1. Evoluzione genetica di HIV ed evoluzione clinica della malattia AIDS: due aspetti correlati?
Carlo Federico Perno

2. Why does a Virus evolve? A virus needs to evolve to:
Infect different cell types
Rapidly become resistant to otherwise highly effective antiviral drugs
Evade the immune system

3. Virus transmission Obstacles for virus transmission:
Natural
Host genetics
Host Immune system
Viral replication rate
Etc
Artificial
Vaccines
Passive immunity
Antiviral drugs
A continuous evolution allows viruses to achieve new characteristics able to overcome these obstacles, and be successful in their replication effort 3

4. …SURVIVE !!! Why does a Virus evolve? A virus needs to evolve to:
Infect different cell types
Rapidly become resistant to otherwise highly effective antiviral drugs
Evade the immune system
…SURVIVE !!!

5. Consequences
- The most fit virus, with the highest chances to survive, does not kill the host or, at minimum, kills the host in a long run
- To be selected and expanded, it kills the host at a rate lower than other viruses of the same species
- Ex: HIV subtype A vs Subtype D

6. Does HIV-1 genetic diversity have an effect on clinical progression?
Subtype C vs. subtype A, P at log-rank = 0.2
Subtype D vs. subtype A, P at log-rank = 0.05
1: J Infect Dis Apr 15;195(8): Epub 2007 Mar 2. Links
HIV-1 subtype D infection is associated with faster disease progression than subtype A in spite of similar plasma HIV-1 loads.
Baeten JM, Chohan B, Lavreys L, Chohan V, McClelland RS, Certain L, Mandaliya K, Jaoko W, Overbaugh J.
Department of Medicine, University of Washington, Seattle, WA, USA.
We investigated the effect of human immunodeficiency virus type 1 (HIV-1) subtype on disease progression among 145 Kenyan women followed from the time of HIV-1 acquisition. Compared with those infected with subtype A, women infected with subtype D had higher mortality (hazard ratio, 2.3 [95% confidence interval, ]) and a faster rate of CD4 cell count decline (P=.003). The mortality risk persisted after adjustment for plasma HIV-1 load. There were no differences in plasma viral load by HIV-1 subtype during follow-up. HIV-1 subtype D infection is associated with a >2-fold higher risk of death than subtype A infection, in spite of similar plasma HIV-1 loads.
Analysision 145 HIV-1 infected Kenyan women followed from the time of HIV-1 acquisition.
“HIV-1 Subtype D is associated with faster disease progression than Subtype A in spite of similar plasma HIV-1 loads”
Baetan JM , JID 2007 6

7. How does a virus evolve? Evolution = genetic variation
The steps in virus evolution are:
generation of diversity through mutation, recombination, and genome segment reassortment in multipartite genomes
competition among the generated variants
selection of those mutants showing the largest phenotypic advantage in a given environment

8. FUNDAMENTALS OF MOLECULAR EVOLUTION

9. Evolution is the unifying theory of biology
“All the organic beings which have ever lived on this earth have descended from some one primordial form”
Charles Darwin
From this idea, each characteristic of a species could be the result of a peculiar evolutionary history:
Peacock’s ancestors
The number and the sequences of his genes
The catalytic ability of his enzymes
His needs
The structure of his cells
His environmental fitness
... This is his evolutionary history
Evolution is the unifying theory of biology
“Nothing in biology makes sense except in the light of the evolution”
Theodosius Dobzhansky

10. In biology a mutation is a randomly derived change to the nucleotide sequence of the genetic material of an organism.
Non lethal mutations accumulate within the gene pool and increase the amount of genetic variation. The abundance of some genetic changes within the gene pool can be reduced by natural selection, while other "more favorable" mutations may accumulate and result in adaptive evolutionary changes.

