| |
Viremic and Virologically Suppressed HIV Infection Increases Age-Related Changes to Monocyte Activation Equivalent to 12 and 4 Years of Aging, Respectively
|
| |
| |
Download the PDF here
J Acquir Immune Defic Syndrome JAIDS May 1, 2015
"Although HIV infection is known to be associated with innate immune dysfunction and the premature appearance of age-related changes, the extent to which age and HIV contribute to these changes in HIV-positive individuals, and whether these effects are completely reversed by cART, has remained unclear. We report that viremic HIV-positive individuals display a similar level of monocyte activation to HIV- controls aged 12 years older, and although viral suppression significantly reduced the level of biomarkers such as CXCL10, VS HIV-positive individuals still show an equivalent level of monocyte activation 4 years earlier. Strikingly similar estimated effects of HIV in year equivalents were observed for a range of soluble and cellular markers of monocyte/innate immune activation, supporting the validity of this analysis. These findings suggest that in addition to the effects of aging, viremic and VS (virologically suppressed) HIV infection have an additional burden of innate immune activation equivalent to up to 12 and 4 years of normal aging, respectively."
"Although it has been speculated that HIV may accelerate biological aging, this is likely an oversimplification of the diverse and incompletely understood mechanisms that contribute to immunological aging. The data presented here indicate that HIV infection heightens age-related changes such that levels of monocyte activation reflect those observed in HIV- individuals who are significantly older and that this effect is only partially ameliorated by cART. Given the established associations between these activation biomarkers and inflammatory disease, this likely confers increased disease risk, although this requires investigation. Determining the pathological consequences of these changes, and the mechanism underlying them, is required to prevent inflammatory diseases in the increasingly aging VS HIV-positive population."
"To quantify the effect of HIV infection on monocyte activation and estimate the age at which HIV-positive individuals reach a given level of monocyte activation, we used a similar regression model as above but where age was the outcome of the regression equation........In summary, these data indicate that viremic HIV-positive individuals exhibit levels of monocyte activation markers similar to those seen in seronegative controls aged up to 12 years older, and although viral suppression significantly reduces this effect, VS HIV-positive individual still experience levels of monocyte activation similar to those observed in controls aged up to 4 years older."
"Table 2 shows the impact of HIV infection in addition to age-related changes on each parameter and indicates that proportions of intermediate monocytes were significantly increased and the proportions of classical monocytes consequently decreased, in viremic HIV infection (P = 0.002 and 0.015, respectively, column "Pvalue vs HIV-") but not in VS individuals. In contrast, levels of soluble age-related monocyte activation markers CXCL10, neopterin, and sCD163 were significantly elevated in both viremic (P ≤ 0.001 for all, column "P value vs HIV-") and VS individuals (P < 0.001 for all). Although it was not significantly altered by age, sCD14 was increased in both viremic and VS HIV-positive individuals as compared with controls (P ≤ 0.001 for both)."
------------------------
Viremic and Virologically Suppressed HIV Infection Increases Age-Related Changes to Monocyte Activation Equivalent to 12 and 4 Years of Aging, Respectively
Angelovich, Thomas A. MSc*,; Hearps, Anna C. PhD*,; Maisa, Anna BSc*; Martin, Genevieve E. MBBS*; Lichtfuss, Gregor F. PhD*,; Cheng, Wan-Jung MSc*; Palmer, Clovis S. PhD*,; Landay, Alan L. PhD; Crowe, Suzanne M. MBBS*, Jaworowski, Anthony PhD*,,#
*Centre for Biomedical Research, Burnet Institute, Melbourne, Australia;School of Applied Sciences, RMIT University, Melbourne, Australia;Department of Infectious Diseases, Monash University, Melbourne, Australia;School of Medical Sciences, University of New South Wales, Sydney, Australia;Department of Microbiology and Immunology, Rush University Medical Center, Chicago, IL;Infectious Diseases Unit, Alfred Hospital, Melbourne, Australia; and#Department of Immunology, Monash University, Melbourne, Australia.
Abstract
Background: Chronic inflammation and immune activation occur in both HIV infection and normal aging and are associated with inflammatory disease. However, the degree to which HIV influences age-related innate immune changes, and the biomarkers which best reflect them, remains unclear.
