See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/48378467

Virus Dynamics: Mathematical Principles of

Immunology And Virology

Article January 2001

Source: OAI

CITATIONS READS

643 544

2 authors:

Martin A Nowak Robert M May

Harvard University University of Oxford
539 PUBLICATIONS 61,461 CITATIONS 508 PUBLICATIONS 68,028 CITATIONS

SEE PROFILE SEE PROFILE

Available from: Robert M May

Retrieved on: 17 October 2016
 Collaborators
Virus dynamics
] Robert May (Oxford)
] Sebastian Bonhoeffer (Zurich)
] Dominik Wodarz (Seattle)
] Marc Lipsitch (Harvard)
] Alun Lloyd (Princeton)
] George Shaw (Birmingham, Alabama)
Martin Nowak
] Andrew McMichael (Oxford)
Institute for Advanced Study ] Charles Bangham (London)
Princeton ] Jeff Lifson (Washington)

HIV is a retrovirus

Virus RNA
How fast does HIV
reproduce in vivo ?
Reverse
transcription
mRNA

Provirus DNA

1994: Treatment leads to a rapid

protease decline in virus load
inhibitors
and
quantitative
PCR Anti-viral treatment

Virus load

HIV
T1/2=1-3 days

Time
George Shaw

1
 Virus dynamics with
Virus dynamics treatment

k k


+
d u a d u a

Uninfected Virus Infected Uninfected Virus Infected

target cell target cell target cell target cell

The basic model

of virus dynamics Anti-viral treatment

Uninfected cells x& = dx xv Uninfected cells x& = dx xv

Infected cells y& = xv ay Infected cells y& = xv ay
Free virus v& = ky uv Free virus v& = ky uv

Micro-epidemiology
within infected host

Virus declines as Latently infected cells

Anti-viral treatment ue at aeut Anti-viral treatment

v(t ) = v*
Free virus half-life ua Virus load
T1/2=1-3 days
Virus load Infected cell T1/2=10 days
half-life: 1-3 days

Time
Time

2
 An extended model
of virus dynamics HIV-1 half-lives
] Productively infected cells : 1-3 days
Uninfected cells x& = dx xv
Productively
] Latently infected cells : 10 days
infected cells y& 1 = q 1 xv a 1 y 1 + y 2 ] Defective provirus : 100 days
Latently
infected cell y& 2 = q 2 xv a 2 y 2 y 2 ] Free virus : hours
Cells with
defective provirus y& 3 = q 3 xv a 3 y 3
Free virus
v& = ky 1 uv

HIV-1 half-lives HIV-1 half-lives

] Productively infected cells : 1-3 days ] Productively infected cells : 1-3 days
] Latently infected cells : 10 days ] Latently infected cells : 10-100 days
] Defective provirus : 100 days ] Defective provirus : 100 days
] Free virus : hours ] Free virus : hours

HIV eradication requires 1-3 years of HIV eradication requires >10 years of
effective therapy. effective therapy and is most likely
impossible.

What kills productively Comparing HIV and HBV

infected cells? dynamics:

] viral cytopathicity ] Half-life of productively infected cells:

] CTL responses ] HIV: 1-3 days
] HBV: 10-100 days

Note that all patients have very similar decay slopes

corresponding to half-lifes of 1-3 days.

3
Viral cytopathicity leads to a constant
half-life despite different CTL activity

What is the mechanism

of HIV disease progression?
Infection Cell becomes Cell produces
a target for CTL new virions and dies

The experiment is biased toward those cells that

produce plasma virus. CTL can greatly reduce
virus production without affecting the half-life.

Evolution of virulence HIV-1: clinical profile

] The closest relatives of HIV-1 and HIV-2 1000 <2 to >15 years
are SIVs.
CD4
] All SIVs appear to be apathogenic in their 0
natural hosts.
] SIV can be transferred to other species, Virus
where it induces AIDS.

Primary phase Asymptomatic phase AIDS

Time

A mechanism of disease
HIV-1: clinical profile progression

1000 <2 to >15 years ] .. has to explain why the steady state of
CD4 virus dynamics (with a timescale of days)
0
shifts over many years.
Why is there such a long and
Virus variable asymptomatic phase? ] 2 possibilities:
\ the immune system changes
\ the virus changes
Primary phase Asymptomatic phase AIDS

Time

4
 HIV is a quasispecies Evolution toward disease
] Viral replication is error prone.
] Escape from immune responses
] HIV reverse transcriptase and RNA 4 ] Faster replicating, more aggressive strains
polymerase have error rates of about 10
] Broader cell tropism
] The virus population in any one patient is
extremely heterogeneous.
Virus
] HIV can escape from immune load
Diversity threshold
responses.

Time

Antigenic variation Antigenic variation

virus mutant i v&i = rvi pxi vi Total virus load is proportional to antigenic diversity.
i = 1,..., n
immune response
x&i = cvi bxi br
v := i vi = n
against mutant i

Each mutant goes to equilibrium: cp

br r
vi = xi =
cp p

Add new mutants over time.

