Immune Dysfunction in the Virologically Controlled HIV Patie

background image

1

Causes and consequences of

immune dysfunction in the

virologically controlled HIV

patient

background image

2

Morbidity and mortality in the virologically

controlled patient

Patients who have achieved long-term viral load control on HAART are

now increasingly experiencing premature onset of diseases associated

with aging, such as

cardiovascular (CV) disease

non-AIDS cancers

liver disease

renal disease

These serious non-AIDS diseases are being observed at an excess rate

compared to the general population

Persistent HIV-induced immune dysfunction appears to be associated

with this excess

Engels EA. AIDS 2009; 23:875-885.

background image

3

Morbidity and mortality in the virologically

controlled patient

This slide set summarizes the following topics:

Changes in the patterns of morbidity and mortality in HIV-infected

patients since the introduction of HAART

Excess risk of serious disease and death vs the general population,

despite virologically effective HAART

The association between CD4+ T-cell count and chronic immune

activation with risk of serious non-AIDS disease/death

Limitations of HAART in restoring immune function

Potential mechanisms that may play a role in causing serious non-

AIDS diseases, such as CV disease, non-AIDS cancers and liver

disease

Promising directions for ongoing and future research

background image

4

Mortality and morbidity in the

modern HAART era

background image

5

Adapted from Lohse N, et al. Ann Intern Med 2007;146:87–95.

P

ro

b

a

b

il

it

y

o

f

s

u

rv

iv

a

l

Pre-HAART
(1995–1996)

Early HAART
(1997–1999)

Survival from age 25 years (N=3,990)

1

0.75

0.5

0.25

0

25

30

35

40

45

50

55

60

65

70

Age, years

Late HAART
(2000–2005)

Population
controls

Patients in the HAART era have at least a 10-year

shorter expected survival than age- and gender-

matched controls

background image

6

Life expectancy of HIV-infected patients in the

HAART era remains shorter than that of the

general population*

Cohort(s)

Country

Patient

population
Time period

Impact of HIV infection on survival

APROCO and

Aquitaine

1

France

2,435 HIV-

infected
patients
1997–2005

• Age- and gender-adjusted overall

mortality 7-fold higher in HIV+ vs general

population

DHCS

2

Denmark

3,990 HIV-

infected
patients
1995–2005

• Median survival after age 25 years for

HIV+ was only 20 years vs ~51 years for

general population

CASCADE

3

Europe,
Canada,
Australia

7,680 HIV-

infected
patients with
known dates of
seroconversion
Up to 2006

• Similar mortality to the general

population in the initial 5 years from

seroconversion, but a significant mortality

excess in HIV patients over the longer term

(5.2% in the first 10 years after

seroconversion among patients 15–24

years of age)

* HIV-negative age- and gender-matched controls

1. Lewden C, et al. JAIDS 2007; 46:72-77.

2. Lohse N, et al. Annals of Internal Medicine 2007; 46:87-95.

3. Porter K, et al. 15

th

CROI 2008; Abstract 14.

background image

7

Survival on HAART is similar to the general

population only if CD4+ counts are ≥500

cells/mm

3

for ≥6 years

Lewden C, et al. J Acquir Immune Defic Syndr 2007; 46:72–77.

* Standardized mortality ratio =
(adjusted for age and gender)

APROCO and Aquitaine cohorts

S

ta

n

d

a

rd

iz

e

d

m

o

rt

a

li

ty

r

a

ti

o

*

Years with CD4+ T-cell count ≥500

cells/mm

3

Mortality in HIV-infected patients

Mortality in general population

0

0.5

1

1.5

2

2.5

3

3.5

4

0

1

2

3

4

5

6

7

N=2,435

background image

8

An emerging model of HIV-induced immune

dysfunction

Continued

Uncontrolled viral

replication and

CD4+ depletion

Death from

AIDS/OIs

H

A

A

R

T

D

e

cr

e

a

se

d

v

ir

a

l

lo

a

d

/

in

cr

e

a

se

d

C

D

4

+

T

-c

e

ll

co

u

n

t

U

n

d

e

te

ct

a

b

le

v

ir

a

l

lo

a

d

Pre-HAART era

1-4

Current HAART era

5,6

Shift to non-AIDS

related causes of

morbidity/

mortality e.g.

Heart disease

Non-AIDS

cancers

Liver disease

V

ir

a

l

re

p

li

ca

ti

o

n

/G

A

LT

d

e

st

ru

ct

io

n

In

cr

e

a

se

d

v

ir

a

l

lo

a

d

/d

e

cr

e

a

se

d

C

D

4

+

T

-c

e

ll

c

o

u

n

t

Initial HIV

Infection

1. Selik RM, et al. AIDS 2002; 29:378–387. 2. Baker RK, et al. JAIDS 2006; 43:27–34. 3. Smit C, et al.

AIDS 2006; 20:741–749.

4. Centlivre M, et al. AIDS 2007; 21:1–11. 5. Jacobson LP, et al. Curr HIV/AIDS Rep 2004; 1:74–81. 6. Aiuti

F, et al. AIDS Rev 2006; 8:88–97.

background image

9

HOPS cohort: Deaths in the HIV patient

population are increasingly due to non-AIDS-

related causes

* P<0.001 1996 vs 2004

Palella FJ, Jr., et al. J Acquir Immune Defic Syndr 2006; 43:27–34.

