Figure 1.
Partially ordered set (poset) and induced genotype lattice for mutations associated with resistance to ZDV.
A. Poset of resistance development to ZDV. Vertices represent mutations and directed edges represent the order constraints of mutation accumulation. We observed the clustering of thymidine analog mutations (TAMs) along the two classical TAM-1 and TAM-2 pathways that is well-known under ZDV therapy [37], [38]. The left arm of the poset (mutations 41L, 215Y and 210W) is the TAM-1 pathway, while the right arm (mutations 67N, 70R and 219Q) is the TAM-2 pathway. B. Genotype lattice of mutants induced by the poset of mutations in A. The vertices represent the genotypes that are compatible with the poset in A. Predicted levels of phenotypic resistance are color-coded (green, fully susceptible; red, highly resistant). Please see Supplementary Table S2 in Supporting Information for the waiting times of mutations.
Figure 2.
Fitness costs, resistance factors, and selective advantages of mutants arising under ZDV therapy.
A. Estimated fitness costs (normalized by setting fitness cost of wild type to 0), B. Resistance factors (normalized by setting resistance factor of wild type to 1), on a logarithmic scale, and C. Estimated selective advantages (normalized by setting selective advantage of wild type to 1) of ZDV mutants. In A, B and C, the x-axis depicts the number of mutations. The TAM-1 mutants (joined by blue solid lines) are to the left and TAM-2 mutants (joined by red solid lines) are to the right of the wild type. The mixed mutants (joined by black solid lines) are to the left of the TAM-1 mutants. The TAM-1, TAM-2 and mixed mutants are separated by vertical dashed lines.
Table 1.
Estimated fitness costs for ZDV mutants.
Figure 3.
Abundance of the 70R mutation and mutant genotypes with 70R under ZDV therapy.
A. Absolute abundance (in numbers) of the 70R mutation. B. Relative abundance of the 70R mutation in the viral population. The transient appearance and eventual fixation of the mutation 70R can be seen. C. Absolute abundance (in numbers) of mutant genotypes containing the mutation 70R. The absolute abundance of a certain mutation is calculated by adding all mutant genotypes containing the mutation.
Figure 4.
Partially ordered set and induced genotype lattice for mutations associated with resistance to IDV.
A. Poset of the continuous time conjunctive Bayesian network for resistance development to IDV. B. The genotype lattice of mutants induced by the poset in A. The vertices represent the genotypes that are compatible with the poset in A. The predicted levels of phenotypic resistance are color-coded (green = fully susceptible, red = highly resistant). Please see Supplementary Table S2 in Supporting Information for the waiting times of mutations.
Figure 5.
Fitness costs, resistance factors and selective advantages of mutants arising under IDV therapy.
A. Estimated fitness costs (normalized by setting fitness cost of wild type to 0), B. Resistance factors, on a logarithmic scale (normalized by setting resistance factor of wild type to 1), and C. Estimated selective advantages (normalized by setting selective advantage of wild type to 1) of IDV mutants. In A, B and C, the x-axis depicts the number of mutations. Black crosses represent the values for the different mutant genotypes, while the blue solid line represents the average of fitness costs, resistance factors and selective advantages across all mutant genotypes with a given number of mutations.
Table 2.
Estimated fitness costs for IDV mutants.
Figure 6.
Treatment outcome with ZDV+IDV dual therapy.
A. Genotypic reasons of treatment failure were assessed in terms of mutations present at point of virological failure. For different combinations of drug efficacies and
, the different genotypic reasons of failure are shown in different colours. The treatment outcome could be a) failure with mutations resistant to both ZDV and IDV, (b) failure with mutations resistant only to ZDV, (c) failure with mutations resistant only to IDV, (d) failure with wild type, and (e) no detection of failure. B. Viral load (in copies RNA/ml) under ZDV+IDV therapy with
= 0.75 and
= 0.90. The blue line shows the total viral load, while the red dashed line depicts the wild type. The horizontal black dashed line represents the detection threshold used (500 copies/ml).
Figure 7.
Two stage mechanistic model of in vivo HIV-1 infection dynamics [6].
Target cells TU (T-cells) and MU (macrophages) can be infected by infective viruses (with effective infection rate constants
and
), resulting in early stage infected cells
and
, respectively. Infection can also be unsuccessful after fusion of the virus, rendering the cell uninfected and thereby eliminating the virus (
).
and
can also possibly return to uninfected states by destruction of essential viral proteins or DNA prior to integration (
).
cells can enter into a latent state
(with probability
) that can get re-activated with a rate constant
. Integration of viral DNA in the host genome proceeds with reaction rate constant
in the T-cells and
in the macrophages, resulting in late stage infected T-cells
and macrophages
, respectively. The infected
cells release new viruses (
) and non-infective (
) viruses (with rate constants
and
, respectively) while the infected
cells release new infective and non-infective viruses (with rate constants
and
, respectively). Target cells TU and MU are produced by the immune system at constant rate with rate constants
and
, respectively.
,
,
,
,
and
can be cleared by the immune system with reaction rate constants
,
,
,
,
and
, respectively. Viruses are cleared by the immune system with a rate constant
. Mutations are modelled to occur at the stage of integration of the viral DNA. The incorporation of the various drug classes is indicated by the inhibition of corresponding processes: EI/FI - entry/fusion inhibitors, NRTI/NNRTI - nucleoside/non-nucleoside reverse transcriptase inhibitors, InI - integrase inhibitors, PI/MI - protease/maturation inhibitors.