BIOCHEMICAL MARKERS OF CEREBRAL
MALARIA
Cerebral malaria (CM) forms part of
the spectrum of severe malaria, with a case fatality rate ranging from 15% in
adults in southeast Asia to 8.5% in children in Africa. Clinical signs of
acidosis carry a higher risk of death but nevertheless CM accounts for a
significant proportion of malaria mortality, as well as the potential for
neurological deficits in survivors. The standard clinical definition of CM
centers on a state of unarousable coma partnered with the presence of malaria
infected red blood cells in the peripheral circulation and a lack of other
potential causes of coma such as other infections or hypoglycemia. More
recently, ophthalmic observations of retinopathy have been added to this
definition in both adults and children to increase the specificity of the
clinical diagnosis. Most observations of the pathophysiology of disease come
from postmortem observations of Plasmodium falciparum (Pf) infections, which
are thought to account for the vast majority of CM cases, and show a common
feature of vascular sequestration of infected erythrocytes (IE) in the brain.
Plasmodium falciparum can cause a diffuse encephalopathy known as cerebral
malaria (CM), a major contributor to malaria associated mortality. Despite
treatment, mortality due to CM can be as high as 30% while 10% of survivors of
the disease may experience short- and long-term neurological complications. The
pathogenesis of CM and other forms of severe malaria is multi-factorial and
appear to involve cytokine and chemokine homeostasis, inflammation and vascular
injury/repair. Identification of prognostic markers that can predict CM
severity will enable development of better intervention.
Biomarkers have been used to diagnose
and prognosticate the progress and outcome of many chronic diseases such as
neoplastic and non communicable diseases. However, only recently did the field
of malaria research move in the direction of actively identifying biomarkers
that can accurately discriminate the severe forms of malaria. Malaria continues
to be a deadly disease, killing close to a million people (mostly children)
every year. One life-threatening complication of malaria is cerebral malaria
(CM). Studies carried out in Africa have demonstrated that even with the best
treatment, as high as 15–30% of CM patients die and about 10–24% of CM
survivors suffer short-or long-term neurological impairment. The transition
from mild malaria to CM can be sudden and requires immediate intervention.
Currently, there is no biological test available to confirm the diagnosis of CM
and its complications. It is hoped that development of biomarkers to identify
CM patients and potential risk for adverse outcomes would greatly enhance
better intervention and clinical management to improve the outcomes. We review
here what is currently known regarding biomarkers for CM outcomes.
Biomarkers in disease management
A biomarker is a substance or a
characteristic that can be objectively measured and evaluated as an indicator of
normal biologic processes, pathogenic processes
or responses to a therapeutic
intervention. Biomarkers have been used to diagnose and prognosticate the progress
and outcome of many diseases. They can be used to support disease management
and control in many different aspects:
1) to diagnose a disease,
2) to identify “at risk” individuals,
3) to stratify patients depending on disease severity,
4) provide prognosis of the disease,
5) to assess disease severity,
6) to provide some guidance in the treatment and management
of a disease and eventually
7) to identify at risk patients for
long term complication after the manifestation of
a particular illness.
Biomarkers can be found in any
biological fluids such as serum, plasma, urine, cells or they can be biological
products such as metabolites, cytokines or genetic markers. An example of an
ideal biomarker is the measurement of hemoglobin as an indicator of anemia.
This is a simple, quantitative and inexpensive procedure, therefore allowing
for wide usage. Biomarkers have become valuable for the diagnosis and prognosis
of many chronic diseases such as cancer, diabetes, autoimmune diseases and
HIV/AIDS. However, the use of biomarkers in parasitic infectious diseases is
limited. Biomarkers have been used to diagnose and prognosticate the progress
and outcome of many chronic and acute infectious, metabolic and non
communicable diseases. The transition from MM to the severe
forms of malaria can be sudden and
requires immediate intervention. Therefore, the use of biomarkers to
risk-stratify severe malaria patients would greatly enhance patient care and
assist in appropriate management of health care resources. In addition,
biomarkers that identify asymptomatic cases, who might have levels of
parasitemia undetectable by light microscopy or other conventional testing
methods, will be crucial for monitoring the elimination of malaria reservoirs
from endemic populations. Therefore, biomarkers for malaria management are
clearly a necessary tool for malaria control programs.
