IMMUNODEPRESSION ACQUISE EN REANIMATION

Outils diagnostiques et perspectives thérapeutiques

DESC de Réanimation2011

Guillaume Monneret

Laboratoire d’Immunologie CellulaireHospices Civils de Lyon

Hôpital E. Herriotguillaume.monneret@chu-lyon.fr

Hôpitaux de Lyon

- Septic Syndromes : leading cause of death in ICU- 3rd cause of death after cardiovascular diseases and cancers- 2005 US figures : 800 000 cases / year- 2001 F figures : 70 000 cases / year

Sepsis as a starting point : Septic syndromes an increasing and significant healthcare challenge

NEJM 2003

Constant rise for many years

200,000

400,000

600,000

800,000

1,000,000

1,200,000

1,400,000

1,600,000

1,800,000

2001 2025 2050

Year

100,000

200,000

300,000

400,000

500,000

600,000

Severe Sepsis Cases

US Population

Sep

sis

Cas

es

To

tal

US

Po

pu

lati

on

/1,0

00

- Better care of co-morbidities- Increased longevity

future

Hospitalization rate nearlydoubled from 1993 to 2003

Population-based mortality rate rose by two thirds.

Wenzel 2002 - N Engl J Med

Incr

ea

sed

seve

rity

MultipleOrganfailure

DefinitionAssociation of an infection and a systemic inflammatory response syndrome (SIRS)

Onset :germ(s) + site of infection + local inflammation

Amplification ?Lack of modulation ?

Pathophysiology :uncontrolled inflammatory

response

organ

systemic

DefinitionAssociation of an infection and a systemic inflammatory response syndrome (SIRS)

=> Sepsis is not caused by the infection itself but by the host response to this infection

(“The germ is nothing, the terrain is everything” - Pasteur L. - 1895)

Decreased arterial pressureShock

Multiple organ failure

UncontrolledInflammatory

response

Decreased arterial pressureShock

Multiple organ failure

UncontrolledInflammatory

response

Emergency symptomatic treatment :AntibiotherapyAgressive vascular resuscitationVasoactive agents

Cytokines : TNF, IL-1, IL-6, IL-12Activation, amplification

Chemokines : IL-8, MIP-1, MCP-1Mobilize and activate leukocytes

Vasoactive hormones, Lipid mediators :PAF, PG’s, LK, TX, ADM, ANP…Multiple actions

Oxygen radicals :superoxide, nitric oxideAntimicrobial propertiesregulation vascular tone

« Cytokine storm »=> « Mediator storm » > 300 released mediators (Marshall, Nature Rev. 2003)

Extensive description for many years…..

Uncontrolledinflammatory response

Anti-inflammatory drugs

Failure of clinical trials testing anti-inflammatory therapies

38

35

39

45

40

40

37

39

…

381203433Total

35

36

50

41

38

40

35

…

2010

755

870

4132

688

514

1267

…

4

2

2

8

2

3

9

…

Anti-endotoxine

Anti-bradykinine

Anti-PAF

Anti-TNF

R solubles TNF

AINS

Stéroïdes

…

Mortality (%)

Placebo DrugNumber of

patientsNumber of

studiesDrug

Zeni et al, Crit Care Med, 1997

Reasons for such a perfect disaster

1. Inappropriate animal models of sepsis

2. Heterogeneity of septic population (germs, virulence factors, co-morbidities, inflammatoryresponse, site of infection….)

1. Pathophysiology largely unknown

> 70 % total mortality

Resultant immune response at the systemic level = immunosuppression

Pro-inflammatory Response

Anti-inflammatory Response

Time

Simplified description of systemic pro- and anti-inflammatory immune responses over time after septic shock

Similar mechanisms each time SIRS occurs:trauma, surgery, pancreatitis, burns……

Onset ofAnti-inflammatory-based RCT

1. Inappropriate animal models of sepsis2. Heterogeneity of septic population3. Pathophysiology largely unknown

Summarized view of sepsis-induced immunosuppression

immune functions

Innate Immunity

Monocyte / DCanergy

Adaptive Immunity

Lymphocyteanergy

IL-10(and soluble mediators)

Apoptosis(different mechanisms)

Endotoxintolerance

PNN chemotaxis and phagocytosisMonocyte deactivation:

- antigen presentation capacity- pro-inflammatory cytokine production- CX3CR1

dendritic cells nbr

Profound lymphopeniaTh2 shiftLymphocyte anergy apoptosis % Treg

Epigeneticregulation

Consequences

- Decreased clearance of initial infection (Togersen 2009)- Increased nosocomial infections- Viral reactivation

=> May directly contribute to mortality

Diagnostic?