1997

1996

1998

1999

2000

2001

2002

0

5

10

15

20

25

30

35

40

45

2003

2004

13

27

27

15

34

36

39

3
6

43

Year

N=6,945

P

e

rc

e

n

ta

g

e

o

f

d

e

a

th

s

a

tt

ri

b

u

te

d

t

o

n

o

n

-A

ID

S

-r

e

la

te

d

c

a

u

s

e

s

background image

10

SMART study: Most deaths were due to

non-AIDS-defining events

SMART Study Group. NEJM 2006; 355: 2283–96.

0

5

10

15

20

25

30

Unknown

Opportunistic disease

Non-AIDS cancer

CV disease

Infection

Hepatic complications

Renal complications

Substance abuse

Accident/violent death

COPD

Hematological

Digestive system

CNS disease

Respiratory disease

Percent of deaths

N=5,472 (85 deaths)

background image

11

Mortality data from other cohorts confirm

increasing non-AIDS causes of death in HIV

patient population

D:A:D (Europe, Australia, USA)

1

23,441 HIV-infected patients, followed

up 1999–2004

855 of 1,246 deaths (69%) were non-

AIDS related

1. Weber R, et al. Arch Intern Med 2006; 166:1632–1641. 2. Hessamfar-Bonarek M, et al. 15th CROI 2008; Poster 666.

3. Lewden C, et al. Int J Epidemiol 2005; 34:121–30. 4. Bonnet F, et al. Clin Infect Dis 2009; 48:633–9.

AIDS-

related

Other

Liver-

related

CV or other

heart

diseases

Non-AIDS

cancers

33.8

%

31.1

%

14.5

%

11.0

%

9.4%

2005 Mortalité survey (France)

2–4

1,042 HIV-infected patients

The proportion of deaths due to non-

AIDS causes increased vs the 2000

survey

Leading non-AIDS causes of death in

2005 were

non-AIDS cancers

hepatitis C

CV disease

AIDS related

Non-AIDS related

53

64

47

36

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

2000

2005

background image

12

Increasing incidence of non-AIDS

hospitalizations confirmed in UK and US clinics

Brighton and Sussex University Hospitals (UK)

1

More than half (80/150) of HIV patient hospital admissions in 2006 were

non-HIV related

Most admissions in the pre-HAART era were due to an AIDS-defining illness

HOPS (USA)

2

7,155 HIV-infected patients in 10 US clinics

1. Soni S, et al. 14th BHIVA Conference 2008; Abstract P128.

2. Buchacz K, et al. AIDS 2008; 22:1345–1354.

%

h

o

s

p

it

a

li

za

ti

o

n

s

AIDS opportunistic infections

Chronic end-organ disease conditions

31.0

9.5

7.2

14.3

0

5

10

15

20

25

30

35

1994-1996

2003-2005

background image

13

Coronary heart disease incidence in young HIV-

infected patients is higher than in non-HIV-

infected counterparts

Currier JS, et al. J Acquir Immune Defic Syndr 2003; 33:506–512.

California Medicaid population 1994–2000

(N=3,083,209)

*Relative risk (HIV infected vs HIV uninfected), P<0.05

Non-HIV infected (N=3,054,696)

HIV infected (N=28,513)

Age category

6.46

5.55

7.79

4.29

6.12

3.69

2.78

2.26

1.51

1.60

0.76

1.64

0.11

0.77

Men

25–34*

18–24*

35–44 45–54* 55–64* 65–74* 75+

0

1

2

3

4

5

6

7

8

9

In

c

id

e

n

c

e

p

e

r

1

0

0

p

e

rs

o

n

-y

e

a

rs

25–34*

18–24*

35–44* 45–54 55–64* 65–74

75+

0

Women

7.42

8

In

c

id

e

n

c

e

p

e

r

1

0

0

p

e

rs

o

n

-y

e

a

rs

6.30

4.81

5.43

6.49

4.55

2.58

2.21

1.03

1.72

0.42

0.64

0.16

0.40

1

2

3

4

5

6

7

background image

14

Cohorts show an excess risk of non-AIDS

morbidity in HIV-infected patients vs the

general population

Cohort/Patient

population/Time period

Impact of HIV infection on risk

SHCS (Switzerland)

1

7,304 HIV-infected patients*

1985–2002

Non-AIDS cancers

17-fold Hodgkin’s lymphoma

Italian Registry of AIDS

2

12,104 HIV-infected patients

Up to 1998

Non-AIDS cancers

including anal cancer, lung cancer, brain tumors, Hodgkin’s

lymphoma, and leukemias

Linked population-based

AIDS and cancer registry

data (USA)

3

57,350 HIV-infected patients

1996–2002

Non-AIDS cancers

including Hodgkin’s lymphoma (standardized incidence ratio [SIR]:

5.6), lung cancer (2.6), and liver cancer (2.7)

ASD and HOPS (USA)

4

54,780 HIV-infected patients

1992–2003

Non-AIDS cancers

including anal cancer (SIR: 42.9), vaginal cancer (21.0), Hodgkin’s

lymphoma (14.7), liver cancer (7.7), lung cancer (3.3), and melanoma

(2.6)

VACS (USA)

5

33,420 HIV-infected veterans**

1997–2004

Liver disease and renal disease

* vs general population standardized for, age, gender, and reporting period
** vs age-, gender-, and race-matched HIV-uninfected veterans

