Relevance of lipopolysaccharide levels in HIV-associated neurocognitive impairment: the Neuradapt study
- First Online:
- Cite this article as:
- Vassallo, M., Dunais, B., Durant, J. et al. J. Neurovirol. (2013) 19: 376. doi:10.1007/s13365-013-0181-y
- 178 Views
Contributory factors to HIV-associated neurocognitive disorders (HAND) have been shown to include age, co-morbid infections, medication toxicity, virological, genetic and vascular mechanisms, as well as microbial translocation of lipopolysaccharide (LPS), which is suspected to trigger monocyte activation and increase trafficking of infected cells into the brain. In this study, our aim was to assess the degree of neurocognitive impairment in a group of randomly selected HIV-infected patients and investigate potential risk factors, including LPS plasma levels. Furthermore, we evaluated the relevance of LPS as a potential marker for screening patients with mild neurocognitive impairment. LPS plasma levels were compared among patients with HAND and those with no HAND. As LPS has also been shown to be elevated in hepatitis C co-infection, the analysis was stratified according to the presence or not of hepatitis C virus (HCV) co-infection. Differences between groups were evaluated using chi-square tests and Kruskal–Wallis non-parametric tests. Stepwise logistic regression was performed to identify independent risk factors for HAND in the subgroups of HCV-positive and negative patients. A p value <0.05 was considered significant. Analyses were conducted using SPSS® software. From December 2007 to July 2009, 179 patients were tested (mean age 44, 73 % male, 87 % on treatment, 30 % HCV co-infected, median CD4 504/ml and 67 % with viral load below 40 copies/ml). HAND was identified in 40/179 patients (22 %), the majority displaying asymptomatic neurocognitive impairment or mild neurocognitive disorder. Univariate analysis showed that age, illicit drug use, hepatitis C co-infection, prior AIDS-defining events, CD4/CD8 ratio and LPS plasma levels were significantly associated with HAND. The median LPS level was 98.2 pg/ml in the non-HAND group versus 116.1 pg/ml in the HAND group (p < 0.014). No differences were found in LPS values between subgroups of impairment. There was a clear association between LPS levels and HAND in the HCV-positive group (p = 0.036), while there was none in the HCV-negative group (p = 0.502). No difference in degree of hepatic fibrosis was found between the HAND and non-HAND groups. In conclusion, LPS levels were associated with HAND in the HCV-positive group, while, in the HCV-negative group, age and pro-viral DNA were the only variables independently associated with HAND. There was no difference in degree of liver disease as predicted by score of fibrosis between HAND and non-HAND groups. The role of HCV co-infection and higher LPS levels in the pathogenesis of HAND in patients with viral suppression on treatment requires further investigation.
KeywordsLipopolysaccharideHIV-associated neurocognitive disordersAntiretroviral therapy
Over recent years, the course of HIV-associated neurocognitive disorders (HAND) has been altered by the introduction of combination antiretroviral therapy (cART), while the incidence of HIV-associated dementia (HAD) has decreased dramatically. However, the global prevalence of HAND remains high (20–50 %), with a majority of patients presenting with mild and often asymptomatic cognitive impairment (Mc Arthur et al. 2010; Robertson et al. 2007; Woods et al. 2009; Heaton et al. 2010). According to the American Academy of Neurology (AAN) 2007 revised criteria, the two most frequent categories of impairment are asymptomatic neurocognitive disorders (ANI), in which initial symptoms are absent, and mild neurocognitive disorders (MND), in which symptoms are subtle (Antinori et al. 2007). Contributory factors to HAND have been shown to include virological, genetic and vascular mechanisms, age, co-morbid infections and medication toxicity (Goodkin et al. 2009). Recent findings also suggest that microbial translocation is a cause of systemic immune activation in HIV-infected patients and that increased immune activation may be linked to the development of HAD (Ancuta et al. 2008). Indeed, HIV-induced disruption of mucosa-associated lymphoid tissue results in translocation of microbial products across the intestinal mucosa into the peripheral circulation, producing high plasma levels of lipopolysaccharide (LPS), a component of the cell wall of gram-negative bacteria that is suspected to trigger monocyte activation and to increase trafficking of infected cells into the brain (Brenchley et al. 2006; Epple et al. 2009). Furthermore, Shiramizu et al. showed that individuals with HAD had a higher HIV-DNA viral load in PBMC than HIV-infected patients with normal cognition, suggesting that the primary source of HIV-DNA may be activated monocyte/macrophage cells (Shiramizu et al. 2005).
