HIV Infects Bronchial Epithelium and Suppresses Components of the Mucociliary Clearance Apparatus

Recurrent lung infections and pneumonia are emerging as significant comorbidities in the HIV-infected population in the era of combination antiretroviral therapy (cART). HIV infection has been reported to suppress nasal mucociliary clearance (MCC). Since the primary components driving nasal MCC and bronchial MCC are identical, it is possible that bronchial MCC is affected as well. Effective MCC requires optimal ciliary beating which depends on the maintenance of the airway surface liquid (ASL), a function of cystic fibrosis transmembrane conductance regulator (CFTR) activity and the integrity of the signaling mechanism that regulates ciliary beating and fluid secretion. Impairment of either component of the MCC apparatus can compromise its efficacy and promote microbial colonization. We demonstrate that primary bronchial epithelium expresses HIV receptor CD4 and co-receptors CCR5 and CXCR4 and can be infected by both R5 and X4 tropic strains of HIV. We show that HIV Tat suppresses CFTR biogenesis and function in primary bronchial epithelial cells by a pathway involving TGF-β signaling. HIV infection also interferes with bronchial epithelial cell differentiation and suppresses ciliogenesis. These findings suggest that HIV infection suppresses tracheobronchial mucociliary clearance and this may predispose HIV-infected patients to recurrent lung infections, pneumonia and chronic bronchitis.


Introduction
MCC is a primary innate defense mechanism of mammalian airways that protects the host from the noxious effects of airborne pathogens, pollutants and allergens [1]. The MCC apparatus consists of the cilia, a protective mucus layer, and a periciliary Airway surface liquid (ASL) layer to optimize ciliary beating [2]. Abnormalities in any compartment of the mucociliary system can compromise mucus clearance leading to mucus impaction and consequently, chronic bacterial infection [3][4][5]. The height of the ASL layer lining the airway surfaces is crucial for mediating MCC rates [6] and is tightly regulated by CFTR [7]. CFTR dysfunction can have a pronounced effect on ASL depth as well as ciliary beating and can contribute directly to a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 microbial colonization. Bacterial pneumonia and COPD are the most prevalent lung comorbidities in people living with HIV [8,9]. Lung infections are exacerbated in HIV-infected smokers [10,11] and this could be due to the ability of cigarette smoke to independently attenuate both CFTR function [12,13] and ciliary beating. HIV-infected patients show abnormalities in their nasal MCC apparatus [14,15]. However, nasal Clefflux and CBF is often measured as a barometer of overall airway MCC health [16][17][18]. Hence it is possible that tracheobronchial mucociliary clearance is affected as well. HIV has also been recovered from cellfree bronchoalveolar lavage fluid [19] suggesting that it can directly mediate its effects in the airway.
In our earlier report we have demonstrated that TGF-β signaling and cigarette smoke (via TGF-β signaling) suppresses CFTR biogenesis and function [12]. HIV Tat can induce TGF-β signaling in different cells types [20][21][22] possibly by binding to a Tat responsive element in the TGF-β1 promoter [22,23]. In this study, we will show that HIV Tat also suppresses CFTR biogenesis and function via TGF-beta signaling. Tat can be expressed and secreted from infected cells even in presence of cART due to a protein transduction domain. Secreted Tat can be taken up by bystander cells where it can have pleotropic effects. [24,25]. Considering reports that HIV can productively [26] or non-productively [27] infect epithelia, we will show that primary human bronchial epithelium expresses canonical HIV receptors and can be infected with HIV. Its effects on the Mucociliary clearance apparatus will be discussed.