1. Clifford GM, et al. J Natl Cancer Inst 2005; 97:425–432.

2. Dal Maso L, et al. Br J Cancer 2003; 89:94–100.

3. Engels EA, et al. Int J Cancer 2008; 123:187–194.

4. Patel P, et al. Ann Intern Med 2008; 148:728–736.

5. Goulet JL, et al. Clin Infect Dis 2007; 45:1593–1601.

background image

15

Mortality and morbidity in the modern HAART

era: Summary

Despite virologic suppression, patients in the HAART era still have at

least a 10-year shorter expected survival than age- and gender-

matched controls

1

Expected survival is closely associated with CD4+ T-cell counts

if patients on HAART reach CD4+ T-cell counts of ≥500 cells/mm

3

for ≥6 years,

survival can be similar to the general population

2

Mortality and morbidity in the HIV patient population are

increasingly due to non-AIDS-related causes,

3–8

e.g.

liver-related diseases

cardiovascular/other related heart diseases

non-AIDS cancers

Several cohorts show an excess risk of non-AIDS morbidity in HIV-

infected patients vs the general population

9–13

1. Lohse N, et al. Ann Intern Med 2007;146:87–95. 2. Lewden C, et al. J Acquir Immune Defic Syndr. 2007;46:72–77.

3. Palella FJ, Jr., et al. J Acquir Immune Defic Syndr 2006; 43:27–34. 4. SMART Study Group NEJM 2006; 355: 2283–96.

5. Weber R, et al. Arch Intern Med 2006; 166:1632–1641. 6. Hessamfar-Bonarek M, et al. 15th CROI 2008; Poster 666.

7. Lewden C, et al. Int J Epidemiol 2005; 34:121–30. 8. Bonnet F, et al. Clin Infect Dis 2009; 48:633–9.

9. Clifford GM, et al. J Natl Cancer Inst 2005; 97:425–432. 10. Dal Maso L, et al. Br J Cancer 2003; 89:94–100.

11. Engels EA, et al. Int J Cancer 2008; 123:187–194. 12. Patel P, et al. Ann Intern Med 2008; 148:728–736.

13. Goulet JL, et al. Clin Infect Dis 2007; 45:1593–1601.

background image

16

Association of immune response

with serious non-AIDS diseases

background image

17

Immunologic response to HAART is an

independent predictor of clinical outcome

1–3

A lack of immunologic response

3–9 months after HAART initiation

in patients who respond

virologically was shown to be

associated with adverse long-term

clinical outcomes

1

Higher on-treatment CD4+ T-cell

counts have been associated with

lower rates of non-AIDS diseases

and AIDS

2

Mortality rates for AIDS-defining

cancers were 20.1 vs 0.1 per 1,000

patient-years and for non-AIDS-

defining cancers were 6.0 vs 0.6

per 1,000 patient-years when the

current CD4+ T-cell count was <50

vs >500 cells/mm

3

, respectively

3

1. Tan R, et al. J Acquir Immune Defic Syndr 2008; 47:553–558.

2. Baker JV, et al. AIDS 2008; 22:841-848.

3. Monforte A, et al. AIDS 2008; 22:2143–2153.

background image

18

Proportion of deaths from non-AIDS causes is

increasing, but at CD4+ T-cell counts below

normal

Bonnet F, et al. Clin Infect Dis 2009; 48:633–9.

Lewden C, et al. JAIDS 2008; 48; 590-8.

2000 (N=964
deaths)

2005 (N=1,042
deaths)

42

205

227

338

337

156

Most recent CD4+ T-

cell count prior to

death (2005)

0

10

20

30

40

50

% of deaths

AIDS

Non-AIDS

cancer

Liver

related

Cardiovascular

Suicide

Non-AIDS-

defining

infection

-11%

+6%

+2%

+1%

+1%

-3%

Difference in %

deaths between 2000

and 2005

background image

19

SMART

D:A:D

CASCADE*

FIRST

VACS

D:A:D

CASCADE

FIRST

Aquitaine

SMART

D:A:D

CASCADE

FIRST

SMART

FIRST

VACS

Cohorts/clinical studies suggest lower CD4+ T-cell

counts are associated with higher rates of non-

AIDS clinical events

Cardiovascular

disease

Non-AIDS cancer

Liver Disease

Renal disease

Goulet JL, et al. Clin Infect Dis 2007; 45:1593–1601.

Weber R, et al. 12th CROI, Abstract 595.

CASCADE Collaboration, AIDS 2006; 20:741–749.

Baker JV, et al. AIDS 2008; 22:841–848.

Weber R, et al. Arch Intern Med 2006; 166:1632–1641.

Bruyand M, et al. 15th CROI, Abstract 15.

El Sadr WM, et al. N Engl J Med 2006; 355:2283–2296.

* Cardiovascular disease or type 2 diabetes;

Vascular disease

Causes of
morbidity/
mortality:

Studies
suggestin
g
associatio
n with
lower
CD4+
count:

background image

20

Current CD4+ T-cell count

(cells/mm

3

)

Philips A . AIDS 2008; 22:2409–2418.

