Supplementary info for Griffiths et al.
1) Immunological methods
a) Simultaneous infection experiments
We extracted plasma by centrifugation from 5l tail-blood samples diluted 1:1 in
heparin. We measured total IgG2a concentration by sandwich ELISA using Nunc
maxisorp immunoplates (Thermo Fisher Scientific) coated with 50l/well of
capture antibody (BD Pharmingen 553446) at 2g/ml in 0.06M carbonate buffer
and incubated overnight at 4 °C. We blocked non-specific binding with 100l/well
5% milk powder (Marvel) in Tris Buffered Saline with 1% Tween (TBST) for 2h at
37 °C. At this stage and after each subsequent step, we washed plates 5 times
with TBST. We added plasma samples in a series of doubling dilutions from
1/100 -1/ 800 in 50l/well TBST diluent. We used purified IgG2a antibodies (BD
Pharmingen 553459) to create a standard curve (highest concentration
200g/ml). Following incubation for 2h at 37 °C, we added biotinylated detection
antibody (BD Pharmingen 553388) diluted in TBST with 0.5% Foetal Calf Serum
(FCS; Sigma) at 2g/ml. Plates were incubated for 1h at 37 °C before we added
100l/well of a 1/8000 dilution of Extravadin Peroxidase: TBST-5%FCS. Plates
were incubated for 30mins at 37 °C. Following the final TBST washes, we
washed plates in distilled H2O and added TMB substrate (100l/well). After the
colorimetric change developed at room temperature in the dark, we stopped it
with 100l/well of 1M HCl. The optical density (OD) was measured at 450nm on
an Emax Precision Microplate Reader. Comparison of sample OD to the
standard curve generated IgG2a measurements in g/ml.
b) Sequential infection experiments
ELISAs were performed as described in Fairlie-Clarke et al. (2010), with the
target antigen Pcc lysate used at 2ug/ml (see below). Antibody titres were
calculated as the reciprocal of the greatest dilution at which optical density (O.D)
was greater than the mean plus 3 standard deviations of the O.D values
observed for control mouse sera at 1/200 dilution.
Plasmodium chabaudi lysate (Pc lysate) is a crude preparation of malariainfected RBCs taken with a heparinised syringe from animals with 20% or higher
parasitaemia and then purified to remove host proteins. Whole blood was diluted
in 10ml RPMI media and underlain by 15ml of 67% Percoll (Sigma-Aldrich). Cells
were separated in the Percoll gradient by centrifugation at 6450g for 10 min and
the interface containing parasitised RBC was harvested. Cells were washed
twice with 50ml RPMI and antibodies were cleaved from cells by incubation with
trypsin at 37 °C for 10 mins. Cells were washed twice in RPMI and resuspended
in 10ml of 0.05% saponin (Sigma-Aldrich) to lyse RBC membranes. Parasite
cells were then centrifuged at 6450g for 10 min and washed with PBS until the
supernatant was clear. The pellet was resuspended in 1ml sterile PBS and the
parasites lysed by 5 cycles of freeze (-80 °C) thaw.
2) Analyses of peaks and troughs in RBC dynamics
The average RBC densities in Pcc-infected and Pcc-Nb coinfected groups of
mice differ across the 21 days of infection, with Pcc-infected mice having more
severe anemia ten days post infection (Hoeve et al. 2009). Two indicators of the
severity of malaria infection are peak parasite density and peak anemia. We
hypothesise that these indicators will be delayed in coinfected mice. We looked
at the relationships among the level and timing of peak anemia (minimum count
of total RBCs), and the level and timing of peak parasite density (maximum % of
malaria-infected RBCs). We analysed these using mixed models where the
response variable was the day of the peak, predictor variables were the level of
the peak (covariate) and the treatment group (two level factor: Pcc or Pcc-Nb),
and the random effect was experiment.
Of all the Pcc-infected mice, data from 149 individuals culled on or after 20 days
p.i were used to quantify the maximum density of infected RBCs (peak parasite
density) and minimum density of total RBCs (peak anemia) until 20 days p.i. (78
Pcc, 71 Pcc-Nb).