13. Does occurrence of mutations mean their necessary selection and appearance (fixation) in circulating virus strains? NO

14. Different amino acid, protein mutated
Same amino acid, same protein
Silent Mutation
Synonymous
Point Mutations
Substitution of a Nucleotide
Early stop of protein
Mutation
Non Sense
A C G T 4 nucleotides
X X X 1 codon = 1a.a.
43 = 64 codons -> 20 a.a.
Mutation
NON synonymous
Different amino acid, protein mutated 14

15. What is the relation between the HIV tropism and its evolution15
Objective
To examine tropism after ART failure using a clinically validated genotypic tropism assay in a large sample of treatment- experienced patients
HIV-1 tropism was assessed at baseline and virologic failure over 48 weeks in patients receiving an optimized background regimen (OBT) with a maraviroc placebo (PBO) in the MOTIVATE 1 and 2 studies1
1Gulick et al. N Engl J Med. 2008;359:
Svicher et al. 2009

16. Rate of successful V3 sequencing16
Rate of successful V3 sequencing
V3 sequencing was successful for 87 (80.5%) out of 108 samples
83 out of 87 have also the trofile result available. For the remaining 4,
Trofile failed to assess viral tropism.
All of them resulted R5 with Trofile at screening
Svicher et al. 2009

17. Tropism at treatment failure
~90% of patients with R5 tropism at screening had an R5 result at treatment failure. 17
Screening Tropism
Tropism at treatment failure
False positive rate 5% R5 X4
78 (89.7)*
3 (3.4)
1 (1.2)
5 (5.7)
Viral tropism has been predicted by Geno2Pheno algorithm at a false positive rate of 5% using 87 V3 sequences
*P< 0.001

18. Analysis of 65 patients where R5-usage is maintained 18
Analysis of 65 patients where R5-usage is maintained
both at screening and at failure
Svicher et al. 2009

19. Despite tropism stability (R5 both at screening and at failure), the majority (23/35) of V3 positions shows higher entropy at failure than at screening suggesting viral evolution despite tropism stability 19
Change in entropy from screening to failure
V3 positions
The analysis was performed in the sub set of 65 patients with R5-tropism at baseline and failure (using geno2pheno at FPR of10%).
The Shannon entropy was calculated for each V3 position following the formula:
H(i) = - Σ P(si) log P(si) (where s=A,S,L,… for the 20 amino acids Ala, Ser, Leu, . . .).
The difference between entropy at screening and at failure at each V3 position is reported in the graph.
Svicher et al. 2009

20. Increased genetic diversity between screening and failure 20
Increased genetic diversity between screening and failure
significantly correlates with higher duration of treatment
Spearman Correlation
between Genetic Diversity
and Therapy Duration
Rho
P Value
0.2001
0.003
Rho is the Spearman's rank correlation coefficient. Rho, ranging from to 1.00, is a measure of the strength and direction of the association between two variables. A positive coefficient indicates that the variables X and Y increase in a correlated manner.
The genetic distance (mean number of substitutions per site) of V3 sequences from screening to failure for each patient and treatment duration were used to calculate Rho. The analysis was performed in the sub-set of 65 patients with R5-tropism at baseline and failure (using geno2pheno algorithm at FPR of 10%).
Svicher et al. 2009

21. Despite the high natural genetic variability of V3, the frequency of tropism switches remains limited 21
- An accumulation of synonymous substitutions was observed from screening to failure in all 87 patients (100%)
- An accumulation of non-synonymous (amino acidic) substitutions was observed in 26/87 patients (28.9%)
- Tropism switches were observed in only 4 patients (3 from R5 to X4, 1 from X4 to R5) [4/87, 4.6%]
The analysis was performed on all 87 patients on study (using geno2pheno algorithm at FPR of 5%).
Average observation from baseline to virological failure was 150 days

22. Switches from R5 to X4 usage
is mainly driven by a shift of viral species
R5 at screening
X4 at failure
R5 at screening
X4 at failure
R5 at screening
X4 at failure
The analysis was performed on the 3 samples resulting R5 at screening and X4 at failure using geno2pheno algorithm at both 5% and 10%. Genetic distance is the mean number of substitutions per site.
The high genetic distance values and the high number
of amino acid substitutions from screening to failure support
the shift from an R5-using strain to an X4-using strain
Svicher et al. 2009

