Abstract
Objectives
One of the goals of antiretroviral therapy (ART) is to attenuate HIV-induced systemic immune activation and inflammation. We determined the dynamics of biomarkers of immune activation, microbial translocation and inflammation during initial ART with a nucleos(t)ide-sparing regimen of darunavir/ritonavir plus raltegravir. We also evaluated associations between these biomarkers and the virological response to the regimen.
Methods
We determined baseline and week 24 and 48 levels of CD4+ and CD8+ T cell activation (% HLA-DR+/CD38+), interleukin-6 (IL-6), interferon-γ-inducible protein-10 (IP-10), soluble CD14 (sCD14), D-dimer and lipopolysaccharide. Associations between the biomarkers at baseline were assessed using Spearman's rank correlation. The Wilcoxon signed rank test analysed changes from baseline. Comparisons between groups were made using the Wilcoxon rank sum test, and Cox proportional hazards models assessed predictors of virological failure (VF).
Results
Assays were completed on 107 of 112 subjects after excluding five subjects who had only baseline samples. The subjects included were 94 (88%) men with a median age of 37 years, a median baseline CD4 count of 261.5 cells/mm3 and a median baseline viral load (VL) of 75 876 copies/mL. Subjects with a baseline VL >100 000 copies/mL had higher baseline T cell activation, IL-6, IP-10, sCD14 and D-dimer. These biomarkers declined during treatment (P < 0.05). Although subjects who experienced VF had higher baseline CD4+ T cell activation (P = 0.035), only baseline VL independently predicted VF (hazard ratio for >100 000 versus ≤100 000 copies/mL was 4.5–5.6, P ≤ 0.002).
Conclusions
Darunavir/ritonavir plus raltegravir attenuated immune activation, inflammation and microbial translocation. T cell activation remained higher in subjects with VF than those without. Baseline VL >100 000 copies/mL remained the primary driver of VF.
Keywords: nucleos(t)ide sparing, soluble CD14, microbial translocation
Introduction
In the ACTG A5262 trial, treatment-naive HIV-1-infected patients received a nucleos(t)ide-sparing regimen of 800/100 mg of darunavir/ritonavir once daily plus 400 mg of raltegravir twice daily for 52 weeks.1 This regimen was effective in most patients, but 26% experienced protocol-defined virological failure (VF) by week 48, approximately two-thirds of whom had a viral load (VL) between 50 and 200 copies/mL at VF. Baseline VL >100 000 copies/mL strongly predicted VF.
One of the goals of antiretroviral therapy (ART) is to attenuate the systemic immune activation and inflammation induced by HIV.2 Whether these factors and microbial translocation independently influence virological outcomes has not been clearly demonstrated. We investigated the dynamics of biomarkers of immune activation, microbial translocation and inflammation in the A5262 trial and explored the relationships between these biomarkers and VF.
Materials and methods
Peripheral blood mononuclear cells were isolated by density centrifugation from whole blood samples collected pre-entry and at weeks 0 (entry), 24 and 48 of therapy. Samples were cryopreserved and analysed in batches at the end of the study. T cell activation (percentage of CD4+ and CD8+ T cells coexpressing HLA-DR and CD38) was measured by flow cytometry.3 To assay soluble biomarkers of immune activation and inflammation, frozen plasma in EDTA anticoagulant or frozen serum was thawed once and analysed in batch. Interleukin-6 (IL-6), interferon-γ-inducible protein-10 (IP-10) and soluble CD14 (sCD14) were measured in plasma by ELISA (R&D Systems, Minneapolis, MN, USA). D-dimers were measured using the Asserachrom D-DI immunoassay (Diagnostica Stago, Asnieres, France). For measurement of bacterial lipopolysaccharide (LPS), serum was diluted to 10% or 20% with endotoxin-free water and then heated at 85°C for 15 min to denature the proteins. LPS levels were then quantified using a modified Limulus Amebocyte Lysate (LAL) assay (Lonza, Walkersville, MD, USA).
Baseline was defined as the average of the pre-entry and entry values except for LPS, for which the baseline was the value obtained at entry (in a confirmed fasting state). Associations between biomarkers at baseline were assessed using Spearman's rank correlation. The Wilcoxon signed rank test analysed changes from baseline; comparisons between groups were made using the Wilcoxon rank sum test, and Cox proportional hazards models assessed predictors of VF. Tests were performed using a 5% level of significance with no adjustment for multiple testing. Ethics committees at each research site approved the study. Each participant provided written informed consent.
