Detection of Specific Metabolic Signatures by Nuclear Magnetic Resonance or Metabolomic NMR in Correlation With Markers of Inflammation or Immune Activation and the Risk of Developing Neurocognitive Disorders - Longitudinal Retrospective Study Nested in the ANRS12225 PEDIACAM III Cohort

Context

Worldwide, more than three million children are living with HIV, 90% of them in sub-Saharan Africa. With the generalisation of antiretroviral treatment (ART), most children living with HIV from birth or children exposed to HIV in utero are now adults (4). Cognitive disorders and dementia associated with HIV are well known in adults. Neurological damage has also been well described in children whose mothers transmitted HIV to them (MTCT), in particular the specific and severe HIV encephalopathy affecting 10-15% of infants living with HIV before the era of multi-drug antiretroviral therapy (5, 6). The very early administration of triple therapy-type ART to infants living with HIV considerably reduced infection-related mortality and co-morbidities, but exposed them very early and over the long term to a treatment whose neurological toxicity in association with mitochondrial dysfunction has been reported, particularly in exposed children not living with HIV (7). Recent publications on prospective studies of neurocognitive, neuromotor or visual monitoring of HIV+ children on treatment or HIV-exposed children from different paediatric cohorts in developing countries have shown that, overall, children exposed to HIV in utero and not living with HIV have the same cognitive development as unexposed children not living with HIV (1, 8). In contrast, children living with HIV with or without ART have a higher prevalence of cognitive problems than untreated children. These disorders may appear from the age of 2-3 years, be found around the age of 6-7 years and especially in adolescence (9, 10). A study of a cohort of HIV-positive adolescents on ART showed that their overall cognitive development appeared to be similar to that of their HIV-positive peers. However, the study showed that the trajectory of executive functioning appeared to be different in HIV+ adolescents on treatment, which could be explained by previous brain damage (8). Another study of HIV-positive children on ART shows that these children have altered cortical and subcortical structures and regional brain connectivity.

These brain defects may contribute to deficits in their neurocognitive functions. This study has not been continued longitudinally to confirm the persistence or otherwise of such deficits (9).

On the other hand, the causes of the appearance of neurodevelopmental disorders are thought to be linked to the virus, since even if the viral load has decreased under triple therapy, the virus can continue to activate the immune system and cause inflammation that could have an impact on children's developing brains throughout their growth. A very recent study shows that ART taken as prophylaxis for one year has no effect on the growth and neurodevelopment of precocious children (11). On the other hand, the causes of the onset of neurodevelopmental disorders are thought to be linked to the virus, because even if the viral load has decreased under triple therapy, the virus can continue to activate the immune system and cause inflammation that could have an impact on children's developing brains throughout their growth. A very recent study shows that ART taken as prophylaxis for one year has no effect on the growth and development of the brain.

A. HIV, Inflammation, Immune Activation and Neurodevelopment The brain is particularly vulnerable in utero and during infancy and early childhood, and damage during these critical periods has the potential to cause long-term damage (2, 12, 13). Several neurodevelopmental disorders have been linked to immune activation and inflammation in early life, including autism spectrum disorders (ASD), schizophrenia, cerebral palsy, epilepsy, cognitive impairment and depression (2, 14). Maternal immune activation increases pro-inflammatory cytokines such as IL-6 in the foetal brain and leads to long-lasting changes in brain cytokines during postnatal development. Most cytokines remain present in the serum and brain throughout development and can alter the development and behaviour of the CNS (15-18). In addition, the passage of inflammatory cytokines such as IL-17 across the placenta can cause neuroinflammation in the foetus, leading to neurodevelopmental disorders very early in the infant's life (19). Viral infections, and HIV in particular, cause various types of brain damage during the prenatal and postnatal periods in HIV+ children born to HIV-positive mothers (7, 9). The virus itself can settle in the brain, causing direct damage. However, even though ART started very early after the birth of these infants succeeded in considerably reducing mortality and the circulating viral load, immune activation and inflammation remain present in children living with HIV (20). In addition, soluble plasma markers such as CD163 have recently been correlated with neurocognitive impairment in children living with HIV treated early (21). Finally, long-term ART could be considered toxic to the developing brain, but recent studies have shown that starting treatment early in infancy through to adulthood does not affect the brain (8).