HIV Tat suppresses CFTR biogenesis and function via TGF-β signaling
NHBE ALI cultures were treated with Tat (heat inactivated Tat as control). Several groups have reported HIV Tat in the serum in the sub-nanomolar range [28,29] and Tat may also bound to endogenous glycosaminoglycans and heparin sulfates in vivo which cannot be measured [30] We have used a concentration of Tat at a dose found in the serum and also used by others studies [31][32][33][34]. After forty-eight hours, cells were lysed and total RNA was extracted and analyzed for changes in TGF-β and CFTR mRNA by qRT-PCR. Tat increased TGF-β1 mRNA levels with a concomitant decrease in CFTR mRNA levels ( Fig 1A). To determine if decrease in CFTR mRNA translates to a corresponding decrease in CFTR function, NHBE redifferentiated at the ALI were similarly treated with Tat, forty-eight hours post-initial treatment, cells were mounted in Ussing chamber and CFTR specific short circuit current was determined by adding albuterol as reported [12]. Separately, to determine if Tat exerts its effect via TGF-β signaling, another set of Tat treated NHBE ALI cultures were pretreated with anti-TGFBR2 before Tat addition. As seen in Fig 1B, HIV Tat mediated suppression of CFTR mRNA leads to a concomitant suppression of CFTR function. Anti-TGFBR2 antibody was able to rescue HIV Tat mediated suppression of CFTR function suggesting that HIV Tat mediates suppresses CFTR via TGF-β signaling. The extent for mRNA suppression with Tat and TGF-β (previous report [12]) directly relates to the extent of suppression of CFTR function suggesting that suppression of CFTR biogenesis is responsible for suppression of CFTR function.

Bronchial epithelial cells express canonical HIV receptors and can be infected with HIV
Expression of HIV receptors and co-receptors has been demonstrated in other epithelial cells [26] which can be productively [26] or non-productively [27] infected. We first analyzed expression of CD4, CCR5 and CXCR4 in NHBE by qRT-PCR (CFTR used for comparison) and western blot (3 different lungs). As seen in Fig 2A and 2B, NHBE redifferentiated at the ALI expressed all canonical HIV receptors. While detectable RNA levels were lower for CCR5, protein levels were comparable with CD4 and CXCR4. This difference in detectable RNA and protein levels for CCR5 could be due to inherent CCR5 mRNA instability due to a pseudoknot structure in the mRNA that promotes non-sense mediated decay [35] and tightly regulates CCR5 expression. To determine if we observe similar expression in vivo, bronchial brushings from HIV-infected patients was analyzed for expression of HIV receptors and co-receptors. We observed a similar pattern of expression in bronchial brushings with maximal expression observed for CD4 followed by CXCR4 and CCR5 ( Fig 2C).
Next, we determined if HIV could infect NHBE redifferentiated at the ALI. Fully differentiated NHBE ALI cultures were infected basolaterally and apically with the X4-tropic HIV (HIV IIIB) or the prototypic R5-tropic HIV (HIV BaL) for 16 hours. At designated time points, culture supernatants were collected and analyzed for p24 production by ELISA. To exclude measuring any p24 in the supernatant from residual input virus, sixteen hours post-infection, cells were washed four times with PBS apically and basolaterally and the final wash (Day 0) was collected and analyzed for p24. As seen in Fig 2D, no detectable p24 is observed on day 0 after washes suggesting that p24 levels on subsequent time points could only have been due to infection of NHBE cells. Maximal p24 production is observed on day 3 with both X4-tropic and R5-tropic HIV that declines progressively until Day 9. On Day 9, total RNA was isolated from these cells and analyzed for HIV cell-associated RNA using Taqman probe (life technologies cat # PA03453409_s1). As seen in Fig 2E we   We analyzed expression of CD4, CCR5 and CXCR4 in NHBE by qRT-PCR. Total RNA was isolated from NHBE ALI cultures and expression of CD4 (Applied Biosystems # HS01058407-m1), CXCR4 (Applied Biosystems # HS00607978-s1) and CCR5 (Applied Biosystems # HS99999149-s1) was determined by qRT-PCR using Taqman probes. CFTR (Applied Biosystems # HS00357011-m1) expression was determined for comparison. NHBE ALI cultures express all canonical HIV receptors. Panel B: Western blot analysis from three different lungs demonstrates that NHBE ALI cultures show comparable expression of all HIV receptors. While RNA levels were significantly lower for CCR5 than CD4 and CXCR4, protein levels were comparable with CD4 and CXCR4. This could be due to inherent CCR5 mRNA instability due to a pseudoknot structure in the mRNA that promotes non-sense mediated decay [35]. Panel C: To determine if bronchial epithelial cells express HIV receptors and co-receptors in vivo, Bronchial brushings were analyzed for expression of CD4, CCR5 and CXCR4 by qRT-PCR. Expression patterns of the three receptors were similar to that observed in our ex vivo model with maximal expression of CD4 followed by CXCR4 and significantly lower CCR5 expression validating the physiological relevance of our ex vivo NHBE ALI culture model. Panel D and E: NHBE ALI cultures can be infected with both R5 and X4-tropic strains of HIV. NHBE cultures redifferentiated at the Air-Liquid Interface were infected apically and basolaterally with either HIV IIIB (X4-tropic) or HIV BaL (R5-tropic) strains. After 16 hours cells were washed apically and basolaterally with PBS four times to remove any residual input virus. The fourth wash was collected for p24 analysis and measured as Day 0 to confirm that all input virus had been removed. No p24 was detected in the 4 th wash (Day 0) confirming that all input virus had been removed. We observe an initial spike in p24 output on Day 3 for both HIV IIIB and HIV BaL infections that declines gradually till Day 9 (panel D). Experiments were terminated and total RNA was isolated and cell associated HIV RNA was quantitated by qRT-PCR using Taqman probe (Applied Biosystems # PA03453409-s1 cultures were stained for the pan-leukocyte antigen marker CD45 and analyzed by flowcytometry. NHBE cells were found to be CD45 negative confirming that the primary bronchial cells do not have any contaminating immune cells (S1 Fig). These data suggest that p24 output and viral RNA are due to infection of primary bronchial epithelial cells.
HIV establishes long-term restricted infection in primary bronchial epithelial cells Next we tried to determine if HIV was capable of establishing long-term infection in NHBE ALI cultures. We infected NHBE cultures redifferentiated at the ALI with HIV IIIB and HIV BaL as described before. Culture supernatants were collected for p24 analysis for an initial eight days and then over long-term at 21, 25, 41, 45 and 50 days. Experiments were terminated on Day 50 due to excessive cell death observed in infected NHBE ALI cultures. Total RNA was recovered and analyzed for HIV RNA by qRT-PCR. As seen in Fig 3A and 3B, a spike in p24 is observed that declines by day 8. However, stable levels of detectable p24 are observed up to 50 days post-infection. qRT-PCR for HIV RNA analysis 50-days post-infection, shows cell-associated viral RNA in all HIV BaL infected and two out of three HIV IIIB infected NHBE ALI cultures ( Fig 3C). The single HIV IIIB infected culture in which we were not able to detect cell associated viral RNA was due to very low levels of RNA that were recovered from this culture due to excessive cell death.  of integrated HIV was detected using the 2-step PCR approach described by Brussel and Sonigo [41]. This approach specifically detects only integrated viral DNA. Simultaneously another set of similarly infected cells were tested for presence of episomal viral DNA using the approach described by Buzon et. al. [42]. As seen in Table 1 we detected both integrated and episomal HIV DNA.

