- Open Access
T-bet expression in CD8+ T cells associated with chronic hepatitis B virus infection
© Fan et al. 2016
Received: 9 November 2015
Accepted: 20 January 2016
Published: 25 January 2016
The mechanisms leading to virus-specific CD8+ T cell dysfuction in chronic hepatitis B virus (HBV) infection remain to be elucidated. Our study focused on the role of transcription factor T-bet in HBV infection because it is a crucial regulator of T cell immunity.
We assessed the expression of T-bet along with PD-1, IFN-γ and perforin, in HBV-specific CD8+ T cells from resolved acute hepatitis B (rAHB) patients, chronic hepatitis B (CHB) patients, as well as asymptomatic HBV carriers (ASCs). We observed dynamic changes of T-bet, PD-1, IFN-γ and perforin in acute stage and recovery stage of acute hepatitis B (AHB).
Comparing with other cohorts, HBV-specific CD8+ T cells from rAHB demonstrated a superior ability in T-bet, IFN-γ and perforin expression, but an inferior ability in PD-1 expression. In the CHB group, the level of T-bet has a linear relationship with the level of PD-1, IFN-γ and HBV DNA, respectively. A lower expression of T-bet and PD-1 was observed in ASCs when compared with CHB. A higher expression of T-bet, PD-1, IFN-r and perforin was observed in acute stage when compared with the recovery stage of AHB.
Our results suggest that expression of T-bet may influence the function of HBV-specific CD8+ T cells and thus can be an attractive target for modulation to improve HBV-specific immunity in CHB.
Both the quantity and quality of the adaptive antiviral immune response influence the clinical outcome of HBV infection . A multi-specific and vigorous T cell response is present in acute hepatitis B (AHB) patients who have successfully cleared HBV infection. However, in chronic HBV infection including CHB and ASCs, the T cell immune responses are weak and oligoclonal . HBV-specific CD8+ T cells play important roles in clearance of HBV infection and in the control of HBV replication . A sustained and potent CD8+ T cell response to HBV antigen is associated with resolved acute HBV infection, but not with chronic HBV infection .
Chronic HBV infection acquired perinatally or in early childhood is believed to progress through distinct phases: immune tolerant, immune active and inactive carrier. ASC means a patient is in immune tolerant phase, while CHB means a patient is in immune active phase . The constant existence of viral antigens intrinsic in chronic infection may lead to loss of function in antigen-specific T cells, decreased production of interleukin- 2 (IL-2), tumor necrosis factor (TNF), interferon gamma (IFN-γ) and chemotactic factor β, and gradually impaired proliferation, decreasing cytotoxicity as well as diminishing ability of T cell survival which eventually result in T cell exhaustion .
Immunoregulation is centrally involved in T cell exhaustion. These negative pathways can be grouped into three main categories: cell surface inhibitory receptors (such as PD-1), soluble factors (such as IL-10), and immunoregulatory cell types (such as regulatory T cells). In addition, several specific transcriptional pathways have been implicated in T cell exhaustion. Particularly, the transcription factor T-bet is centrally involved in CD8+ T cell exhaustion .
T-bet, which belongs to the T-box transcription factor family, is encoded by the Tbx21 gene and expressed in many immune cells, showing an extensive immunoregulatory function . T-bet sustains the effector function of CD8+ T cells through various mechanisms, such as regulation of CD8+ T cell proliferation, suppressed expression of inhibitory receptors such as PD-1, and promotion of IFN-γ and perforin secretion . Dysfunction of CD8+ T cells has been found in mice with defective Tbx21 and chronic infection occurred after LCMV infection . Chronic HIV infection leads to decreased T-bet expression in HIV-specific CD8+ T cells [11, 12]. Moreover, T-bet has been implicated as an anti-tumor regulator and pathogenic factor for autoimmune diseases [13, 14].
Currently, the correlation between the impaired regulation of T-bet and the development of chronic HBV infection has not been established. A study by Kurktschiev PD et al. assessed the influence of T-bet on CD8 + T cells activity observed during acute and chronic HBV infection . In our study, patients with HBV infection were further divided into three groups: AHB, CHB and ASCs. We have determined T-bet expression in these groups and examined its correlation with the clinical outcome. Our observations suggest a central role of T-bet in regulating different immune states of HBV infection.