Antigenic variation Antigenic variation of HIV

virus mutant i
v&i = vi (r pxi qz ) virus mutant i
v&i = vi (r pxi qz )
specific specific
immune response x&i = cvi bxi i = 1,..., n immune response x&i = cvi bxi uvxi i = 1,..., n
cross reactive
immune response
z& = kv bz cross reactive
immune response
z& = kv bz uvz
Virus load: brn Virus load: brn
v= v=
cp + kqn cp (ru kq)n
n

5
 The diversity threshold model
Antigenic variation of HIV has 3 possible outcomes

Virus load: brn 1. Disease after long asymptomatic period.

v= kq < ru < kq + cp
cp (ru kq)n
Diversity threshold: 2. Indefinite virus control.
ru < kq
cp
nc =
ru kq 3. Immediate disease.
kq + cp < ru

Immune responses to Immune responses to

multiple epitopes multiple epitopes

Immunodominance

Antigenic variation in
Multiple epitope theory presence of multiple epitopes

v&ij = vij (rij pi xi q j y j )

x&i = ci vi* + xi (ci vi* b)
y& j = k j v* j + y j (k j v* j b)

6
Antigenic variation in Antigenic variation in
presence of multiple epitopes presence of multiple epitopes

a new mutant arises

Diversification in the
immunodominant epitope

Antigenic variation in Antigenic variation in

presence of multiple epitopes presence of multiple epitopes

Partial shift in immunodominance, Partial shift in immunodominance,

no response to new variant response to new variant
loss of response to old variant

Antigenic variation in Antigenic variation in

presence of multiple epitopes presence of multiple epitopes

Complete shift in immunodominance

loss of old variant
Immune response against variable epitope
selects for viral diversity.

7
Antigenic variation in Immune responses to
presence of multiple epitopes multiple epitopes

Immunodominance
Immune response against conserved epitope breadth of the response is related to immune memory
selects against viral diversity.
Dominik Wodarz

HIV disease progression The virus will return if

according to this model therapy is withdrawn

] There is a highly dynamic balance Anti-viral treatment

between the virus and the immune
system with rapid virus turnover. Virus load

] The evolutionary adaptation of the virus in Detection

individual patients is the mechanism of limit
disease progression.
Time

A new theory of CTL memory

Is it possible to treat and

help the patients immune ] Long lived CTL responses can eliminate
system to gain control of virus infections or reduce virus load to low
the virus? levels.

8
 Cytotoxic T lymphocytes
CTL HIV

] HIV kills CD4 cells which are needed for CTL

Helper cells memory.
CD4
] Failure to establish a CTL memory response
leads to persistent infection, high virus load and
help
rapid disease progression
Killer cells
] A good CTL memory response leads to virus
CTL memory
CD8 CTL elimination (rare ?) or low virus load and slow
is characterized by highly kill
disease progression
responsive, long-lived cells

Infected cell

HIV replication and

HIV: rate of disease progression establishment of memory

Fast progressors: high virus load No CTL Memory

CTL Memory No CTL Memory

CTL memory makes the difference. CTL Memory

Slow progressors: low virus load

Initial rate of viral spread

HIV replication and Treatment during

establishment of memory primary infection

No CTL Memory No treatment Treatment

CTL Memory No CTL Memory

CTL Memory

Vaccination or early treatment CTLp Virus

Initial rate of viral spread SIV: Jeff Lifson HIV: Bruce Walker

9
SIV infection, no treatment 4 weeks of treatment ; re-challenge

Virus
CD4 response

CD4 response

Virus

Time (weeks) Time (weeks)

SIV primary infection Treatment during

without treatment chronic HIV infection

Virus Virus load in the first

week of infection Treatment Treatment with drug holiday(s)
is correlated with
set-point
set-point is correlated with
survival.

Time
CTLp Virus
Jeff Lifson: 12 monkeys, 12 authors

Anti-viral treatment and

immunotherapy HIV therapy

Immunotherapy ] For primary infection: Use vaccination

and early treatment to reduce the initial
viral growth rate and bring patients into a
state of long term non-progression.
] For chronic infection: Use treatment
and immunotherapy to switch patients
into a state of long term non-progression.
CTLp Virus

10
Summary
] HIV dynamics
] Disease progression
] CTL memory / virus control
Three possible mechanisms
of HIV disease progression
] Evolution of the virus
] Slow break-down of the immune system
] Accumulation of opportunistic infections
Collaborators
] Robert May (Oxford)
] Sebastian Bonhoeffer (Zurich)
] Dominik Wodarz (Seattle)
] Marc Lipsitch (Harvard)
] Alun Lloyd (Princeton)
] George Shaw (Birmingham, Alabama)
] Andrew McMichael (Oxford)
] Charles Bangham (London)
] Jeff Lifson (Washington)
11