The subject of the present invention is a method for the early
5 prediction of severe dengue or hemorrhagic dengue using protein
markers.
Over the past 30 years, dengue, a viral disease transmitted by
urban hematophagous mosquitoes of the Aedes genus has worryingly
10 spread throughout the world. It is currently a real public health
problem for more than one hundred countries located in the subtropical
zone, particularly in the Pacific Nest, South America and South-East
Asia zones. The emergence of the disease is largely due to the
population explosion and to chaotic urbanization. Climatic
15 abnormalities also play a not insignificant role.
In this respect, dengue could emerge in the western regions of
the world which until now have been spared the virus. Thus, Aedes
albopictus, one of the vectors of the disease, has recently been found
in the North of Italy and in the South of France. Most recently,
20 autochthonous cases of dengue have been recorded in the South of
France. It is estimated that close to three billion people are exposed
to the risks of dengue. Close to one million hospitalizations are
registered yearly and there have been thousands of deaths. Children
are the main victims of the disease.
25 The dengue virus is a single-stranded, positive-polarity
enveloped RNA virus of the family Flaviviridae. The genome of the
virus (11 000 nucleotides) encodes a polyprotein of approximately
3400 amino acids which undergoes co- and post-translational cleavage
which results in structural proteins (C, prM, E) and non-structural
30 proteins (NS1, NS2A, NS2B, NS3, NS4A, NS4B, NS5). There are 4 viral
serotypes (DVl to DV4), which can coexist in endemic zones. There is
approximately 70% sequence homology between the various serotypes.
Infection by a given serotype confers long-term immunity for this
serotype. Cross-protection lasts only a few months: reinfection is
35 therefore possible with a different serotype.
Infection begins with a bite from a mosquito infected with one of
the dengue viruses. Incubation, the period during which the virus
replicates in the blood without hor,lever giving rise to any symptoms,
generally lasts from 4 to 10 days. The first signs occur after the
incubation period.
In its conventional form ("conventional" dengue fever: DF) ,
dengue is characterized by sudden-onset hyperthermia accompanied by
5 one or more of the follov~ing symptoms: shivering, headaches, joint
and/or muscle pain, nausea and vomiting. A rash may also appear,
generally on around the 5th day of symptoms. This acute febrile stage,
which corresponds to the viremic phase, generally lasts from 3 to 5
days (extremes: 2 to 7 days). More than 95% of cases will have no
10 signs of severe illness and will recover with no complications in
under 7 days.
In 2 to 4% of cases, the patient may develop a critical phase
characterized by a more or less severe plasma leakage syndrome and an
increased hematocrit level, leading to dengue hemorrhagic fever: DHF.
15 This phase typically (but not necessarily) appears at the time of
defervescence, around the 4th or 5th day. It is generally brief (24 to
48 h) but may develop into a severe form characterized by major
hemorrhagic manifestations, a state of shock and/or the failure of one
or more organs. Development into a severe form is most often signalled
20 by one (or more) warning sign (s) , such as :
- fever (temperature of greater than 39°C) after the 5th day;
- intense abdominal pain, persistent diarrhea, uncontrollable
vomiting and complete refusal of food;
- edemas and/or minor effusion;
25 - bleeding of mucous membranes which does not stop
automatically;
- pronounced lethargy or restlessness;
- thronhocytopenia;
- signs of hemoconcentration.
30 In the most severe cases, the leaking of plasma can lead to
deadly hypovolemic shock (Dengue Shock Syndrome: DSS) if the patient
is not rapidly treated. Rare but deadly hepatic and neurological
involvement is also associated with the severity of the disease. The
mortality rate, which is variable according to epidemics, can reach 5%
35 of established DHF cases. This rate can increase up to 20% without
hospital care or appropriate treatments.
To simplify, these severe cases will be referred to as severe
dengue in the remainder of the description, as opposed to conventional
dengue, DF.
90% of cases of severe dengue occur during secondary infection
5 with a heterologous serotype, and 10% during primary infection,
generally in infants aged from 6 months to 1 year. There are several
factors which influence the severity of the infection, such as the
factors of the host, serotype and genotype of the virus, the order and
time between successive infecting viruses, the quality and quantity of
10 cross-reactive antibodies and the CD4/CD8 response. Studies have shown
a correlation between viral load and severity of the disease. The
exact causes of the occurrence of severe dengue are, however, still
not known. Up until now, no specific determining factor for virulence
has been demonstrated. Furthermore, since there is no vaccine against
15 the dengue virus, the only treatments available are symptomatic
treatments. Consequently, it is important to be able to monitor
epidemics and to predict severe cases for appropriate hospital care.