Clinical characteristics of the three patient groups
AHB n = 9
CHB n = 18
ASCs n = 15
Age (years, mean ± SD)
33.67 ± 5.32
42.45 ± 4.67
20.22 ± 3.71
Gender ( male/female)
Median HBV DNA level (log10 IU/mL)
ALT (units/L)mean ± SD
1773.11 ± 1038.47
523 ± 775.40
19.43 ± 5.47
Subjects were selected according to previously described criteria  as the following: AHB was defined as acute onset of nonspecific flu-like symptoms and jaundice in previously healthy persons with peak alanine aminotransferase (ALT) elevation 10 times above the upper limit of normal, and was confirmed by concomitant detection of hepatitis B surface antigen (HBsAg), HBV DNA, or anti-hepatitis B core IgM antibody (anti-HBc-IgM). rAHB was confirmed by seroconversion of hepatitis B surface antibodies (anti-HBs). CHB was defined by detection of HBV DNA or HBsAg for more than 6 months with ALT fluctuations. ASCs were defined as HBsAg positive, ALT and aspartate aminotransferase (AST) within the normal range for more than 3 visits, and a history of HBV infection. Patients with other possible causes for chronic liver damage, such as alcohol use, drug use, congestive heart failure and autoimmune diseases, and pregnant women were also excluded from this study.
The experiments in this study were carried out under the guidance of moral standards described in Declaration of Helsinki and International Ethical Guidelines for Biomedical Research Involving Human Subjects by Council for International Organizations of Medical Sciences (CIOMS), with the approval of ethics committee in First Affiliated Hospital of Harbin Medical University (approval ID: ChiCTR-CCC-14004949). An informed consent was signed by all study subjects.
Synthetic peptides, pentamers, and cytokines
Recombinant HBV core antigen (HBcAg) covering the overall protein sequence of HBV genotype D was purchased from ProSpec (NJ, USA). Human leukocyte antigen (HLA) restricted peptide HBV core antigen 18–27 (FLPSDFFPSV and FLPSDFFPSI, HBV c 18–27) was purchased from Proimmune (Oxford, UK). HBcAg and HBV c 18–27 were used for the in vitro stimulation of HBV-specific CD8+ T cells. HBV c 18–27 was detected by PE-labeled MHC-I restricted pentamers (Proimmune, Oxford, UK). Recombinant human IL-2 (PeproTech, NJ, USA) was used for stimulation experiments.
Monoclonal antibodies for flow cytometry
FITC anti-human HLA-A2 (BioLegend, San Diego, CA, USA), PE-Cy7 anti-human/mouse T-bet (eBioscience, San Diego, CA, USA), APC anti-Human CD8a (eBioscience, San Diego, CA, USA), FITC anti-human CD279 (PD-1) (BioLegend, San Diego, CA, USA), PerCP anti-CD14 (eBioscience, San Diego, CA, USA), APC-eFluor®780 anti-CD19 (eBioscience, San Diego, CA, USA) and 7-AAD (BD Biosciences, San Diego, CA, USA) were used for flow cytometry. Isotype control was used for each antibody. The Foxp3/Transcription Factor Staining Buffer Set Kit (eBioscience, San Diego, CA, USA) was used for intracellular staining according to the manufacturer’s instructions.
HLA-A2 genotype detection
Screening for HLA-A2 was performed by staining peripheral blood mononuclear cells (PBMCs) with a FITC-labeled mouse anti-HLA-A 2 and isotype control (BD, Biosciences, San Diego, CA, USA).
Isolation of PBMCs
PBMCs were isolated from fresh heparinized blood using Ficoll-Hypaque density gradient centrifugation and were either analyzed directly or resuspended in medium for stimulation of PBMCs.
For HBV-specific CD8+ T cells expansion, PBMCs were cultured for 10 days in RPMI 1640 medium containing 2 mM l-glutamine, 1 mM sodium pyruvate, 100 U/ml of penicillin, 100 μg/ml of streptomycin and 5 % human type AB serum. PBMCs were seeded at a density of 1 × 106/ml in 24-well plates, and 1 ml medium was used in each well. In the cytokine-stimulated groups, IL-2 (20 IU/ml) was added on day 0. The antigen-stimulated groups received 5 μg/ml of antigen on day 0 and were re-stimulated with the same dose of antigen on day 10. HBcAg (5ug/ml;) and HBV c 18–27 (FLPSDFFPSV, 5ug/ml; FLPSDFFPSI, 5ug/ml) were used for stimulation. After stimulation, cells were prepared for flow cytometry by cell surface staining and intracellular staining .