However, the role of microbial translocation in mild forms of HAND is less well known. Moreover, hepatitis C infection has also been shown to be associated with microbial translocation and elevated LPS levels (Ancuta et al. 2008; French et al. 2013; Rotman and Liang 2009). In this study, our aim was to assess the degree of neurocognitive impairment in a group of randomly selected HIV-infected patients regularly followed in our infectious diseases department and to investigate the role of LPS plasma levels among this population.
Study design and participants
Neuradapt© is a prospective study investigating the prevalence of HAND among HIV-1 infected patients followed in Nice University Hospital. According to the American Academy of Neurology 2007 revised criteria, the two most frequent categories of impairment are ANI, in which initial symptoms are absent, and MND, in which symptoms are subtle (Antinori et al. 2007). Exclusion criteria included active opportunistic infection, a change in psychotropic therapy within the last 3 weeks or any neurological history. Patients were asked to provide informed consent. The study was approved by the Montpellier ethics committee.
ANI, involving at least two cognitive domains and documented by a performance of at least one SD below the mean on NP tests, without interference in everyday functioning. The asymptomatic characteristics of impairment were defined by the instrumental activity of daily living short version battery and by interviewing the patient and his/her family.
MND, involving at least two cognitive domains and documented by a performance of at least one SD below the mean on NP tests, with mild interference in daily functioning.
HAD, involving at least two cognitive domains and documented by a performance of two SD below the normative mean on NP tests, with marked interference in daily functioning.
Those with HAND (i.e. ANI, MCD or HAD)
Those without HAND.
Demographic parameters and background measurements
In order to correlate results to NP test group, the following parameters were recorded for each patient: age, gender, education, co-morbid conditions (hypertension, smoking, dyslipidemia, illicit drug use and diabetes), use of psychotropic molecules (benzodiazepines, antidepressants, carbamates and antiepileptic drugs), inclusion and nadir CD4 T cell count, CD8 T cell count, CD4/CD8 ratio, HIV-RNA viral load (Abbott real-time PCR), HIV-DNA viral load (Abbott real-time PCR), viral hepatitis markers, type of antiretroviral therapy upon inclusion and 2010 CNS penetration effectiveness score for cART (Letendre et al. 2008).
Moreover, for patients with hepatitis C virus (HCV) and/or HBV co-infections, markers of liver fibrosis were taken into account, i.e. liver biopsy or non-invasive procedures (Fibroscan© and Fibrotest©).
Quantification of plasma LPS
LPS was determined in heat-inactivated plasma using a commercially available limulus amoebocyte lysate assay (LAL) test (Lonza, Basel, Switzerland) according to the manufacturer's instructions. Briefly, samples diluted 1/5 with endotoxin-free water were heated at 70 °C for 10 min to inactivate interfering plasma components. After centrifugation and incubation with the LAL reagent and the chromogen, samples were measured in duplicate at 405 nm in a photometric plate reader, and background, if present, was subtracted.
For those patients who provided a plasma sample, LPS plasma levels were compared in patients with HAND (ANI, MND and HAD) to those with no HAND. Median and interquartile range (IQR) for LPS values were determined and plotted for the total study population and for HCV-infected and non-infected patients. As LPS has also been shown to be elevated in hepatitis C co-infection, the analysis was stratified according to the presence or not of HCV co-infection. Differences between groups were evaluated using chi-square tests and Kruskal–Wallis non-parametric tests. Stepwise logistic regression analyses were performed to identify independent risk factors for HAND in the subgroups of HCV-positive and negative patients. A p value <0.05 was considered significant. Analyses were performed using SPSS® software.