The methods currently used to diagnose dengue do not make it
possible to predict the development of severe dengue. At the very most,
20 serological methods make it possible to distinguish between primary
and secondary infections and molecular methods make it possible to
detect the virus and to carry out serotyping [I, 2, 3, 41.
The present invention provides a solution to the problems
presented above by means of a method which allows both early and
25 specific detection of proteins in a blood sample making it possible to
predict patients developing severe dengue. Indeed, the inventors found,
surprisingly, that proteins from the host were expressed more or- less
abundantly (overexpressed/uriderexpressed) in cases of patients
developing severe dengue, compared to the amount or expression thereof
30 in cases of patients remaining with conventional dengue (that is to
say not developing severe dengue) in blood samples consisting, for
example, of plasma. Most particularly, they have demonstrated for the
first time and completely unexpectedly that olfactomedin 4 (OLFM4) is
overexpressed in the case of patients developing severe dengue and
35 thus constitutes a marker for predicting severe dengue.
Thus, a subject of the present invention is a method for
predicting, in vitro, the probability of a patient developing severe
dengue, based on a hlood sample, wherein:
a) the quantity in said blood sample of at least one marker, which
is olfactomedin 4, is determined,
b) the quantity of olfactomedin 4 determined in step a) is compared
with a reference quantity of said marker obtained from a group of
individuals who have been diagnosed with non-severe dengue, wherein,
if the quantity of olfactomedin 4 determined in step a) is greater
than the reference quantity established in step b), it is predicted
that the patient will develop severe dengue.
According to the method of the invention, it is also possible to
10 determine, in step a), the quantity in the blood sample of at least
one other marker chosen from platelet factor 4 and a2-macroglobulin or
the respective quantities of the two markers and, in step b), the
quantity of the marker or of the two markers of step a) is compared
with a reference quantity obtained from a group of individuals who
15 have been diagnosed with non-severe dengue and, if the quantity of
platelet factor 4 and/or a2-macroglobulin determined in step a) is
less than the reference quantity established in step b), it is
determined that the patient will develop severe dengue.
The invention also relates to a kit for the in vitro prediction of
20 severe dengue, comprising:
- a binding partner for olfactomedin 4,
- a binding partner for the dengue virus NS1 protein.
The kit may also comprise a binding partner for platelet factor 4
(PF4) and/or a binding partner for a2-macroglobulin (A2M).
25
Definitions
The term "blood sample" is intended to mean k~h0le blood, serum
and plasma.
The term "group of individuals who have been diagnosed with non-
30 severe dengue", used to determine the reference quantity of the marker
of interest, is intended of course to mean that the group of
individuals has not developed severe dengue. Thus, in step b) the
quantity of platelet factor 4 determined in step a) is compared with a
reference quantity of said marker obtained from a group of individuals
35 who have been diagnosed with dengue without having developed severe
dengue.
The term "binding partner" is intended to mean, for example,
receptors, antibodies, antibody fragments, antibody analogs and any
other ligand capable of binding to a protein.
The binding-partner antibodies are, for example, either polyclonal
5 antibodies or monoclonal antibodies.
The polyclonal antibodies may be obtained by immunization of an animal
with the appropriate immunogen, follovred by recovery of the desired
antibodies in purified form, by taking the serum of said animal, and
separation of said antibodies from the other serum constituents,
10 especially by affinity chromatography on a column to which is bound an
antigen specifically recognized by the antibodies.
The monoclonal antibodies may be obtained by the hybridoma
technique, the general principle of which is summarized below.