Methods and Results: We measured established innate immune aging biomarkers in 309 individuals including 88 virologically suppressed (VS) and 52 viremic (viral load ≤ and >50 copies per milliliter, respectively) HIV-positive individuals. Levels of soluble (ie, CXCL10, soluble CD163, neopterin) and cellular (ie, proportions of inflammatory CD16+ monocytes) biomarkers of monocyte activation were increased in HIV-positive individuals and were only partially ameliorated by viral suppression. Viremic and VS HIV-positive individuals show levels of age-related monocyte activation biomarkers that are similar to uninfected controls aged 12 and 4 years older, respectively. Viremic HIV infection was associated with an accelerated rate of change of some monocyte activation markers (eg, neopterin) with age, whereas in VS individuals, subsequent age-related changes occurred at a similar rate as in controls, albeit at a higher absolute level. We further identified CXCL10 as a robust soluble biomarker of monocyte activation, highlighting the potential utility of this chemokine as a prognostic marker.
Implications: These findings may partially explain the increased prevalence of inflammatory age-related diseases in HIV-positive individuals and potentially indicate the pathological mechanisms underlying these diseases, which persist despite viral suppression.
INTRODUCTION
The advent of effective combination antiretroviral therapy (cART) has made HIV infection a manageable chronic condition for HIV-positive individuals who maintain long-term viral suppression. However, as this population ages, it is becoming increasingly apparent that even virologically suppressed (VS) HIV-positive individuals suffer ongoing immune activation and dysfunction, which is not reflected by traditional biomarkers of HIV disease such as viral load and CD4+ T-cell count. HIV infection induces immunological changes that are characteristic of the normal aging process including increased inflammation (as indicated by inflammatory markers such as TNF and IL-6, reviewed in Deeks et al1), an expansion of senescent T cells with shortened telomeres,2,3 and an increase in markers of adaptive and innate immune activation.4 Beyond being simply biomarkers of aging or HIV infection, many of these age-related parameters are of clinical significance and are increasingly being shown to be predictive of inflammatory age-related morbidities that HIV-positive individuals are at heightened risk of developing.1,5 Markers of inflammation such as IL-66 and high-sensitivity C-reactive protein7 are associated with increased risk of cardiovascular disease (CVD) in HIV-positive individuals and the general population (reviewed in Hearps et al8). Monocytes/macrophages play critical roles in the pathogenesis of inflammatory disease such as CVD,9 and we have previously shown that monocyte activation markers increase during aging10 and in both VS and viremic HIV infection.5,11 Elevated levels of monocyte activation biomarkers including the proportion of inflammatory CD16+ monocytes and levels of soluble (s)CD163 and sCD14 (shed from activated monocytes) are associated with indices of CVD including noncalcified plaques,12 arterial inflammation,13 coronary artery calcification,14 and progression of carotid intima-media thickness (a surrogate clinical measure of atherosclerosis).15
Although adaptive immune parameters such as CD4+ T-cell counts may be useful prognostic markers in untreated HIV infection, it is becoming increasingly clear that innate markers of inflammation and monocyte activation may more accurately predict morbidity and mortality than T-cell parameters in cART-treated individuals.16-18 However, the relative contributions of both aging and HIV infection on these immunological changes are unclear.
Studies investigating the link between monocyte activation and clinical outcomes in HIV infection use various cellular and soluble markers of monocyte activation including the proportions of monocyte subsets (particularly the "inflammatory" CD16+ subsets), expression of activation markers such as CD11b or tissue factor, and plasma levels of molecules either shed from or secreted by activated monocytes/macrophages (eg, sCD14, sCD163, CXCL10, MCP-1). Importantly, it remains unclear which of these parameters best indicates monocyte activation and to what extent these biomarkers may overlap and indicate common underlying immunological mechanisms.
Here, we evaluated a combined dataset of commonly tested biomarkers associated either directly or indirectly with monocyte activation to (1) identify the most robust and indicative markers of monocyte activation and (2) quantitate the individual effects of aging and viremic and VS HIV infection on changes to monocyte activation, which are thought to underpin development of inflammatory diseases in these populations.
