gives small plaques relative to avirulent NB5o92[MBI/
Vero-5). Further, the carcase of a mouse harvested 7 days after inoculation with the mixture
of ratio Io:~o 6"n virulent to avirulent was processed and the infectivity titrated in day old
mice. All infected mice died, but for most comparable dilutions of this material the AST
was greater than for SI]NB5o92]MBIz controls (Table 7), indicating that avirulent virus
was still present in the harvest. Mice inoculated with Io °'2 LDs0 of the material died with an
AST of 4"9 + o-2 days. This survival time was shorter than the 5-6 + o.2 days recorded when
SI/NB5o9z[MBIz comprised 25 % of a similar inoculum, (mixture Io°: 1o6"6; Table 8), suggesting that in the course of replication an enrichment of at least IOS'°-fold had occurred in
favour of S~[NBso92[MBI2.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 14 Jun 2017 21:10:15
68
w.P.
TAYLOR
AND
I. D.
MARSHALL
Table 8. Average survival times for day old mice inoculated i.c. with mixtures containing a
virulent R R V strain (SI/NB5o92/MBI2) and an avirulent strain (NB5o92/MBI/Vero-5)
during passage in monolayers of Vero cells
D o s e ratio o f virulent
to avirulent R R V in
original i n o c u l u m
A S T + s.e. (days) with Vero cell passage
~
Vero-2
Vero-4
Io 7 : Io 6.o
IO6 : IO 6.6
1o5 : to ~6
lO 4 : to ~~
IO3 : IO~~
Unpassaged
4"8 + o - I
5"6+0"2
6"3 _+0-2
6'3 _+o-z
9'2_+0"2
4"3 _+0"3
4-6-+0'2
5"o_+ o'o
7"5 ---0'3
9"0_+0'4
4-0+0"2
5"2-1-O"1
5"7 _+0"3
6"7 _+0"4
8"3+0'3
I C :0
4"0_+0"I
4'7_+O'3
3"7+O'2
Table 9- Comparative results in mice of different ages inoculated with clones selected
from R R V strain NB6oz4 at different passage levels in day old mice
M o u s e passage
level/clone
number
i]i
x/2
3]/
3]2
6/I
6]2
Jolt
IO[Z
Age of
mice
IDs0
~o
morbidity
~
mortality
9I
97
9
I3
A S T _+s.c.
(days)
day old
day old
5"8
8"3
day old
I week
day old
I week
7"5
6"t
7"8
7"2
] oo
40
IOO
IO0 ( P < o ' o I ) *
I oo
4
ioo
60 ( P < o ' o I )
day old
I week
2 weeks
day old
I week
2 weeks
8"3
8"7
8"6
7"7
7"2
7"2
IOO
IOO
7
Ioo
]oo
59 ( P < o ' o t )
IOO
62
o
Ioo
93 ( P < o ' o 5 )
25 ( P < o ' o 0
6.8 + 0-2
I3'6_+o'5
day old
I week
2 weeks
3 weeks
day old
I week
2 weeks
3 weeks
8"9
8.2
8"4
7"4
8"5
8"1
7"9
6-2
IOO
Ioo
IOO
97
Ioo
IO0
1oo
47 ( P < o ' o 5 )
IOO
IOO
Ioo
47
IOO
I00
27 ( P < o ' o I )
o
4.2-+ o-I
6-8-+ 0.2
8"8 + o . 2
11 "4 -- 0"5
5.9+0.2 (P<o.ool)
9"0_+0'2 ( P < o ' o o I )
12"3_+o'4
9'0
II'8+_o'6
7"3 -+ 0.2
7"I +0"3
12"3_+o'3
5"9_+o'2 ( P < o . o I )
II'I_+O" 4 ( P < o ' o o l )
I2"7+o'4
* All values for P are for differences between clones o f the s a m e passage level.
Contrasting results were obtained in cell cultures. The first 5 mixtures in the above experiment (Table 7) were inoculated on to Vero monolayers and passed a further four times.
At passage levels zero, two and four, undiluted harvests from each input mixture were
inoculated into day old mice, all of which died. The AST results are shown in Table 8. The
range of survival times obtained after four passages was not substantially different from that
obtained with the unpassaged mixtures, implying that the proportion of viruses in each
mixture had remained constant during growth in Vero cells.
Characterization of clones from R R V strain NB6o24 passaged in day old mice
A series of clones was obtained from the plaque assay end points (in Vero cells) with
NB6o24 stocks of the first, third, sixth and tenth passage levels in day old mice (Fig. I).