Cell surface staining and intracellular staining for flow cytometry
2–3 × 106 PBMCs were stained with MHC-I pentamers according to the manufacturer’s instructions. After staining with viability dyes and antibodies specific for surface markers, cells were fixed with intracellular fixation buffer and permeabilized with permeabilization buffer. After the permeabilization step, cells were stained with intracellular markers. For ex vivo staining, at least 200,000 PBMCs were collected for pentamer + CD8+ T cells as they were low frequency cells. Samples were acquired on a FACSCanto II flow cytometer (BD, Biosciences, San Diego, CA, USA). Data were analyzed with FlowJo 9.6.1 software (Tree Star). Gating strategy excluded monocytes (CD14+), B-lymphocytes (CD19+), and dead cells (7-AAD+) by a dump channel.
Enzyme-linked immunosorbent assay (ELISA)
PBMCs were cultured for HBV-specific CD8+ T cell expansion with stimulation by HBV antigen and peptides for 10 days. Cells were then seeded into 96-well plates at a density of 1 × 105 cells in 100 μl RPMI 1640 medium in each well, and then stimulated again with HBV c 18–27 for 18 h. Subsequently, the cells and supernatant were collected. A human IFN-γ embedded ELISA kit (Dakewe Biotech, Beijing, China) and a human perforin embedded ELISA kit (Dakewe Biotech) were used according to the manufacturer’s instructions. Data were detected with a microplate reader and analyzed thereafter.
All data were analyzed using Prism 5.0 software (GraphPad, CA, USA). Data were expressed as the median with range or mean ± SEM. The Kruskal-Wallis Test was performed to test the differences between the three groups, and the difference of every two groups was tested using the Mann–Whitney test (median) or t test (mean). Spearman’s rank correlation test was used for correlation analysis. P-values of P < 0.05 were considered significant.
HBV-specific CD8+ T cell responses
T-bet expression level in HBV-specific CD8+ T cells
PD-1 expression level in HBV-specific CD8+ T cells
Induced IFN-γ and perforin levels in HBV-specific T cells
Relationships between T-bet, PD-1, IFN-γ and HBV DNA in CHB
Dynamic changes of T-bet, PD-1, IFN-γ and perforin in AHB
The poor T-bet expression in virus-specific CD8+ T cells may be a central regulation factor for T cell exhaustion . We compared in vitro amplification of HBV-specific CD8+ T cells isolated from rAHB and CHB patients and ASCs, under stimulation by viral antigens. Similar to previous study , we found dysfunction and low response of HBV-specific CD8+ T cells in the CHB patients and ASC patients comparing with those of rAHB patients. The expression of T-bet in the HBV-specific CD8+ T cells of CHB and ASCs were significantly lower than those of rAHB patients. The T-bet expression was lowest in the HBV-specific CD8+ T cells of ASCs. It has been demonstrated that overexpression of T-bet can promote T cell response, while suppression of T-bet expression can inhibit T cell response . Our data showed expression of T-bet was in accordance with HBV specific CD8+ T cell response. Similar results concerning T-bet expression were observed in HIV infected patients . Our data suggest that T-bet might regulate the level of the HBV-specific CD8+ T cell response, and low T-bet levels might be an important factor leading to exhaustion of cytotoxic T cells in chronic hepatitis B. In the CHB group, the level of T-bet had a negative correlation with the level of HBV DNA, suggesting T-bet expression is vital for HBV clearance. The study by Kurktschiev PD et al. compared T-bet expression between AHB and CHB . Our study further showed that the level of T-bet was lower in ASCs than in CHB. We propose that lower expression of T-bet may be a factor of immune tolerance and higher expression of T-bet may trigger a biological process leading to loss of immune tolerance.