Results
Biomarker levels at baseline
Assays were completed on 107 of 112 subjects after excluding five subjects who had only baseline samples. The subjects included were 94 (88%) men with a median age of 37 years, a median baseline CD4 count of 261.5 cells/mm3 and a median baseline VL of 75 876 copies/mL. At baseline, median (IQR) CD4+ T cell activation was 11% (7%–20%) while median CD8+ T cell activation was 39% (29%–51%). Median values for IL-6, LPS, IP-10, sCD14 and D-dimer were 1.38 pg/mL (0.81–2.37 pg/mL), 21 pg/mL (10–35 pg/mL), 517 pg/mL (349–929 pg/mL), 2.42 × 106 pg/mL (1.96–3.08 × 106 pg/mL) and 202 ng/mL (111–362 ng/mL), respectively. Patients with baseline VL >100 000 copies/mL had higher baseline CD4+ and CD8+ T cell activation as well as higher IL-6, IP-10, sCD14 and D-dimer levels than those with VL ≤100 000 copies/mL (all P ≤ 0.035). Only LPS levels did not differ between the two baseline strata of VL (P = 0.90) (Table 1).
Table 1.
Immunological and inflammatory markers [median (IQR)] by baseline HIV-1 RNA
| Marker | Totala | Baseline HIV-1 RNA (copies/mL) |
|||
|---|---|---|---|---|---|
| ≤100 000b | >100 000c | P value* | |||
| CD4+/HLA-DR+/CD38+ % | baseline | 11 (7, 20) | 9 (5, 15) | 18 (9, 23) | <0.001 |
| week 24 | 5 (3, 9) | 5 (2, 6) | 8 (5, 12) | <0.001 | |
| week 24 Δ | −6 (−11, −2)# | −4 (−9, −2)# | −8 (−13, −3)# | 0.017 | |
| week 48 | 3 (2, 7) | 3 (2, 5) | 5 (3, 8) | 0.003 | |
| week 48 Δ | −7 (−14, −4)# | −5 (−11, −3)# | −11 (−16, −5)# | 0.006 | |
| CD8+/HLA-DR+/CD38+ % | baseline | 39 (29, 51) | 36 (26, 50) | 46 (33, 55) | 0.014 |
| week 24 | 16 (9, 23) | 11 (8, 19) | 18 (10, 27) | 0.007 | |
| week 24 Δ | −22 (−32, −13)# | −20 (−30, −13)# | −25 (−34, −14)# | 0.374 | |
| week 48 | 10 (7, 16) | 9 (6, 13) | 11 (9, 21) | 0.031 | |
| week 48 Δ | −27 (−37, −16)# | −24 (−33, −16)# | −30 (−39, −20)# | 0.084 | |
| IL-6 (pg/mL) | baseline | 1.38 (0.81, 2.37) | 1.11 (0.70, 2.01) | 1.73 (1.25, 3.15) | 0.003 |
| week 24 | 1.17 (0.65, 1.76) | 1.06 (0.65, 1.67) | 1.19 (0.66, 1.91) | 0.848 | |
| week 24 Δ | −0.32 (−0.97, 0.18)# | −0.14 (−0.71, 0.28)# | −0.61 (−2.08, −0.02)# | 0.040 | |
| week 48 | 1.17 (0.63, 1.66) | 1.13 (0.53, 1.66) | 1.21 (0.65, 1.69) | 0.526 | |
| week 48 Δ | −0.37 (−0.96, 0.22)# | −0.22 (−0.69, 0.26)# | −0.42 (−1.64, 0.12)# | 0.299 | |
| LPS (pg/mL) | baseline | 21 (10, 35) | 21 (10, 33) | 20 (10, 38) | 0.899 |
| week 24 | 26 (12, 35) | 26 (10, 35) | 25 (12, 36) | 0.800 | |
| week 24 Δ | 0 (−9, 12) | 1 (−8, 13) | 0 (−10, 10) | 0.557 | |
| week 48 | 22 (6, 34) | 16 (5, 34) | 25 (12, 33) | 0.478 | |
| week 48 Δ | −1 (−12, 8) | −3 (−14, 5) | 4 (−11, 12) | 0.176 | |
| IP-10 (pg/mL) | baseline | 517 (349, 929) | 425 (282, 673) | 690 (459, 1184) | <0.001 |
| week 24 | 180 (140, 326) | 175 (136, 332) | 194 (144, 280) | 0.642 | |
| week 24 Δ | −296 (−555, −144)# | −185 (−395, −98)# | −481 (−968, −260)# | <0.001 | |
| week 48 | 172 (123, 249) | 173 (122, 267) | 166 (123, 246) | 0.737 | |