HIV infects NHBE cells and undergoes reverse transcription to generate proviral DNA
To confirm that HIV undergoes reverse transcription followed by gene expression, NHBE cells redifferentiated at the ALI were infected with the fluorescent RGH-WT HIV reporter reported by Dahabieh et. al. [43]. This virus expresses EGFP under the control of HIV LTR and mCherry under the control of a CMV promoter inserted in the env ORF. The virus is envand can be pseudo-typed with any envelope. Due to a lack of Env expression, it undergoes only a single round of infection. RGH-WT was enveloped with either R5-tropic env (pBaL.01) or X4-tropic envelope (pHXB2 env) in HEK 293T. 10ngs p24 equivalent of both viruses was used to infect NHBE ALI cultures and infection was allowed to proceed for 12 days. Even though the virus expresses both GFP and mCherry we decided to focus on GFP expression alone as this is directly under the control of HIV LTR and would be more representative of LTR mediated transcription. Moreover CMV promoter is not ideal for bronchial epithelial cells as it is rapidly silenced in these cells [44][45][46]. As seen in Fig 4, GFP expression is observed in NHBE ALI cultures infected with RGH-WT virus enveloped with R5 and X4 envelope on Days 6 and 12. On Day 12 post-infection, significant cell death leading to disruption of epithelial barrier was observed in infected NHBE cells (Fig 4B). These data indicate that HIV RNA is reverse transcribed upon entry into NHBE ALI cultures to proviral DNA and expresses GFP from the HIV LTR. Since the RGH-WT virus is incapable of productive infection, any cell  [50]. As seen in Fig 6 the thickness and integrity of the redifferentiated bronchial epithelium is greatly diminished in cells infected with HIV. In areas where the epithelium had formed, ciliogenesis was greatly diminished. These data indicate that HIV also suppresses ciliogenesis by infecting undifferentiated cells.