The axis of PD-1 and its ligand is a major inhibitory receptor pathway involved in CD8+ T cell exhaustion , and PD-1 can be suppressed by T-bet in chronic murine LCMV infection . IL-12 can induce T-bet and decrease PD-1 expression level in chronic hepatitis B and restore IFN production . In our study, PD-1 expression was higher in the HBV-specific CD8+ T cells of CHB patients comparing with those of rAHB patients, and a higher PD-1 expression was coexisting with a lower HBV-specific CD8 + T cell response. The PD-1 expression level was positively correlating with the level of HBV DNA in CHB patients. This suggests that PD-1/PD-L1 pathway may inhibit HBV clearance by inhibiting HBV-specific CD8+ T cells. A similar PD-1 expression was observed in HIV infected patients . It was noted that the PD-1 expression was lower in the ASC group than in the CHB group. One reason is that ASC patients is in immune tolerant phase, and PD-1 expression is lower in immune tolerant phase than in immune active phase . Another reason is that the ASC patients we chose were younger than the CHB patients, and PD-1 expression is lower in the young adult patients than in the older cohort from chronic HBV infected patients .
As a ‘master regulator’ of cell-mediated immunity, T-bet participates in regulation of genes encoding effector molecules in immune cells, such as IFN-γ, perforin, and granzymes . CD8+ T cells participate in clearance of intracellular viruses upon production of IFN-γ, and can lyse target cells with perforin and granzymes . We observed a high level expression of IFN-γ and perforin in the rAHB group and a low level expression of these genes in the CHB and ASC groups, with a positive correlation between T-bet and IFN-γ production in the CHB group. These results indicate that low T-bet expression in chronic HBV infection might lead to impaired production of IFN-γ and perforin. Because the transcription factor eomesodermin also controls production of IFN-γ and peforin in effector CD8 + T cells, further studies of the roles of eomesodermin in chronic HBV infection is necessary .
We followed up the AHB patients to determine differences in the acute stage and recovery stage of AHB. Our results showed that both T-bet and PD-1 had higher levels along with IFN-γ and perforin in the acute stage than in the recovery stage. A previous study also indicated that at clinical onset of acute HBV infection, PD-1 was significantly up-regulated and subsequently led to the functional suppression of HBV-specific CD8+ T cells, and that following disease resolution, HBV-specific effector CD8+ T cells developed into memory T cells. During this period, the dynamic PD-1 decrease was numerically correlated with the reduction of HBV-specific CD8+ T cell frequency . Another study showed that the effect of T-bet on PD-1 expression was modest during acute infection but became greater during chronic infection . We propose that in the acute stage, a high level of T-bet expression can promote amplification of CD8+ T cells and result in virus clearance, and that a high level of PD-1 might lead to rapid apoptosis of effector cells. So, the balance between T-bet and PD-1 in the acute phase is critical to virus clearance and pathological damage control.
Concerning the potential mechanism underlying low expression of T-bet in virus specific CD8+ T cells in patients with chronic HBV infection, Smith et al. suggested that during constant infection, antigen presenting cells might influence the transcription of transcription factors in virus specific T cells, such as T-bet . EJ Wherry et al. suggested that constant virus infection led to a continuous conversion from T-bethigh progenitor cells to Eomeshigh cells in virus specific CD8+ T cells, which eventually led to the disappearance of T-bethigh cells . Previous studies have shown that inflammatory signals (e.g., IL-12) not only enhance T-bet expression but may also repress Eomes and thus play a dominant role in regulating memory/effector T-cell potential . The IL-21–induced cytotoxic T cell (CTL) function is T-bet dependent, because T-bet deficiency results in defective IL-21–dependent cytotoxicity in CD8+ T cells in vitro and in vivo . Therefore, it is possible that lack of IL-12 or IL-21 results in low T-bet expression in chronic hepatitis B.
The expression of T-bet is closely related to exhaustion of T cells in chronic hepatitis B, and T-bet might be a critical factor leading to chronic HBV infection. Further understanding of T-bet function may provide new therapeutic approaches for chronic hepatitis B. We observed different levels of T-bet and PD-1 expression between CHB and ASC patients, and these may explain the different immune status during chronic HBV infection.
This study was supported by the grant from the National Key Technologies Research and Development Program of China during the 11th Five-year Plan Period (2008ZX10002-006), and the National Key Technologies Research and Development Program of China during the 12th Five-year Plan Period (2012ZX10002007).