Firstly, an animal, generally a mouse, is immunized with the
15 appropriate immunogen, and the B lymphocytes of said animal are then
capable of producing antibodies against this antigen. These antibodyproducing
lymphocytes are then fused with "immortal" myeloma cells
(murine in the example) so as to give rise to hybridomas. Using the
heterogeneous mixture of cells thus obtained, a selection of the cells
20 capable of producing a particular antibody and of multiplying
indefinitely is then carried out. Each hybridoma is multiplied in the
form of a clone, each resulting in the production of a monoclonal
antibody of which the recognition properties with respect to the
protein may be tested, for example, by ELISA, by one-dimensional or
25 two-dimensional Western blotting, by immunofluorescence, or by means
of a biosensor. The monoclonal antibodies thus selected are
subsequently purified, especially according to the affinity
chromatography technique described above.
The monoclonal antibodies may also be recombinant antibodies
30 obtained by genetic engineering, using techniques well known to those
skilled in the art.
The term "antibody analogs" is intended to mean biological and/or
chemical compounds which have the same binding capacities as the
antibodies or antibody fragments or similar binding capacities. In
35 particular, the antibody analogs include small proteins which, like
antibodies, are capable of binding to a biological target thus making
it possible to detect it, to capture it or quite simply to target it
within an organism or within a biological sample. The fields of
application of these antibody analogs are virtually as vast as those
of antibodies. By way of example, mention may be made of the
ano of it ins^"', which are small proteins sold by the company Affilogic.
The binding partners specific for the desired protein can be used
5 as a capture reagent, as a detection reagent or as capture and
detection reagents.
The visualization of the immunological reactions, i.e. the
protein/binding partner binding, can be carried out by any means of
detection, via labeling, of the binding partner.
10 The term "labeling" is intended to mean the binding of a label
reagent capable of generating a detectable signal, i .e. a compound, a
substance or a particle which can be detected by visual, fluorescent
or instrumental means.
A nonlimiting list of these label reagents consists of:
15 - metal or alloy particles, such as colloidal gold particles,
polymer particles, such as colored latex particles,
- magnetic particles,
- fluorescent molecules,
- chemoluminescent molecules.
20 By way of example of immunological tests as defined above, mention may
be made of "sandwich" and "competition" methods.
Figures :
The majority of the figures illustrate the validation of the results
25 by a quantitative ELISA assay carried out on individual samples taken
from patients during the acute febrile phase of the disease, before
defervescence. The patients having remained with conventional dengue
are denoted DF and the patients having then developed severe dengue
are denoted SevD. In all cases, the reading is carried out at an
30 optical density (OD) of 450 nm. The results were obtained on samples
with different geographical origins: Columbia and Cambodia. On the
graphs obtained (Graphpad Prism software, V 4 . 0 3 ) , the median
calculated is represented by a horizontal line. The box illustrates
the values encompassing 50% of the individuals. The maximum and
35 minimum values are also illustrated. The values taken into account
correspond to the mean of two independent tests carried out in
duplicate.
Figure 1 illustrates the presence of virus in the fractions purified
from plasma from patients but not in the control (C), demonstrated
with Western blotting using a monoclonal antibody directed against the
viral protein E.
5 Figure 2 illustrates the results obtained for the quantitative
assaying by means of an ELISA assay of the OLFM4 marker on plasma
samples from Colombian patients.
Figure 3 illustrates the results obtained for the quantitative
assaying by means of an ELISA assay of the OLFM4 marker on plasma
10 samples from Cambodian patients.
Figure 4 illustrates the results obtained for the quantitative
assaying by means of an ELISA assay of the PF4 marker on plasma
samples from Colombian patients.
Figure 5 illustrates the results obtained for the quantitative
15 assaying by means of an ELISA assay of the PF4 marker on plasma
samples from Cambodian patients.
Figure 6 illustrates the results obtained for the quantitative
assaying by means of an ELISA assay of the A2M marker on plasma
samples from Cambodian patients.