METHODS
Participant Recruitment
HIV-positive individuals were recruited from the Alfred Hospital Infectious Diseases Unit with informed consent and institutional review board approval.
Blood collected into EDTA-coated tubes was used for fresh whole-blood immunophenotyping of monocytes, whereas plasma was prepared and stored for future analysis as previously described.5 Healthy controls were recruited from the community. This study used data from newly recruited individuals (n = 31) plus biomarker data from previously published cross-sectional cohort studies.5,11,19,20
Experimental Procedures and Statistical Analysis
Monocyte subsets and soluble plasma markers of innate immune activation/inflammation including CXCL10, neopterin, sCD163, sCD14, and LPS were determined as previously described.5,11,19 To evaluate the continuous effect of aging on the biomarkers tested, data from cross-sectional studies were combined with additional patient recruitment for analysis. Linear regression analysis was used to identify monocyte biomarkers that were significantly altered with age and determine the influence of sex, current smoking status, and either viremic or VS HIV infection on this relationship. Spearman's correlation analysis was used to determine the association between individual monocyte activation biomarkers. To quantitatively estimate the impact of viremic and VS HIV infection on monocyte activation in the context of aging, linear regression modeling was used to estimate the age difference between VS and viremic HIV-positive individuals with HIV- controls for a given level of each biomarker.
Differences in coefficients due to HIV status were analyzed using a Student t test (P < 0.05 considered significant), whereas differences in slope between sample groups were determined by comparing models fitted with or without an interaction term (likelihood ratio test). Where the slope differed between groups, the linear equations were solved for a given parameter level taking into account the interaction term. Linear regression analysis was performed using Stata v.11 (StataCorp), and correlations were assessed using GraphPad Prism (version 6.02).
RESULTS
Data from a total of 309 participants were analyzed including 169 HIV-seronegative controls [median age (range): 48 (20-84) years, 67.5% male], 88 VS HIV-positive individuals with undetectable viral load [<50 RNA copies per milliliter, 97.7% receiving cART, median age: 48 (20-70) years, 81.8% male], and 52 viremic HIV-positive individuals with detectable viral load [VL > 50 copies per milliliter (range: 80 to >100,000 copies per milliliter), 19.2% receiving cART, median age: 40 (19-63) years, 82.7% male]. Clinical and demographic information is detailed in Table 1.
To identify markers of monocyte activation that altered significantly with age, we performed linear regression modeling on all biomarker data and adjusted for HIV infection, sex (which we have previously shown influences a number of these biomarkers10), and current smoking status. Plasma levels of CXCL10, sCD163, neopterin, and the proportions of monocyte subsets (defined as classical, CD14++CD16-; intermediate, CD14++CD16+; and nonclassical, CD14+CD16++) were all significantly altered with age independent of HIV infection (P < 0.05 for all, see Table S1, Supplemental Digital Content, http://links.lww.com/QAI/A642), confirming and extending our previous findings using age as a categorical variable that these parameters represent biomarkers of age-related monocyte activation.10 Although routinely used as biomarkers of monocyte activation in HIV studies, neither sCD14 nor LPS levels were significantly associated with age in this study. Sex affected the relationship between a number of the biomarkers analyzed (including monocyte subsets, CXCL10, sCD163 and LPS, see Table S1, Supplemental Digital Content, http://links.lww.com/QAI/A642) and age, confirming the need to adjust for sex in these analyses, although adjustment for smoking did not alter the relationship between any of the biomarkers and age in the control cohort (data not shown).
We have previously shown in a small cross-sectional study that younger HIV-positive individuals (aged < 45 years) exhibit higher levels of monocyte activation than age-matched controls, and this effect is only partially ameliorated by viral suppression.5 We thus used linear regression to extend these findings in an expanded cohort of HIV-positive and HIV-seronegative individuals aged 20-84 years using statistical analysis that allowed appropriate adjustment for the effect of sex and smoking status (due to the increased prevalence of current smokers in many HIV-positive cohorts including ours; Table 1). We also sought to determine whether the impact of HIV on monocyte activation is amplified at older ages. Biomarkers that were significantly altered with age in controls showed a similar trend in HIV-positive individuals (data not shown).
