HIV induces airway basal progenitor cells to adopt an inflammatory phenotype

Despite the introduction of anti-retroviral therapy, chronic HIV infection is associated with an increased incidence of other comorbidities such as COPD. Based on the knowledge that binding of HIV to human airway basal stem/progenitor cells (BC) induces a destructive phenotype by increased MMP-9 expression through MAPK signaling pathways, we hypothesized that HIV induces the BC to express inflammatory mediators that contribute to the pathogenesis of emphysema. Our data demonstrate that airway BC isolated from HAART-treated HIV+ nonsmokers spontaneously release inflammatory mediators IL-8, IL-1β, ICAM-1 and GM-CSF. Similarly, exposure of normal BC to HIV in vitro up-regulates expression of the same inflammatory mediators. These HIV-BC derived mediators induce migration of alveolar macrophages (AM) and neutrophils and stimulate AM proliferation. This HIV-induced inflammatory phenotype likely contributes to lung inflammation in HIV+ individuals and provides explanation for the increased incidence of COPD in HIV+ individuals.


Culture of primary human airway basal cells.
For the studies with normal, HIV‾ basal cells exposed in vitro to HIV, normal human airway basal cells (Lonza, Walkersville, MD) were seeded at 3000 cells/cm 2 into plastic flasks and maintained in BEGM medium supplemented with 1% penicillin/streptomycin (Life Technologies, Grand Island, NY), 0.5% amphotericin B (Life Technologies) and 0.1% gentamicin (Sigma, St Louis, MO) in a 5% CO 2 , 37 °C humidified incubator with media replaced every 2 to 3 days. Once the cells reached 70% confluence they were harvested with 0.05% trypsin-EDTA (Life Technologies) for 5 min at 37 °C, with the reaction stopped by addition of HEPES supplemented with 15% fetal bovine serum (FBS, Life Technologies). www.nature.com/scientificreports/ Production of infectious HIV stocks. HIV-1 stocks (X4-tropic NL4-3 and R5-tropic AD8) were generated by calcium phosphate transfection of HEK293T cells (MBS mammalian transfection kit, Agilent Technologies, Santa Clara, CA) with the proviral vector, pNL-4-3, replication-defective pNL4-3-luc and pNLAD8 (NIH AIDS Research and Reference Reagent Program). Culture supernatants were collected after 48 h posttransfection and passed through 0.45 μm filters. Mock supernatants from non-transfected 293 T cells were collected for each preparation. Both mock and viral supernatants were prepared with the same number of 293 T cells and culture medium except no proviral plasmid in mock during transfection. Mock and viral supernatants were concentrated using LentiX reagent (Clontech, Mountain View, CA) and the viral pellet was resuspended in tissue-culture graded PBS after centrifugation. Quantification of the HIV p24 capsid protein was determined by HIV-1 p24 ELISA (Immuno Diagnostics, Woburn, MA) according to manufacturer's instructions. The volume of viral supernatant added to the BC was in the range of 1:30 -1:60 dilution in BC culture (i.e., not more than 34 µl per ml of culture medium).

Cytokine gene expression by TaqMan PCR.
To assess the effect of HIV on basal cell cytokine gene expression, basal cells (5 × 10 4 ) were exposed to increasing HIV input (p24 from 5 to 200 ng/ml in 300 μl) for 2 days. Heat-inactivated HIV was used as a control. Total RNA was extracted using Trizol reagent (Invitrogen) and the aqueous phase was purified using an RNAEasy MinElute RNA purification kit (Qiagen). RNA concentration was determined using a NanoDrop ND-100 spectrophotometer (NanoDrop Technologies). First-strand cDNA was synthesized from 0.5 µg of total RNA using TaqMan Reverse Transcription Reagents with random hexamer as primer (Applied Biosystems). All samples were analyzed in triplicate at cDNA dilution of 1:10. All reactions were run on an Applied Biosystems Sequence Detection System 7500 and relative expression levels determined using the dCt method with 18S ribosomal RNA as an endogenous control. The primers were obtained from Applied Biosystems, including IL-8 (Hs00174103_m1), GM-CSF (Hs00929873_m1), ICAM-1 (Hs00164932_m1), IL-1β (Hs00174097_m1) and 18s RNA (Hs99999901_s1). All primers were optimized for TaqMan PCR and referenced in recent studies 56-60 . Western analysis. BC  . Following the primary antibody incubation, membranes were washed three times for 5 min each with PBST, followed by incubation with an anti-rabbit, anti-goat or anti-mouse antibody conjugated to horseradish peroxidase (1:10,000; Santa Cruz Biotechnology) in PBS containing 3% BSA for 1 h at room temperature with shaking. The membranes were then washed again three times for 5 min with PBST and twice with PBS, and antibodies visualized after the addition of ECL Western analysis detection reagents (GE Healthcare Biosciences) by exposure to X-ray film.