20
Example 1: Characterization of the samples
15 Colombian plasma samples positive for dengue were selected, from
which 8 originate from patients remaining with conventional dengue
without developing severe dengue (patients/samples referred to
25 hereinafter as DF or conventional dengue) and 7 originate from
patients having then developed severe dengue (patients/samples
referred to hereinafter as SevD or severe dengue). The various plasmas
were grouped together, composing respectively a pool of conventional
dengue DF plasma and a pool of severe dengue SevD plasma for those
30 having developed severe dengue. All the plasmas were taken during the
acute febrile phase of the disease, before the critical phase, from
patients having a secondary infection. The serotypes concerned were
serotypes 1, 2 and 3. All the patients having developed severe dengue
were hospitalized and had signs of hemorrhaging. No comorbidity was
35 reported [51. All the plasmas were verified as being NS1-positive
(Platelia dengue kit, Bio-rad) and the viral load was also verified by
Q-RT-PCR with a commercially available kit (PrimerDesign) following
supplier instructions: the mean number of viral RNA copies was
estimated at 4 x 10' and 4.1 x lo7 for the conventional dengue (DF)
pools and severe dengue (SevD) pools, respectively. The pools composed
correspond to a volume of approximately 2 ml of plasma. Before
purification, the plasma mixtures were centrifuged for 5 mins at 1000
5 x g and at 4°C so as to remove the impurities present in the sample
and to obtain clarified samples.
The plasma selection criteria are described in table 1 below. The
samples were taken after appearance of symptoms. .
Table 1
10 These plasma sample pools were then purified to obtain virus-enriched
fractions as described below.
Example 2: Purification of the samples
All the steps are carried out at 4°C. The clarified samples are
15 supplemented with 8 ml of cold pH8 PBS (PBS8), then centrifuged for 2h
at 41 000 x g in an Optima L90 ultracentrifuge (Beckman). The rotor
used is the SW41 rotor (Beckman). After centrifugation, the
supernatant is removed and the viral pellet obtained is resuspended in
200 microliters of PBS8 then loaded onto a discontinuous gradient
20 cornposed of 5 ml of 60% (w/w) sucrose in PBS8 and 5 ml of 20% (w/w)
sucrose in PBS8. After renewed centrifugation for 2 h at 41 000 x g, a
virion-enriched ring located at the interface between the two sucrose
solutions is taken off with a pipette, diluted 10 times with PBS8 and
finally centrifuged one last time for 2 h at 41 000 x g. The pellet
25 obtained is resuspended in 200 microliters of PBS8.
This resuspension is then purified using an insoluble polyelectrolyte,
Viraffinity (BioSupportGroup, USA). For this purpose, 200 microliters
of an MN buffer (60 mM ME'S pH 6.5, 150 mM NaC1) are added to the viral
suspension along with 100 microliters of Viraffinity. The mixture is
incubated for 5 min at room temperature then centrifuged for 10 min at
1000 x g, following supplier instructions. The supernatant is removed
and the polymer pellet is rinsed 3 times with 200 microliters of MN
5 buffer. The viral proteins are recovered by heating the polymer for 5
min / 70°C in the presence of 50 microliters of a buffer containing
SDS (Novex InVitrogen) then centrifugation for 5 min at 1000 x g.
The presence of the virus in the final .samples was verified by
immunoblotting with a monoclonal antibody directed against the
10 envelope protein of the dengue virus (E protein). As illustrated in
figure 1, strong signals at 60 KDa and 120 KDa, corresponding
respectively to the monomeric and dimeric forms of the envelope
protein, are specifically detected by the monoclonal antibody in the
plasma pools. On the contrary, the envelope protein was not detected
15 on a control corresponding to a pool of healthy (non-dengue) plasma
purified in the same way as has been described above.
Example 3: Identification of the specific proteins for each plasma
pool, conventional dengue DF and severe dengue SevD, by mass
20 spectrometry (MS)
Method:
The viral preparations and the controd sample obtained according to
example 2 are deposited on a non-denaturing polyacrylamide gel and
migrated until the proteins penetrate into the gel, in order to
25 desalify the sample. The band containing the proteins is excised
manually then washed three times in a buffer containing 50%
acetonitrile then finally dried in 100% acetonitrile. The gel is then
rehydrated in a 7% HZ02 solution before being washed again. A solution
of trypsin diluted in 25 mM NH4HC03 is then added for hydrolysis at
30 37°C overnight. The peptides thus obtained are extracted by 15 minute
sequential extractions with 30 microliters of 50% acetonitrile, 30
microliters of 5% formic acid and 30 microliters of 100% acetonitrile.
These sequential extractions are mixed, dried under vacuum and
resuspended in a solution containing 5% acetonitrile and 0.1%
35 trifluoroacetic acid. After quantification of the samples, a defined
quantity of peptides is analyzed by nano liquid chromatography coupled
together with mass spectrometry (Ultimate 3000, Dionex and LTQ--
Orbitrap VelosPro, Thermo Fisher Scientific). The results are acquired
by virtue of the Xcalibur software (Thermo Fisher) and automatically
converted by the Mascot Daemon V2.2 software (Matrix Science).