Table 2 shows the impact of HIV infection in addition to age-related changes on each parameter and indicates that proportions of intermediate monocytes were significantly increased and the proportions of classical monocytes consequently decreased, in viremic HIV infection (P = 0.002 and 0.015, respectively, column "P value vs HIV-") but not in VS individuals. In contrast, levels of soluble age-related monocyte activation markers CXCL10, neopterin, and sCD163 were significantly elevated in both viremic (P ≤ 0.001 for all, column "P value vs HIV-") and VS individuals (P < 0.001 for all). Although it was not significantly altered by age, sCD14 was increased in both viremic and VS HIV-positive individuals as compared with controls (P ≤ 0.001 for both). We also determined whether the slope of the regression curve, which indicates the rate at which these parameters change with age, differed between HIV-positive and control individuals (column "P value slope"; Table 2). The rate of change for nonclassical monocytes was altered in HIV-positive individuals; however, the overall effect of HIV infection on this parameter was not significant. There was a significant difference in the slope of the regression curve between viremic and control individuals for neopterin (P = 0.012, Table 2; Fig. 1), indicating age-related changes to neopterin occur at a greater rate in those with uncontrolled HIV replication. However, no such differences in slope were detected for soluble biomarkers in VS HIV-positive individuals (P > 0.05, Table 2; Fig. 1; Note similar slope of the regression curves for VS HIV-positive and uninfected individuals), indicating that although HIV increases the overall level of these markers, subsequent age-related changes occur at the same rate in VS HIV-positive individuals as in uninfected controls. These data confirm that HIV significantly alters age-related changes to monocytes independent of sex and smoking and that viral suppression associated with cART does not normalize the levels of soluble monocyte activation markers.
To assess the level of redundancy of each marker and determine whether 1 or more markers could be used to accurately predict other markers of monocyte activation, we next analyzed the associations between each of the monocyte activation markers. CXCL10 levels significantly correlated with all other soluble monocyte activation markers measured except sCD14 and also with proportions of classical and intermediate monocyte subsets (Table 3). Neopterin levels also correlated significantly with a number of soluble markers (CXCL10, sCD163, and sCD14) and intermediate but not classical monocyte subsets. In this cohort, sCD14 showed a significant association with only neopterin and not with monocyte subsets. Taken together, these data indicate that CXCL10, and to a lesser extent neopterin, levels are indicative of other soluble monocyte activation markers and monocyte subset proportions.
To determine whether CXCL10 also correlated with other commonly measured cellular markers of monocyte activation, in a subset of participants, we measured the association between CXCL10 levels and expression of functional/activation markers including CD38,21 which is upregulated on monocytes from HIV-positive individuals irrespective of cART (unpublished data, Hearps AC, 2013). We found that CXCL10 levels correlated significantly with expression of the activation markers CD11b and CD38 and the procoagulant protein tissue factor on classical and intermediate monocyte subsets (Table 3). The sCD14 levels did not correlate with any of these markers, whereas neopterin and sCD163 correlated with CD11b and tissue factor levels only. Taken together, these data suggest that CXCL10, but not sCD14, is a robust and easy to measure plasma marker of monocyte activation that can be used to indicate other soluble and importantly the more difficult to analyze cellular markers of monocyte activation under both healthy and inflammatory states.
To quantify the effect of HIV infection on monocyte activation and estimate the age at which HIV-positive individuals reach a given level of monocyte activation, we used a similar regression model as above but where age was the outcome of the regression equation. We selected CXCL10 to use in this analysis, having established it above as a robust biomarker of age and HIV-related monocyte activation and an accurate indicator of other monocyte activation markers. In this analysis, the coefficient for the viremic or VS HIV adjustment term indicated the difference in age between HIV-positive individuals and uninfected controls for any given level of CXCL10. In VS HIV-positive individuals, the value of the coefficient was -4.0 [95% confidence interval (CI): -9.1 to 1.1, see Table S2, Supplemental Digital Content, http://links.lww.com/QAI/A642], indicating that VS HIV-positive individuals exhibit the same level of CXCL10 as seronegative controls aged approximately 4 years older. The coefficient for viremic HIV-positive individuals indicated a more dramatic effect, in that these individuals reached this level of activation 12.0 (95% CI: -18.9 to -4.9) years earlier than uninfected controls. We validated this statistical approach using other biomarkers shown to be elevated in HIV infection and found strikingly similar coefficient values (ranging from -9.7 to -12.9 and -1.8 to -3.7 years for viremic and VS HIV infection, respectively, see Table S2, Supplemental Digital Content, http://links.lww.com/QAI/A642). In summary, these data indicate that viremic HIV-positive individuals exhibit levels of monocyte activation markers similar to those seen in seronegative controls aged up to 12 years older, and although viral suppression significantly reduces this effect, VS HIV-positive individual still experience levels of monocyte activation similar to those observed in controls aged up to 4 years older.
