icon-    folder.gif   Conference Reports for NATAP  
  19th Conference on Retroviruses and
Opportunistic Infections
Seattle, WA March 5 - 8, 2012
Back grey_arrow_rt.gif
CROI 2012: Studies on Inflammation in HIV
  David H. Shepp, MD
Associate Professor of Medicine
Hofstra North Shore-LIJ School of Medicine
North Shore University Hospital - Manhasset, NY

The increasing use of combination antiretroviral therapy (ART) worldwide has led to a dramatic decline in AIDS-related illnesses in people infected with HIV. During the past decade, non-AIDS illnesses, including cardiovascular disease (CVD), non-AIDS infection and cancer, and liver failure, have become the most common cause of death. Despite gains due to ART, life expectancy, remains shorter for HIV+ individuals than in the general population. Along with traditional risk factors such as smoking, obesity, diabetes, and dyslipidemia, chronic immune activation and inflammation contribute to the pathogenesis of these conditions. Inflammation is present in untreated HIV and improves but does not fully resolve with long-term ART. The causes of persistent immune activation are not fully understood. Microbial translocation due to persistent immunologic injury in gut-associated lymphoid tissue is a major contributor. Other possible causes include chronic co-infections such as CMV, and persistent, low-level HIV infection that eludes ART. Understanding the causes and consequences of chronic inflammation in HIV and developing ways to control it are important areas of contemporary HIV research. As a result, numerous abstracts at CROI 2012 focused on inflammation.

Consequences of Inflammation - CVD. The relation between CVD and inflammation in well established both in HIV and the general population. Soluble CD14 (sCD14) is a marker of monocyte/macrophage activation, which is pathogenically important in the development of atherosclerotic plaque. sCD14 is also used as a marker of microbial translocation, since CD14 is a major receptor for bacterial lipopolysaccharide (LPS). Kelisides et al. (abstr. 122) presented a study of HIV+ individuals (n=55) and matched HIV-negative controls (n=36), all at fairly low risk for CVD, who had baseline biomarkers measured and were followed longitudinally with measurement of carotid intima-media thickness (cIMT), an established correlate of atherosclerotic vascular disease. In an multivariable analysis adjusting for many potentially confounding variables, baseline levels of sCD14 and LPS were strongly and independently correlated with the rate of increase of cIMT over 96 weeks of follow-up. This correlation was seen in the HIV+ but not the HIV-negative group.

Other Complications. Three presentations examined the relation between inflammatory biomarkers and complications not previously well studied: hypertension, peripheral sensory neuropathy (PSN), and chronic kidney disease (CKD). Investigators from Norway measured LPS and sCD14 before ART in a subset of normotensive patients (n=42) from a larger prospective cohort study of hypertension in HIV (Manner IW, abstr. 814). In multivariable analysis adjusting for other potential risk factors, the risk of developing of hypertension during follow-up was strongly and independently associated with baseline levels of either LPS or sCD14.

The pathogenesis of painful sensory in HIV is not well understood. Reduced use of dideoxy-nucleosides has greatly lessened the severity of the problem in patients on ART, but symptomatic and subclinical cases still occur. Pathologic studies suggest monocyte activation is present in dorsal nerve root ganglia. Investigators from the CHARTER study sought evidence of monocyte activation in 78 patients with or without a diagnosis of PSN by measurement of sCD14 in the CSF (Ellis R, abstr. 495). All were virologically suppressed on ART. Levels of sCD14 were significantly higher in those with PSN while soluble TNF receptor 2, an activation marker not specific for monocytes, was similar in both groups. However, the basic characteristics of groups with and without PSN differed considerably and no adjusted analysis was done to determine the independence of the association, possibly due to the small study size. This limits the strength of the observed association.

Although an uncommon cause of death, development of CKD is a relatively common morbidity in HIV+ patients on ART. The role of inflammation as a risk factor has not been extensively evaluated. Neuhaus et al. measured hsCRP, IL-6 d-dimer at baseline in a large cohort (n=3403) drawn from the SMART and ESPIRIT studies (abstr. 830). CKD was defined by GFR decline criteria. Information on proteinuria, needed for a more comprehension definition of CKD, was not available. After adjustment for multiple confounding variables, none of the markers correlated with CKD.