Proposed
biomarkers for cerebral malaria
Phase 1 of biomarker identification starts with some understanding
of the pathogenesis of the disease and an exploratory phase in which different
factors that can clearly discriminate different levels of disease severity are
identified. The pathogenesis of CM is still not well understood although it is
clearly multifactorial,
involving sequestration of P. falciparum infected erythrocytes to brain
vasculature, triggering inflammatory cytokine responses and apoptotic pathway
leading to a breach and dysfunction of the blood brain barrier, tissue damage
and repair. In an attempt to understand the pathogenesis of CM, our previous
studies in Ghana and India for the first time revealed a striking association
between the chemokine interferon inducible protein-10 (CXCL10-) and CM
severity, suggesting that CXCL10 may be a biomarker for CM severity. We believe
that the management of CM will require at least three main categories of
biomarkers: early biomarkers, diagnostic biomarkers and prognostic biomarkers.
Early/predictive
biomarkers
Only a subset of malaria patients develops CM with the rest
developing asymptomatic or mild forms of the disease, or other forms of SM such
as SMA. Risk factors for CM include age less than 10 years and living in
malaria endemic regions. However, it is not known if other unknown risk factors
exist that would facilitate the identification of “at risk” patients during
hospital visits before they succumb to CM. Early/predictive biomarkers would
allow heath care workers to stratify febrile patients into those at risk for CM
and those who are likely to develop MM or other forms of SM. The “at risk”
patients would then be started early on the appropriate treatment and on any
adjunctive measures available or referred to better health facilities as required.
To date, no such biomarkers exist and clearly more studies are required in this
area.
Diagnostic
biomarkers
Diagnostic biomarkers are biomarkers that can be used to
categorically identify CM patients. Currently, the diagnosis of CM relies on
clinical indicators characterized mainly by
(1) unarousable coma (no
localizing response to pain) that persists for more than six hours after
experiencing a generalized convulsion;
(2) Presence of asexual
forms of P. falciparum;
and
(3) Exclusion of other
causes of encephalopathy (e.g. viral, bacterial).
Other clinical algorithms and indicators have been investigated.
One of these is malaria retinopathy, whose usefulness has been tested in
different parts of Africa and India. Many of these studies have shown malaria
retinopathy to be a good diagnostic indicator capable of distinguishing CM
patients from patients with other causes of coma in parasitemic comatose
patients. In addition, several serological factors have been identified that
are differentially expressed in CM patients compared to either MM patients or
healthy controls. Our group and others have demonstrated that distinct profiles
of cytokine and/or chemokine are associated with discrete clinical
manifestation of falciparum
malaria. For example, increased levels of CXCL10, sFas, sTNF-R2,
IL-8, IL-1ra, and decreased levels of RANTES and vascularendothelial growth
factor (VEGF) were all found in CM patients and not in MM or health control
cases. The endothelial regulators, angiopoietin-1 (ANG-1) and ANG-2 were also
shown to be differentially expressed in CM compared to MM patients. Some of
these discriminatory factors have also been observed in cerebral spinal fluid
(CSF) samples. Our studies in Ghana demonstrated that higher levels of CSF
CXCL10 and lower levels of VEGF were found in CM patients compared toMMor
healthy controls. Many of the factors discussed above can be viewed as being in
the phase 1 of biomarker development during which preclinical exploration of
potential biomarkers is performed andmarkers capable of discriminating the severity
of the disease are identified. Perhaps the only serological factors analyzed
for their predictive value for CM diagnosis, to date, are the endothelial
regulators, ANG-1 and ANG-2 and the chemokine CXCL10. Angiopoietins are known
to regulate the maintenance of vascular integrity and were found to be differentially
regulated in two different populations of CM patients from Thailand and Africa.
The levels of ANG-1 were significantly reduced in CM patients compared to
healthy controls and MM patients and, conversely, ANG-2 levels were
significantly elevated in CM patients compared to control groups. Receiver Operating
Curve (ROC) analyses were performed to test the sensitivity and specificity of
these factors, as discriminatory biomarkers for the accurate diagnosis of CM.
ANG-1, ANG-2 and the ratio of ANG-2/ANG-1 were all shown to accurately
discriminate between CM and MM patients. Indeed, ANG-1 levels were
independently associated with CM even in a multivariate logistic regression
model. In the same study, TNF levels were shown to be significantly increased
in CM compared to MM patients in the Thai adult population, but this factor did
not discriminate well in the Uganda children. This finding demonstrates the
importance of testing the utility of any potential biomarker using different
populations and different age groups as the sensitivity and specificity might
be affected by the transmission intensity and host immunity.
4.3.