Searching is then carried out on the Swissprot and Trembl databases
5 via Mascot 2.2. Each experiment was carried out twice, independently.
The proteins were identified by the EDyP Service laboratory (CEA
Grenoble, France) .
10
Results :
The viral envelope protein E was repeatedly identified in the samples
containing virus. The predominantly identified peptide sequence is
GWGNGCGLLFKG. This result confirms the presence of the virus in the
15 purified fraction.
For the proteins of cellular origin, identified by proteomics on
purified plasma pools, the results obtained are summarized in tables
2a and 2b. In these tables, only those proteins having a variance of
less than 25% for the number of peptides found from one experiment to
20 the other have been considered. Similarly, for the severe dengue
sample, a number of peptides of greater than 2 was required. According
to these criteria, 189 proteins were finally selected. These proteins
are described in tables 2a and 2b below. A ratio of "number of
peptides in severe dengue (SevD) sample" / "number of peptides in
25 conventional dengue (DF) sample" (SevD/DF) could be calculated for the
majority of these proteins (cf. table 2a). Some proteins were only
identified in the SevD sample (cf. table 2b); in this case, the
SevD/DF ratio could not be calculated.
30 Table 2a
Accession
number
PO9871
PO7225
PO1008
Q16610
Protein name
Mean
number
of
peptides
DF
e s t e r a s e )
Mean
number
of
peptides
SevD
Vitamin K-dependent p r o t e i n S
Antithrolnbin-111 (ATIII) (Serpin C1)
E x t r a c e l l u l a r matrix p r o t e i n 1
(Secretory component p85)
SevD/DF
peptides
ratio
1
3
1
4
9 . 5
3
4.00
3.17
3.00

PO2649 Apolipoprotein E - (Apo-E) --4
Alpha-1-antitrypsin (Alpha-1
PO1009 - protease inhibitor) (Serpin Al) - 8.5
Galectin-3-binding protein (Basement
Q08380 membrane autoantigsn p105) 12
Putative uncharacterized protein
Q6N092 DKFZp686K18196 (Fragment) 12
4.5
9.5
13
13
1.13 .
1.12
1.08
1.08

P u t a t i v e uncharac
ICO In-
P16452
PI1277
PO2549
4 . 2 (P4.2) -5.5 -1 -
3
1.5
S p e c t r i n b e t a chain, e r y t h r o c y t e
(Beta-I s p e c t r i n )
S p e c t r i n alpha chain, e r y t h r o c y t e
(Erythroid a l p h a - s p e c t r i n )
0.18
0.13
0.05
23
33
Table 2b
Accession
5 Example 4: Confirmation ELISA
Method:
So as to confirm the mass spectrometry results, specific quantitative
ELISAs were carried out in duplicate on individual plasmas. The
proteins selected and tested, from those identified in tables 2a/2b,
10 are those with a severe dengue (SevD) / conventional dengue (DF) ratio
of greater than or equal to 1.33 and less than or equal to 0.75 with a
mean number of peptides of greater than 1 for each sample and a
potential link to dengue pathogenesis. This first screening made it
possible to only assay those proteins most of interest. According to
15 these criteria, the following proteins were selected:
ceruloplasmin,
protein S,
properdin complement factor,
secretory component p85,
- complement Clr protein,
complement Cls protein,
angiotensin,
factor 11,
. CFB,
anti-factor VIII,
serum amyloid P-component,
- olfactomedin 4 (OLFM4),
thrombospondin,
platelet factor 4 ( P F 4 ) ,
. complement Clq protein,
moesine, and
- complement C8 protein.
Multimerin-1, apolipoprotein B-100 and von Willebrand factor were also
assayed.
It should be noted that these proteins are predominantly elements of
15 the coagulation pathway or the complement cascade.
These ELISAs were carried out by virtue of commercially available kits
(USCN, China), following supplier instructions. Statistical analyses
(Mann-Whitney test and ROC/AUC curve) were carried out by means of
GraphPad Prism V4.03 software.
