HIV-positive persons experience variable rates of disease progression in the absence of antiretroviral therapy (ART). HIV elite controllers spontaneously control viral replication, and maintain preserved CD4+ T-cell counts. Elite control has been previously associated with improved CD8+ T cell  and natural killer (NK) cell function  and lower levels of proinflammatory cytokines , however the exact underlying mechanisms of elite control remain unknown.
Tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) is an immune effector protein which induces apoptotic cell death of cancerous or infected cells . TRAIL can bind four membrane-bound receptors, but only TRAIL-R1 (DR4) and TRAIL-R2 (DR5) induce programmed cell death [5–8]. TRAIL, TRAIL-R1, and TRAIL-R2 are interferon-stimulated genes, and are expressed following Toll-like receptor (TLR) stimulation, interferon signaling [9,10], and proinflammatory cytokine signaling [11–14].
TRAILshort is a splice variant of TRAIL that acts as dominant negative TRAIL antagonist . TRAILshort is detectable in HIV-infected cell cultures and HIV-positive patient plasma , and is produced by both HIV-infected and uninfected cells [15,16]. TRAILshort is secreted in extracellular vesicles, causing TRAIL resistance in both cis and trans. Because TRAILshort expression causes TRAIL resistance , it is possible that TRAILshort expression may promote HIV persistence. Furthermore, as patients with elite control of HIV have decreased tonsillar mRNA for IFN-α, TRAIL, and TRAIL-R2 compared with those with progressive disease , we questioned whether levels of TRAILshort correlate with HIV reservoir size and HIV disease phenotypes.
All patient samples [peripheral blood mononuclear cells (PBMCs) and plasma] were obtained with written, informed consent following institutional review board approvals from the Mayo Clinic College of Medicine and Science, University of California San Francisco, and the Karolinska Institutet. All participants were adults.
The discovery cohort was obtained from the UCSF SCOPE cohort (NCT00187512) and baseline characteristics are listed in Table 1. The validation cohort was obtained from the Swedish InfCareHIV cohort at the Karolinska Institutet, and patient characteristics were described elsewhere [18,19]. Definitions were as follows: elite controllers: known HIV-positivity for more than a year and at least three consecutive viral load below 75 copies/ml over 1 year (and all previous viral load below 1000 copies/ml), known HIV-1 positivity at least 10 years, and minimum two viral load-measurements (≥90% of all viral loads <400 copies/ml). Viremic progressors had HIV-1 plasma RNA levels more than 10 000 copies/ml. For long-term ART (lART), the mean duration of suppressive treatment was 17 years (range: 13–20) without any detectable viral rebound. Patients were followed up every 6 months for routine virological testing.
Next generation RNA-sequencing (RNA-seq) data processing and analysis
RNA-seq was performed on cryopreserved PBMCs from a random subset of patients from the elite controller and viremic progressor cohorts. Total RNA was extracted using Qiagen RNeasy Plus Universal mini kit (Qiagen, Hilden, Germany) and quantified using Qubit 2.0 Fluorometer (Life Technologies, Carlsbad, California, USA) and RNA integrity confirmed with Agilent TapeStation (Agilent Technologies, Palo Alto, California, USA). RNA library preparation, sequencing reaction, and initial bioinformatics analysis were conducted at GENEWIZ, LLC. (South Plainfield, New Jersey, USA). Raw sequence data (.bcl files) generated from Illumina HiSeq was converted into fastq files and demultiplexed using Illumina's bcl2fastq 2.17 software (San Diego, California, USA). One mis-match was allowed for index sequence identification.
Isoform expression quantification from RNA-seq data was performed with salmon  using default settings, with reference genome GRCh38 (v92). Differential expression analysis was performed with DESeq2  using estimated counts obtained from salmon, and all downstream analysis and visualization was performed in (R Core Team; R Foundation for Statistical Computing, Vienna, Austria). P values between cohorts were derived from DESeq2 and represent the Benjamini–Hochberg-corrected significance (P.adj). Gene ontology analysis (GO Enrichment Analysis, Gene Ontology Consortium) was used to identify differential gene expression in the elite controller vs. viremic progressors group.