The day on which infected RBCs peaked was earlier the more infected cells
there were (Fig. S1A), with the peak occurring one day sooner for every
additional 0.73 x 109 mL-1 infected cells (F1,141=7.2, p<0.01). This hastening of
peak density of infected cells started from a lower intercept for Pcc-only infected
mice; the difference was equivalent to peak parasite density occurring 0.77 days
sooner in Pcc-only than Pcc-Nb mice (triangles and dashed line below circles
and solid line in Fig. S1A; F1,141=11.5, p<0.001). The slopes for the two
treatment groups were not significantly different (likelihood ratio test F1,141=0.5,
p>0.45). Therefore, coinfection with Nb did not alter the slope of Pcc peak-bytime linear association, but did delay peak Pcc parasite density.
The higher the total RBC density, the later the peak anemia (Fig. S1B); each
additional 1.11 x 109 mL-1 RBCs delayed anemia by one day (F1,146=41.8,
p<0.001). Pcc and Pcc-Nb groups had different intercepts, with peak anemia
occurring 0.6 days later in coinfection mice (F1,146=4.8, p<0.05). There was no
significant difference between the slope of the Pcc and Pcc-Nb groups in their
total RBCs (likelihood ratio test F1,1=0.02, p>0.88). Therefore, coinfection with Nb
did not alter the slope of the RBC-by-time linear association, but did delay peak
total RBC density (Fig. S1).
As well as calculating the single minimum total RBC density across the 21 days
p.i., we separately calculated the minimum total RBC density in the first week
versus the two subsequent weeks. The two minima were positively associated
(Fig. S2); mice that were more anemic in the first week were also more anemic in
the later two weeks. For every additional billion total RBCs per mL at the
minimum in the first week there were an additional 0.27 x 109 mL-1 in the
minimum in the second and third weeks (p<0.01). Minima of total RBCs in both
weeks were lower by 0.44 x 109 mL-1 in Pcc infected mice than Pcc-Nb
coinfected mice (Fig. S2, p<0.001), but the slopes were not significantly different
from one another (p>0.05). Coinfection with Nb therefore reduced total RBC
density in both the first and subsequent weeks p.i., but the slope of linear
association between early and late anemia was not altered by Nb coinfection.
B
20
Day of minimum total RBC density
Day of maximum infected RBC density
A
14
12
10
8
18
16
14
12
10
8
0.5
1.0
1.5
2.0
2
Maximum infected RBC density (109mL! 1)
3
4
5
6
Minimum total RBC density (109mL! 1)
5.0
4.5
4.0
3.5
3.0
2.5
2.0
Minimum total RBC density weeks 2-3 (109mL! 1)
Fig. S1 Density and day p.i. of maximum infected (A) and total (B) RBCs in mice
with just Pcc (triangles) and Pcc-Nb coinfection (circles).
6.5
7.0
7.5
8.0
8.5
9.0
9
Minimum total RBC density week 1 (10 mL
9.5
!1
)
Fig. S2 Level and timing of anemia in the first week and following two weeks (C)
for mice with Pcc-Nb coinfection (circles, solid line), or Pcc-only infection
(triangles, dashed line).
Mice with more reticulocytes had fewer infected RBCs (Fig. S3). In Pcc-Nb
infected mice each percentage increase in reticulocytes was associated with 15
million fewer infected RBCs per mL of blood (β=-0.015±0.001SE, F1,684=401).
The slope of this negative association was modified by treatment (significant
treatment:reticulocyte interaction, F1,684=8.85) so that Pcc-only mice had 5 million
fewer infected RBCs for every percent increase in reticulocytes (β=0.0052±0.001). There was a slight difference in the intercept such that Pcc mice
had 130 million more infected RBCs per mL than Pcc-Nb mice at 0%
reticulocytes (β=-0.13±0.06, F1,684=0.45).
While this result suggests that anemia does suppress Pcc parasites, we cannot
conclude from this that bottom-up regulation is stronger in Pcc-only mice than
Pcc-Nb mice as we have not taken into account the previous day’s availability of
infected and uninfected red blood cells (see analyses in the main text).
Billion infected cells per mL blood
1.5
Pcc-Nb
Pcc
1.0
0.5
0.0
0
10
20
30
40
50
60
% reticulocytes
Fig. S3 Density of Pcc-infected RBCs against percentages of young RBCs
(reticulocytes) by infection treatment. For ease of visualisation we present a
single datapoint per mouse corresponding to the minimum Effective Propagation
Number in the second week p.i. (Pcc-only n=66, Pcc-Nb n=45).