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
- Liaw YF, Chu CM. Hepatitis B virus infection. Lancet. 2009;373:582–92.View ArticlePubMedGoogle Scholar
- Guidotti LG, Chisari FV. Immunobiology and pathogenesis of viral hepatitis. Annu Rev Pathol. 2006;1:23–61.View ArticlePubMedGoogle Scholar
- Sobao Y, Tomiyama H, Sugi K, Tokunaga M, Ueno T, Saito S, et al. The role of hepatitis B virus-specific memory CD8 T cells in the control of viral replication. J Hepatol. 2002;36:105–15.View ArticlePubMedGoogle Scholar
- Boni C, Fisicaro P, Valdatta C, Amadei B, Di Vincenzo P, Giuberti T, et al. Characterization of hepatitis B virus (HBV)-specific T-cell dysfunction in chronic HBV infection. J Virol. 2007;81:4215–25.PubMed CentralView ArticlePubMedGoogle Scholar
- Della Corte C, Nobili V, Comparcola D, Cainelli F, Vento S. Management of chronic hepatitis B in children: an unresolved issue. J Gastroenterol Hepatol. 2014;29:912–9.View ArticlePubMedGoogle Scholar
- Virgin HW, Wherry EJ, Ahmed R. Redefining chronic viral infection. Cell. 2009;138:30–50.View ArticlePubMedGoogle Scholar
- Wherry EJ. T cell exhaustion. Nat Immunol. 2011;12:492–9.View ArticlePubMedGoogle Scholar
- Lazarevic V, Glimcher LH. T-bet in disease. Nat Immunol. 2011;12:597–606.View ArticlePubMedGoogle Scholar
- Kao C, Oestreich KJ, Paley MA, Crawford A, Angelosanto JM, Ali MA, et al. Transcription factor T-bet represses expression of the inhibitory receptor PD-1 and sustains virus-specific CD8+ T cell responses during chronic infection. Nat Immunol. 2011;12:663–71.PubMed CentralView ArticlePubMedGoogle Scholar
- Hatton RD, Weaver CT. Immunology. T-bet or not T-bet. Science. 2003;302:993–4.View ArticlePubMedGoogle Scholar
- Hersperger AR, Martin JN, Shin LY, Sheth PM, Kovacs CM, Cosma GL, et al. Increased HIV-specific CD8+ T-cell cytotoxic potential in HIV elite controllers is associated with T-bet expression. Blood. 2011;117:3799–808.PubMed CentralView ArticlePubMedGoogle Scholar
- Ribeiro-dos-Santos P, Turnbull EL, Monteiro M, Legrand A, Conrod K, Baalwa J, et al. Chronic HIV infection affects the expression of the 2 transcription factors required for CD8 T-cell differentiation into cytolytic effectors. Blood. 2012;119:4928–38.View ArticlePubMedGoogle Scholar
- Zhu Y, Ju S, Chen E, Dai S, Li C, Morel P, et al. T-bet and eomesodermin are required for T cell-mediated antitumor immune responses. J Immunol. 2010;185:3174–83.View ArticlePubMedGoogle Scholar
- Rangachari M, Mauermann N, Marty RR, Dirnhofer S, Kurrer MO, Komnenovic V, et al. T-bet negatively regulates autoimmune myocarditis by suppressing local production of interleukin 17. J Exp Med. 2006;203:2009–19.PubMed CentralView ArticlePubMedGoogle Scholar
- Kurktschiev PD, Raziorrouh B, Schraut W, Backmund M, Wachtler M, Wendtner CM, et al. Dysfunctional CD8+ T cells in hepatitis B and C are characterized by a lack of antigen-specific T-bet induction. J Exp Med. 2014;211:2047–59.PubMed CentralView ArticlePubMedGoogle Scholar
- Webster GJ, Reignat S, Brown D, Ogg GS, Jones L, Seneviratne SL, et al. Longitudinal analysis of CD8+ T cells specific for structural and nonstructural hepatitis B virus proteins in patients with chronic hepatitis B: implications for immunotherapy. J Virol. 2004;78:5707–19.PubMed CentralView ArticlePubMedGoogle Scholar