Among 1,963 patients followed in Nice from December 2007 to July 2009, 322 were randomly selected, and 256 provided informed consent. Among these, LPS levels were available for 179 patients (70 %). Median age was 44(IQR: 39–51) years, 73 % were male, 87 % were on cART, and 67 % had a viral load below 40 copies/ml. The median CD4 cell count was 504/mL (IQR: 352–702), and 30 % were hepatitis C co-infected. HAND was identified in 40/179 patients (22 %), distributed as follows: ANI 22/40 (55 %), MND 14/40 (35 %) and HAD 4/40 (10 %). The main impairments in the HAND group included speed of information processing (22 patients), recall memory (20 patients), working memory (20 patients), motor/psychomotor speed (12 patients) and attention/concentration (11 patients). No difference in terms of impaired cognitive domain was observed between ANI, MCD and HAD (data not shown). Only four patients had a learning memory deficit. Patients without HAND numbered 139 (78 %); among these, 14 patients had a NP test performance score at least two SD below the mean in only one cognitive domain, thus, not meeting HAND-defining criteria. Their main impairments concerned were speed of information processing and attention. Four patients in the HAND group and three in the non-HAND were diagnosed as suffering from depression; however, none had significant functional limitations requiring deferral of NP testing, according to the AAN criteria, so that they were not excluded from the analysis.
Patient demographic characteristics and background measurements: univariate analysis of risk exposure to HIV-associated neuro-cognitive disorders
N (%) or median (IQR)
N (%) or median (IQR)
Number of patients
Prior AIDS-defining event
Duration of HIV infection (years)
Hepatitis C co-infection
F3–F4 stage hepatic fibrosis
12 (9.2 %)
HIV-RNA (log10 copies/ml) at inclusion
CD4 count (cells/mm3)
CD8 count (cells/mm3)
Nadir CD4 count (cells/mm3)
Pro-viral DNA (log10 copies/ml) at inclusion
Antiretroviral CNS penetration
Effectiveness (CPE) score
No antiretroviral treatment
Regimens with CPE score <8
Regimens with CPE score ≥8
Figure 1 shows median and IQR LPS levels in the non-HAND and HAND patient groups. The median LPS level was 98.2 pg/ml (IQR, 73.5–118.2) in the non-HAND group versus 116.1 pg/ml (IQR, 79.3–149.3) in the HAND group (p < 0.014). Mean LPS values for the impairment subgroups were as follows: ANI = 122 ± 61 pg/ml, MND = 131 ± 8 pg/mL and HAD = 129 ± 45 pg/ml. No significant difference was found between these groups.
No correlation was found between LPS and either HIV-RNA viral load or CD4 T cell count. However, LPS values were higher in patients with hepatitis C co-infection (113 (IQR, 88–139) vs. 96 pg/ml (IQR, 68–116), p = 0.002) and in those with a history of illicit drug use (110 (IQR, 77–134) vs. 98 pg/ml (IQR, 73–23), p = 0.08), although, this did not reach statistical significance. LPS did not correlate with HCV viral load (data not shown).
Patient characteristics according to HCV serological status for the HAND and non-HAND groups (% or median (IQR))
HCV + (n = 54)
HCV − (n = 125)
HAND (n = 18)
Non-HAND (n = 36)
HAND (n = 22)
Non-HAND (n = 103)
HIV duration (years)
Prior AIDS-defining event
Pro-viral DNA (log10 copies/ml)**
HIV-RNA (log10 copies/ml)
CD4 T cell
CD8 T cell
In this study, we investigated the role of circulating LPS in neurocognitive impairment in a group of randomly selected HIV-infected patients regularly followed at Nice University Hospital, exploring the role of LPS plasma levels as a possible contributing factor for HAND and, consequently, the potential for LPS levels to be used as a marker of HAND. Based on standard neurocognitive tests, impairment was diagnosed in 22 % of patients. A significant association was found between higher LPS levels and HAND, but only among HCV co-infected patients.