No clinical sign => biological monitoring

1. Functional testing2. Soluble mediators3. Cellular approach4. Genomics

outcomes: - mortality - occurrence of nosocomial infections

- Because this directly measures ex vivo the capacity of a cell population to respondto an immune challenge, functional testing theoretically represents the method of reference

- Monocyte capacity to release TNF in response to LPS challenge- Lymphocyte proliferation in response to recall antigens or mitogens- Phagocytosis, chemotaxis….

- Time consuming (days of incubation for lymphocyte proliferation)- Home-made protocols => difficult to standardize

=> Not suitable for routine monitoring

- They remain essential to gain insights in the understanding of pathophysiology and to assess the validity of surrogate markers

1. Functional Testing

2. Soluble mediators

- In septic shock, > 300 released mediators- Both pro- and anti-inflammatory mediators are elevated : not informative- A panel of markers is likely more desirable (or at least a ratio)

=> If a single one : IL-10- Potent immunosuppressive cytokine- Many studies have identified it as the most informative - Standardized measurement

20

40

60

80

100

120

140

160

180

200

IL-10 pg/ml

1-2 3-4 5-7 8-15

**

**** **

Non-survivors 20

40

60

80

100

120

140

TNF pg/ml

1-2 3-4 5-7 8-15

Non-survivors

High IL-10 is associated with mortality – not TNF

2004

=> Correlation with mortality but no data on nosocomial infections

3. Cellular phenotyping

- monocytes- lymphocytes

Monocytes

Gold standard = Decreased mHLA-DR

- mHLA-DR expression level = Integrated Σ of the effects of numerous mediatorsIt is the true reflection of what force dominates at any given time-point

- Decreased mHLA-DR correlates with decreased functional testing (TNF release, Ag presentation)

Why using flow cytometry ? => example of mHLA-DR

100 % 0 %

Ting et al.

(n = 120 septic shock patients)

30

40

50

0 - 48 h 48 – 96 h

% H

LA-D

R +

mon

ocyt

es

Survivors

Non-survivors

p : 0.2 p < 0.001

Monneret et al., Intensive Care Med 2006

(Control values > 90-100 %)

Low HLA-DR predicts mortality

Survival curves stratified onmHLA-DR at 30 % at days 3-4

(n = 120 patients)

Sex (F)

Age > 64 years

IGS II (onset) > 49

Type of admittance (surgery vs med.)

Comorbidity (≥ 1)

Type of infection (noso. vs commu.)

Infection site (pulm., abdo., others)

SOFA (≠ J1J2 vs J3J4) > 0

HLA-DR (J1-J2) < 30 %

HLA-DR (J3-J4) < 30 %

-

-

6.14

-

4.34

-

-

6.58

-

8.81

-

-

1.3 – 28.4

-

1.0 – 18.5

-

-

1.5 – 28.6

-

1.9 – 40.4

-

-

0.02

-

0.05

-

-

0.01

-

0.005

Odds ratio 95 % CI p

Multivariate analysis : mHLA-DR is an independent predictor of mortality(after adjustment for usual clinical confounders)

9-fold increased risk of death with mHLA-DR < 30 %

- Multivariate analysis (including usual confonding factors):SOFA, SAPSII, Intubation, catheterization- Competitive risks

Does mHLA-DR predict nosocomial infections in other ICU contexts ?

HLA-DR expression and soluble HLA-DR levels in septic patients after traumaDitschkowski et al. Ann. Surg. 1999

Minor injuries

Severe injuries + secondary sepsis

Severe injuries withoutsecondary sepsis

Normalvalues

Recovery slope(> 1.2)

AUC : 0.8

Low mHLA-DR predicts nosocomial infections after Day 15

ROC Curves Analysis for the predictionsecondary infections

(AUC = 0.9)

According tosecondary infection(n = 29 non infected

n = 24 infected)

Dendritic cells

Persisting low circulating myeloid dendritic cells numberis associated with the development of nosocomial infections after septic shock

Days after shock

Pene, Chiche et al. (Société Réanimation Langue Française 2009)

Lymphocytes

Castelino et al.