Discussion
In this manuscript, We demonstrate for the first time that HIV infects both differentiated and undifferentiated primary bronchial epithelial cells. Upon infection, the viral RNA undergoes reverse transcription to express viral proteins Tat and p24. Tat suppresses CFTR biogenesis and function via TGF-β signaling. Infection of differentiating NHBE ALI cultures leads to suppression of ciliogenesis. Hence HIV infection can suppress ASL depth maintenance while at the same time decreasing ciliogenesis. People living with HIV show abnormalities in their nasal MCC apparatus [14,15]. However, nasal Clefflux and CBF is often measured as a barometer of overall airway MCC health [16][17][18]. Our data suggest that infection of primary bronchial epithelium could be responsible. Effective MCC requires the combined action of ASL depth maintenance and ciliary beating. In normal conditions, apical nucleotides (ATP and its metabolites) are important for hydrating airway surfaces (e.g. see [51]). ATP binds to purinergic G-protein coupled receptors, leading to activation of Ca 2+ dependent Clchannels and CFTR. Normal mucociliary function fails when the capabilities of the cAMP and ATP/ Ca 2+ -mediated apical Clefflux fail, leading to airway surface dehydration (e.g. see [52]). Since optimal ASL depth allows optimal ciliary beating any ASL dysregulation will also manifest as impaired ciliary beating. Likewise suppression of ciliogenesis or inhibition of signaling mechanisms that regulate ciliary beating can affect MCC as well. Attenuation of MCC leads to mucus impaction. The accumulated mucus entraps bacteria and promotes chronic bacterial infection [3]. TGF-β signaling upregulated in smoker and in chronic airway diseases suppresses CFTR biogenesis and function [12]. This could explain recurrent lung infections in people living with HIV. We first looked at the effects of HIV Tat given reports that HIV Tat induces TGF-β1 expression in a number of cell types [20][21][22]. Expression of HIV Tat cannot be suppressed by antiretrovirals [53][54][55][56]. Its protein transduction domain allows its secretion by infected cells and uptake by bystander cells where it mediates pleotropic effects [30, [57][58][59]. Circulating Tat in the serum of HIV patients can reach nanomolar concentrations [28,29]. We used a dose that