Prognostic biomarkers
More often than not, appropriate treatment of CM leads to complete
patient recovery. However, an unacceptably high number (up to 30%) of CM
patients die regardless of receiving the recommended treatment and up to 24% of
CM survivors develop neurological complications and cognition problems
post-recovery. Prognostic biomarkers to predict CM outcomes may be life-saving
as the “at risk” patients can be given the necessary interventions and/or
adjunctive therapies to prevent the adverse CM outcomes. These biomarkers can
be divided into two main groups: those that can predict neurological sequelae
post recovery and those that predict CM mortality. To date, some potential biomarkers
that can be used to predict CM adverse outcomes have been suggested although
their utility needs to be validated.
4.4.
Biomarkers to predict risk of developing neurological sequelae post-recovery
Identifying persons at risk for neurological sequelae will assist
in providing the necessary remedy at discharge or as early as is necessary and
possible. For example, a recent study demonstrated that computerized cognitive
training for children who survived CM provided some benefit on some
neuropsychological and behavioral functions even long after the malaria episode.
Therefore, this remedy, or others, can be administered to those patients at
risk for the development of neurological deficit post-recovery. Neuroprotective
factors such as erythropoietin (EPO) are critical in the brain repair process
and are therefore important for recovery from any brain insult. A recent study
using African children demonstrated that high plasma EPO levels were associated
with a 70% reduction in the risk of being discharged with neurological sequelae.
EPO levels above 200 units/liter were independently associated with about an
80% reduction in risk of developing neurological sequelae in a matched
analysis. The beneficial effects of EPO occur locally in the tissue or the
brain via EPO receptors found in neurons, astrocytes, microglia and endothelial
cells [48], and previous studies have shown that EPO can actively translocate
across the blood brain barrier. From the point of view of biomarker
development, serological EPO levels would provide a usable biomarker for
development of neurological sequelae with CM patients having < 200U/liter (or another cutoff found to be
more appropriate) of plasma EPO being at risk. Validation of this hypothesis is
required using large studies in different endemic regions and using different
age groups.
4.5.
Biomarkers to predict fatal CM
While a biomarker to predict fatal CM would serve for prognosis,
it would be of more value if interventions or adjunctive therapies are
available to save the patients. Using these biomarkers, clinicians can provide
the additional intervention before it is too late. Inflammatory cytokines
induce brain injury and cause apoptotic death of brain tissue components and
several studies have shown that some of these are associated solely with CM mortality.
In a study conducted in India, we evaluated the role of the ratio of some apoptotic
and angiogenic factors in discriminating between the different malaria disease
groups. Increased levels of CXCL10, sTNF-R2 and sFas were associated with
disease severity being highest in the CM non-survivors. The role of CXCL10 in
CM mortality was further confirmed using a murine model of CM in which, CXCL10
and CXCL9 (both CXCR3 ligands) were shown to be highly induced in the brains of
mice infected with P.
Berghei ANKA. Mice deficient in these ligands were protected from
experimental CM-related death and accordingly, CXCL10 and CXCL9 knockout mice
were shown to be partially protected from CM associated death [50]. Elevated
plasma levels of CXCL10 and CXCL4 were tightly associated withCM mortality, and
ROC analysis revealed that these chemokines can discriminate CM-non survivors
from MM (p < 0.0001)
and CM-survivors (p < 0.0001)
with an area under the curve (AUC) = 1. In addition, other studies revealed
that CXCL10 independently predicted severe and fatal CM as elevated levels of
CXCL10 expression in the CSF and peripheral blood plasma were observed in CM
patients who died compared to CM survivors. Other factors shown to influence CM
outcome include circulating levels ofVEGF, a factor known to protect neuronal
compartments from injury and death, ANG-1 and ANG-2/ANG-1 ratio. CM patients who
died had the lowest level of VEGF compared to CM-survivors. The study by Jain
et al. Further demonstrated that the mean and median ratios of CXCL10/VEGF,
sTNFR2/VEGF, and sFas/VEGF increased as the disease severity increased, with
the highest ratio occurring in the CM-non survivor group. ANG-1 levels at
presentation were also associated with higher risk of mortality in African
children, while ANG-2/ANG-1 ratios were higher in those patients who
subsequently died of CM. All these studies point to the fact that these
serological factors can be used to predict fatal CM.
Quantification of Cerebral Malaria
Biomarker
The detection, semi-quantification,
and clinical use of the Plasmodium falciparum histidine-rich protein-2
(PfHRP-2) as a parasite antigen biomarker for cerebral malaria had been
investigated.