ELISA.
To quantify IL-8 and GM-CSF levels released by BC, culture supernatants were harvested from basal cell cultures and spun at 10,000 rpm to remove cell debris. IL-8 and GM-CSF levels in the culture supernatants were measured using ELISA kits (RayBiotech, Norcross, GA), according to the manufacturer's instructions.

Alveolar macrophage (AM) migration. Human alveolar macrophages from normal nonsmokers (n = 3)
were obtained from bronchoalveolar lavage (BAL) from healthy nonsmokers as previously described 47 . BAL fluid was filtered through 2 layers of gauze and centrifuged at 1200 rpm for 5 min, 4 °C. Cells were washed twice in RPMI 1640 containing 10% FBS, 50 U/ml penicillin, 50 μg/ml streptomycin, and 2 mM glutamine (Invitrogen, Carlsbad, CA). Cells were suspended in complete medium, and cell viability was estimated by trypan blue exclusion and expressed as a percentage of the total cells recovered. Total cell number was determined by counting in a hemocytometer. Cells recovered from BAL fluid were seeded in 6-well plastic culture dishes (10 6 per ml) and alveolar macrophages were purified by 24 h adherence at 37 °C in a 5% CO 2 humidified incubator. Nonadherent cells were removed by thorough washing with RPMI 1640. AM represented ≥ 95% of cells 47,61-63 . Chemotaxis was measured using CytoSelect 24-well cell migration assay kit (pore size: 5 μm, Cell Biolabs). Briefly, alveolar macrophages (3 × 10 5 cells in 200 μl) in serum-free RPMI were plated in the upper compartment overnight. Cells were washed with sterile PBS twice and 200 μl of serum-free RPMI was added to the upper chamber. BC-conditioned media (200 μl) were diluted with serum-free RPMI in 1:1 (total volume: 400 μl) added to the lower compartment. For blocking experiments, BC-conditioned media was pre-incubated with anti-IL-8 neutralizing antibody (clone # 6217) and corresponding mouse IgG1 control (clone # 11,711; both at 2 μg/ml, endotoxin level < 0.10 EU per 1 μg of the antibody by the LAL method; R & D Systems) for 30 min at 37 ºC prior to addition to the lower chamber. The chemotactic chambers were incubated for 3 h at 37 ºC in humidified incubator with 5% CO 2 . Following termination of the assay, the bottom of polycarbonate filters were removed from the chambers and rinsed with 200 μl of sterile PBS to dislodge the migratory cells underneath the filter to the lower chamber. The numbers of macrophages that had migrated from the top to the bottom of the polycarbonate filter were counted by light microscopy. Neutrophil Isolation Kit, StemCell Technologies, Cambridge, MA). Briefly, 5 ml of fresh blood collected in a K2EDTA tube was incubated with 250 μl of isolation cocktail and 250 μl of Rapid Sphere (StemCell Technologies) at 23 °C, 5 min. The sample was topped up to 12 ml by adding Mg 2+ and Ca 2+ -free PBS containing 1 mM EDTA and mixed gently. The sample tube was placed in a Big Easy magnet (StemCell Technologies) and incubated for 5 min to allow unwanted cells attached to the magnet. The sample tube was inverted, and enriched cell suspension was poured into a new tube for another round of magnetic separation. A pure neutrophil cell suspension (CD16 + and CD66b + ) was obtained after three rounds of separation and purity was > 95% as assessed by flow cytometer (BD Fortessa; see Supplemental Fig. 9). Chemotaxis was measured using the CytoSelect 24-well cell migration assay kit (pore size, 3 μm, Cell Biolabs). Freshly isolated neutrophil (3 × 10 5 cells in 200 μl per insert) in serum-free RPMI were plated in the upper compartment. BC-conditioned media (100 μl) were diluted with serum-free RPMI in 1:4 (total volume: 400 μl) added to the lower compartment. For blocking experiments, BC-conditioned media were pre-incubated with mouse IgG 1 control and anti-IL-8 neutralizing antibody (both at 2 μg/ml, R & D Systems) for 30 min, 37ºC prior to addition to the lower chamber. The chemotactic chambers were incubated for 1 h, 37ºC in humidified incubator with 5% CO 2 . Following termination of the assay, the bottom of polycarbonate filters were removed from the chambers and rinsed with 200 μl of sterile PBS to dislodge the migratory cells underneath the filter to the lower chamber. The numbers of macrophages that had migrated from the top to the bottom of the polycarbonate filter were counted by light microscopy.