20 Each candidate marker was assayed on individual plasma samples. These
samples are plasma samples taken during the acute febrile phase of the
disease (viremic phase), these samples either originating from
patients having remained with conventional dengue DF, without
developing severe dengue, or from patients having developed severe
25 dengue SevD. All the patients had secondary dengue. Only serotypes 1,
2 and 3 were represented (no serotype 4). These samples originated
from 1 'Universidad Industrial de Santander (Bucaramanga, Colombia) [ 5 ]
or from the Institut Pasteur in Cambodia (Phnom-Penh) . The latter were
part of a prospective study carried out in agreement with the local
30 ethics committee. The characteristics of the two sampling sources are
given in tables 3 and 4 belori. The samples were collected after
appearance of symptoms.
Table 3: Colombian plasma samples
Table 4: Cambodian plasma simples
ns: non-significant p value
-Resul~t s : -
For the majority of the ELISA-assayed markers, no difference in plasma
concentration was observed between the DF and SevD plasmas, whether
5 Colombian or Cambodian (p>0. 1) .
On the other hand, for two markers, the results make it possible to
clearly distinguish those patients who then developed severe dengue
SevD from those who remained solely with conventional dengue without
developing severe dengue. The first marker is OLFM4 (olfactomedin 4).
10 For the Colombian samples, the plasma concentration of the marker is
higher in the SevD samples compared to the DF samples (p=0.07; cf.
figure 2) .
This is confirmed on the Cambodian samples with an extremely
significant difference in concentration (p<0.0003) and a median that
15 is more than twice as high for the SevD samples compared to the Dl?
samples (figure 3). The AUC is 0.858 (95% CI = 0.7307-0.985) . The ROC
curve made it possible to determine the best specificity for a
sensitivity close to 100%. The results are summarized in table 5: for
this marker and for a sensitivity close to 95%' a specificity of
20 greater than 72% is reached.
Table 5
The second marker is PF4 (platelet factor 4)
For a sensitivity of:
For the Colombian samples, a difference in plasma concentration in
25 favor of the DF samples is observed (p<0.001) (figure 4). The AUC is
0.88 (95% CI: 0.7305-1).
The best specificity i s :
This is confirmed for the Cambodian samples for which there is a
significant difference in plasma concentration in favor of the DF
samples (p< 0.0001) (figure 5). The AUC is 0.94 (95% CI = 0.87-1). The
30 ROC curve made it possible to determine the best specificity for a
sensitivity close to 100%. The results are summarized in table 6 below:
for this marker and for a sensitivity of 95%, a specificity of close
to 78% is reached.
5 Table 6
In parallel, another marker, a-2 macroglobulin (A2M) was identified
from unpurified Cambodian plasma samples by a SILAC-type differential
proteomic method (Stable Isotope Labelling by Aminoacids in cell
10 Culture) [6]. The identification of this third marker is described in
the following examples.
-
PF4
Example 5: Characterization of the samples
The composition of each plasma pool or group used in this experiment
15 is summarized in table 7. All the Cambodian plasmas selected to
compose the pools were taken during the acute febrile phase of the
disease, before the critical phase, from patients having a secondary
infection. The serotype concerned was serotype 1. All the SevD
patients were hospitalized and had signs of hemorrhaging. No
20 comorbidity was reported. All the plasmas were verified as being NS1-
positive (Platelia dengue kit, Bio-rad) and the viral load was also
verified by Q-RT-PCR with a commercially available kit following
supplier instructions. The pools composed correspond to a final volume
of approximately 2 ml of plasma. The plasma groups are inactivated
25 beforehand with heat (56"C/20 minutes) then preclarified by
centrifugation for 5 mins at 1000 x q and at 4"C, so as to remove
impurities present in the sample.
Table 7
For a sensitivity of:
100
9 5
The best specificity is:
6 1
77.9
-
Serotype
> --
DF (n=6)
DVl
-
SevD (n=6)
-
DV1
Secondary dengue YES
I I
Age in years (mean)
NS1 positive/virus positive
Severity grade 3-4
-
Comorbidity NO NO
I I
Example 6: Differential proteome analysis
The method used is a semi-quantitative proteomic method of SILAC type
5 (Stable Isotope Labelling by aninodeids in cell Culture) [6] developed
by Pronota (Ghent, Belgium) using the MA~~termind~p"la tform and
carried out on conventional dengue DF or severe dengue SevD plasma
groups. Each group is composed of a mixture of 6 samples, as detailed
in example 5.