Mortality. Elevated plasma markers of activation of the innate immune system and the coagulation cascade, such as hsCRP, IL-6, sCD14 and d-dimer have been associated with an increased risk of death in patients initiating ART. T-cell activation markers also have been correlated with mortality in untreated patients. It is not known which markers have the best predictive value or are directly involved in the pathogenesis of these illnesses. Hunt et al. (abstr. 278) measured a variety of biomarkers in patients who died from non-violent causes and in matched controls from two longitudinal cohort studies (SOCA and SCOPE). Markers of innate immunity (IL-6, hsCRP, sCD14, sTNFR-1)[from Jules: hsCRP & IL6 are common inflammation markers], coagulation (d-dimer) and intestinal epithelial integrity (intestinal fatty acid binding protein [I-FABP], zonulin), all were stronger correlates of mortality than were markers of T-cell activation markers.

Immune Recovery. Previous studies have shown that one of the consequences of higher levels of T-cell activation is poorer CD4+ cell recovery on ART. This observation was extended by Zhang et al (abstr. 274) who looked at 120 early participants in the ALLRT cohort, which provides long-term follow-up of participants in ACTG trials. After 10-15 years of suppressive ART, those with higher levels of CD8+ activation had significantly lower CD4+ counts (about 100 cells less) than those with lower activation levels.

Differential Effects of ART. Contemporary ART offers a choice among many regimens that are highly effective and well tolerated. However, little is known about differences among regimens in their ability to reduce and potentially normalize the chronic inflammation. Several abstracts began to shed some light on this very important and inadequately studied area of antiretroviral research.

Raltegravir. The effects of switching to raltegravir (RAL) or remaining on ritonavir-boosted protease inhibitors (PIs) were compared in a subset of patients in the SPIRAL study, a randomized open-label 48 week study in virologically suppressed patients (Martinez E, abstr. 834). At the end of the study, patients switching to RAL had significantly lower levels of hsCRP, IL-6, MCP-1, osteoprotegerin and TNF-alpha relative to those staying on PIs. Insulin and d-dimer levels were also lower, as were lipids, while markers of endothelial function did not differ. A second study (EASIER) involved switching from enfuvirtide to RAL (Silva E, abstr. 840). After 24 weeks of randomized therapy, all participants received RAL. IL-6, hsCRP, and d-dimer declined at week 24 then stabilized in the RAL arm, while levels were unchanged until week 24 then declined in the group that initially remained on PIs then switched. A third study intensively measured markers of immune activation, microbial translocation and T-cell exhaustion in 15 treatment-naive patients initiating RAL-containing ART. Comparison was made to historical controls who had received a similar duration of non-RAL ART, and to HIV-negative controls (Pallikkuth S, abstr. 277). At 24 weeks, levels of activated CD4+ and CD8+ T-cells, PD-1+/CD4+ and CD8+ T-cells, plasma LPS and sCD14 were all significantly reduced from baseline, to levels that were significantly lower than those in the historical controls. However, activated CD8+ and LPS levels remained higher than in HIV-negative controls.

Other Antiretrovirals. Maraviroc (MVC) is an antiretroviral that targets the host cell surface receptor CCR5, the major co-receptor for HIV entry. The natural ligands are the cc-chemokines, so inhibiting CCR5 may have immunomodulatory effects. Vascular inflammation, especially monocyte activation, plays a critical role in endothelial dysfunction and the pathogenesis of atherosclerosis. The naturally occurring deletion in CCR5 gene in humans and animal studies of CCR5 inactivation may reduce atherosclerosis. Previous studies have examined immune activation during MVC intensification of ART with conflicting results. In a randomized, placebo controlled trial, Hsue et al.(abstr. 123) examined the effect of MVC intensification on flow-mediated dilation (FMD) of the brachial artery, a marker of endothelial dysfunction and CVD risk. Fifty-two patients virologically controlled on ART were studied. Compared to placebo, and in contrast to expectations, 24 weeks of MVC intensification had no effect on FMD, and sCD14 and sCD163, markers of monocyte activation, were increased. MVC also increased plasma levels of the cc-chemokine MIP-1-beta, which may mediate monocyte activation via alternate receptors.