Plasma profiling of TRAILshort
Plasma profiling of TRAILshort was performed by antibody bead array as previously described . Assays were performed in duplicates in two independent assays. Sample-by-sample variation within each assay plate was determined by probabilistic quotient normalization quotient  and adjusted for plate effects using Multi-MA . The values of two assays were averaged from the two measurements after centering. All samples tested passed quality control in the antibody bead array and were included in the analysis.
Plasma profiling of soluble tumor necrosis factor-related apoptosis-inducing ligand
Plasma profiling of soluble TRAIL (sTRAIL) was performed by proximity extension assay (PEA) and part of Olink Immuno-oncology panel (Olink Bioscience AB, Uppsala, Sweden) . Protein analysis is reported as normalized protein expression levels, an arbitrary unit. The correlation of TRAIL by PEA and ELISA (R&D Systems, Minneapolis, Minnesota, USA) showed good correlation (Spearman r: 0.695, P < 0.0001) .
Descriptive statistics are presented as means ± SD unless otherwise noted. Parametric or nonparametric statistical tests were used as appropriate and are listed in the respective figure legends and tables. Statistical significance was accepted when P less than 0.05. Statistical analysis was performed using GraphPad Prism 6 (GraphPad, Inc., San Diego, California, USA).
Low TRAILshort expression is associated with elite control of HIV infection in vivo
In vivo, T-cell number reflects the cumulative effect of T-cell losses and of T-cell production and proliferation. We reasoned that an HIV-positive individual with low levels of TRAILshort would have enhanced killing of both HIV-infected and HIV-uninfected cells, but that the production of new, uninfected CD4+ T cells would be expected to counter CD4+ T losses, and dilute the number of HIV DNA positive cells, which would be reminiscent of the elite controller phenotype. Thus, we compared TRAILshort expression between elite controller patients (undetected HIV-1 RNA levels persisting for >1 year without ART) and viremic progressors (HIV-1 RNA viral load >10 000 copies/ml) (Table 1). Elite controllers were significantly older than viremic progressors (51 ± 10 vs. 41 ± 11 years, P < 0.001), and had higher baseline CD4+ T cell counts (991 ± 426 vs. 479 ± 237 cells/μl, P < 0.001), and significantly lower levels of cell-associated HIV-1 DNA in PBMCs than viremic progressors (82 ± 128 vs. 1572 ± 1628 copies/million cells, P < 0.001, Fig. 1a).
We performed RNA-seq on PBMCs from a random subset of viremic progressors (n = 15) and elite controllers (n = 14). Consistent with our hypothesis, elite controllers had significantly fewer TRAILshort transcripts per million (P = 0.002, Fig. 1b) and lower full-length TRAIL transcripts per million in PBMCs compared with viremic progressors (P = 0.001, Fig. 1c). We validated these findings in a second independent validation cohort of elite controllers (n = 19). TRAILshort transcripts (P = 0.06, Fig. 1d) and Full-length TRAIL transcripts were lower in elite controllers compared with viremic progressors (P = 0.004, Fig. 1e). We also assessed whether the ratio of full-length TRAIL to TRAILshort differed between elite controllers and viremic progressors, and elite controllers had lower TRAILshort/ TRAIL ratios compared with viremic progressors (Fig. 1f).