B. Immunometabolism, inflammation and HIV Studies of the links between metabolism and the functioning of the immune system have proliferated in recent years, highlighting the contribution of metabolism to the development, fate and behaviour of immune cells. A large part of energy metabolism involves the mitochondria, which engages in the Krebs cycle, oxidative phosphorylation, aerobic glycolysis, Bêta-oxidation of fatty acids and glutaminolysis. This energy production is accompanied by the production of metabolites essential for signalling and cellular function. Metabolites are the substrates and products of the various chemical reactions that take place in all living systems. By analysing the relative variations in the quantity of metabolites, it is possible to characterise the metabolic pathways involved in different biological processes, such as inflammation.

In addition, disruption of metabolic pathways can alter immune cell fate, modulate immunity and generate chronic inflammation. A large body of literature shows clear links between metabolic defects and inflammation in diseases such as cancer, colitis, kidney disease and autoimmune diseases (22-24). Metabolites such as certain amino acids may be able to control the TH17/Treg balance in autoimmune diseases (25, 26). The immune response to bacterial or viral infections also has an impact on certain metabolic pathways. The tuberculosis model is currently being extensively studied (27). In the case of AIDS, the arginine metabolic pathway is thought to be closely linked to inflammation during HIV infection (28).

C Metabolic analysis of plasma for inflammatory signatures Analysing the metabolic profile of a sample involves systematically quantifying all the metabolites present in a tissue, cells or biological fluid (plasma, serum, cerebrospinal fluid or urine) to determine specific or abnormal signatures in a group of samples compared with a control group. Two main analytical techniques are used to study metabolomics: mass spectrometry (MS) and NMR. The pioneering technique is MS, which is highly sensitive and can detect a large number of metabolites. SdM is often used in the targeted approach of metabolomics, in which the compounds involved are already known. Generally, NMR spectroscopy, although less sensitive than mass spectroscopy, is simple to use, fast, reproducible and enables absolute quantification of the main metabolites in a mixture, without any separative techniques that could alter the results. This is a great advantage when analysing large long-term cohorts, with several study groups, and brings strong statistical power to the analyses. NMR and SdM are complementary techniques. NMR is less expensive and easier to use for obtaining global information on metabolism in a non-targeted way, and is suitable as a first-line technique for exploratory metabolic studies.

A non-targeted metabolomic study carried out on plasma from treated patients living with HIV, untreated patients living with HIV and a control group of healthy patients revealed specific and abnormal metabolic signatures in the group of patients living with HIV. This analysis was associated with a proteomic study of inflammatory and neurological biomarkers, the combined analyses of which indicated a high risk of development of inflammatory and neurological diseases in patients living with VIH on ART (29, 30). A very recent longitudinal metabolomic study performed on plasma from patients living with VIH associated with immune reconstitution/inflammatory syndrome revealed specific metabolic signatures affecting lipids and specific amino acids. This abnormal signature was correlated with inflammatory biomarkers such as IL-12p70 and IL-18 (31).

D. Preliminary results. Metabolomic NMR analysis of samples from children living with HIV.

As part of a collaboration between Dr. Laforge Mireille (A team) and Dr. Gildas Bertho (E team), a feasibility study was carried out on frozen plasma samples from children living with HIV in France whose immune responses had been studied by Dr. Florence Buseyne. These children had been studied in the early years of antiretroviral treatment (32-35). For 5 children, two samples taken approximately 2 years apart and corresponding to an uncontrolled viral load (HIV RNA supérieur à 500 copies/ml in the absence or failure of treatment, the 'before' group) and a controlled viral load (HIV RNA supérieur à 500 copies/ml after initiation or modification of treatment, the 'after' group) were selected and analysed.

This pilot study demonstrated the feasibility of the technique on samples from infected children and the possibility of generating two completely separate groups with distinct metabolic signatures at the two sampling points before and after control of HIV replication by ART. The analysis was carried out using the metabolite profile (Metabo) and the lipoprotein profile (Lipo), taking into account different statistical calculation models.

Hypothesis The study proposes to retrospectively include three groups of children from the ANRS 12225 - PEDIACAM III paediatric cohort in Cameroon, based on their history of HIV infection, ART, virological and immunological follow-up and the existence or not of neurological disorders defined according to appropriate thresholds. The study will be carried out on the biobank, which consists of plasma samples stored from the first months of life and followed longitudinally up to the present day, when the oldest child is 14 years old. The investigators of this cohort carried out psychological tests and showed that children living with HIV presented a higher frequency of neurocognitive disorders compared with children exposed to the virus born to HIV-positive mothers and children in an uninfected control group born to HIV-positive mothers.