Several large cohort studies have found an association between abacavir (ABC) use and increased risk of myocardial infarction or other cardiovascular diseases, but the findings are controversial because other studies, including some randomized trials have been negative. Identification of a plausible mechanism for increased risk could help clarify the controversy. ACTG 5202 randomized treatment-naive patients to one of four commonly used ART regimens: either tenofovir (TDF)/emtricitabine (FTC) or ABC/lamiviudine (3TC) combined with either efavirenz (EFV) or atazanavir/ritonavir (ATV/r). In a substudy (n=269), McComsey et al. measured a battery of inflammatory and endothelial markers at baseline, weeks 24 and 96 (abstr. 835). Most markers studied (TNF-alpha, sTNFR-I, sTNFR-II, sICAM-1, sVCAM-1) declined at 24 and 96 weeks and did not differ among the four study arms. hsCRP rose with either ABC/3TC or EFV, and the difference between ABC/3TC and TDF/FTC was significant at both 24 and 96 weeks. This study was conducted without the requirement for HLA-B5701 testing but the differences in hsCRP levels did not appear to be attributable to patients on ABC/3TC who had hypersensitivity reactions. The difference between EFV and ATV/r in hsCRP was significant only at week 24. IL-6 levels fell more at week 24 with TDF/FTC than with ABC/3TC but not week 96. A second smaller study examined a cross-section of patients who originally participated in randomized trials of ABC (n=46) vs. TDF(n=72). In this study, after adjustment for multiple confounding variables, there were no significant differences in hsCRP, d-dimer or in endothelial function as assessed by FMD (Wohl D, abstr. 838).

Possible differences between lopinavir/ritonavir or EFV-containing ART on markers of microbial translocation were assessed in a subset (n=71) of a larger randomized trial conducted in treatment-naive scandinavian patients (Barqasho B, abstr. 836). After 72 weeks of treatment, sCD14 and anti-flagellin antibodies declined significantly, without differences between regimens. I-FABP, a marker of intestinal mucosal disruption and cell death, increased from baseline but the change only achieved statistical significance in the EFV arm. No reason for the observed discordances between certain markers was apparent.

Adjunctive Therapies . Although studies like those discussed above suggest some ART regimens may improve inflammation better than others, ART alone may not fully normalize inflammation in all patients. The search for therapies to use in conjunction with ART begins with agents that have demonstrated both anti-inflammatory properties and safety in other disease states. One such group of drugs are statins, which reduce CVD to a greater extent than would be predicted by lipid lowering effects alone, reduce immune activation and pro-inflammatory cytokine levels and may confer benefits in rheumatoid arthritis and pneumonia. Previous studies in HIV have shown reduced T-cell activation in patients not on ART and a cohort analysis has suggested a possible survival advantage. ACTG investigators sought evidence for a reduction in clinical events among statin users in the ALLRT cohort (Overton ET, abstr. 124). The study evaluated 3600 patients and over 15,000 patient-years of follow-up. Patients initiating statins were found to have a 19% reduction in the composite end-point of total clinical events, a trend that was not statistically significant. Among the individual end-points examined, there was a significant reduction among statin users in non-AIDS cancers. While this study does not provide good evidence for statins as adjunctive therapy in HIV, it does not negate the existing data that suggest benefits. Controlling for confounding variables in a non-randomized cohort analysis is very challenging, especially when using a composite end-point of multiple diseases with numerous known risk factors. Also, some statins are more potent than others. Information on the specific agents and dosage used was not provided. Of note, no benefit of statins was found for CVD events, suggesting a major limitation of the study.

Other classes of commonly used drugs that may have anti-inflammatory properties are ACE inhibitors and angiotensin receptor blockers (ARBs). A small placebo-controlled pilot study randomized patients virologically suppressed on ART to lisinopril, pravastatin, both or neither (Baker J, abstr. 825). After 4 months, hsCRP, TNF-alpha and diastolic blood pressure were reduced with lisinopril compared to placebo. Pravastatin had no effect on inflammatory markers or lipids, consistent with it's relatively low potency among statins. Pravastatin has been favored among statins for use in HIV, primarily because of its lack of interaction with commonly used antiretrovirals. However, studies of other more potent statins at appropriately adjusted doses may be of greater interest.

Telmisartan is an ARB reported to have beneficial effects on monocyte activation and visceral abdominal fat in HIV-negative subjects. A small uncontrolled pilot study gave telmisartan to HIV+ subjects with lipohypertrophy and good virologic suppression on ART (Lake J, abstr. 826). After 24 weeks, the monocyte activation markers sCD14 and sCD163 were not improved and were, in fact, found to increase. The reason for the apparent difference in patterns seen in this study and those reported in HIV-negatives are not clear.