1. CASCADE Collaboration. AIDS 2006; 20:741–749. 2. Weber R, et al. Arch Intern Med 2006; 166:1632–1641.

R

a

te

p

e

r

1

0

0

p

a

ti

e

n

t-

y

e

a

rs

(9

5

%

C

l)

CASCADE

1

(N=7,680 [ART-

naive];

1,938 deaths)

Non-AIDS causes

1.6

1.2

0.8

0.4

0.0

D:A:D

2

(N=23,441;

1,246 deaths)

≥500

200–

349

All causes

0.0

1.6

1.2

0.8

0.4

CASCADE and D:A:D: Higher rates of death

from non-AIDS causes at lower CD4+ T-cell

counts

350–

499

≥500

200–

349

350–

499

background image

21

CASCADE: Lower risk of death from serious

non-AIDS diseases if CD4+ T-cell counts are

higher

CASCADE Collaboration. AIDS 2006; 20:741–749.

N=7,680 ART-naive patients

1,938 deaths (1997–2003)

Category

Hazard ratio* (95% CI)

* per 100 cells/mm

3

higher latest CD4+ T-cell count

† including accidents, suicide, overdose

(Un-)intentional

Cardiovascular disease/
diabetes mellitus

Unknown

Organ failure

Hepatitis/liver related

Other infections

Non-AIDS cancer

0.86

0.81

0.9

0.83

0.95

0.89

0.8
6

0.5

1.5

Lower risk

Higher risk

background image

22

Baker JV, et al AIDS 2008; 22:841–

848.

Category

Patients

Hazard ratio (95% CI)

Liver

CV

Renal

Cancer
(non-AIDS)

14

24

14

32

0.1

1

10

Lower risk

Higher risk

0.87
1.02

0.85
0.91

0.63
0.71

0.81
0.82

Univariate

Multivariate

FIRST study: Risk for non-AIDS diseases is lower

in HIV-infected patients who have higher CD4+ T-

cell counts

* per 100 cells/mm

3

higher latest CD4+ T-cell count; multivariate

hazard ratios adjusted for latest HIV RNA and baseline
covariates: age, sex, race/ethnicity, prior AIDS, and HBV and
HCV co-infection

N=1,397 ARV-naive
patients

Hazard ratio* (95% CI)

background image

23

Philips A . AIDS 2008; 22:2409–2418.

Non-AIDS cancer:

Lower

risk consistently

associated with

higher

CD4+ T-cell counts

* Adjusted hazard ratio per 100 cells/mm

3

higher latest

CD4+ T-cell count for the association between latest CD4+ T-
cell count and risk of non-ADIS cancers or deaths from non-
AIDS cancers

Adjusted hazard ratio

(95% CI)*

Number

of

events

Includes

non-fatal

32

255

46

49

FIRST

(N=1,39

7)

D:A:D

(N=23,44

1)

CASCADE

(N=7,680)

SMART

(N=5,472)

0.6

0.7

0.8

0.9

1.0 1.1 1.2 1.3 1.4 1.5 1.6

Yes

Yes

No

No

background image

24

14

27

18

0.4

0.5

0.6

0.7

0.8 0.9 1.0

1.2

1.5

Number

of

events

Includes

non-

fatal

Yes

Yes

No

Renal disease:

Lower

risk consistently

associated with

higher

CD4+ T-cell counts

Adjusted hazard ratio

(95% CI)*

Philips A . AIDS 2008; 22:2409–2418.

* Adjusted hazard ratio per 100 cells/mm

3

higher latest

CD4+ T-cell count for the association between latest
CD4+ T-cell count and risk of renal disease or deaths from
renal disease

FIRST

(N=1,39

7)

D:A:D

(N=23,44

1)

SMART

(N=5,472)

background image

25

SMART

(N=5,472)

D:A:D

(N=23,4

41)

Cardiovascular disease:

Lower

risk

consistently associated with

higher

CD4+ T-cell

counts

Philips A . AIDS 2008; 22:2409–2418.

* Adjusted hazard ratio per 100 cells/mm

3

higher latest

CD4+ T-cell count for the association between latest CD4+
T-cell count and risk of cardiovascular disease or deaths
from cardiovascular disease

FIRST

(N=1,3

97)

CASCADE

(N=7,680)

24

855

0.6

0.7

0.8

0.9

1.1

1.0

1.3 1.4

Number

of

events

Includes

non-

fatal

Yes

Yes

No

Adjusted hazard ratio

(95% CI)*

36

145

Yes

1.2

1.5 1.6

background image

26

FIRST

(N=1,39

7)

D:A:D

(N=23,44

1)

CASCADE

(N=7,680)

SMART

(N=5,472)

Number

of

events

Includes

non-

fatal

Liver disease:

Lower risk

consistently

associated with

higher

CD4+ T-cell counts

Adjusted hazard ratio

(95% CI)*

Philips A . AIDS 2008; 22:2409–2418.

* Adjusted hazard ratio per 100 cells/mm

3

higher latest

CD4+ T-cell count for the association between latest CD4+
T-cell count and risk of liver disease or deaths from liver
disease

14

301

40

25

0.6

0.7

0.8

0.9 1.0

1.1 1.2 1.3 1.4 1.5 1.6

No

Yes

No

Yes

background image

27

R

a

te

s

p

e

r

1

0

0

p

a

ti

e

n

t-

y

e

a

rs

(9

5

%

C

I)

Liver-related death rates in D:A:D: Higher in

patients with lower CD4+ T-cell counts

Weber R, et al. Arch Intern Med 2006;

166:1632–1641.

Overall

rate

Number of events

22

21

44

52

21

21

Latest CD4+ T-cell count

Bars show death rates per 100 patient-years of follow-
up, stratified by latest CD4+ T-cell count

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

<50

50–99

100–199 200–349 350–499

≥500

N=23,441

background image

28

SMART study: Levels of inflammation and

coagulation markers are associated with risk

of death

Kuller LH, et al. PLOS Medicine 2008; 5:

e203.