23. CONSEQUENCES
C. If a mutation with lower fitness remains fixed, we obtain a minority species (called quasispecies), that may become predominant if the environment changes Ex 1. Antiviral pressure that selects for a viral strain with lower sensitivity to drugs 2. Immunological pressure by a vaccine that selects for an escape mutant not neutralized by the immune system In the case of viruses, this switch in predominance may take days (not millennia!!) - Selection of strains resistant to antiviral drugs 23

24. Re-emergence of the most fit R5-virus
Tropism
at Failure:
D/M
Re-Emergence
of R5!!
Baseline Tropism:
Designated R5
Maraviroc
in Suboptimal
Regimen
Stop
Maraviroc
X4 HIV
Not Detected at <4% R5 X4
D/M
Non-functional clone
Lewis M, et al. 16th IHIVDRW, Abstract 56. 24

25. Clinical Case: Id 186 - Patient infected with HIV-1 B subtype
Age:
Sex: M
Risk Factor:
Not known
CDC stage: C3
GRT during therapy interruption
GRT under antiretroviral treatment
GRT March ’02 (ARV: 3TC d4T ABC LPV/r)
PR: L10I M36V M46L I54V L63P A71T V82A N88D L90M I93L
RT: M41L E44E/D D67N L74L/V V118V/I M184V G190G/E/Q/R L210W T215Y K219K/N G333E
GRT September ‘02
PR: L63P V77I I93L
RT: G333E
GRT March ’05 (ARV: 3Tc d4T LPV/r)
PR: L10I K20R L33F M36M/I/V M46I I54V L63P A71T G73G/A V82A N88D L90M I93L
RT: M41L E44E/D D67D/N V118I M184V L210W T215Y G333E
GRT May ‘06
PR: L63P V77I I93L
RT: G333E
GRT January ’08 (ARV: AZT 3TC ABC DRV/r)
PR: L10I K20R V32I L33F M36I K43T M46I I47V I54V L63P A71T G73A/T I84V N88D L90M I93L
RT: M41L E44D D67N V118I M184V L210W T215Y G333E

26. Virus under drug pressure selects, among thousands of quasispecies present in the body, the virus strain with the greatest fitness in that environment
Wild type strain (the most fit) without drugs
Highly mutated (resistant) strain in the presence of drugs
No chances of winning the battle until viral replication is sharply decreased/nullified

27. CONSEQUENCES
The replication is a necessary prerequisite for occurrence and appearance of mutations
Without replication, no mutation
By decreasing the replication rate of a virus, we dramatically decrease its ability to escape immune system and antiviral drugs

28. BUT……
……If a mutation produces a variant with low fitness, and/or this mutation is not fixed, this new variant disappears Ex. Loss of viral species 28

29. Population Dynamics Allele Frequency Time 1 fixed mutation
polymorphism maintained
Allele Frequency
lost mutation
Time 29

30. Effective sample size: the genetic bottleneck
Population dynamics of alleles
Bottleneck event
Mutation
event
Coalescence
time
Each different symbol represent a different allele.
A mutation event in the sixth generation gives rise to a new allele. The figure illustrates fixation and loss of alleles during a bottleneck event, and the concept of coalescence time (tracking back the time to the most recent common ancestor of the gray individuals).
N: population size.
Adapted from The Phylogenetic Handbook 2009, M Salemi and AM Vandamme 30

31. PRACTICAL CONSEQUENCE
By reducing the sample size of a species (bacteria, viruses, etc) we dramatically reduce the chance that the species mutates and thus escapes pressure by chemotherapy and/or immune system:
- Success of antivirals and antibiotics despite a small remaining number of microorganisms
- Success of vaccines against viruses with low mutational rate
- Insuccess of vaccines against highly mutating viruses
- Insuccess of vaccines targeted against genes with high rate of mutations