Effect of BC on alveolar macrophage proliferation. Freshly isolated alveolar macrophages (5 × 10 4 ) in serum-free RPMI were plated in 96 well for 2 h to allow cell attachment. Cells were washed with PBS twice and incubated with complete RPMI. To assess the effect of BC-released mediators on AM proliferation, BrdU Cell Proliferation ELISA kit was used (Abcam). Conditioned media from untreated, HIV-treated and heat-inactivated HIV-treated BC (20 μl) were diluted in complete RPMI in 1:5 (total volume: 100 μl) and added to AM. For blocking experiment, BC-conditioned media were pre-incubated with mouse IgG 1 control and anti-GM-CSF neutralizing antibody (both at 10 μg/ml, R & D Systems) for 30 min at 37ºC prior to addition to AM culture. After 3 days of incubation, cells were pulsed with BrdU for 16 h. BrdU (a thymidine analog) was incorporated into the DNA of dividing cells. To detect incorporated BrdU, cells were fixed, permeabilized and incubated with detector anti-BrdU monoclonal antibody for 1 h. Unbound antibody was washed away and horseradish peroxidase-conjugated goat anti-mouse antibody was added which bound to the detector antibody. The horseradish peroxidase catalyzed the conversion of the chromogenic substrate tetra-methylbenzidine (TMB) from a colorless solution to a blue solution and turned yellow after the addition of stopping reagent. The intensity of colored product was proportional to the amount of BrdU incorporated in the cells. The colored reaction product is quantified using a spectrophotometer set at a dual wavelength of 450/550 nm. Each sample was run in triplicate.

Statistical analysis.
Unless otherwise mentioned, data were expressed as mean ± SD from three independent experiments. Statistical analysis was carried out and the data were subjected to statistical analysis using unpaired two-tailed Student's t-tests with unequal variance. Values of p < 0.05 were considered significant.

HIV induces human airway BC to express and release inflammatory mediators.
With the knowledge that the lung remains a reservoir of HIV + in HAART-treated individuals despite low blood HIV levels [20][21][22][23]25,26,28 , we hypothesized that HIV may interact with BC and induce the BC to release inflammatory mediators. We have previously shown that some HIV bind to BC, activating the BC through heparan sulfate proteoglycan receptors, but HIV does not replicate in normal human airway BC 46 . HIV binding to the BC initiates a cascade of events mediated through MAPK/ERK signaling pathways and induces increased expression and secretion of matrix metalloproteases (MMP), including MMP-9, capable of inducing tissue destruction 46 .