IL-6

(inflammation marker)

1

1

4.5

12.4

0

5

10

15

20

25

30

35

40

45

<25th

percentile

25–49th

percentile

50–74th

percentile

≥75th

percentile

O

d

d

r

a

ti

o

(

9

5

%

C

I)

d

e

a

th

D-dimer

(coagulation marker)

1

8.3

12.6

41.2

0

50

100

150

200

250

<25th

percentile

25–49th

percentile

50–74th

percentile

≥75th

percentile

Quartiles of biomarker level at baseline

N=255

baseline

plasma

samples

background image

29

SMART study: Levels of inflammation and

coagulation markers are associated with risk

of CV disease/death

Kuller L, et al. CROI 2008. Abstract 139.

P-values are for adjusted odds ratios for the 4th vs 1st quartile of
each biomarker at baseline, estimated using logistic regression

IL-6

2

0

1

2

3

4

5

6

O

d

d

r

a

ti

o

(

9

5

%

C

I)

C

V

d

is

e

a

s

e

/d

e

a

th

D-dimer

2.8

P=0.003

P=0.06

2.8

P=0.002

Amyloid P

N=499 stored

baseline plasma

samples

background image

30

Association of immune response with serious

non-AIDS diseases: Summary

Evidence suggests that incomplete immune restoration in HIV

patients on HAART is, at least in part, accountable for the excess in

non-AIDS-related morbidity and mortality

1–4

Lower CD4+ T-cell counts are associated with higher rates of non-

AIDS clinical events,

1–4

including those associated with

non-AIDS cancers

renal disease

cardiovascular disease

liver disease

Higher baseline levels of markers of inflammation and coagulation

are associated with higher risk of death and cardiovascular disease

in HIV-infected patients

5

1. Baker JV, et al. AIDS 2008; 22:841–848. 2. Weber R, et al. Arch Intern Med 2006; 166:1632–1641.

3. Friis-Moller N, et al. AIDS 2003; 17:1179–1193. 4. Philips A . AIDS 2008; 22:2409–2418.

5.KullerLH, et al. PLOS Medicine 2008; 5: e203.

background image

31

Limitations of HAART in restoring

immune function

background image

32

Limitations of HAART in restoring immune

function

The immune deficits related to HIV infection are only partially

resolved on current HAART regimens compared to HIV-negative

subjects, despite sustained virologic suppression

1,2

CD4

+

T-cell counts remain below normal for many patients

3–5

The same pathologic processes of immune activation and

immune deficiency persist, albeit at reduced levels

6

1. Hunt PW, et al. J Infect Dis 2003; 187: 1534-43. 2. Anthony, KB, et al. JAIDS 2003; 33: 125-133.

3. Kaufmann GR, et al. AIDS 2000; 14: 959-69. 4. Kaufmann GR, et al. Clin Infect Dis 2005; 41: 361-72.

5. Gras L, et al. JAIDS 2007; 45: 183-92. 6. Wainberg M & Clotet B. AIDS Pat Care STD 2007; 21: 609-620.

background image

33

0

0

100

200

300

400

500

600

700

800

900

6

12

18

24

30

36

42

48

54

60

P<0.001

Months after HAART

initiation

C

D

4

+

T

-c

e

ll

c

o

u

n

t

(c

e

ll

s

/m

m

3

)

All patients

(N=293)

Complete CD4+ T-cell responders

(N=188)

Incomplete CD4+ T-cell responders

(N=105)

Even after 5 years of effective HAART, many

patients fail to obtain normal CD4+ T-cell

counts

Adapted from Kaufmann GR, et al. Clin Infect Dis 2005; 41:361–372.

background image

34

Patients with <500 cells/mm

3

after 4 years of

HAART are unlikely to achieve ≥500 cells/mm

3

in the long term

Kelley C, et al. Clinical Infectious Diseases 2009;

48:787–94.

Months of HAART

1.00

0.7

5

0.5

0

0.2

5

0.0

0

P

ro

b

a

b

il

it

y

o

f

im

m

u

n

o

lo

g

ic

r

e

s

to

ra

ti

o

n

0

1
0

2
0

3
0

4
0

5
0

6
0

7
0

8
0

9
0

100 110 120 130 140 150

N=216

N=76

N=74

Year 4 CD4+ T-cell

count:

<350

cells/mm

3

350–500

cells/mm

3

>500 cells/mm

3

background image

35

HAART is associated with reduced immune

activation, but levels remain high vs HIV-

negative subjects

Adapted from Hunt PW, et al. J Infect Dis 2003; 187:1534–1543.