Seventeen transcripts correlated with TRAILshort expression (Pearson Correlation Coefficient more than 0.8, P < 0.05, Fig. 1g), and four were expressed less in elite controllers compared with viremic progressors [TLR7, P = 0.005; phospholipid scramblase 1 (PLSCR1), P = 0.000; N-myc and STAT interactor (NMI), P = 0.000; and GTPase, immunity associated protein family member 4 (GIMAP4), P = 0.004; Fig. 1h]. Of note these genes cluster in two biologic pathways: regulation of IFN-α biosynthetic process (P = 7.58E − 07, FDR = 1.19E − 02); regulation of IFN-β biosynthetic process (P = 1.49E − 06, FDR = 1.17E − 02), supporting our previous experimental data that type I interferons, as well as TLR7 and TLR8 agonists, promote TRAILshort expression . Other genes associated with reduced TRAILshort in elite controllers were as follows: PLSCR1, NMI (N-myc and STAT interactor), and GIMAP4, and these have been previously associated with HIV and/or execution of cell death [26–32].
Decreased levels of tumor necrosis factor-related apoptosis-inducing ligand and TRAILshort proteins are associated with elite controllers
TRAILshort can be shed from producing cells within microvesicles and confer TRAIL-resistance upon neighboring non-TRAILshort-producing cells . Full length TRAIL, on the other hand, is not present in plasma, yet it can be cleaved and the cleaved soluble form can induce apoptosis in neighboring TRAIL-sensitive cells [33,34]. As TRAILshort and sTRAIL are detectable in plasma from HIV-positive persons [15,34], we compared expression of these proteins between elite controller and viremic progressor. Because CD4+ T cells are a source of TRAILshort , we normalized TRAILshort and sTRAIL concentration to CD4+ T-cell number (Supplemental Table 1, http://links.lww.com/QAD/B487). Elite controllers had lower plasma TRAILshort concentration than individuals with chronic HIV infection (P < 0.001), primary HIV infection (P = 0.002), and patients on lART (P = 0.002) (Fig. 1i). Significantly, CD4+-normalized plasma TRAILshort concentration decreased significantly after initiating ART during primary HIV-1 infection (P = 0.002, Fig. 1j) or chronic HIV-1 infection compared with pre-ART levels (P = 0.015, Fig. 1k), consistent with ART-induced reduction of Type 1 interferon expression and gene signature . Similar results were observed for plasma sTRAIL (Fig. 1l–n), for which eight samples did not pass the quality control in the PEA and were excluded from the analysis. All together, these data corroborated the decreased gene expression of TRAILshort and full-length TRAIL noted above.
Elite controllers are able to maintain CD4+ T-cell counts and control viral replication in the absence of ART , and this phenotype has been associated with improved HIV-1-specific CD8+ T cell and NK cell function [1,2] and lower levels of proinflammatory cytokines . Here, we show an association of low TRAILshort expression with lower HIV viral load and lower HIV reservoir size in HIV-positive persons, which is consistent with our previously published data . This association could result from two nonmutually exclusive mechanisms. First, low TRAILshort expression could result in enhanced killing of HIV-infected cells thereby causing fewer productively infected cells to produce fewer HIV progeny virions, resulting in lower viral replication. A second possibility is that reduced viral replication leads to less inflammation and less TLR activation, in turn causing less TRAILshort production. This model is supported by our data showing that ART induced reductions in HIV replication causes reduced TRAILshort expression within individual patients. It is notable however that TRAILshort expression is lower in elite controllers compared with chronically HIV-infected patients on long-term, suppressive ART (Fig. 1i), suggesting that both mechanisms are likely involved.
We note that the gene expression levels of TRAILshort in PBMCs differ slightly in magnitude between the two cohorts studied (Fig. 1b and d). We hypothesize that there may be underlying genetic or environmental factors contributing to those differences between the two populations, which were geographically diverse. These differences are worthy of future study. However, despite this, the relative difference in TRAILshort gene expression between elite controllers and viremic progressors was similar between the two cohorts.