HIV infects undifferentiated NHBE cells and inhibits ciliogenesis. Panel A:
NHBE ALI cultures were infected as described before on the day they were placed under Air-liquid culture conditions. Uninfected lung matched cultures were similarly differentiated as controls. At designated time points, culture supernatants were collected and analyzed for p24. Undifferentiated NHBE ALI cultures can support HIV infection. Panel B: On day 18 when ciliogenesis was observed in Lung-matched uninfected controls, experiments were terminated and total RNA was analyzed by qRT-PCR for presence of HIV RNA. HIV RNA was detected in both HIV IIIB and HIV BaL infected NHBE cells. n = NHBE ALI cultures from 3 different lungs. * = significant (p < 0.05). Panel C: Another set of cells were infected similarly was On Day 18 post-infection when ciliogenesis was observed in lung matched uninfected controls, Experiments were terminated, cells were fixed and stained for cilia as described by us earlier [50] and visualized using a Zeiss fluorescence microscope with high resolution Axiocam 506 mono microscope camera (Zeiss, Germany). HIV infection of NHBE ALI cultures impairs epithelial layer formation and interferes with ciliogenesis. Primary bronchial epithelium is pseudostratified. In a healthy well-differentiated epithelium, all ciliated cells will not be in the same plane. Hence some cilia are sharply focused in a particular plane while others, out of focus appear blurred. In infected cells, we do not observe the formation of the pseudostratified epithelium and the epithelial integrity is also impaired hence ciliated cells appear in a single plane. correlates with Tat found in the serum and also used in other studies [29,[31][32][33][34]. Our observations demonstrate that HIV Tat suppresses CFTR biogenesis and function. A neutralizing antibody to TGFBR2 rescued the effect of HIV Tat on CFTR function suggesting that Tat mediates its effects on CFTR via TGF-β signaling.
Our results demonstrate that primary human bronchial epithelial cultures redifferentiated at the ALI express canonical HIV receptors CD4, CCR5 and CXCR4 and can be infected with both R5-tropic and X4-tropic HIV. This infection could be blocked by respective CCR5 and CXCR4 inhibitors. Gundavarapu et. al. [60] have demonstrated expression of CXCR4 but not CCR5 mRNA in NHBE cells by qRT-PCR. Although we were able to detect very low levels of CCR5 mRNA in our cultures, western blot analysis showed that CCR5 protein levels were comparable with CD4 and CXCR4. This could be due to inherent CCR5 mRNA instability due to a pseudoknot structure that functions in combination with microRNAs to promote nonsense mediated decay [35]. It is possible that expression of CCR5 is tightly regulated in bronchial epithelial cells or epithelial cells in general. Indeed both HeLa and Vero cells used in that study are of epithelial origin. Although virus production by NHBE cultures was lower than that reported for cultures of primary macrophages and T cells, it was within the range of values that have been reported for other non-immune primary cells [61,62].
We followed HIV infection over 50 days and observed that initially, p24 output declines rapidly but then stabilizes in the long-term. Most studies with epithelial cells have demonstrated these kinetics. One of the explanations for this decline in p24 could be due to a combined effect of failed integration events and death of infected cells. While all classes of multiply-spliced, singly spliced and unspliced viral mRNA transcripts can be observed from episomal proviral DNA [53,55], the proportion of these transcripts varies greatly with multiply splice transcripts being abundant while levels of unspliced transcripts being significantly lower [53,55]. This is due to low levels of Rev that are transcribed from unintegrated DNA [54]. Lack of rev limits nuclear export of full-length HIV transcripts, which in turn limits the levels of gag and env expression. Hence, we are able to observe low levels of p24 in the culture supernatants up to Day 50. RNA analysis by qRT-PCR Day 50 post-infection demonstrates cell-associated HIV RNA in both X4-infected and R5 infected NHBE ALI cultures. Upon entry, HIV undergoes reverse transcription to generate proviral DNA that is transcriptionally active as shown by qPCR analysis (Table 1) as well as using fluorescent HIV virus [43] (Fig 4A and 4B). In cells infected with the fluorescent virus, we observed significant cell death Day 12. Since the virus is env -, it is only capable of single round of infection suggesting that cell death could be due to expression of HIV genes. HIV-infected NHBE ALI cultures also secrete functionally active Tat that was able to transactivate Tat inducible GFP in co-cultured HEK-293 cells transfected with LTRshGFPpA. Together our observations demonstrate that NHBE ALI cultures can be infected with HIV and express viral proteins like Tat to suppress CFTR biogenesis and function. Our data agree with some observations of Brune et. al., [36] in that we also observe that HIV impairs lung epithelial integrity and we are able to detect viral RNA in infected cells. However we have also been able to demonstrate viral entry, reverse transcription and gene expression in primary human bronchial epithelial cells redifferentiated at the ALI. Our data differs from observations of Brune et al. in that they were unable to detect proviral DNA. This could be due to several differences in our experimental approach as well as culture conditions. Our NHBE ALI cultures are redifferentiated following a single expansion (Passage 1) after being isolated from the lung while Brune et. al. have primarily used commercially available cells which may have been expanded more than once before being provided to the consumer. Indeed we have observed that repeated passages alter the epithelium in subtle ways. We have observed decreased number of ciliated cells and lower CFTR activity when cells are passaged more than once (data not shown). We have used a different approach to detect proviral DNA.
It is known that genomic DNA isolation methods and presence of high molecular weight genomic DNA can inhibit PCR [63] precluding detection of low levels of integrated or episomal viral DNA. We used a 2-step qPCR approach reported earlier to detect integrated HIV DNA [41]. For episomal DNA We have used plasmid miniprep isolation protocols to exclude genomic DNA. These methods enrich our DNA of interest before the qPCR step and improve PCR sensitivity [64,65].
Finally we demonstrate that HIV can also infect undifferentiated bronchial epithelial cells and impairs lung epithelial integrity and ciliogenesis. Most clinical studies have reported undetectable levels of HIV in patients on cART. Hence de novo infection of primary bronchial epithelial cells may be suppressed in presence of cART. However recurrent lung infections and other chronic inflammation associated with cigarette smoke or inhaled substance abuse can lead to recruitment of infected immune cells. These cells along with infected bronchial epithelial cells can serve as HIV reservoirs in the lung and lead to bursts of HIV Tat. In addition, these studies involve rigorous follow-up by research coordinators to minimize the incidence of missed doses. Hence it is unrealistic to assume that all patients will have undetectable levels of HIV all the time. The situation is exacerbated in people living with HIV who also abuse drugs as they intentionally miss cART doses when planning drug use [66,67]. Under these circumstances increased Tat expression or de novo infection of bronchial epithelial cells leading to disruption of bronchial epithelial integrity, decreased ciliogenesis and inhibition of CFTR biogenesis and function. In conclusion, this study demonstrates that HIV can infect NHBE ALI cultures and this results in suppression of components of the MCC apparatus by Tat mediated CFTR suppression, epithelial disruption and decreased ciliogenesis.