To evaluate the hypothesis that HIV induces normal BC to acquire an inflammatory phenotype that may contribute to this inflammatory process, BC were exposed to increasing HIV NL4-3 (p24 from 5 to 200 ng/ml) for 48 h in vitro. Total RNA and culture supernatants were analyzed for cytokine gene expression and release. We focused on GM-CSF, IL-8, IL-1β and ICAM-1 as these mediators are relevant to COPD pathogenesis and are found at increased levels in epithelial lining fluid from HIV + individuals 35,63 and HIV + BC-conditioned media when cells were cultured ex vivo (Fig. 1A). TaqMan quantitative PCR showed HIV induced upregulation of GM-CSF, IL-8 and IL-1β in a dose-dependent fashion (p < 0.05 as compared to no viral input; Supplemental Fig. 3). BC were also treated with HIV or heat-inactivated HIV for 48 h. When compared to untreated and heat-inactivated HIV, HIV significantly induced expression of IL-1β ( Fig. 2A), IL-8 (Fig. 2B), ICAM-1 (Fig. 2C)  www.nature.com/scientificreports/   Fig. 6). The corresponding cytokine levels were evaluated in BC-conditioned media by Western analyses (Fig. 3). Secreted and activated form of IL-8 and IL-1b were quantified in this study. Consistent with the TaqMan gene expression data, there were increased levels of IL-1β, IL-8, ICAM-1 and GM-CSF in HIV (both NL4-3 and AD8, lanes 3 and 5; Fig. 3) when compared to untreated and heat-inactivated HIV-treated BC respectively (lanes 1, 2 and 4; Fig. 3).  www.nature.com/scientificreports/ HIV induces BC to release mediators that induce migration of AM and neutrophils. We next examined whether the mediators released by BC could functionally influence other cells in the lung such as AM and neutrophils, cells that are important components in the pathogenesis of emphysema/COPD 70,71 . Consistent with the gene expression data, there was an increased secretion of IL-8 from HIV-exposed BC vs untreated (4.35 vs 0.86 ng/ml, p < 0.0002) and vs heat-inactivated HIV (4.35 vs 1.32 ng/ml, p < 0.0005) as confirmed by ELISA (Fig. 4A). Heat-inactivated HIV did not have any effects on IL-8 release (Fig. 4A).
To assess if HIV-BC released mediators could attract inflammatory cells, AM and neutrophil migration assays were carried out using HIV-BC conditioned media. Human AM from normal healthy nonsmokers were plated on the upper chamber of transwell and incubated with conditioned media from untreated BC, heat-inactivated HIV-treated BC and HIV-treated BC in the lower chamber, and the number of migratory cells were quantified after 3 h of incubation. There was an increase in the number of migratory AMs in HIV-treated BC as compared to untreated control (p < 0.005; Fig. 4B) and heat-inactivated HIV (p < 0.01; Fig. 4B).
Chemoattractant studies with normal AM and normal neutrophils demonstrated that the IL-8 released by BC exposed to HIV was functional, capable of attracting normal human (Fig. 4B) and neutrophils (Fig. 4D). With the restriction of the COVID-19 pandemic, we had difficulties to obtain fresh AM from one individual to set up all treatment groups in a single experiment. Using cells from other individuals, pretreatment of HIVtreated BC-conditioned media with anti-IL-8 antibody significantly blocked migration of AM when compared to pretreatment of HIV-treated BC-conditioned media with the mouse IgG 1 control (p < 0.003; Fig. 4C). Similar to AM, there was an increased number of migratory neutrophils in HIV-treated BC conditioned medium as compared to untreated BC (p < 0.001; Fig. 4D) and compared to heat-inactivated HIV-BC (p < 0.005; Fig. 4D). In separate experiments using cells from different nonsmokers, pretreatment of HIV-BC conditioned media with anti-IL-8 significantly inhibited neutrophil migration as compared to HIV-BC conditioned media pretreated with mouse IgG 1 control (p < 0.001; Fig. 4E). Same findings were observed from a complete experiment including all the controls and treatment groups using neutrophils from one individual (Supplemental Fig. 7). In addition, we assessed neutrophil migration using conditioned medium from BC of HIV‾ and HIV + nonsmoker cultured ex vivo. Compared to HIV‾ nonsmoker BC, HIV + BC released chemotactic factors which increased neutrophil migration, suggesting that HIV + BC exhibited inflammatory phenotypes (p < 0.0005; Supplemental Fig. 8).

HIV-induced BC release mediators that stimulate AM proliferation. Based on the cytokine array
and Western analyses, there was an increased release of GM-CSF level in HIV-reprogrammed BC. This was confirmed by ELISA, with increased GM-CSF in the supernatants of HIV-exposed BC vs untreated (9 vs 45 pg/ml; p < 0.01; Fig. 5A) and vs heat-inactivated HIV (8.4 vs 45 pg/ml; p < 0.01; Fig. 5A). BC exposed to heat-inactivated HIV did not have any increased GM-CSF in the supernatant (Fig. 5A).