As TRAILshort is a splice variant of the TRAIL full length mRNA, it stands to reason that factors regulating the transcription of TRAIL, such as interferon and Toll-like receptor signaling, also regulate TRAIL short gene expression, as we have previously shown . Therefore, that both TRAIL and TRAILshort gene expression are similarly lower in elite controllers compared with viremic progressors is not surprising. It is of note, though, that the ratio of TRAILshort to TRAIL message is lower in elite controllers compared with viremic progressors (Fig. 1f), suggesting a posttranscriptional difference in either mRNA splicing or stability. This is worthy of future investigation, but is beyond the scope of this study.
A previous study failed to show a difference in serum TRAIL between elite controllers and HIV-positive persons on ART , whereas in our study elite controllers had lower plasma soluble TRAIL (normalized to CD4+ cell count, P = 0.011, Fig. 1l). This discrepancy could be due to several factors. First, circulating TRAIL was measured by different assays, with Jacobs et al. using a multiplex cytokine assay on serum samples. Second, that study only included women participants, whereas our study had both men and women. This suggests the potential for sex based differences in TRAIL expression in HIV-positive persons, which will be important to study in the future.
There are some notable viremic patients who do not express high levels of TRAILshort. This is consistent with our previously published data as well . To determine if there is a definitive subgroup within viremic patients who do not express TRAILshort is of great interest to our group, and will be studied in the future. However, we acknowledge that elite control status is likely a result of multiple immunologic, and potentially virologic, factors, to which TRAILshort contributes. Therefore, the minor individual exceptions to otherwise significant group differences are not contrary to our hypothesis.
The novel association between elite control of HIV and reduced expression of TRAILshort, which is an antagonist of the immune surveillance effector molecule TRAIL, highlights the importance of immune based cytotoxicity pathways in immune control of HIV, and suggests that host-cell production of TRAILshort represents a homeostatic response intended to limit the degree of cell killing induced by HIV infection.
The authors would like to thank the persons living with HIV who graciously participated in the studies described. A.C.P. and N.W.C. were supported through the Mayo Clinic Foundation. U.N. is supported by the Swedish Research Council establishment grant (2017-01330). S.R.L. was supported by the National Institutes of Health (NIH) Delaney AIDS Research Enterprise (DARE U19 AI096109 and UM1 AI126611-01) and the National Health and Medical Research Council (NHMRC) of Australia (NHMRC program grant and practitioner fellowship to S.R.L.). A.D.B was supported by the National Institute of Allergy and Infectious Diseases of the National Institutes of Health grant numbers (AI110173 and AI120698). H.-P.K. is a Markey Molecular Medicine Investigator and received support as the inaugural recipient of the José Carreras/E. Donnall Thomas Endowed Chair for Cancer Research and the Fred Hutch Endowed Chair for Cell and Gene Therapy. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the article.
Author contributions: Conceptualization, A.C.P., N.W.C., S.N., and A.D.B.; Methodology, S.N., E.G.R., G.D.B., H.P.K, and A.D.B; Investigation, A.C.P., N.W.C., S.N, A.P.C., N.K., M.S., G.D.B., and E.P.; Writing – Original Draft, A.C.P., N.W.C., S.N., and A.D.B.; Writing – Review and Editing, all authors; Funding Acquisition, U.N., S.R.L., and A.D.B.; Resources, U.N., A.S., S.D., and H.P.K.; Supervision, N.W.C. and A.D.B.
Conflicts of interest
One or more of the investigators associated with this project and Mayo Clinic have filed patents on TRAILshort, and therefore have a potential financial conflict of interest in technology used in the research and that the investigator(s) and Mayo Clinic may stand to gain financially from the successful outcome of the research.