Cell culture
All experiments were performed with primary human bronchial epithelial cells, re-differentiated at the air-liquid interface (ALI). ALI cultures were prepared as described by Fulcher and Randall [68,69] and adapted by us [12,50]. Cells were isolated from appropriately consented donors whose lungs were found unsuitable for transplantation for reasons unrelated to airway disease and provided by the University of Miami Life Alliance Organ Recovery Agency. Since the material was obtained from deceased individuals with minor, de-identified information, its use does not constitute human subjects research as defined by CFR 46.102. A signed consent of each individual (or legal healthcare proxy) for donation of the lungs for research is on file with the Life Alliance Organ Recovery Organization covers research use of this material. Unless otherwise specified, experiments used cells from non-smokers to not confound the findings. The primary cultures undergo mucociliary differentiation at the ALI reproducing the in vivo morphology and key physiologic processes to regenerate the native bronchial epithelium ex vivo [68,69]. Epithelial cells from bronchial brushings were directly expanded at the ALI and used for experiments.
All methods were carried out in accordance with the approved guidelines. Additionally, stored epithelial cell brushes were used for this study. The material was obtained with clinical and demographic data and everything else was de-identified to protect confidentiality. All participants provided written informed consent. The study was approved by the University of Pittsburgh IRB.

Electrophysiology experiments
Ussing chambers were used to determine CFTR activation [12]. The input resistance of each filter was measured by application of 1 mV bipolar pulses of 2s duration. Amiloride (10 μM) was added apically to eliminate Epithelial sodium channel (ENaC) influences. Following treatment with albuterol short circuit current (Isc) was allowed to stabilize. Experiments were terminated by adding CFTR inh 172 (20 mM) to confirm that changes in I SC were due to CFTR activation.

Detection of HIV proviral DNA
NHBE ALI cultures were infected with HIV IIIB (X4-tropic) or HIV BaL (R5-tropic) strains. Genomic DNA was isolated and integrated HIV DNA was determined by a 2-step qPCR method described by Brussel and Sonigo [41]. For episomal DNA, we used a modification of the method described by Buzon et. al. [42]. DNA was extracted using a Qiagen miniprep kit to isolate only episomal forms of DNA while excluding the genomic DNA. This bacterial plasmid DNA isolation method has been the most commonly used method for isolating 2-LTR circles from infected cells [64,65,70]. Eliminating high molecular weight nucleic acids from the sample, which might inhibit PCR [63], could enrich the episomal DNA and also improve PCR sensitivity. Copy numbers were determined by comparative Ct method using logarithmic dilutions of plasmid pNL4-3 to generate standard curves.

Western blot analysis
Total protein was loaded onto a gel and run at 100V. Protein was transferred to PVDF membrane. Following transfer, blot was blocked in 5% milk then incubated in CD4 primary antibody (1: 500; Sigma-Aldrich), CXCR4 (1:1000; Thermo scientific) or CCR5 (1:1000; Thermo scientific). Blot was incubated in an anti-rabbit secondary antibody diluted to a concentration of 1:2500. Bands were detected in Chemidoc using Enhanced Chemiluminescence Reagents (ECL; Bio-Rad Laboratories).

Data analysis
Unless otherwise stated, experiments were repeated three times using primary bronchial epithelial cells from at least three lung donors. Statistical significance was evaluated using unpaired t tests for two groups or ANOVA followed by Tukey Kramer honestly significant difference test for multiple comparisons as appropriate. P < 0.05 was considered significant. MABC002F, Emd Millipore) were also analyzed. Data were acquired with Amnis 1 Flow-Sight 1 Imaging Flow Cytometer. Staining and image collection were carried out according to manufacture's protocol. Images for compensation were collected with compensation beads (552843, BD Biosciences) labeled with the same antibody. Once images were captured, compensation and analysis was carried out with IDEAS 1 image analysis software. From all events collected, first single cells were gated, then FITC+ cells were gated from histogram of intensity of FITC for 10 4 or higher. The bar graphs represent the percentage of gated cells.