| week 48 Δ | −344 (−649, −141)# | −255 (−447, −75)# | −522 (−999, −304)# | <0.001 | |
| sCD14 (×106 pg/mL) | baseline | 2.42 (1.96, 3.08) | 2.11 (1.71, 2.90) | 2.75 (2.16, 3.16) | 0.002 |
| week 24 | 2.03 (1.63, 2.37) | 2.02 (1.57, 2.28) | 2.06 (1.69, 2.63) | 0.377 | |
| week 24 Δ | −0.46 (−0.97, 0.08)# | −0.21 (−0.93, 0.12)# | −0.71 (−1.09, −0.25)# | 0.046 | |
| week 48 | 1.98 (1.65, 2.43) | 2.00 (1.62, 2.41) | 1.95 (1.68, 2.56) | 0.815 | |
| week 48 Δ | −0.41 (−1.01, 0.06)# | −0.22 (−0.70, 0.12)# | −0.71 (−1.24, −0.07)# | 0.021 | |
| D-dimer (ng/mL) | baseline | 202 (111, 362) | 160 (105, 332) | 266 (140, 513) | 0.035 |
| week 24 | 136 (73, 258) | 118 (70, 255) | 147 (82, 260) | 0.331 | |
| week 24 Δ | −45 (−130, −12)# | −34 (−113, −2)# | −64 (−170, −25)# | 0.036 | |
| week 48 | 130 (78, 251) | 144 (85, 255) | 129 (74, 232) | 0.566 | |
| week 48 Δ | −62 (−148, −12)# | −47 (−100, 11)# | −83 (−212, −39)# | 0.013 | |
aDepending on assay result availability, n ranges from 96 to 107.
bDepending on assay result availability, n ranges from 52 to 59.
cDepending on assay result availability, n ranges from 44 to 48.
*Exact Wilcoxon rank sum test comparing baseline HIV-1 RNA ≤100 000 and >100 000 copies/mL.
#Exact Wilcoxon signed rank test comparing changes from baseline: P < 0.05.
The baseline log10 VL correlated inversely with the baseline CD4+ T cell count (r = −0.44) and directly with CD4+ T cell activation (r = 0.39), CD8+ T cell activation (r = 0.28), IL-6 (r = 0.31), IP-10 (r = 0.43) and sCD14 (r = 0.33) (all P ≤ 0.003). A weaker direct correlation was present with D-dimers (r = 0.19, P = 0.05) and none with LPS (r = 0.01, P = 0.95).
Changes in the levels of biomarkers during treatment
Among the entire study population, all assessed biomarkers except LPS declined significantly from baseline to weeks 24 and 48 (Table 1). Median LPS levels at entry (21 pg/mL) and at weeks 24 (26 pg/mL) and 48 (22 pg/mL) were not significantly different (all P > 0.1).
The group with a baseline VL >100 000 copies/mL experienced a greater decrease in CD4+ T cell activation from baseline to week 24 (−8% versus −4%, P = 0.017) and week 48 (−11% versus −5%, P = 0.006) than did the group with a baseline VL ≤100 000 copies/mL. Despite the greater decline, CD4+ T cell activation remained higher in those with a baseline VL >100 000 copies/mL at week 24 (8% versus 5%, P < 0.001) and week 48 (5% versus 3%, P = 0.003). CD8+ T cell activation was also higher in the high VL group at weeks 24 (18% versus 11%, P = 0.007) and 48 (11% versus 9%, P = 0.031), although we did not detect a significant difference between the two groups in the decline in CD8+ T cell activation from baseline to weeks 24 (P = 0.374) and 48 (P = 0.084). Patients with a baseline VL >100 000 copies/mL had greater falls in most assessed markers on treatment, but absolute levels of IL-6, LPS, IP-10, sCD14 and D-dimer at weeks 24 and 48 were similar in the two baseline VL groups.