0

10

20

30

HIV infected

untreated

(N=13)

HIV infected

treated

(N=99)

HIV

uninfected

(N=6)

P

e

rc

e

n

ta

g

e

o

f

a

c

ti

v

a

te

d

C

D

4

+

T

c

e

ll

s

P<0.001

P<0.001

0

10

20

30

HIV infected

untreated

(N=13)

HIV infected

treated

(N=99)

HIV

uninfected

(N=6)

P

e

rc

e

n

ta

g

e

o

f

a

c

ti

v

a

te

d

C

D

8

+

T

c

e

ll

s

P<0.001

P<0.001

CD4+

CD8+

background image

36

Amyloid A

Amyloid P

C-reactive protein

(CRP)

Interleukin 6 (IL-6)

Tumor necrosis

factor (TNF-α)

Soluble TNF

receptors (sTNF-

R1, sTNF-R2)

Adiponectin

E-selectin

Granulocyte-

macrophage colony-

stimulating factor

(GM-CSF)

IL-10

Monocyte

chemotactic protein

(MCP1)

Myeloperoxidase

(MPO)

D-dimer

Prothrombin

fragment 1+2

(F1.2)

Many biomarkers of immune activation,

inflammation, and endothelial activation have

been identified

These markers may provide additional prognostic information on disease

progression and treatment response

Inflammation

Endothelial activation and

damage

Coagulation

Soluble

intercellular

adhesion

molecule and

cell adhesion

molecule

(sICAM-1 and

sVCAM-1)

Tissue-type

plasminogen

activator (t-PA)

and activator

inhibitor (t-PAI-

1)

von Willebrand

factor (vWF)

Kuller L, et al. 15th CROI 2008. Abstract 139. Ross A, et al. 15th CROI 2008; Abstract 949.

Ross A, et al. 15th CROI 2008; Abstract 954. Calmy A, et al. 15th CROI 2008; Abstract 140.

background image

37

Markers of inflammation, such as C-reactive

protein, may persist at elevated levels despite

HAART

* P<0.001 vs HIV uninfected
** P<0.001 vs HIV infected, untreated

Adapted from Kristoffersen US, et al. 15th CROI 2008; Poster 953.

N=115 HIV-infected patients

N=30 HIV-uninfected matched controls

40,000

35,000

30,000

25,000

20,000

15,000

10,000

5,000

0

P

la

s

m

a

c

o

n

c

e

n

tr

a

ti

o

n

o

f

h

s

C

R

P

(

n

g

/m

L

)

HIV uninfected HIV infected,

untreated

HIV infected,

3 months of

HAART

HIV infected,

12 months of

HAART

**

*

background image

38

Limitations of HAART in restoring immune

function: Summary

The immune deficits related to HIV infection are only partially resolved

on current HAART regimens compared to HIV-negative subjects,

despite sustained virologic suppression

1,2

Even after 5 years of effective HAART, many patients fail to obtain

normal CD4+ T-cell counts

3

Patients with <500 cells/mm

3

after 4 years of HAART are unlikely to achieve

≥500 cells/mm

3

during long-term follow-up

4

Initiation of HAART is associated with significant reductions in immune

activation, but levels remain elevated vs HIV-negative subjects

1

Biologic markers of immune activation, inflammation, and endothelial

activation may provide additional prognostic information on disease

progression and treatment response

Markers of inflammation, such as C-reactive protein, may persist at elevated

levels despite HAART

5

1. Hunt PW, et al. J Infect Dis 2003; 187: 1534-43. 2. Anthony, KB, et al. JAIDS 2003; 33: 125-133.

3. Kaufmann GR, et al. Clin Infect Dis 2005; 41:361–372. 4. Kelley C, et al. Clin Infect Dis 2009; 48:787–94.

5. Kristoffersen US, et al. 15th CROI 2008; Poster 953.

background image

39

Potential mechanisms behind

premature incidence of serious

non-AIDS diseases in HIV-infected

patients

background image

40

Chronic immune activation and inflammation

may contribute to elevated risk of non-AIDS

diseases

Impaired CD4+ T-cell reconstitution and immune dysfunction may

underlie the premature incidence of non-AIDS conditions in HIV-

infected patients

Processes that may contribute to persistent immune deficits

include:

Very early loss of CD4+ T cells from the gastrointestinal tract

1-3

Disruption of the gastrointestinal mucosal barrier, leading to

microbial translocation

4

HIV antigen persistence, resulting in ongoing immune

stimulation

5

Chronic, generalized immune activation

6

Fibrosis and activation-induced damage to lymphatic tissue

7

1. Li Q, et al. Nature 2005; 434:1148–1152. 2. Mattapallil JJ, et al. Nature 2005; 434:1093–1097.

3. Brenchley JM, et al. J Exp Med 2004; 200:749–759. 4. Brenchley JM, et al. Nat Med 2006; 12:1365–1371.

5. Hunt P, et al. J Infect Dis 2003; 187:1534–43. 6. El Far M, et al. Curr HIV/AIDS Rep 2008; 5:13–19.

7. Schacker T, et al. J Clin Invest 2002; 110:3–113.

background image

41

Acute HIV infection is characterized by a rapid

and massive loss of the body’s CD4+ T cells

Plasma viremia

Blood CD4+ T cells

Body CD4+ T cells

Primary

infection

±Acute HIV syndrome
Wide dissemination of virus
Seeding of lymphoid organs

R

e

la

ti

v

e

m

a

g

n

it

u

d

e

Time after initial infection

Weeks

Death

Opportunistic

diseases

Clinical latency

Constitutional

symptoms

Years

0

3

6

9

12

1

2

4

5

7

8

10 11+

3

6

9

background image

42

The bulk of early T-cell loss appears to occur in

GALT

Most of the body’s CD4+ T-cells are located in gut-associated

lymphoid tissue (GALT)

1

The majority of GALT CD4+ T cells express CCR5 and are

permissive to HIV infection

2

GALT is the principal target of the acute phase of HIV replication

and associated T-cell loss

3-5

This phase results in lasting damage to GALT

1. Veazey RS & Lackner AA. PLoS Med 2006; 3:e515. 2. Appay V & Sauce D. J Pathol 2008; 214:231–241.

3. Li Q, et al. Nature 2005; 434:1148–1152. 4. Mattapallil JJ, et al. Nature 2005; 434:1093–1097. 5. Brenchley JM, et al. J Exp Med 2004; 200:749–759.

background image

43

Even with early HAART, depletion of CD4+ T cells

in GALT may persist despite normalization in

peripheral blood

Mehandru S, et al. PLoS Med 2006; 3:e484.