Our working hypothesis is that the early HIV infection of children born to infected mothers may have disrupted metabolic pathways generating metabolic defects that will leave a signature persisting throughout the child's life. This infection must also have induced inflammation and/or immune activation, which is capable of impacting the developing brain, leaving sequelae that could be characterized by the appearance of neurodevelopmental disorders later in the child's life. It is possible that metabolic defects can amplify inflammation and make it chronic. In addition, the virus can be found in the brain and have a direct impact on it, as well as impacting metabolic pathways by creating neuroinflammation, which would be harmful to neurological development. In addition, the children were put on ART early after birth, with major clinical benefit. However, it is not known whether this early treatment over a prolonged period has any toxicity potentially associated with neurodevelopmental disorders.

Research objectives and innovative nature of the research

The aim of this research is to characterize and compare the signatures of early metabolic abnormalities in relation to the development of neurocognitive disorders detected in HIV-infected, HIV-exposed and uninfected children over the age of 5, and uninfected children born to HIV-infected or uninfected mothers. This retrospective study has multiple interests. I) This research, based on a classification of children with neurocognitive disorders, will enable markers of early metabolic and immunological defects to be validated retrospectively. II) In addition, this initial research, which will be carried out longitudinally at several points, will make it possible to monitor these markers over time, showing their persistence and confirming the diagnosis.

To meet this objective, the research will be nested in the ANRS 12225 PEDIACAM cohort, which is distinguished by its bio-bank of plasma samples collected for the three groups of children mentioned above, with longitudinal follow-up from the first months of the children's lives (with points before and after the start of ART treatment) up to the age of 14.

The project is divided into three work packages (WP). WP1. Metaboscreen: Metabolomic NMR analysis of plasma in search of specific metabolic signatures at the level of metabolites and lipoproteins in the different study groups.

WP2. Immunoscreen: Elisa and Luminex analysis of the same study groups and plasma samples for markers of inflammation and/or immune activation in order to correlate metabolic defects with immune system abnormalities. An analysis of mitokines, FGF21 and GDF15, markers of defects in energy metabolism and production of reactive oxygen species (ROS), will also be carried out in parallel (3).

WP3. NeuroDev: Research into the association between metabolic and immunological signatures and the development of neurocognitive disorders, longitudinal correlation study of early and late study points with suggested risk stratification.

The innovative nature of the project lies in 1) the specific features of the ANRS 12225 Pediacam III cohort, with its unique large cohort comprising a group of HIV-infected children treated early with ARVs, control groups and longitudinal plasma samples covering a long period of 14-15 years; 2) the use of cutting-edge technologies such as metabolomic NMR to search for markers that can be correlated with immune system dysfunction and neurodevelopmental disorders. The IVDr (In Vitro Diagnostic for research) NMR (Nuclear Magnetic Resonance) technique applied to the metabolomic analysis of biological fluids is highly reproducible and enables rapid global analysis of the metabolites present in a complex mixture.

The platform enables rapid analysis of a large number of samples and its reproducibility, enables longitudinal analysis in humans. Absolute quantification in blood of the 40 main metabolites and 140 lipoprotein parameters will be carried out by the platform. The team will thus be able to determine the possible presence of metabolic signatures in blood. The E team will be able to provide guidance to project partners on the metabolic pathways involved, with a view to modulating them for therapeutic purposes. 3) In parallel, an analysis of markers of inflammation, immune activation and mitokines, a marker of defects in energy metabolism, will be carried out on the same samples, enabling a correlation to be made between the metabolomic NMR data, the immune system and cognitive disorders.

Study population

This study will concern children included and followed up in the ANRS 12225 Pediacam III cohort. Several criteria were used to constitute the population of this study:

- Any infected or uninfected child in the ANRS 12225 Pediacam III cohort who participated in the ANRS 12322 - PediacamDev study on the evaluation of neurocognitive development and who had the results of the KABC II test (IPM and INV).

Neurocognitive status was defined on the basis of the results of the KABC II test given by two indices (IPM and INV) which are in fact scores. Any child will be considered cognitively impaired for IPM and INV index scores below the thresholds (mean of the index - 1σ; defined on the basis of data from uninfected children born to HIV-negative mothers).