32. - The case of smallpox virus - The case of influenza virus32

33. Evolutionary abilities of Variola
Variola has a single linear double stranded DNA genome of 186 kilobase pairs.
Some studies showed the presence of a low mutation rate.
A similar situation is present in other component of orthopoxviruses genus.
Number of SNPs found in different couples of viral isolates
Variola virus
Isolated compared
Year isolation
SNPs* among genomes
ETH72_16 vs ETH72_17
1972
AFG70_vlt4 vs SYR72_119
1970, 1972 1
SYR72_119 vs PAK69_lah
1972, 1969
SYR72_119 vs IRN72_tbrz
Adapted form Li et al., *Single Nucleotide Polymorphisms
Variola is lacking of great evolutionary potential 33

34. Smallpox Vaccine
Its history is strictly connected to the birth of modern vaccinology.
Variolation
1796: Edward Jenner.
1977: last case of smallpox. 34

35. Such good result was due to…
the biological characteristics of the organism,
vaccine technology,
surveillance and laboratory identification,
effective delivery of vaccination programmes and international commitment to eradication.
Smallpox virus has no host reservoir outside humans!!. 35

36. The case of influenza virus

37. Pandemic severity index
The evolutionary power of
antigenic shift
The last known flu pandemics
Name of pandemic
Subtype involved
Pandemic severity index
Asiatic (Russian)
H2N2 NA
Spanish
H1N1 5
Asian 2
Hong Kong
H3N2
“Swine”

38. Genesis of “Swine flu” H1N1 virus
Avian flu virus
(unknown subtype)
Human H3N2 flu virus
Swine H1N1 flu virus
H1N1 “Swine flu” virus
Classical swine flu virus
H1N1
H3N2 swine virus

39. The reservoir hosts act as variability source for the new evolutionary steps of flu virus

40. The evolutionary novelty of “Swine flu” virus
… gene sequences collected from the USA for swine flu (subtype H1N1) in the year 2009 are
evolutionarily widely different form the past few years sequences…the 2009 sequences are
evolutionarily more similar to the most ancient sequence reported in the NCBI database collected
in (Sinha et al., 2009)

41. Eradication possible (and obtained indeed!!)
CONSEQUENCES
Smallpox: Low rate of polymerase errors + lack of animal reservoir (even in the presence of high replication rate) =
Eradication possible (and obtained indeed!!)
Flu: High rate of polymerase errors + presence of multiple animal reservoirs (+ high rate of recombination) =
Eradication impossible
New vaccine required every year

42. Resistance to anti-HIV drugs is the most elegant, and practically relevant, example of the consequences of viral evolution

43. What about the effect of resistance
on clinical outcomes? 43

44. HIV-1: Drug resistance development
Toxicity
No Adherence
Bioavailability
Patients’ Metabolism
Reservoir
It’s important to detect resistant quasispecies before the treatment starting or as soon as possible during treatment 44

45. Percentage of patients’ viruses who had RTI resistance mutations at t0 and of those who acquired such mutations from t0 to t1 by specific mutation/drug class
In patients kept on the same virologically failing cART regimen for a median of 6 months, there was considerable accumulation of drug resistance mutations
RTI resistance at t0
RTI resistance from t0 to t1
Cozzi-Lepri et al., AIDS 2007 45

46. Poor survival in drug-class multi-resistance
Logistic regression
Multivariate P
CD4 time-dependent
(per 50 cells increase)
0.79 ( )
0.006
Plasma HIV-RNA time-dependent
(per 1 log10 increase)
1.14 ( )
0.539
Previous AIDS
2.29 ( )
0.098
LPV after GRT
0.57 ( )
0.302
3 drug class multi-resistance (DCMR)
12.29 ( )
<0.001
0 DCMR
1 DCMR
2 DCMR
P at log-rank <0.001
*disponibile solo per 155 nel 1° gruppo, 190 nel 2° gruppo
3 DCMR
Zaccarelli, AIDS 2005 46