Effects of Pregnancy . To prevent mother to child transmission, all HIV+ pregnant women not already on ART should initiate treatment. Although it is very effective, ART is associated with an increased risk of pre-term delivery (PTD), especially with PI-containing regimens. The mechanism is not known. Normal pregnancy results in significant immunologic alterations, including a shift in the functional profile of T-helper lymphocytes from a pro-inflammatory state (called Th1) to an anti-inflammatory/pro-allergic reaction state (called Th2). This shift helps prevent the mother from immunologically rejecting the fetus, as she might a transplanted organ, and is important for maintenance of pregnancy. The anti-inflammatory Th2 cytokine IL-10 has previously been identified as an important mediator of full-term pregnancy. Untreated HIV tends to resemble a Th2 state. ART shifts the T-helper profile back toward a Th1 state. To investigate the relationship between the immunologic changes and PTD, Short et al. (abstr. 1055) made serial measurements of cytokine production in a cohort of HIV+ and HIV-negative pregnant women. All cytokines tended to be higher in HIV+ vs HIV- women. Th1 cytokines did not appear to vary consistently according to type of ART. However, IL-10 production was significantly lower throughout pregnancy in women receiving multi-class ART (PI or NNRTI) compared to those receiving NRTI-only regimens (combination or monotherapy). Six of 51 HIV+ women, all receiving PIs, and 1 of 12 HIV-negative women experienced PTD. Most women receiving multi-class ART were on it throughout pregnancy, while most women receiving NRTI-only started in the second trimester, limiting the comparability of the data. Also, virologic suppression was likely more complete in the multi-class ART recipients. Therefore, some of the observed differences may reflect more effective ART rather than specific drug class effects. However, the low levels of IL-10 found in those on multi-class ART, predominantly with PIs, provides a plausible mechanism for the increase in PTD seen in HIV+ women on ART in pregnancy.

Some women without other strong indications for ART may choose to discontinue it after delivery. Treatment interruptions in other patient groups have resulted in a surge inflammatory biomarkers, an associated increased risk of disease events, and are not recommended. To determine the effect of pregnancy, delivery and ART discontinuation on inflammatory biomarkers, Hoffman et al (abstr. 1056) studied hsCRP, IL-6 and d-dimer in late pregnancy (>20 weeks), at delivery and 6 weeks postpartum. All 3 markers rose substantially from late pregnancy to delivery, then dropped rapidly postpartum to levels similar to or below those found in late pregnancy. This pattern was seen in women who did or did not continue ART after delivery. D-dimer, but not IL-6 or hsCRP declined somewhat less in those who discontinued ART. Substantial early declines in inflammatory markers in those discontinuing ART suggests the immunologic effects of termination of pregnancy can override the usual effects of ART discontinuation.

Summary. The large amount of data on inflammation in HIV at CROI 2012 has brought some new insights and suggests future avenues of research, although there are few definitive conclusions that can influence clinical management. There was further evidence that CVD is driven in part by chronic inflammation. Hypertension and possibly painful sensory neuropathy may be added to list of medical complications that are influenced by chronic inflammation in patients on ART. CKD, at least as defined by GFR declines alone, may be influenced mainly by other factors. Activation of both the innate and adaptive immune systems drive chronic inflammation in HIV, but markers of innate immunity appear to have a stronger predictive value for mortality. Poorer CD4+ T-cell recovery is linked to higher levels of immune activation and that effect may persist even after many years of ART. Comparisons of the effects of various antiretrovirals on inflammation are starting to appear. Early results suggest possible advantages for raltegravir and tenofovir/emtricitabine. Larger comparative studies of regimens combining these agents would be of great interest. Studies to determine if similar effects are seen with the new, investigational drugs in the integrase class, elvitegravir and dolutegravir, and investigational tenofovir prodrug GS7340 should also be conducted. Studies of adjunctive therapies to reduce inflammation are very preliminary. Potent statins and possibly ACE inhibitors or ARBs warrant further evaluation. An ongoing ACTG randomized trial is comparing the short term effect on inflammatory biomarkers of statins, ART or both. Ultimately, any beneficial reductions in biomarkers of inflammation seen with either specific ART regimens or adjunctive therapy will require correlation with clinical outcome. Differences in cytokine profile seen in pregnancy with different ART regimens provide a potential mechanism for the increased risk of pre-term delivery in HIV. The sharp decline in inflammation seen in the postpartum period could mean ART discontinuation is less problematic during this period but longer follow-up and monitoring for clinical events are needed.