Characterization of virulence was made by titrations in mice of selected ages (Table 9).
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 14 Jun 2017 21:10:15
Adaptation studies with Ross River virus
69
10
7
t~
o
6
o
5
/
20
40
60
Time after inoculation (h)
80
100
Fig. 2. Growth of cloned RRN strains in muscle of day old mice. In terms of ascending virulence
these rank as follows: clone 6z (O--O), clone 64 (©--©), clone LI/NB5o92/MBI2 (A--A),
clone Io/I (11--11).
There was no distinction between clones I/t and I/2 isolated from first mouse brain
passage. In spite of a 3oo-fold difference in infectivity neither clone produced ~oo
morbidity, and mortality rates of both were low. Clones selected at the third passage, 3/I
and 3/2, were both more virulent than I/I or ~/2, as seen by the IOO ~ mortality rates in day
old mice. Further, a distinction was possible between 3/i and 3/2; for with 3/2 both morbidity and mortality in I-week mice were significantly higher.
At the sixth passage the two clones selected again differed in virulence, 6/2 being more
virulent than 6/1 in terms of AST in day old and I-week old mice, morbidity in 2-week old
mice, and mortality in both I- and 2-week old mice. In terms of virulence, clone 6/I appeared
identical with the previously isolated clone 3/2.
Clones Io/I and I0/2 could be differentiated by AST in day old and z-week old mice,
morbidity in 3-week old mice, and mortality in 2-week old mice, clone Io/I being the more
virulent. The only means of distinguishing clones I0/2 from the earlier clone, 6•2, lay in the
higher morbidity rate in 2-week old mice found for I0/2 (P < o-oi).
While the differentiation of clones 6/2 and Io/2 was probably not significant, the sum of
evidence indicated that as adaptation proceeded stocks retained a sub-population heterogeneity with respect to virulence, and, that the level of virulence of the constituent subpopulations tended to increase with passage.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 14 Jun 2017 21:10:15
70
W.P.
T A Y L O R A N D 1. D. M A R S H A L L
Replication of R R V clones in 24 h mice
Comparative estimates of virus growth were made by s.c. inoculation of ioo iDs0 of four
clones of differing virulence into 4 randomized litters of day old mice. Clones 62 and 64 were
from NB6oz4/field (Table 6). Clone IO/t was from the Ioth passage stock of NB6o24,
(Table 9), and clone LI was a large plaque-forming variant isolated from NB5o92/MB12 and
similar in virulence to the SI clone from the same parent stock (Table 7). In terms of ascending virulence these clones could be arranged as follows: clone 62, clone 64, clone LI and
clone lO/I.
At various intervals two mice from each group were bled and a pool made of the entire
hind-quarter musculature. Samples were ground to lO ~ w/v suspensions and titrated on
Vero monolayers. Results are shown in Fig. 2. With each virus a period of exponential
growth occupied the first 2o to 30 h, although the rates of growth appeared to follow
virulence ranking. The least virulent virus, clone 62, did not increase in titre beyond 30 h
and the levels may have declined during the subsequent 66 h. By 72 h all remaining clones
had reached peak levels after a period of less rapid growth. Mice infected with clones Io/I
and LI did not live beyond this time. Ultimate virus titres retained the same ranking as
growth rate and innate virulence.
DISCUSSION
In terms of lethality for adult rabbits, the virulence of strains of myxoma virus was shown
to be correlated with the average survival times of the infected rabbits (Fenner & Marshall,
1957). Attempts to apply these parameters to measure the virulence for mice of strains of
RRV derived during selection by passage led to several difficulties. Differences between
attenuated strains of RRV can be detected in terms of the survival times of infected newborn
mice, but these strains do not kill older mice; highly virulent strains, on the other hand,
produce virtually identical average survival times in newborn mice, but can often be readily
differentiated by their effects in mice I to 4 weeks old. We therefore titrated RRV strains in
groups of mice of different ages and then combined the estimates of infectivity, morbidity,
mortality, average survival times and age resistance in these groups to give a comprehensive
measure of the relative virulence of the virus strains.
The present experiments clearly identify two selection processes with contrasting effects on
RRV during serial passage; cell cultures select for lower virulence in mice, and young mice
select for higher virulence.