• Availability of plasma samples, never thawed (an essential criterion for the NMR-metabolomic study) and in sufficient quantity (500 micro litre; for NMR-metabolomic analysis and ELISA analysis of inflammatory markers, activation and mitochondrial defects (mitokines) at different times during longitudinal follow-up: at 6 (±1.5) and 12 (± 3) months, and at 108 (± 6) months.
• The first two points were considered to be early in relation to the earliness of the infection and the start of ART. The point at 6 months of age was chosen mainly for the availability of sufficient samples for our study and above all to homogenise the groups studied in terms of the ART criterion. The 12-month time point is important for analysing the stability of the markers of interest and their persistence over time.
• The third point considered as 'late', at 108 months, corresponds on average to the age at which the psychological tests were carried out and evaluated, leading to the stratification established in the table below.

By combining the criteria described above, the three groups in our study will be made up of the following numbers and divided according to the results of the KABC II as shown in the table below

Group 1: HIV-infected children on ART (N=62) Group 2: HIV-uninfected children born to HIV-positive mothers (N=31) Group 3: HIV-uninfected children born to HIV-negative mothers (N= 35)

With 3 sampling points on different dates for each child, the total number of samples to be processed will be N= 384 samples.

Type of study

This is a retrospective study nested within the ANRS 12225 Pediacam III cohort. The data to be analysed will be those collected previously or analysed from biobank samples already available at various points during the follow-up.

Biological analyses

WP 1 NMR-Metabolomic analysis HIV-infected infants (N=62) started ART at the age of <6 months, uninfected children divided into 2 groups, exposed children born to HIV+ infected mothers (N=31) and children born to HIV- infected mothers (N=35). The plasma of the children followed up at three stages will be analysed first by NMR-metabolomics (500 micro litre) and the remaining volume will be dedicated to the rest of the analyses planned (WP2).

WP2 ELISA and Luminex analysis of inflammatory markers, immune activation and defects in energy metabolism In HIV-infected adults, monocyte infiltration of the brain is the main mechanism of the neurological disorders observed. From acute infection onwards, the systemic biomarker of monocyte activation - sCD163 - is associated with neurocognitive performance (32); this association between plasma levels of the sCD163 molecule and neurological status persists in treated subjects whose viremia is undetectable (33, 34). In HIV-infected children and adolescents, several studies have reported associations between lower levels of neurodevelopment and higher markers of monocyte activation and inflammation - including sCD163 and CRP (35, 39-41); conversely, the antiviral chemokine MIP-1β is positively associated with neurodevelopment scores (39).

The analytes that will be assayed in plasma to assess immune activation and inflammation reflect different processes: inflammation (CRP, IL-6, IL-18 and IL-18BPA), monocyte activation (sCD14, sCD163), chemokines associated with IFN-α (CXCL10), the TNF-α receptor family (sTNR2), lymphocytes (TNF-α, IL-21, IL-8, MIP-1β), and intestinal permeability (iFABP).

With regard to the search for associations between plasma metabolites and markers of immune activation and inflammation, the feasibility of the project and the relevance of the analytes selected are demonstrated by the correlations observed in metabolomics studies in adults infected with HIV-1 (31). In parallel, assays of mitokines, signalling molecules produced in response to a defect in energy metabolism (GDF-15, FGF-21, Humanin) will also be carried out (3).

WP3 correlation with the 3 different axes, metabolic, immunological and neurocognitive.

All the data obtained will be analysed together, taking into account metabolic defects, which will be correlated with inflammatory and activation markers, and neurocognitive impairment disorders (risk group and impairment group).

Conduct of the study The study will be conducted in France, mainly thanks to the availability of an NMR-metabolomics platform (MetaboParis-Santé) capable of carrying out the study. This platform is unique at national level and is attached to the University of Paris. Because of the importance of the samples and their scientific value, maximum optimisation of their use will be put in place to answer the various questions raised by the project. The samples will be shared between the NMR-metabolomics platform at the UFR Biomédicale des Saints-Pères, Dr Florence Buseyne's laboratory at the Institut Pasteur and Prof. Pierre Gressens' laboratory at Hôpital Robert Debré, where Mireille Laforge heads a research team on the role of immunometabolism in infectious diseases and inflammation. The NMR-metabolomics data will be extracted and analysed directly by Dr Gildas Bertho from the NMR team at the University of Paris and shared with the teams involved in the project. The same applies to the results of the other analyses, which will be shared with all the partners. All the data produced in France will be stored in the UMR 1141 laboratory and then transferred to Cameroon to the laboratory of Dr Mathurin Cyrille TEJIOKEM - Epidemiology and Public Health Department, Centre Pasteur du Cameroun, in order to cross-reference them with the neurocognitive data and carry out correlation analyses between the various metabolomic, immune and neurocognitive studies. Dr Mathurin Cyrille TEJIOKEM will be in charge of the methodology and statistical analyses to produce correlation indices for these studies.