AEI,

untreated

(N=32)

0

20

P

e

rc

e

n

ta

g

e

o

f

a

c

ti

v

a

te

d

C

D

4

+

T

c

e

ll

s

40

60

80

AEI, treated

up to 1 year

(N=7)

AEI, treated

1–3 years

(N=7)

AEI, treated

3–7 years

(N=8)

HIV

uninfected

(N=18)

Peripheral blood

Mucosa

AEI = acute/early HIV infected

background image

44

Microbial translocation, reflected in increased

plasma LPS levels, may stimulate systemic

immune activation

Brenchley JM, et al. Nat Med 2006; 12:1365–1371.

(N=50)

(N=59)

(N=37)

(N=47)

P<0.0001

P<0.0001

P<0.0001

P<0.0001

P=0.842

200

100

0

P

la

s

m

a

L

P

S

(

p

g

/m

L

)

Uninfected

Acute/early

Chronic

AIDS

LPS = lipopolysaccharide

background image

45

With HAART, LPS levels are only partially

reduced, suggesting persistent microbial

translocation

Brenchley JM, et al. Nat Med 2006; 12:1365–1371.

P=0.0107

0

P

la

s

m

a

L

P

S

(

p

g

/m

L

)

50

100

150

Progressors,

untreated

(N=68)

Progressors,

48 weeks HAART

(N=28)

Uninfected

(N=47)

P=0.0026

background image

46

Higher levels of structural damage to lymphoid

tissue are associated with impaired CD4+ T-

cell reconstitution

Schacker T, et al. J Clin Invest 2002; 110:3–113.

N=7

700

600

500

400

300

200

100

0

800

900

1000

C

h

a

n

g

e

i

n

p

e

ri

p

h

e

ra

l

C

D

4

+

T

-c

e

ll

c

o

u

n

t

Percent area collagen

2

4

6

8

10

12 14

16

18

20

background image

47

Premature “aging” of immune system:

32-

40

● Decreased thymic function

● Deregulation of hemataopoiesis

● Decreased vaccine response

● Lymphoid tissue fibrosis

Immune activation

and

inflammation

14-21

Biomarkers:

TNF-α

19,22,23

D-dimer

24-26

hs-CRP

24,27,28

IL-6

23,24

CD4+ T-cell

depletion

29-31

Excessive CD4+

T-cell

turnover

30,31,53

HIV persistence in

the GALT/lymphoid

tissue.

7-9

Bacterial

translocation.

10

Viral coinfections

(eg. HCV, CMV,

EBV)

11-13

Uncontrolled viral

replication

Death from

AIDS/OIs

H

A

A

R

T

D

e

cr

e

a

se

d

v

ir

a

l

lo

a

d

/

in

cr

e

a

se

d

C

D

4

+

T

-c

e

ll

co

u

n

t

U

n

d

e

te

ct

a

b

le

v

ir

a

l

lo

a

d

Pre-HAART era

1-4

Current HAART era

5,6

V

ir

a

l

re

p

li

ca

ti

o

n

/G

A

LT

In

cr

e

a

se

d

v

ir

a

l

lo

a

d

/d

e

cr

e

a

se

d

C

D

4

+

T

-c

e

ll

c

o

u

n

t

Initial HIV

Infection

Uncontrolled viral

replication

Death from

AIDS/OIs

H

A

A

R

T

D

e

cr

e

a

se

d

v

ir

a

l

lo

a

d

/

in

cr

e

a

se

d

C

D

4

+

T

-c

e

ll

co

u

n

t

U

n

d

e

te

ct

a

b

le

v

ir

a

l

lo

a

d

Pre-HAART era

1-4

Current HAART era

5,6

In

cr

e

a

se

d

v

ir

a

l

lo

a

d

/d

e

cr

e

a

se

d

C

D

4

+

T

-c

e

ll

c

o

u

n

t

An emerging model of HIV-induced immune

dysfunction

Liver disease

12,49,50-

52

● Accelerated Hep C

● Hepatic failure

Non-AIDS

cancers

44-48

● Lung

● Hodgkin’s

lymphoma

Heart disease

1,42,43

● Atherosclerosis

● Increased CIMT

AIDS-related or

other-cause

morbidity/mortalit

y

1-3

HIV persistence in

the GALT/lymphoid

tissue.

7-9

Bacterial

translocation.