Gard et al. (I973), in a study of the epizootiology of RRV in the Nelson Bay area, noted an
increase in the virulence of NB5o92 during 12 mouse brain transfers. A similar alteration was
achieved when the isolate NB6o24 was passaged in the muscle of day old mice. In our
adaptation series maximum virulence appeared to be attained in about Io passages. While
passage is often terminated when a convenient level of virulence is reached, Smithburn &
Haddow (1944) observed increases in virulence when Semliki Forest virus was carried past
the 4oth passage in adult mice, and Cole & Wisseman (1969) could distinguish between
Dengue virus stocks passaged 33 and I25 times using a combination of suckling and
weanling mice. On the other hand, Bradish, Allner & Maber (i 972) using a range of animals
as test hosts, and an avirulent strain of SFV, noted a sharp and apparently final increase in
the virulence of the virus for mice after only two i.c. passages in weanling hamsters. Adaptation may be more rapid in young than old mice, as Meiklejohn, Stacey & Simpson (1949)
found that Dengue virus grew better in the brain of 4 to 6 day old mice than in adults, and
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 14 Jun 2017 21:10:15
Adaptation studies with Ross River virus
7~
subsequenty Meiklejohn, England & Lennette 0952) adapted field strains to mice of similar
age while failing with 2I to 18 day old animals.
Although not examined in depth, the gross similarity in clinical signs and gross pathology,
and the growth potential of clone ~o[I in striated muscle, indicate that muscle passaged virus
retains the same tissue tropisms previously described for unadapted virus (NB5o92/MB4
MOS3) and brain passaged virus (T48/MBI2) (Murphy et al. I973). In view of the predilection shown by RRV for skeletal muscle it was perhaps unrealistic to expect route of passage
to influence the character of the adapted virus, although Smithburn 0949) was able to
modify Rift Valley fever virus from a pantropic nature when transmitted i.p., to a variant
with neurotropic properties when passaged by the i.c. route.
Both NB6oz4/field and NB5o9//MBI may be regarded as unadapted stocks, as both
correspond with the normal vertebrate-invertebrate cycles by which arboviruses are maintained in nature. Cloning experiments with these pools indicated the presence of subpopulations of low yet distinct virulence, however it is suggested that the observed alteration
in mice did not proceed simply by the selection of a pre-existing virus. Firstly, the NB6oz4
passage series included an initial terminal dilution and, secondly, experiments with mixtures
of virulent and avirulent strains of RRV indicated the ease with which a virulent virus could
have been detected were it present in unadapted material. From these considerations it
appears that sequential increases in virulence during passage in mice depends upon mutational changes occurring during virus replication, and subsequent selection of mutants of
successively higher virulence; a process that would be favoured by the high virus inputs
employed further to the terminal dilution. Supporting evidence was obtained when clones
from successive passage levels were tested and found to be increasingly more virulent, although sub-population heterogeneity was still apparent at any one passage level. Overgrowth
by the more virulent mutants appeared to be related to growth characteristics; in infant
mouse muscle the more virulent virus strains replicate more rapidly than less virulent, perhaps allowing preferential seeding into the blood stream and secondary replication sites and
leading to higher final titres.
While direct experimental evidence was not obtained, it is possible that attenuation for
mice during serial passage of RRV in cell cultures is mediated in a similar manner. Certainly
Scherer, Ellsworth & Ventura (197 ~)were able to show that the mouse-avirulent TC- 83 strain
of VEE virus adsorbed and replicated more rapidly in BHK zI cells than the mouse-virulent
parent strain of the virus. In our own observations with virus mixtures undergoing passage in
Vero cells it was seen that the mouse-virulent partner, although not displaced, gained no
selectional advantage in this system. Selection of avirulent variants of RRV during passage
was not as consistent or as rapid in BHK21 cells as in Vero cells.
The skilled technical assistance of Mrs Sylvia Hirsch and Miss Elspeth Davidson is gratefully acknowledged.
REFERENCES
BRADISH, C. ~., ALLNrR, K. & MABER, H. B, (1971) The virulence of original and derived strains of Semliki
Forest virus for mice, guinea-pigs and rabbits. Journal of General Virology I2, I4I-I6O.
BRADISH, C. •., ALLNER, K. & MABER, H. B. 0972). Infection, interaction a n d t h e expression o f v i r u l e n c e by
defined strains of Semliki Forest virus. Journal o f General Virology x6, 359-372.
BURGE, B. W. & PrEFFERKORN, E. R. (I966). Complementation between temperature-sensitive mutants of
Sindbis virus. Virology 3o, 214-223.
BURNET, r. M. (1960). In Principles of Animal Virology, 2nd edition, chapter 17, pp. 412-4t7. New York:
AcademicPress.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 14 Jun 2017 21:10:15
72
W.P.