Samples Collection The Centre Pasteur du Cameroun will prepare the plasma samples according to the above criteria and the procedures adapted by Dr Tagnouokam Paul Alain. The package will be sent to the Inserm UMR 1141-NeuroDiderot laboratory and then transported to the NMR metabolomics platform at the University of Paris where they will be stored at -80 degrees. An organization will be put in place to maximize the use of the samples. NMR analysis is rapid (1 hour per sample), from thawing the sample, preparing it, measuring it in the machine and generating the data. At the end of the biological NMR analyses, the rest of the plasma will be refrozen directly in 2 distinct parts and sent back to the different laboratories at the Pasteur Institute and the Robert Debré Hospital for further ELISA analyses.

Biospecimens Description Plasma samples from children followed in the ANRS 12225 Pediacam III cohort will be used in this study. These never-thawed plasma samples (primary criterion for the NMR-metabolomics study) are available in sufficient quantity (500 micro litre ) for NMR-metabolomics analysis and ELISA analyzes of inflammatory markers, activation and mitochondrial defects (mitokines) at different times of longitudinal follow-up: at 6 (±1.5) months and 12 (± 3) months, and at 108 (± 6) months. With 3 sampling points on different dates for each child, the total number of samples to be processed will be N= 384 samples.

Eligibility criteria

• Any child living or not with HIV from the ANRS 12225 Pediacam III cohort having participated in the ANRS 12322 - PediacamDev study on the evaluation of neurocognitive development and having the results of the KABC II test (IPM and INV).
• Availability of plasma samples, never thawed (essential criterion for the NMR-metabolomic study) and in sufficient quantity (500 micro litre; for NMR-metabolomic analysis and ELISA analysis of inflammatory markers, activation and mitochondrial defects (mitokines) at different times during longitudinal follow-up: at 6 (±1.5) and 12 (± 3) months, and at 108 (± 6) months.
• Availability of plasma samples, never thawed (essential criterion for the NMR-metabolomic study) and in sufficient quantity (500 micro litre).

Expected results and outcomes Many studies have been carried out in cohorts of adults living with HIV as well as children living with HIV on neurological disorders. But very few studies have made the link between metabolism, inflammation/immune activation and neurocognitive disorders in infants and children in countries with limited resources. The Pédiacam cohort offers the unique opportunity in Africa to carry out a study in a population of HIV-infected children matched to a population of uninfected children born to HIV+ mothers and uninfected children born to HIV- mothers. Concerning the NMR technique, the preliminary results are very encouraging for searching for metabolic signatures specific to the different study groups.

These early metabolic and inflammatory markers could possibly be correlated with the development of neurocognitive disorders in the group of HIV+ children and allow early detection and management of these disorders. This study, which will be carried out on 400 samples, can be quickly supplemented subsequently by other samples already selected to complete the longitudinal monitoring of the analysis. It may even include new samples taken prospectively from children who are now 15 years old.

The expected results and their impact could be as follows:

1. Verification of the presence of metabolic defects early from the first months of life (comparison between the different study groups independently of the correlation with neurocognitive disorders) and proposal for more in-depth study for the affected metabolic pathways in order to understand the mechanism and try to propose therapy strategies
2. The demonstration of a probable correlation between these metabolic defects and other inflammatory markers or immune activation and especially with neurocognitive disorders. This is of great importance because it is not only a question of confirming the link between the three processes but, on the one hand, of validating early markers for neurocognitive disorders and, on the other hand, of being able to intervene early to prevent the development of these disorders in the child's brain.
3. longitudinal monitoring, even if it only covers 3 points in this pilot study, will be very informative on the development time of the metabolic defect and its persistence, which will give even more strength to this marker.
4. if the validity of these markers is confirmed, they will be studied in children currently aged 15 years to assess whether there is a possibility of intervention via the modulation of the affected metabolic pathways allowing induce beneficial effects on the brain even late.
5. The extension of our study to other ANRS cohorts such as the PROMISE-M&S cohort (responsible for Professor Ph. Vande Perre/Dr JP Molès, collaborator of this project on the mitokines dosage part) which carried out an assessment of the neurodevelopment of children exposed uninfected at the age of 7 using four tests, KABC-II, MABC-2, TOVA and SDQ25.

The expected results will be of great importance and will lay the foundations for studies on a larger scale and also concerning other infectious diseases potentially associated with neurodevelopmental anomalies. More specifically, for children infected with HIV they will be able to guide specific early care and monitoring.