The examination of the cerebrospinal fluid (CSF) in cerebral malaria cases typically shows normal parameters with normal protein, glucose, and low cell count and hitherto a CSF biomarker for cerebral malaria has yet to be identified.
An international team of scientists led by those from the Johns Hopkins Bloomberg School of Public Health (Baltimore, MD, USA) collected 73 CSF samples from Tanzanian children, 6 months to 9 years of age with P. falciparum parasitaemia. The CSF samples from nine Tanzanian children, who were microscopy negative for P. falciparum in peripheral blood, were also collected at the same site from 1994 to 1995. An additional 24 CSF samples were selected from archived CSF samples obtained from patients with neuroimmunological or neuroinflammatory conditions to use as controls.
The rapid diagnostic test (RDT) Binax NOW Malaria Test (Alere; Waltham, MA, USA), which detects both PfHRP-2 and panspecies aldolase was performed on the CSF samples. An immuno-polymerase chain reaction (PCR) was optimized and the real-time PCR amplifications were performed in a C1000 Thermal Cycler (Bio-Rad; Hercules, CA, USA). Western blots analysis was performed to compare some plasma and CSF samples.
The PfHRP-2 was detected in archival CSF samples from cerebral malaria patients from Tanzania both by a newly developed sensitive and specific immuno-PCR in 72 of 73 samples, and by rapid diagnostic tests in 62 of 73 samples. The geometric mean PfHRP-2 CSF concentration was 8.76 ng/mL with no significant differences in those who survived with 9.2 ng/mL, those who died with11.1 ng/mL, and those with neurologic sequelae with 10.8 ng/mL. The cerebral biomarker CSF PfHRP-2 was undetectable in all aparasitemic endemic and non-endemic control samples.
It was concluded that the widely used RDTs for malaria may be useful in CSF detection of PfHRP-2 in patients with cerebral malaria. A total of 62 of 73 samples had at least 1 ng/mL PfHRP-2 concentration detectable, an amount well above the 100 pg limit of detection threshold, if 100 µL is used for the RDT. The CSF PfHRP2 is proportional to plasma such that detection in the CSF is reflective of elevated plasma or whole blood PfHRP2 seen in many studies of cerebral malaria. The study was published June 30, 2014, in the American Journal of Tropical Medicine and Hygiene
The examination of the cerebrospinal fluid (CSF) in cerebral malaria cases typically shows normal parameters with normal protein, glucose, and low cell count and hitherto a CSF biomarker for cerebral malaria has yet to be identified.
An international team of scientists led by those from the Johns Hopkins Bloomberg School of Public Health (Baltimore, MD, USA) collected 73 CSF samples from Tanzanian children, 6 months to 9 years of age with P. falciparum parasitaemia. The CSF samples from nine Tanzanian children, who were microscopy negative for P. falciparum in peripheral blood, were also collected at the same site from 1994 to 1995. An additional 24 CSF samples were selected from archived CSF samples obtained from patients with neuroimmunological or neuroinflammatory conditions to use as controls.
The rapid diagnostic test (RDT) Binax NOW Malaria Test (Alere; Waltham, MA, USA), which detects both PfHRP-2 and panspecies aldolase was performed on the CSF samples. An immuno-polymerase chain reaction (PCR) was optimized and the real-time PCR amplifications were performed in a C1000 Thermal Cycler (Bio-Rad; Hercules, CA, USA). Western blots analysis was performed to compare some plasma and CSF samples.
The PfHRP-2 was detected in archival CSF samples from cerebral malaria patients from Tanzania both by a newly developed sensitive and specific immuno-PCR in 72 of 73 samples, and by rapid diagnostic tests in 62 of 73 samples. The geometric mean PfHRP-2 CSF concentration was 8.76 ng/mL with no significant differences in those who survived with 9.2 ng/mL, those who died with11.1 ng/mL, and those with neurologic sequelae with 10.8 ng/mL. The cerebral biomarker CSF PfHRP-2 was undetectable in all aparasitemic endemic and non-endemic control samples.
It was concluded that the widely used RDTs for malaria may be useful in CSF detection of PfHRP-2 in patients with cerebral malaria. A total of 62 of 73 samples had at least 1 ng/mL PfHRP-2 concentration detectable, an amount well above the 100 pg limit of detection threshold, if 100 µL is used for the RDT. The CSF PfHRP2 is proportional to plasma such that detection in the CSF is reflective of elevated plasma or whole blood PfHRP2 seen in many studies of cerebral malaria. The study was published June 30, 2014, in the American Journal of Tropical Medicine and Hygiene