From lab computer in a university hospital over a 3 month-periodLymphopenia when Ly < 600 / microL=> 1042 patients with lymphopenia

Septicpatients

Septicpatients

Controls

Controls

Lymphocytes from septic patients do not proliferate upon antigen / mitogen challenge

Roth et al.,

An increased circulating percentage of Treg is associatedwith a decreased cell proliferation in septic shock patients

FOXP3 inhibition restores lymphocyte proliferation

Lymphocytes:

No recent data on correlation with motality and nosocomial infections

n = 1211surgical

ICU patients

Skin testing(within 24 h after admission)

Positive30 %

(n = 369)

+/- 23 %

(n = 272)

Negative47 %

(n = 570)

Mortality

9 % (n = 31)

25 % (n = 55)

32 % (n = 184)

Nosocomialinfection

27 % (n = 99)

33 % (n = 91)

42 % (n = 237)

* p < 0.005

Skin testing with 5 antigens (positive if 2 reacted, +/- if solely 1 reacted, negative if none reacted)

**

Cook CCM 2009

4. Transcriptomic approach

(microarrays and qRT-PCR)

Genes codingfor pro-inflammatory immune responseare decreasedwhile genes coding for apoptosisare increased

Immunol Lett 2006. 106 (1) :63-71

SurvivorsNon-survivors

A cluster of 28 genesmostly linked with immunosuppression

differentiated S from NS

Septic shock patients > 48 h after the onset of shock31 patients (10 NS) + 7 patients in a prospective control study (3 NS)

HG-U133A oligonucleotide arrays (Affymetrix) – 14 500 genes

Among them, the decreased expression of the fractalkine receptor CX3CR1 mRNA was the most

interesting because of largest fold change between S and NS ( 8-fold decreased in survivors)

CX3CR1 is mainly expressed on patroller monocytes that are the first to reach the site of secondary infections

to initiate immune response.

Due to decreased chemotaxis(and subsequent decreased inflammatory cytokines release),

the loss of this receptor might have a role inthe development of nosocomial infections

Days after the onset of shock

Sur

viva

l dis

trib

utio

n fr

actio

n

CX3CR1 mRNA > 0.12

CX3CR1 mRNA < 0.12

p = 0.0002

Survival curves stratified on CX3CR1 mRNA level at day 1-2 (cut off: 0.12)

Confirmation cohort (N = 160 septic shock patients) : qRT-PCR in whole blood

0,2

0,4

0,6

0,8

1

1,2

HL

A-D

RB

mR

NA

(ra

tio) p < 0.01

Days 1-3 Days 4-10

NS

S

SNS

Septic shock

Cytokines / HLA-DR assessment by qRT-PCR in whole blood (Paxgen tubes)

Abe R et al.

220 patients20 infected20 matched control group23 genes (PAXGENE) : TNF, IL1, IL10, IL18, SOCS3, CD3, CD69, Perforin, Granulysin, CCR3, KLRK1, IDO1, HLA-DRA, CD74,IL8, CXCL10, CCL3, CXCL1, HMOX1, S100A8, PF4, IL6

Transcriptomic:

For now, few studies but promising resultsAdvantages :- mRNA (all components : intracellular, cell surface, soluble mediators => automated systems)- Panel of markers- MIP-Réa (800 patients)

5. Conclusions and perspectives

Representative examples

10

20

30

40

50

60

70

80

1 3 4 5 6 72

HLA

-DR

(%

+ m

onoc

ytes

)

Days post-shock

Survivor / non-infected

Non-survivor / infected

2 patients with septic shock – community acquired No comorbidity – first sample < 6 hours after the onset of shock

Representative examples

10

20

30

40

50

60

70

80

1 3 4 5 6 72

HLA

-DR

(%

+ m

onoc

ytes

)

Days post-shock

Survivor / non-infected

Non-survivor / infected

2 patients with septic shock – community acquired No comorbidity – first sample < 6 hours after the onset of shock

Therapy ?