1. Shasha D, Karel D, Angiuli O, Greenblatt A, Ghebremichael M, Yu X, et al. Elite controller CD8+ T cells exhibit comparable viral inhibition capacity, but better sustained effector properties compared to chronic progressors
. J Leukoc Biol
2. Malnati MS, Ugolotti E, Monti MC, Battista D, Vanni I, Bordo D, et al. Activating killer immunoglobulin receptors and HLA-C: a successful combination providing HIV-1 control
. Sci Rep
3. Cortes FH, de Paula HHS, Bello G, Ribeiro-Alves M, de Azevedo SSD, Caetano DG, et al. Plasmatic levels of IL-18, IP-10, and activated CD8(+) T cells are potential biomarkers to identify HIV-1 elite controllers with a true functional cure profile
. Front Immunol
4. Cummins N, Badley A. The TRAIL to viral pathogenesis: the good, the bad and the ugly
. Curr Mol Med
5. Pan G, Ni J, Wei YF, Yu G, Gentz R, Dixit VM. An antagonist decoy receptor and a death domain-containing receptor for TRAIL
6. Pan G, O’Rourke K, Chinnaiyan AM, Gentz R, Ebner R, Ni J, et al. The receptor for the cytotoxic ligand TRAIL
7. Walczak H, Degli-Esposti MA, Johnson RS, Smolak PJ, Waugh JY, Boiani N, et al. TRAIL-R2: a novel apoptosis-mediating receptor for TRAIL
. EMBO J
8. Wu GS, Burns TF, McDonald ER 3rd, Jiang W, Meng R, Krantz ID, et al. KILLER/DR5 is a DNA damage-inducible p53-regulated death receptor gene
. Nat Genet
9. Gong B, Almasan A. Genomic organization and transcriptional regulation of human Apo2/TRAIL gene
. Biochem Biophys Res Commun
10. Sato K, Hida S, Takayanagi H, Yokochi T, Kayagaki N, Takeda K, et al. Antiviral response by natural killer cells through TRAIL gene induction by IFN-alpha/beta
. Eur J Immunol
11. Kayagaki N, Yamaguchi N, Nakayama M, Eto H, Okumura K, Yagita H. Type I interferons (IFNs) regulate tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) expression on human T cells: a novel mechanism for the antitumor effects of type I IFNs
. J Exp Med
12. Sedger LM, Shows DM, Blanton RA, Peschon JJ, Goodwin RG, Cosman D, et al. IFN-gamma mediates a novel antiviral activity through dynamic modulation of TRAIL and TRAIL receptor expression
. J Immunol
13. Liu S, Yu Y, Zhang M, Wang W, Cao X. The involvement of TNF-alpha-related apoptosis-inducing ligand in the enhanced cytotoxicity of IFN-beta-stimulated human dendritic cells to tumor cells
. J Immunol
14. Kemp TJ, Elzey BD, Griffith TS. Plasmacytoid dendritic cell-derived IFN-alpha induces TNF-related apoptosis-inducing ligand/Apo-2L-mediated antitumor activity by human monocytes following CpG oligodeoxynucleotide stimulation
. J Immunol
15. Schnepple DJ, Shepard B, Bren GD, Cummins NW, Natesampillai S, Trushin S, et al. Isolation of a TRAIL antagonist from the serum of HIV-infected patients
. J Biol Chem
16. Nie Z, Aboulnasr F, Natesampillai S, Burke SP, Krogman A, Bren GD, et al. Both HIV-infected and uninfected cells express TRAILshort, which confers TRAIL resistance upon bystander cells within the microenvironment
. J Immunol
17. Herbeuval JP, Nilsson J, Boasso A, Hardy AW, Kruhlak MJ, Anderson SA, et al. Differential expression of IFN-alpha and TRAIL/DR5 in lymphoid tissue of progressor versus nonprogressor HIV-1-infected patients
. Proc Natl Acad Sci U S A
18. Zhang W, Morshed MM, Noyan K, Russom A, Sonnerborg A, Neogi U. Quantitative humoral profiling of the HIV-1 proteome in elite controllers and patients with very long-term efficient antiretroviral therapy
. Sci Rep