DISCUSSION
Although HIV infection is known to be associated with innate immune dysfunction and the premature appearance of age-related changes, the extent to which age and HIV contribute to these changes in HIV-positive individuals, and whether these effects are completely reversed by cART, has remained unclear. We report that viremic HIV-positive individuals display a similar level of monocyte activation to HIV- controls aged 12 years older, and although viral suppression significantly reduced the level of biomarkers such as CXCL10, VS HIV-positive individuals still show an equivalent level of monocyte activation 4 years earlier. Strikingly similar estimated effects of HIV in year equivalents were observed for a range of soluble and cellular markers of monocyte/innate immune activation, supporting the validity of this analysis. These findings suggest that in addition to the effects of aging, viremic and VS HIV infection have an additional burden of innate immune activation equivalent to up to 12 and 4 years of normal aging, respectively.
In addition to increasing the absolute level of monocyte activation biomarkers, viremic HIV infection was associated with an accelerated rate of change of the biomarker neopterin with age, indicating a potentiation of age and HIV viremia on factors driving the production of this biomarker. For all other biomarkers, our analysis revealed that HIV infection heightened monocyte activation but did not accelerate the rate of subsequent age-related changes (indicated by parallel regression curves in Fig. 1). This suggests that HIV infection is associated with a higher absolute level of monocyte activation irrespective of age but that age-related changes continue to occur at a rate similar to uninfected controls.
Although cross-sectional studies such as these are a useful, predictive way of estimating likely age-associated effects, the observations made here require confirmation in a longitudinal study of individuals infected at a similar age. However, due to the decades-long follow-up period required to conduct such analyses, a retrospective study design will likely be required.
Initiating cART early after infection (within 6 months) has been shown to reduce the extent of T-cell activation as compared with those who initiate cART later,22although whether early cART initiation would similarly reduce subsequent innate immune activation has not been determined. Duration of HIV infection and viral suppression are both likely to impact the degree of innate immune activation, although the former cannot usually be determined with accuracy in cross-sectional studies. Nadir CD4 T-cell count is often used in VS individuals as a surrogate marker for the extent of immune damage incurred before cART initiation. In our cohort, there was no significant difference in CXCL10 or neopterin levels in VS HIV-positive individuals with nadir CD4 T-cell counts above versus below 200 cells per microliter (P > 0.05 for both, data not shown). The estimated median duration of cART for VS patients in this study was 4.8 years; however, this only included treatment received at the attending clinic, thus we did not adjust for duration of therapy in our study. We consider that the effect of both early cART initiation and duration of therapy on immune activation is best investigated using a contemporary cohort of well-characterized patients receiving optimal cART regimens.
In this analysis, we adjusted for sex, which we have shown affects monocyte activation biomarkers levels,10 and also determined the effect of current smoking status, due to its established effects on inflammation.23 Although smoking status did not alter the relationship between the biomarkers and age in control individuals, it did influence levels in HIV-positive individuals; plasma LPS levels were significantly elevated in viremic and VS HIV-positive individuals as compared with controls before but not after adjustment for smoking (data not shown), demonstrating the necessity to control for smoking status when comparing LPS levels in HIV studies. Although BMI was not significantly different between HIV-positive and seronegative individuals, concurrent illness including hepatitis C coinfection (more prevalent in HIV-positive individuals), sociodemographic differences, and other lifestyle factors may also affect our results.