- Tsai SL, Lee TH, Chien RN, Liao SK, Lin CL, Kuo GC, et al. A method to increase tetramer staining efficiency of CD8+ T cells with MHC-peptide complexes: therapeutic applications in monitoring cytotoxic T lymphocyte activity during hepatitis B and C treatment. J Immunol Methods. 2004;285:71–87.View ArticlePubMedGoogle Scholar
- Joshi NS, Cui W, Chandele A, Lee HK, Urso DR, Hagman J, et al. Inflammation directs memory precursor and short-lived effector CD8(+) T cell fates via the graded expression of T-bet transcription factor. Immunity. 2007;27:281–95.PubMed CentralView ArticlePubMedGoogle Scholar
- Schurich A, Pallett LJ, Lubowiecki M, Singh HD, Gill US, Kennedy PT, et al. The third signal cytokine IL-12 rescues the anti-viral function of exhausted HBV-specific CD8 T cells. PLoS Pathog. 2013;9:e1003208.PubMed CentralView ArticlePubMedGoogle Scholar
- Day CL, Kaufmann DE, Kiepiela P, Brown JA, Moodley ES, Reddy S, et al. PD-1 expression on HIV-specific T cells is associated with T-cell exhaustion and disease progression. Nature. 2006;443:350–4.View ArticlePubMedGoogle Scholar
- Ren YY, Liu YZ, Ding YP, Song G, Li SH, Wang GQ. Immune characteristics of different immune phases in natural course of chronic HBV infection. Hepatogastroenterology. 2013;60:789–95.PubMedGoogle Scholar
- Kennedy PT, Sandalova E, Jo J, Gill U, Ushiro-Lumb I, Tan AT, et al. Preserved T-cell function in children and young adults with immune-tolerant chronic hepatitis B. Gastroenterology. 2012;143:637–45.View ArticlePubMedGoogle Scholar
- Intlekofer AM, Takemoto N, Wherry EJ, Longworth SA, Northrup JT, Palanivel VR, et al. Effector and memory CD8+ T cell fate coupled by T-bet and eomesodermin. Nat Immunol. 2005;6:1236–44.View ArticlePubMedGoogle Scholar
- Sullivan BM, Juedes A, Szabo SJ, von Herrath M, Glimcher LH. Antigen-driven effector CD8 T cell function regulated by T-bet. Proc Natl Acad Sci U S A. 2003;100:15818–23.PubMed CentralView ArticlePubMedGoogle Scholar
- Pearce EL, Mullen AC, Martins GA, Krawczyk CM, Hutchins AS, Zediak VP, et al. Control of effector CD8+ T cell function by the transcription factor Eomesodermin. Science. 2003;302:1041–3.View ArticlePubMedGoogle Scholar
- Zhang Z, Jin B, Zhang JY, Xu B, Wang H, Shi M, et al. Dynamic decrease in PD-1 expression correlates with HBV-specific memory CD8 T-cell development in acute self-limited hepatitis B patients. J Hepatol. 2009;50:1163–73.View ArticlePubMedGoogle Scholar
- Smith C, Elhassen D, Gras S, Wynn KK, Dasari V, Tellam J, et al. Endogenous antigen presentation impacts on T-box transcription factor expression and functional maturation of CD8+ T cells. Blood. 2012;120:3237–45.View ArticlePubMedGoogle Scholar
- Paley MA, Kroy DC, Odorizzi PM, Johnnidis JB, Dolfi DV, Barnett BE, et al. Progenitor and terminal subsets of CD8+ T cells cooperate to contain chronic viral infection. Science. 2012;338:1220–5.PubMed CentralView ArticlePubMedGoogle Scholar
- Takemoto N, Intlekofer AM, Northrup JT, Wherry EJ, Reiner SL. Cutting Edge: IL-12 inversely regulates T-bet and eomesodermin expression during pathogen-induced CD8+ T cell differentiation. J Immunol. 2006;177:7515–9.View ArticlePubMedGoogle Scholar
- Sutherland AP, Joller N, Michaud M, Liu SM, Kuchroo VK, Grusby MJ. IL-21 promotes CD8+ CTL activity via the transcription factor T-bet. J Immunol. 2013;190:3977–84.View ArticlePubMedGoogle Scholar