Associations between VF and levels of assessed biomarkers
In univariate analysis, baseline CD4+ T cell activation was higher among patients who subsequently experienced VF than among those who did not (median 16% versus 10%, P = 0.035). Baseline CD8+ T cell activation, IL-6, LPS, IP-10, sCD14 and D-dimers showed no association with subsequent VF (all P ≥ 0.086). During treatment, there was a consistent pattern in which the group with VF had higher T cell activation than the group without VF. This was observed for CD4+ T cell activation at week 24 (8% in those with VF versus 5% without VF, P = 0.001) and week 48 (6% versus 3%, respectively; P = 0.006); and also for CD8+ T cell activation at weeks 24 (21% versus 12%, respectively; P = 0.015) and 48 (17% versus 10%, respectively; P = 0.010). In contrast, we did not detect a significant association between any of the other assessed biomarkers and VF, except for higher D-dimer levels at week 24 among those with VF (177 versus 119 ng/mL, P = 0.025).
We next compared the change in assessed biomarkers from baseline to weeks 24 and 48 in patients with and without VF. For CD4+ T cell activation, no significant difference was observed between the two groups (all P > 0.05), while for CD8+ T cell activation, the decline in CD8+ T cell activation from baseline was less in patients with VF (P = 0.005 at week 24, P = 0.008 at week 48). No significant difference was detected in the change from baseline to weeks 24 and 48 for IL-6, LPS, IP-10, sCD14 or D-dimer in patients with VF versus those without VF (all P ≥ 0.057).
Because a high proportion of patients showing treatment failure in the A5262 trial had a VL of 51–200 copies/mL at VF, we conducted a subgroup analysis comparing those with a VL ≤200 copies/mL with those with a VL >200 copies/mL at VF (n = 17 versus n = 10, respectively). In this subgroup analysis, those with VL ≤200 copies/mL at VF had a higher baseline CD8+ T cell activation (44% versus 23%, P = 0.039). No difference was detected in baseline CD4+ T cell activation or in IL-6, LPS, IP-10, sCD14 and D-dimer levels between these subgroups (all P ≥ 0.115). CD4+ and CD8+ T cell activation at weeks 24 and 48 was similar in the two groups, as were all other assessed markers except sCD14, which at week 48 showed a higher concentration in the ≤200 copies/mL failure group (2.44 × 106 versus 1.73 × 106 pg/mL, P = 0.027).
Finally, multivariable Cox PH regression models for time to VF evaluating baseline VL and individual baseline biomarkers (CD4+ T cell activation, CD8+ T cell activation, IL-6, LPS, IP-10, sCD14 or D-dimer) showed that baseline VL >100 000 copies/mL was independently predictive of VF (hazard ratio for >100 000 versus ≤100 000 copies/mL was 4.5–5.6, P ≤ 0.002) No independent association was detected between the risk of VF and baseline levels of T cell activation, or with any of the soluble markers when baseline VL was included in the model.
Discussion
In this study, evaluated biomarkers of immune activation, inflammation and microbial translocation were, except for LPS, higher at baseline in patients with VL >100 000 copies/mL. Darunavir/ritonavir plus raltegravir was associated with a significant decline from baseline in biomarkers (IL-6, sCD14 and D-dimer) that are predictive of mortality.4–6 We did not observe a change in LPS with treatment, but are unable to dissect this finding from potential confounders. The median baseline LPS in our study was 21 pg/mL, ∼4-fold lower than that reported in other studies.7,8 This could have limited the sensitivity of the assay to discriminate differences at low levels.