10 These plasma mixtures have been depleted beforehand in the 14 most
abundant plasma proteins by affinity chromatography. The quantity of
proteins recovered in the end was obtained by a colorimetric assay
based on bicinchoninic acid (BCA assay Thermo Fisher Scientific Inc.,
USA) .
15 The ~~ssterminds~t"u dy compared each sample to a reference sample
which groups together all the samples. This method provides
information on the relative levels, and presence or absence, of
peptides/proteins in the severe dengue SevD samples compared to the
conventional dengue DF samples. The differential analysis is carried
20 out by mixing the samples labelled with different isotopes and by
analyzing, by mass spectrometry, each matched peak. The isotope label
is introduced by tryptic hydrolysis which incorporates 2 '0 atoms
("heavy" labeling) onto the C-terminal arginine of a peptide, which
leads to a mass difference of 4 daltons to the same peptide labeled
25 with 160 ("light" labeling). The reference sample is labeled with 160,
whereas the individual samples are labeled with *'0. The MS/MS data are
then submitted to the MASCOT software for identification of the
peptides and proteins in each sample.
6-12 (8.6) 1
YES/YES 1
Mean day of sampling
6-8 (7)
YES/YES
4 3.85
Following MS/MS analysis, more than 250 quantifiable proteins were
identifiable, 10 proteins of which had at least 1 peptide found to be
differential. For each protein identified, the SevD/DF ratio is
calculated as the weighted mean of the coefficients of all the
5 peptides identified for the given protein. Overall, the results showed
a high degree of similarity between the two proteomes and only a few
proteins were found to be expressed differentially. For three of these
ten proteins, the peptides identified are systematically expressed
differentially and have a mean SevD/DF ratio which deviates from 1
10 (see the results given in the following table 8). The three proteins
identified are: a-2 macroglobulin (AZM), complement C3f and heparin
cofactor 2 . These proteins are predominantly elements of the
coagulation pathway or the complement cascade.
15 Table 8
Proteins
a-2 macroglobulin
I I I
Example 7 : Confirmation ELISA
Method:
Accession
number
A2MG - HUMAN
Complement C3f
I I I
20 So as to confirm the mass spectrometry results, specific quantitative
ELISAs were carried out in duplicate on individual plasmas. The
proteins assayed are those identified in example 6: a-2 macroglobulin
(A2M), complement C3 protein, and heparin cofactor 2.
Heparin cofactor 2
These ELISAs were carried out by virtue of commercially available kits
25 (USCN, China), following supplier instructions. Statistical analyses
(Mann-Whitney test and ROC/AUC curve) were carried out by means of
GraphPad Prism V4.03 software.
SevD/DF ratio
---
0.33
C03 - HUMAN
Each candidate marker was assayed on individual samples. These sanlples
were plasma samplcs taken from patienLs during the acute febrile phase
95% CI on
mean of
ratios
0.25-0.44
HEP2 - HUMAN
0.89 0.68-1.17
2.25 1.71-2.95
of the disease (viremic phase). Clinical follo~g-up of the patients
showed that some finally remained with conventional dengue DF without
developing severe dengue, whereas others had developed severe dengue
SevD. All the patients had secondary dengue. Only serotype 1 was
5 represented. These samples originated from the Institut Pasteur in
Cambodia (Phnom-Penh) and were part of a prospective study carried out
in agreement with the local ethics committee. The characteristics of
the sampling source are given in table 9.
Table 9
DF (n=23)
AST (IU/L)
ALT (IU/L) -
Viral load (copies/ml)
SD: standard deviation
ns: non-significant p value
15
Results :
SevD (n=26)
7.3t-0.45
35.6%
3.2
18.86+-0.98
2.37t-0.12
0.31+-0.02
2.8+-0.2
Mean age +- - SD (years)
% male
Sampling (day)
Mean weight +- SD (kg)
Total cholesterol (mmol/l)
23.1-8 0.0043
66.9+-13.4 ns
7.1 lo8+-6 - 2.06 lo0+-2 ns
18/26
Positive tourniquet test
HGB
P value
ns
ns
ns
ns
10.0001
<0.0001
0.1).