T A Y L O R A N D I. D . M A R S H A L L
CHURCHILL, A. E., CHUBB,R. C. & BAXENDALE,E. W. (I969). The attenuation, with loss of oncogenicity, of the
herpes type virus of Marek's disease (strain HPRS-I6) on passage in cell culture. Journal of General
Virology 4, 557-564.
COLE, G. A. & WISSEMAN,C. L. (I969). Pathogenesis of type I Dengue virus infection in suckling, weanling
and adult mice. American Journal of Epidemiology 89, 669-68o.
DOHERTY, R. L., WHITEHEAD, R. H., GORMAN, B. & O'GOWER, A. K. (I963). The isolation of a third group A
arhovirus in Australia, with preliminary observations on its relationship to epidemic polyarthritis
Australian Journal of Science ~6, 183-184.
FAZEKASDE ST. GROTH, S., GRAHAM,D. M. & JACK,I. (I958). The serology of mumps infections. A new source
of antigen and a simplified complement fixation test. Journal of Laboratory and Clinical Medicine 5x,
883-896.
FENNER,F. & CAIRNS,J. (1959). In The Viruses, vol. 3, PP. 225-247. Edited by F. M. Burner and W. N. Stanley.
London: Academic Press.
FENNER, F. & MARSHALL,1. D. (I957). A comparison of virulence for European rabbits (Oryctolagus cuniculus)
of strains of myxoma virus recovered in the field in Australia, Europe and America. The Journal o f
Hygiene 55, 149-I9t.
GARD, G. P., MARSHALL,I. D. & WOODROOFE,G. M 0973) Annually recurrent epidemic polyarthritis and Ross
River virus activity in a coastal area of New South Wales. 2. Mosquitoes, viruses and wildlife. American
Journal of Tropical Medicine and Hygiene 22, 551-56o.
HOSKINS, J. M. (1959). Host-controlled variation in animal virus. In Virus Growth and Variation. Ninth
Symposium of the Society for General Microbiology, pp. 122-I4I. Edited by A. Isaacs and B. W. Lacey.
Cambridge: Cambridge University Press.
IL'ENKO, V. I., PLATONOV, V. G., PROZOROVA,I. N. & SMORODINCEV,A. A. (1968). Material for the study o f
variation in the Malayan Langat virus (TP-2I strains). Bulletin of the Worm Health Organisation 39,
419-424.
MEIKLEJOHN, G., ENGLAND, G. & LENNETTE, E. H. (I952) Propagation of Dengue virus strains in unweaned
mice. American Journal of Tropical Medicine and Hygiene x, 51-58.
MEIKLEJOHN,G., STACY,I. B. & SIMPSON,T. W. (1949). Dengue antibody in South Pacific residents. Proceedings
of the Seventh Pacific Science Congress, Auckland and Christchurch, New Zealand 7, 348-353.
MIMS, C. A., MURPHY,F. A., TAYLOR,W. P. & MARSHALL,I. D. (1973). Pathogenesis of Ross River virus infection
in mice. I. Ependymal infection, cortical thinning and hydrocephalus. Journal of Infectious Diseases I27,
I2I-I28.
MURPHY, F. A., TAYLOR,W. P., MIMS,C. A. & MARSHALL,I. n. (1973). Pathogenesis of Ross River virus infection
in mice. 2. Muscle, heart, and brown fat lesions. Journal of Infectious Diseases x27, 129-138.
REED, L. J. & MUENCH,H. (I938). A simple method of estimating fifty per cent end points. American Journal
of Hygiene 27, 493-497SCHERER, W. F., ELLSWORTH,C. A. & VENTURA,A. K. (I97I). Studies of viral virulence. 2. Growth and absorption curves of virulent and attenuated strains of Venezuelan encephalitis virus in cultured cells. American
Journal o f Pathology 62, 2I 1-219.
SMITHBURN, K. C. (1949). Rift Valley fever; the neurotropic adaptation of the virus and the experimental use
of this modified virus as a vaccine. British Journal of Experimental Pathology 3o, I-I6.
SMITHBURN, K. C. & HADDOW, A. J. (1944). Semliki Forest virus. Isolation and pathogenic properties. Journal
of Immunology 49, 141-157.
TAYLOR, W. P. & MARSHALL,I. n. (1975). Adaptation studies with Ross River virus: retention of field level
virulence. Journal of General Virology 28, 77-83.
(Received 22 October I973)
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Wed, 14 Jun 2017 21:10:15