Cytomegalovirus (CMV) infection significantly impacts immunosenescence associated with aging,24 and CMV seropositivity is ubiquitous in most HIV-positive populations, making it difficult to delineate the discrete effects of CMV and HIV. CMV reactivation can occur during HIV infection and healthy aging, but it is not known how the timing and/or frequency of reactivation impacts on immune activation. For these reasons, CMV status was not adjusted for here, and although these results clearly demonstrate that immune dysfunction occurs in HIV-positive individuals, it is entirely possible that this may result from the combined effects of HIV and CMV.
Studies investigating monocyte activation in HIV infection typically measure plasma levels of sCD14, LPS, or the proportion of monocyte subsets (or a combination of) as biomarkers of monocyte activation. Accurate determination of plasma LPS levels can be influenced by inhibitory binding proteins in plasma25 and a lack of robustness in the analytical assays used, and is therefore a less reliable biomarker. sCD14 is produced after monocyte stimulation and is thus considered indicative of monocyte activation; however, it can act systemically to inhibit LPS responses,26 complicating the interpretation of elevated plasma sCD14 levels. We have shown both here and previously10 that although monocyte activation biomarkers such as CXCL10 and neopterin are increased during aging, sCD14 levels are not, suggesting that these soluble biomarkers may indicate different types of monocyte activation. Indeed, in our analysis, sCD14 levels correlated only with neopterin levels but no other soluble biomarkers or proportions of monocyte subsets. CD16+ monocyte subsets are often considered "inflammatory"27 and thus a marker of inflammation-induced monocyte activation as they produce large amounts of proinflammatory cytokines after stimulation28 and are present at elevated levels in various inflammatory settings,29 including viremic HIV infection. The lack of correlation between sCD14 levels and proportions of inflammatory monocyte subsets shown here suggest that sCD14 levels may only partially predict the extent of monocyte activation and should therefore be interpreted with caution. In contrast, CXCL10 levels were associated with all other soluble biomarkers measured (except sCD14 as discussed above), with proportions of intermediate and classical (although not nonclassical) monocyte subsets, and also with other cellular markers of monocyte activation/coagulation (ie, CD11b, CD38, and tissue factor). These data suggest that CXCL10 can be used to indicate a range of monocyte activation outcomes and is thus a useful biomarker warranting inclusion in studies of monocyte activation. CXCL10 can be produced from many different cell types including neutrophils and endothelial cells after IFNγ stimulation and is therefore not an exclusive marker of monocyte/macrophage activation. However, our data suggest that CXCL10 levels may be a more sensitive marker than monocyte-specific factors such as sCD14 in this setting. Furthermore, monocytes are the primary source of CXCL10 from PBMC in HIV-positive individuals, and plasma CXCL10 levels correlate significantly with CXCL10 mRNA levels in monocytes.30 In the general population, plasma CXCL10 levels are increased in individuals with a range of conditions including colorectal cancer,31 frailty,31,32 hypertension, and coronary heart disease and correlate significantly with traditional CVD biomarkers including CRP, IL-6, and fibrinogen.33,34 Plasma CXCL10 levels are increased in HIV infection (including in HIV controllers35) and are associated positively with T-cell activation and negatively with CD4 T-cell counts.35,36 CXCL10 is a neurotoxic factor that is elevated in the cerebrospinal fluid of HIV-positive individuals and is further increased in those with AIDS-related dementia37,38; further work is required to assess the predictive value of plasma CXCL10 levels in morbidities such as frailty and CVD in HIV infection.
Although it has been speculated that HIV may accelerate biological aging, this is likely an oversimplification of the diverse and incompletely understood mechanisms that contribute to immunological aging. The data presented here indicate that HIV infection heightens age-related changes such that levels of monocyte activation reflect those observed in HIV- individuals who are significantly older and that this effect is only partially ameliorated by cART. Given the established associations between these activation biomarkers and inflammatory disease, this likely confers increased disease risk, although this requires investigation. Determining the pathological consequences of these changes, and the mechanism underlying them, is required to prevent inflammatory diseases in the increasingly aging VS HIV-positive population.
|
|
| |
| |
|
|
|