However, for the Cambodian samples ELISA-assayed for A2M, there was a
5 significant difference in concentration (p<0.0004) and a median
approximately twice as high for the DF samples compared to the SevD
samples (figure 6). The AUC was 0.89 (95% CI = 0.75-1).
For A2M, the ROC curve made it possible to determine the best
specificity for a sensitivity close to 100%. The results are
10 summarized in table 10: for a sensitivity close to 94%, a specificity
of greater than 83% is reached.
Table 10
A2M
93.8 83.3
For a sensitivity of:
100
The best -specificity is:
7 2
LITERATURE REFERENCES
1. SB Halstead . The lancet 2007; 370: 1644-52
2. AS Leong et al. Semin.Diagn.Patho1. 2007; 24(4):227-236
5 3. K. Clyde et al. J. Virol. 2006; 23: 11418-11431
4. Fields Virology, Fifth edition, Knipe DM ed., LWW
5. Villar-Centeno LA et: al. Am. J.Trp.Med.Hyg. 2008; 78: 370-374 6. R.
Drissi et al. FEBS J. 2013

CLAIMS
1. A method for determining, in vitro, the probability of a patient
5 developing severe dengue, based on a blood sample, wherein:
a) the quantity in said blood sample of at least one marker, which
is olfactomedin 4, is determined,
b) the quantity of olfactomedin 4 determined in step a) is compared
with a reference quantity of said marker obtained from a group of
individuals who have been diagnosed with non-severe dengue,
wherein, if the quantity of olfactomedin 4 determined in step a) is
greater than the reference quantity established in step b), it is
determined that the patient will develop severe dengue.
15 2. The method as claimed in claim 1, wherein, in step a), the quantity
of at least one other marker chosen from platelet factor 4 and a2-
macroglobulin is determined in the blood sample and, in step b), the
quantity of said other marker of step a) is compared with a reference
quantity of said other marker obtained from a group of individuals who
20 have been diagnosed with non-severe dengue and, if the quantity of
said at least one other marker determined in step a) is less than the
reference quantity established in step b), it is determined that the
patient will develop severe dengue.
25 3. The method as claimed in claim 1, wherein, in step a), the quantity
of at least two other markers chosen from platelet factor 4 and a2-
macroglobulin is determjned in the blood sample and, in step b), the
quantity of each of the two other markers of step a) is compared with
a reference quantity for each of said other markers obtained from a
30 group of individuals who have been diagnosed with non-severe dengue
and, if the quantity of each of the other markers determined in step a)
is less than the reference quantity of each of the markers established
in step b), it is predicted that the patient will develop severe
dengue.
4. The use of olfactomedin 4 as protein marker for determining, in
vitro, the probability of a patient developing severe dengue, based on
a blood sample.
5 5. The use as claimed in claim 4, also comprising the use of platelet
factor 4 as protein marker for determining, in vitro, the probability
of a patient developing severe dengue, based on a blood sample.
6. The use as claimed in claim 4 or 5, also comprising the use of a2-
10 macroglobulin as protein marker for determining, in vitro, the
probability of a patient developing 'severe dengue, based on a blood
sample.
i
7. A kit for determining, in vitko, the probability of a patient
15 developing severe, comprising a binding partner for olfactomedin 4 and
a binding partner for the dengue virus NS1 protein, said binding
partners being chosen from a receptor, an antibody, an antibody
fragment, an antibody analog or any other ligand.
20 8. A kit for determining, in vitro, the probability of a patient
developing severe, comgrising a binding partner for olfactomedin 4 and
a binding partner for platelet factor 4, said binding partners being
chosen from a receptor, an antibody, an antibody fragment, an antibody
analog or any other ligand.
25
9. A kit for determining, in vitro, the probability of a patient
developing severe, comprising a binding partner for olfactomedin 4 and
a binding partner for a2-macroglobulin, said binding partners being
chosen from a receptor, an antibody, an antibody fragment, an antibody
30 analog or any other ligand.
10. The kit as claimed in claim 9, also comprising a binding partner
for platelet factor 4.
35 11. The kit as claimed in claim 8, 9 or 10, also comprising a binding
partner for the dengue virus NS1 protein.