10

Viral coinfections

(eg. HCV, CMV,

EBV)

11-13

Immune activation

and

inflammation

14-21

Biomarkers:

TNF-α

19,22,23

D-dimer

24-26

hs-CRP

24,27,28

IL-6

23,24

Normalized

lifespan

41

● CD4+ T-cell

counts ≥500 for

≥6yr

Animatic of an emerging model of HIV-induced immune

dysfunction can be viewed in presentation mode

background image

48

Promising directions for ongoing and future

research

Investigational approaches to improve immune

function restoration in HIV-infected patients:

Earlier HAART initiation (at CD4+ T-cell counts >500 cells/mm

3

)

Treatment with valganciclovir to decrease T-cell activation in

HIV/CMV co-infected patients

Novel immunomodulating strategies

synthetic cytokines

recombinant human growth hormone

toll-like receptor agonists

CCR5 antagonism

ClinicalTrials.gov: NCT00264290 and NCT00071240. Sereti I, et al. Blood 2009; [Epub ahead of print].

background image

49

Additional IL-2–associated CD4+ T-cell gains

conferred no clinical benefit

SILCAAT study

1

(N=1,695)

ESPRIT study

2

(N=4,111)

Category

Patients

IL-2 / Control

OD or death
(primary endpoint)

Death

Grade 4
events

113 / 119

80 / 75

199 / 184

0.1

1

2

Lower risk

Higher risk

0.95

1.08

1.10

Hazard ratio (95% CI)

IL-2 vs control

159 / 163

107 / 114

467 / 380

0.95

0.91

1.24

1. Levy CROI 2009 Abstract 90bLB. 2. Losso CROI 2009 Abstract 90aLB.

OD = opportunistic disease; CI = confidence interval

background image

50

CCR5 antagonism was associated with an

additional CD4+ T-cell count gain in treatment-

experienced patient studies

C

D

4

+

T

-c

e

ll

c

o

u

n

t

(c

e

ll

s

/m

m

3

)

CCR5 antagonists (vicriviroc, maraviroc)

Other ARVs (etravirine, enfuvirtide, raltegravir, darunavir,
tipranavir
)

-10

10

30

50

70

90

110

130

150

170

Proportion of patients with HIV RNA <50 copies/mL at week 24

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

Size of circle represents sample size

Wilkin T, et al. 15th CROI 2008; Poster 800.

background image

51

Rationale for immune effects of CCR5

antagonism

Certain host genotypes for CCR5 and its major HIV-suppressive ligand

CCL3L1 appear to affect CD4+ T-cell recovery in long-term HAART,

independent of virologic suppression

Patients with favorable

CCR5Δ32

and CCL3L1l genotypes experience

greater HAART-mediated CD4+ T-cell count increases

1

Transplant studies suggest an immunomodulating role of CCR5 antagonism

Administration of a CCR5 antagonist in an established monkey heart

transplant model attenuated immune activation and postsurgical stress

responses

2

“Natural hosts” of SIV infection have low CCR5 expression on their CD4+

T cells

3

A number of studies are underway to assess the efficacy of maraviroc,

a CCR5 antagonist, in modulating inflammation and immune response

at various phases of HIV infection

1. Ahuja SK, et al. Nat Med 2008;14:413-20.

2. Schröeder C, et al. J Immunol 2007; 179:2289–99. 3. Pandrea I, et al. Blood 2007; 109:1069–76.

background image

52

Summary

While HAART has improved morbidity and mortality rates for the HIV-

infected patient population, an excess risk remains compared to HIV-

negative individuals

1-5

HIV-infected patients appear to have an excess risk for various non-

AIDS conditions vs the general population

1-5

Persistent immune dysfunction may play a central role

6-8

Mechanisms by which HIV increases the risk of serious non-AIDS

disease is an important emerging area for HIV research

Studies are needed to better understand the mechanisms and

processes of immune deficiency and immune restoration in HIV

infection

This in turn may drive development of interventions that maximize

both viral suppression AND immune reconstitution, further reducing

morbidity and mortality in HIV-infected patients

1. Clifford GM, et al. J Natl Cancer Inst 2005; 97:425–432. 2. Dal Maso L, et al. Br J Cancer 2003; 89:94–100.

3. Engels EA, et al. Int J Cancer 2008; 123:187–194. 4. Patel P, et al. Ann Intern Med 2008; 148:728–736.

5. Goulet JL, et al. Clin Infect Dis 2007; 45:1593–1601. 6. Hunt PW, et al. J Infect Dis 2003; 187: 1534-43.

7. Anthony, KB, et al. JAIDS 2003; 33: 125-133. 8. Kaufmann GR, et al. Clin Infect Dis 2005; 41:361–372.


Document Outline


Wyszukiwarka

Podobne podstrony:
Modification of Intestinal Microbiota and Its Consequences for Innate Immune Response in the Pathoge
fitopatologia, Microarrays are one of the new emerging methods in plant virology currently being dev
Issue of Gun Control and Violence As Seen in the U S and
(gardening) Roses in the Garden and Landscape Cultural Practices and Weed Control
Gun Control in the U S New Methods Are Needed
Gun Control in the USA
2000 10 Raid Controllers 15 Scsi Controllers in the Linux Labs
Domjan M , Cusato B Pavlovian feed forward mechanisms in the control of social behavior
Implementing Anti Virus Controls in the Corporate Arena
Jacobson Mind Control in the United States
Design of an Artificial Immune System as a Novel Anomaly Detector for Combating Financial Fraud in t
Mind Control in the 1990 s Neuro Linguistic Programming
Mettern S P Rome and the Enemy Imperial Strategy in the Principate
Early Variscan magmatism in the Western Carpathians
Applications and opportunities for ultrasound assisted extraction in the food industry — A review
In the end!
Cell surface in the interaction Nieznany
Post feeding larval behaviour in the blowfle Calliphora vicinaEffects on post mortem interval estima

więcej podobnych podstron