The mechanisms involved in HAND remain debated. Epple et al. (Epple et al. 2009) showed that HIV-1 can directly disrupt the mucosal epithelial barrier of the gut, leading to microbial translocation via a cytokine-mediated mechanism. The depletion of the gut immune system in HIV-infected patients has been shown to be associated with abnormally high levels of LPS and monocyte expression of activation marker CD16. This immune activation appears to result in CNS injury as suggested by Lyons, who showed that soluble CD14, a marker of monocyte activation, is increased in a group of HIV-infected patients with mild forms of impairment, severe immune-depression and a high rate of virological failure (Lyons et al. 2011). In our study, we found no gradient of LPS plasma levels according to severity of HAND, based on its current definition by the AAN. Although, this could be due to the fact that ours was a cross-sectional study, and we were, thus, unable to measure the degree and duration of immune activation, these results may also negate any direct causality between LPS levels and HAND. Moreover, we found that the association between LPS and HAND in HCV co-infected patients did not appear in HCV-negative patients. This result could indicate that the relation between LPS and HAND reflects the difference in frequency of HCV co-infection or related co-factors such as IDU or substance abuse including alcohol, as shown by Ancuta et al. (Ancuta et al. 2008). On the other hand, although co-infected patients with HAND had higher values of LPS than those without HAND, there was no difference in terms of hepatic disease. Such findings could be in favour of an immunological mechanism rather than a deficit in LPS clearance due to advanced hepatic disease. Similarly, Marchetti et al. found that LPS values did not differ between cirrhotic and non-cirrhotic HCV-infected patients, suggesting that microbial translocation-driven immune activation could be the pathogenic mechanism involved in HIV/HCV co-infection (Marchetti et al. 2012). High LPS values in hepatitis C co-infected patients have been previously reported by Balagopal, while Aronow observed that these patients displayed greater cognitive–motor impairment than mono-infected controls (Balagopal et al. 2008; Aronow et al. 2008). The higher LPS levels we observed in our study thus point to a potentially significant role for the hepatitis C virus in HIV-associated cognitive impairment, via a pathogenic pathway that, to our knowledge, remains unexplained. The immunological hypothesis has been recently suggested by Kushner et al., who showed that co-infected patients have different immune biomarkers than HIV or HCV mono-infected subjects (Kushner et al. 2012) and Rempel et al., who demonstrated that these patients have high levels of monocyte activation and that their immune profile correlates with cognitive impairment. Authors suggest that HCV may trigger monocyte activation by factors probably other than HIV viral load, liver disease and drug or alcohol abuse (Rempel et al. 2013). A similar hypothesis can be drawn from a study conducted by Chattergoon et al., who showed that HIV and HCV activate the inflammasome complex in monocytes, thus stimulating the production of pro-inflammatory cytokines (Chattergoon et al. 2012).
In HCV-negative patients, high levels of HIV pro-viral DNA were associated with HAND. Shiramizu et al. have suggested that cognitive disorders may be linked to the continued presence of intracellular HIV-DNA (Shiramizu et al. 2012). We found no correlation between LPS and viral load or CD4 T cell count. This was also observed by Marchetti et al. (Marchetti 2011). Wallet et al. (Wallet et al. 2010) showed in paediatric patients that microbial translocation is associated with persistent monocyte/macrophage activation, independently of viral replication or T cell activation. Papasavvas et al. (Papasavvas et al. 2009) did not find any correlation between LPS levels and T cell activation markers following treatment interruption.
This study has several limitations, particularly concerning the lack of measurement of monocyte activation markers, which would have been useful to confirm the association between microbial translocation and increased trafficking of infected monocytes in the brain. Moreover, blood samples collected soon after seroconversion would have been helpful to determine whether initially high microbial translocation might predict the onset of HAND. Wide inter-individual variability in circulating LPS was observed in our study, and these shortcomings have also been reported by Marchetti et al. (Marchetti 2011). However, as our patients were randomly selected and displayed satisfactory immunovirological control and mild forms of impairment, our findings may be considered to reflect the majority of cART era patients.
In conclusion, LPS levels were associated with HAND only in the HCV-positive group, while, in the HCV-negative group, age and pro-viral DNA were the only variables independently associated with HAND. There was no difference in degree of hepatic fibrosis predicted by F1–F2 scores between HAND and non-HAND groups. The role of HCV co-infection and higher LPS levels in the pathogenesis of HAND in patients with viral suppression on cART requires further investigation.
We wish to thank the patients who accepted to take part in the Neuradapt study. Dr Eric Fontas provided helpful advice on the statistical analysis.
We also wish to thanks Patrizia Comi.
Study funding: This study received funding from Abbott, Glaxo-Smith-Kline, Boehringer, Gilead, Tibotec and Merck companies. The pharmaceutical companies mentioned above did not participate in the design or the conduct of the study, in the collection, management, analysis and interpretation of data or in the preparation, review or approval of the manuscript. All the phases of the work described above were prepared independently by the authors.