47. Main Findings
In multivariable analyses, patients with drug resistance mutations to ≥ 2 classes during first 2 years of HAART at significantly higher risk of AIDS progression or death
Definition of Resistance
Adjusted RH (95% CI)
P Value
≥1 NRTI mutation
1.52 ( )
.004
≥1 NNRTI mutation
1.95 ( )
.002
≥1 PI mutation (major and minor)
1.50 ( )
Drug-class resistance mutations
1.79 ( )
.0007
Cumulative drug-class resistance (major and minor PI mutations counted)
Virologic failure with no resistance
1.32 ( )
.52
Single-class resistance
1.03 ( )
.90
Double-class resistance
1.55 ( )
Triple-class resistance
1.80 ( )
.005
Cozzi-Lepri A, et al. AIDS. 2008;22:

48. Virus continues to evolve if kept under pressure of failing antiviral therapy.
This may increase cross-resistance, and then decrease chances of efficacy of subsequent drugs and regimens.
In the frame of a correct therapeutic sequencing, first failing therapies should be changed as soon as possible after definition of virological failure. 48

49. Conclusions Viruses represent the best model of evolution on the earth
They mutate in days faster than what humans have ever changed in millennia
Their evolution capacity is function of several factors
The host represents the most important extrinsic factor
Through a proper use of interdisciplinary tools (mathematics, physics, biochemistry, molecular biology, biology, pharmacy, medicine) we can reasonably predict their evolution, and define ways and consequences of the interaction with humans

50. Conclusion (II)
The understanding of viral evolution has major consequences in medicine, of key practical relevance:
Identification of targets for viral vaccines
Definition of potential outcomes of massive vaccinations
Eradication, infection containment, functional cure of infected people
Setting therapeutic strategies against viral infections
Definition of the chances of success of antivirals (resistance testing, antivirograms)
Select therapies with the greatest chances of success (Ex. multiple drugs against viruses with high mutation rate)

51. ACKNOWLEDGEMENTS University of Rome “Tor Vergata” C.F. Perno
INMI “L. Spallanzani
A. Antinori
P. Narciso
C. Gori
R. d’Arrigo
F. Forbici
M.P. Trotta
A. Ammassari
R. Bellagamba
M. Zaccarelli
G. Liuzzi
V. Tozzi
P. Sette
N. Petrosillo
F. Antonucci
E. Boumis
E. Nicastri
U. Visco
P. De Longis
G. D’Offizi
G. Ippolito
and the Resistance Study Group
University of Rome “Tor Vergata”
C.F. Perno
F. Ceccherini Silberstein
V. Svicher
M. Santoro
Bertoli
D. Armenia
S. Dimonte
L. Fabeni
R. Salpini
C. Alteri
V. Cento
F. Stazi
L. Sarmati
M. Andreoni 51

52. L. Sacco University Hospital
ACKNOWLEDGEMENTS
Modena and Ferrara
Infectious Diseases
C. Mussini
V. Borghi
W. Gennari
L. Sighinolfi
F. Ghinelli
University of Padova
G. Palu’
S. Parisi
L. Sacco University Hospital
G. Rizzardini
V. Micheli
A. Capetti
Infectious Diseases, Bergamo
F. Maggiolo
AP. Callegaro
Infectious Diseases Unit Florence
S. Lo Caputo
F. Mazzotta
University of Rome Tor Vergata
Dept. of Mathematics
Livio Triolo
Mario Santoro
University of Turin
G. Di Perri
S. Bonora
University Cergy-Pontoise
LPTM
Thierry Gobron
University of S. Raffaele
A. Lazzarin
M. Clementi
Arca
M. Zazzi
University of Catanzaro
S. Alcaro
A. Artese
Catholic University of Rome, Sacro Cuore
A. De Luca
R. Cauda
San Gallicano Hospital
G. Palamara
M. Giuliani
The I.CO.N.A. Study Group
A. d’Arminio Monforte
M. Moroni 52