10 20 30

days

20

40

60

80

100

Per

cent

sur

viva

l (%

)

Sepsis-induced immune failure(modified CLP models)

CLP (+ antibiotherapy and volume resuscitation)Sham

10 20 30

days

20

40

60

80

100

Per

cent

sur

viva

l (%

)

Sepsis-induced immune failure (modified CLP models)

CLP + Infections with germs (Pseudomonas, Legionella, Candida, Aspergillosis)

CLP (+ antibiotherapy and volume resuscitation) + ØSham + ØSham + germs inoculation

10 20 30

days

20

40

60

80

100

Per

cent

sur

viva

l (%

)

Sepsis-induced immune failure (modified CLP models)

Infectionswith germs

Immune stimulation (IL-7, IL-15, FLT3-L, DC, GITR mAbs…)

p < 0.05

Nascimento et al., Crit Care Med 2010Delano et al., J Immunol 2010Brahmamdam et al, J Leuko Biol 2010Unsinger et al. J Immunol 2010Inoue et al., J Immunol 2010Pene et al., J Immunol 2008Benjamin et al., Blood 2005

CLP (+ antibiotherapy and volume resuscitation) + ØSham + ØSham + germs inoculation

Perspectives

Most ICU patients present with severe immune alterations after injury(sepsis, severe burn, trauma, pancreatitis, major surgery)

=>Immune failure should be considered as an important additional organ failure

=> Septic patients who do not recover normal immune functions are those who die

Biomarkers are readily available to monitor pro-/anti-inflammatory responses (- We now need multicentric clinical studies to validate and reinforce these promising preliminary results

- We need to establish standardized measurement protocols for each potential biomarker)

=> biomarker-based stratification=> individualized and targeted therapy

And beyond immuno-inflammation…..

THERANOSTIC

Anti-inflammatory drugs

Anti-TNF stratified on IL-6 levels.Panacek EA et al., Crit Care Med. 2004 Nov;32(11):2173-82

New Strategies:- Close Monitoring (PCT ?) to detect asap the beginning of infections- Preventive antibacterial therapy (when possible / risk of resistance)- Immunotherapy to restore immune functions (pro-inflammatory drugs)

Sepsis-induced immune dysfunctions

Innate Immunity

Monocyte anergy

Adaptive Immunity

Lymphocyte anergy

Immunotherapy to restore immune functions ?

Immunotherapy to restore immune functions ?

Sepsis-induced immune dysfunctions

Innate Immunity

Monocyte anergy

Adaptive Immunity

Lymphocyte anergy

GM-CSFIFN-g

IL-7

Docke WD et al. Nat Med. 1997;3:678-81

Monocyte deactivation in septic patients:restoration by IFN-gamma treatment

9 patients sepsis severe

HLA-DR < 30 %(2 jours de suite)

Interferon-gamma

Mortalité 33 % Mortalité 58 %

26 patients sepsis severe

HLA-DR < 30 %(2 jours de suite)Non randomisée

Nakos G et al., Crit Care Med 2002;30:1488-1494

Immunoparalysis in patients with severe traumaand the effect of inhaled interferon-gamma

52 trauma patientsHLA-DR at day 3 (BAL)

HLA-DR > 30 % (n = 31)Usual monitoring

HLA-DR < 30 % (n = 21)

Placebo (n = 10)

Infection II : 550 %

Interferon-gamma (n = 11)

Infection II : 19 %

p < 0.05

Infection II : 310 %

Pas d’impact sur mortalité

- No side effects- Significantly shorter time of ventilation (148 vs 208 h, p=0.04)- shorter length of intra-hospital stay (59 vs 69 days)- shorter length of ICU stay (41 vs 52)- 28-day mortality not different (trial not powered for mortality)

GM-CSF

Stratified on ex vivo TNF releasebelow 160 pg/mL

All ways may lead to immune suppression

Vincent JL, Crit Care Med 2005 Cavaillon JM & Annane D, J Endotox Res 2007

To be kept in mindfor next years….

Effective early treatmentscannot avoid a risk

of immune suppression

Short term beneficial effectscan be hidden by deleterious effects

of immune suppressionwhen mortality is assessed at 28 days