19. Zhang W, Ambikan AT, Sperk M, van Domselaar R, Nowak P, Noyan K, et al. Transcriptomics and targeted proteomics analysis to gain insights into the immune-control mechanisms of HIV-1 infected elite controllers
20. Patro R, Duggal G, Love MI, Irizarry RA, Kingsford C. Salmon provides fast and bias-aware quantification of transcript expression
. Nat Methods
21. Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2
. Genome Biol
22. Drobin K, Nilsson P, Schwenk JM. Highly multiplexed antibody suspension bead arrays for plasma protein profiling
. Methods Mol Biol
23. Dieterle F, Ross A, Schlotterbeck G, Senn H. Probabilistic quotient normalization as robust method to account for dilution of complex biological mixtures. Application in 1H NMR metabonomics
. Anal Chem
24. Hong MG, Lee W, Nilsson P, Pawitan Y, Schwenk JM. Multidimensional normalization to minimize plate effects of suspension bead array data
. J Proteome Res
25. Assarsson E, Lundberg M, Holmquist G, Bjorkesten J, Thorsen SB, Ekman D, et al. Homogenous 96-plex PEA immunoassay exhibiting high sensitivity, specificity, and excellent scalability
. PLoS One
26. Kusano S, Eizuru Y. Interaction of the phospholipid scramblase 1 with HIV-1 Tat results in the repression of Tat-dependent transcription
. Biochem Biophys Res Commun
27. Luo W, Zhang J, Liang L, Wang G, Li Q, Zhu P, et al. Phospholipid scramblase 1 interacts with influenza A virus NP, impairing its nuclear import and thereby suppressing virus replication
. PLoS Pathog
28. Yang J, Zhu X, Liu J, Ding X, Han M, Hu W, et al. Inhibition of hepatitis B virus replication by phospholipid scramblase 1 in vitro and in vivo
. Antiviral Res
29. Park SE, Lee MJ, Yang MH, Ahn KY, Jang SI, Suh YJ, et al. Expression profiles and pathway analysis in HEK 293 T cells overexpressing HIV-1 Tat and nucleocapsid using cDNA microarray
. J Microbiol Biotechnol
30. Wang J, Yang B, Hu Y, Zheng Y, Zhou H, Wang Y, et al. Negative regulation of Nmi on virus-triggered type I IFN production by targeting IRF7
. J Immunol
31. Heinonen MT, Kanduri K, Lahdesmaki HJ, Lahesmaa R, Henttinen TA. Tubulin- and actin-associating GIMAP4 is required for IFN-gamma secretion during Th cell differentiation
. Immunol Cell Biol
32. Schnell S, Demolliere C, van den Berk P, Jacobs H. Gimap4 accelerates T-cell death
33. Yang Y, Tikhonov I, Ruckwardt TJ, Djavani M, Zapata JC, Pauza CD, et al. Monocytes treated with human immunodeficiency virus Tat kill uninfected CD4(+) cells by a tumor necrosis factor-related apoptosis-induced ligand-mediated mechanism
. J Virol
34. Herbeuval JP, Boasso A, Grivel JC, Hardy AW, Anderson SA, Dolan MJ, et al. TNF-related apoptosis-inducing ligand (TRAIL) in HIV-1-infected patients and its in vitro production by antigen-presenting cells
35. Fernandez S, Tanaskovic S, Helbig K, Rajasuriar R, Kramski M, Murray JM, et al. CD4+ T-cell deficiency in HIV patients responding to antiretroviral therapy is associated with increased expression of interferon-stimulated genes in CD4+ T cells
. J Infect Dis
36. Migueles SA, Connors M. Long-term nonprogressive disease among untreated HIV-infected individuals: clinical implications of understanding immune control of HIV
37. Jacobs ES, Keating SM, Abdel-Mohsen M, Gibb SL, Heitman JW, Inglis HC, et al. Cytokines elevated in HIV elite controllers reduce HIV replication in vitro and modulate HIV restriction factor expression
. J Virol
2017; 91: e02051–16.