As a group, patients with VF had consistently higher T cell activation during treatment compared with those without VF, suggesting that the magnitude of T cell activation during treatment reflects HIV levels, similar to observations for untreated infection.9 Only baseline VL >100 000 copies/mL remained independently predictive of VF after considering T cell activation or soluble inflammatory biomarkers. T cell activation levels were not significantly different between those with VL ≤200 versus >200 copies/mL at VF, although this analysis was limited by the small sample size. The persistence of heightened activation during ART has been reported even in some patients with VL under 50 copies/mL,10,11 especially those with poor CD4 recovery.12
In conclusion, darunavir/ritonavir plus raltegravir attenuated immune activation, inflammation and microbial translocation. T cell activation remained higher in patients with VF than in those without VF. Baseline VL above 100 000 copies/mL remained the primary driver of VF with this regimen even when accounting for baseline biomarker levels. Larger studies comparing this novel regimen to conventional nucleos(t)ide-containing therapy should further elucidate the potential interactions between T cell activation and VF.
Funding
Awards U01AI068636 and U01AI68634 from the National Institute of Allergy and Infectious Diseases, the National Institute of Mental Health (NIMH) and the National Institute of Dental and Craniofacial Research (NIDCR) supported this study. The study was also supported in part by grants funded by the National Center for Research Resources. Merck provided raltegravir and Janssen Therapeutics provided darunavir.
Transparency declarations
B. T. has served as an advisor for and/or received research support from Janssen, GlaxoSmithKline and ViiV. D. R. K. is a consultant to and/or has received research funding from Abbott, Boehringer-Ingelheim, Bristol-Myers Squibb, Gilead Sciences, GlaxoSmithKline, Merck, Roche and ViiV. J. J. E. is a consultant to Abbott, GlaxoSmithKline, Merck, ViiV and Janssen, and has received research support from GlaxoSmithKline and Merck. All other authors: none to declare.
The funders had no decision-making role in the preparation of this manuscript.
Disclaimer
The content of this manuscript is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of Allergy and Infectious Diseases or the National Institutes of Health.
Acknowledgements
We thank the study volunteers, the staff of ACTG Case Western Reserve and University of Pittsburgh Immunology Specialty Laboratories, and the following A5262 study team members: Barbara Bastow, Sarah W. Read, Jennifer Janik, Debra S. Meres, Lori Mong-Kryspin, Karl E. Shaw, Louis G. Zimmerman, Randi Leavitt (Merck) and Guy De La Rosa (Janssen). We also acknowledge the contributions of the following ACTG site investigators: Karen Coleman and Meredith Rathert [Northwestern University (Site 2701) CTU Grant AI069471]; Edward Seefried and Leticia Muttera [University of California San Diego (Site 701) CTU Grant AI 69432]; Michael F. Para and Heather Harbr [The Ohio State University (Site 2301) CTU Grant AI069474]; Robert Kalayjian and Ann Marie Anderson [MetroHealth Medical Center (Site 2503) CTU Grant AI-069501]; Kerry Upton and Jenna White [Alabama Therapeutics CRS (Site 5801) CTU Grant U01 AI069452]; Pablo Tebas and Aleshia Thomas [University of Pennsylvania (Site 6201) CTU Grant U01-AI-69467-05, CFAR Grant P30-AI-045008-12]; Annie Luetkemeyer and Jay Dwyer [UCSF AIDS CRS (Site 801) CTU Grant 5UO1 AI069502]; Mariea Snell and James Conner [Washington University in St Louis (Site 2101) CTU Grant AI 069495]; Nathan M. Thielman and Jacquelin Granholm [Duke University Medical Center CRS (Site 1601) CTU Grant 5U01 AI069484]; Carl J. Fichtenbaum and Eva Moore [University of Cincinnati (Site 2401) CTU Grant 1U01AI069513]; David Currin and Megan Avots [UNC AIDS Clinical Trials Unit UNC AIDS CRS (Site 3201) CTU Grant 5-U01 AI069423, CTSA Grant UL 1RR 025747, CFAR Grant AI50410]; Roberto C. Arduino and Maria Laura Martinez [Houston AIDS Research (HART) (Site 31473) CTU Grant 1U01AI069503]; Mary Albrecht and Amanda Youmans [Beth Israel Deaconess (Partners/Harvard) CRS (Site 103) CTU Grant U01 AI069472-05]; Debbie Slamowitz and Sandra Valle [Stanford University AIDS CTU (Site 501) CTU Grant AI069556]; Princy N. Kumar and Joseph Timpone [Georgetown University (Site 1008) Grant 5U01AI069494]; Christine Hurley and Roberto Corales [AIDS Care (Site 1108) CTU Grant U01AI069511-02 (as of 2 December 2008), CTSI Grant UL1 RR 024160]; Vicki Bailey and Husamettin Erdem [Vanderbilt Therapeutics CRS (Site 3652) CTU Grant AI-069439, Grant RR-024975]; Sharon Riddler and Sally McNulty [Pittsburgh CRS (Site 1001) CTU Grant 1 UO1 AI 069494-01]; and Barbara Philpotts and Dawn Antosh [Case CRS (Site 2501) CTU Grant AI69501].