To examine if GM-CSF released by HIV-BC could stimulate AM proliferation, freshly isolated AM were exposed to BC-conditioned media from untreated, heat-inactivated HIV and HIV-BC for 3 days. Assessment using BrdU incorporation revealed that HIV-BC conditioned stimulated AM proliferation compared to untreated (p < 0.01; Fig. 5B) and vs heat-inactivated HIV (p < 0.01; Fig. 5B). Pretreatment of HIV-BC conditioned media with anti-GM-CSF significantly inhibited AM proliferation compared to mouse IgG 1 control (p < 0.04; Fig. 5B).

Discussion
With the advent of effective anti-retroviral therapy, the mortality rate of individuals infected with HIV has been significantly declined 2 . However, the increased life span within the HIV + population has been accompanied by an increased prevalence of aging-related chronic disorders, including COPD 3,4,[6][7][8]11,12,72 . The mechanisms of HIV-associated COPD is not fully understood but is clearly associated with chronic lung inflammation, dominated by activated AM, CD8 + T cells and neutrophils [13][14][15]30 . A variety of studies have shown the decline in lung function and pulmonary/airway abnormalities observed in HIV + individuals despite effective HAART treatment 64,[66][67][68][69]73,74 . Previous reports have shown that the lung is a reservoir for HIV despite systemic viral suppression with anti-retroviral therapy [20][21][22][23]25,26,30,31 . Given our recent observation that HIV can bind to heparan sulfate receptors on airway BC, resulting in reprogramming of the BC to up-regulate expression and release of MMP-9 46 , we hypothesized that HIV binding to BC may also induce the BC to an inflammatory phenotype that contributes to the chronic AM/neutrophil lung inflammation that characterizes HIV infection despite anti-retroviral therapy. Our findings demonstrated that small airway BC isolated from HAART-treated HIV + nonsmokers release inflammatory mediators IL-8 IL-1β, ICAM-1 and GM-CSF. Consistent with the ex vivo data, HIV upregulates expression of these inflammatory mediators in normal airway BC in vitro. These HIV-BC derived mediators mediate migration of AM and neutrophils, and proliferation of alveolar macrophages in vitro.

The lung in HIV infection.
The lung serves as a HIV reservoir in both untreated and HAART-treated individuals [22][23][24]75 . Despite systemic viral suppression, HAART does not eliminate HIV reservoirs in compartments such as lung, lymph nodes, colon and the genital tracts 75 . Immunohistochemical study showed that gp120 immunoreactivity was observed in both HIV-infected lungs after HAART treatment 28 . Based on these studies, it appears that HIV-1 may compartmentalize to the area within the lung where HAART is not as effective. In the lung, AM are a major target for HIV-1 20 . HIV preferentially infects AM and impairs AM phagocytic activity 76 . Several studies have been shown that HIV-infected AM release oxidants and various cytokines to trigger immune activation, neutrophil infiltration, and express proteases such as MMPs for degradation of extracellular matrix 20,47,77 . It is conceivable that residual HIV in the lung might infect AM which serve as viral reservoirs to productively replicate HIV in the lung. Newly synthesized viral particles interact with neighboring HIV receptor-expressing cells such as epithelial cells, dendritic cells and lymphocytes. HIV binding to surface heparan sulfate expressed in airway basal cells reprograms BC to acquire destructive phenotypes and induces www.nature.com/scientificreports/ www.nature.com/scientificreports/ MMP-9 upregulation through MAPK signaling pathway 46 . The amount of viral load in the lung compartment of HAART-treated HIV individuals has not been well studied and quantified. In this study, BC were exposed to increasing HIV input (p24 from 5 to 200 ng/ml). HIV upregulates the expression of inflammatory genes in a dose-dependent manner (Supplemental Fig. 3). More importantly, BC isolated from HAART-treated HIV + nonsmokers showed increased secretion of inflammatory mediators ex vivo when compared to HIVnonsmokers (Fig. 1).