References
- 1.Taiwo B, Zheng L, Gallien S, et al. Efficacy of a nucleoside-sparing regimen of darunavir/ritonavir plus raltegravir in treatment-naïve HIV-1-infected patients (ACTG A5262) AIDS. 2011;25:2113–22. doi: 10.1097/QAD.0b013e32834bbaa9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Panel on Antireroviral Guidelines for Adults and Adolescents. Guidelines for the Use of Antiretroviral Agents in HIV-1-Infected Adults and Adolescents. Department of Health and Human Services; http://aidsinfo.nih.gov/contentfiles/lvguidelines/adultandadolescentgl.pdf. (4 April 2012, date last accessed) [DOI] [PubMed] [Google Scholar]
- 3.Jiao Y, Zhang T, Wang R, et al. Plasma IP-10 is associated with rapid disease progression in early HIV-1 infection. Viral Immunol. 2012;25:333–7. doi: 10.1089/vim.2012.0011. [DOI] [PubMed] [Google Scholar]
- 4.Kuller LH, Tracy R, Belloso W, et al. Inflammatory and coagulation biomarkers and mortality in patients with HIV infection. PLoS Med. 2008;5:e203. doi: 10.1371/journal.pmed.0050203. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Sandler NG, Wand H, Roque A, et al. Plasma levels of soluble CD14 independently predict mortality in HIV infection. J Infect Dis. 2011;203:780–90. doi: 10.1093/infdis/jiq118. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Kalayjian RC, Machekano RN, Rizk N, et al. Pretreatment levels of soluble cellular receptors and interleukin-6 are associated with progression in subjects treated with highly active antiretroviral therapy. J Infect Dis. 2010;201:1796–805. doi: 10.1086/652750. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Trøseid M, Nowak P, Nyström J, et al. Elevated plasma levels of lipopolysaccharide and high mobility group box-1 protein are associated with high viral load in HIV-1 infection: reduction by 2-year antiretroviral therapy. AIDS. 2010;24:1733–7. doi: 10.1097/QAD.0b013e32833b254d. [DOI] [PubMed] [Google Scholar]
- 8.Brenchley JM, Price DA, Schacker TW, et al. Microbial translocation is a cause of systemic immune activation in chronic HIV infection. Nat Med. 2006;12:1365–71. doi: 10.1038/nm1511. [DOI] [PubMed] [Google Scholar]
- 9.Cohen Stuart JW, Hazebergh MD, Hamann D, et al. The dominant source of CD4+ and CD8+ T-cell activation in HIV infection is antigenic stimulation. J Acquir Immune Defic Syndr. 2000;25:203–11. doi: 10.1097/00126334-200011010-00001. [DOI] [PubMed] [Google Scholar]
- 10.Robbins GK, Spritzler JG, Chan ES, et al. Incomplete reconstitution of T-cell subsets on combination antiretroviral therapy in the AIDS Clinical Trials Group protocol 384. Clin Infect Dis. 2009;48:350–61. doi: 10.1086/595888. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.French MA, King MS, Tschampa JM, et al. Serum immune activation markers are persistently increased in patients with HIV infection after 6 years of antiretroviral therapy despite suppression of viral replication and reconstitution of CD4+ T-cells. J Infect Dis. 2009;200:1212–5. doi: 10.1086/605890. [DOI] [PubMed] [Google Scholar]
- 12.Lederman MM, Calabrese L, Funderburg NT, et al. Immunologic failure despite suppressive antiretroviral therapy is related to activation and turnover of memory CD4 cells. J Infect Dis. 2011;204:1217–26. doi: 10.1093/infdis/jir507. [DOI] [PMC free article] [PubMed] [Google Scholar]