Our findings demonstrated that HIV-treated BC released mediators that induced AM and neutrophil migration. Ex vivo BC from HAART-treated HIV nonsmokers also released chemotactic factors, for example, IL-8 that induced neutrophil migration. Relevant to COPD pathogenesis, the numbers of both AM and neutrophils significantly increased in sputum and BAL of COPD patients [86][87][88] . Both AM and neutrophils play an important role in airway inflammation. They secrete various serine proteases, including matrix metalloproteinases, proteinase 3, cathepsins and neutrophil elastase which cause mucus hypersecretion and alveolar destruction 70,71,89-92 . Airway epithelium as "inflammatory cells". While HIV-infected AM play a role in mediating lung inflammation associated with HIV infection, little attention has been focused on a role of the airway epithelium in the pathogenesis of HIV associated lung inflammation. It is known that HIV can disrupt lung epithelial integrity 38 , suppresses tracheobronchial mucociliary clearance 93 and that the HIV-1 envelope protein gp120 www.nature.com/scientificreports/ stimulates mucus formation in bronchial epithelial cells via the CXCR4/α7-nicotinic acetylcholine receptor pathway 28 . The present study adds to the concept that airway epithelium BC contribute to the initiation and maintenance of the chronic lung inflammation associated with HIV infection by HIV binding to, and modulating the BC to adapt an inflammatory phenotype that includes increased expression of IL-1β, IL-8, ICAM-1 and GM-CSF. This is consistent with prior studies demonstrating that human differentiated airway epithelial cells have the capacity to contribute to airway inflammation 70,[94][95][96][97] . Airway epithelial cells are capable of releasing a variety of pro-inflammatory mediators including chemokines, cytokines and growth factors through various signaling pathways in response to cigarette smoke and other pathogens such as respiratory syncytial virus and influenza virus 70,[95][96][97] . These mediators play an important role in the pathogenesis of COPD by orchestrating inflammation through recruitment, activation and survival of inflammatory cells 70 .
There is extensive data demonstrating that HIV infection is associated with inflammation in the lung 11,20,33,[35][36][37]42,98 . Compared to HIV‾ nonsmokers, analysis of lung epithelial lining fluid showed that HIV + nonsmokers has demonstrated increased levels of a variety of inflammatory mediators, including GM-CSF, IFN-γ, IL-5, IL-6, IL-13, IL-8, MCP-1, CCL4, RANTES, I-TAC, I-309, IL-23, IL-17, PAI-1 and CD40 compared to HIV‾ nonsmokers 63 . Similarly, another study reported that chronic untreated HIV infection dysregulated the lung cytokine microenvironment and was incompletely restored by long-term ART 99 . Analysis of bronchoalveolar lavage fluid cytokine levels showed a significant increase in IL-8, IL-1β and IL-6 in asymptomatic HIV-infected ART-naive and ART-treated HIV + individuals as compared to healthy individuals 99 . Increased levels of IL-8 have been found in bronchoalveolar lavage from HIV-infected nonsmokers and HIV + subjects with P. carinii pneumonia 35,63,85 . AM from HIV + individuals spontaneously release more GM-CSF and IL-8 than those from normal controls 30,80 , and there is an increased secretion of IL-1β from AM of HIV + individuals as compared to normals 82,100,101 . Similar to our observations, X4 tropic HIV enhance the expression of ICAM-1 in in vitro differentiated airway epithelium 38 . ICAM-1 mediates interaction between polymorphonuclear leukocytes and airway epithelial cells and transmigration of neutrophils and macrophages in the lung 102 . In this case, HIV-induced ICAM-1 promotes immune cell infiltration, activation and local inflammation/damage in the lung. In addition to airway BC, it deserves further investigation that other mature airway differentiated cells including ciliated, mucus and club cells may also contribute to inflammation after HIV exposure.
Several mechanisms of HIV-associated lung diseases have been proposed including oxidative stress, imbalance of proteases/anti-proteases, direct effect of HIV or HIV-related proteins, use of HAART and opportunistic infections 7,11,37 . Using BC isolated from HAART-treated HIV + nonsmokers in this study has limitations to precisely demonstrate if increased inflammatory cytokine expression is solely from HIV interaction with BC. It remains to be determined if other factors such as other HIV proteins, antiviral drugs and opportunistic infections may also contribute to lung inflammation in HIV + nonsmokers. Further studies will be required to better understand the mechanisms of HIV-associated lung disorders.
Taken together, we have identified a novel, unexpected mechanism that HIV induces BC to adopt an "inflammatory phenotypes", releasing mediators that attract and activate alveolar macrophages and neutrophils. These findings explain how lung inflammation in HIV + individuals are generated and perpetuated and why there is an increased incidence of COPD in HIV + individuals in the era of HAART.