We included LTx patients who underwent LTx within our center between September 2003 and November 2012, and of whom pretransplantation serum was available. Prior to transplantation, patients were assessed for transplant eligibility via classical cross-match testing. Pretransplant HLA antibodies were measured via the LABScan 100 flow analyzer (One Lambda, CA, USA) or ELISA (LAT, One Lambda), as described previously (7). All patients were treated with standardized immunosuppressive regime consisting of tacrolimus, basiliximab, prednisolone, and mofetil mycophenolate. Patients depicted as being at risk for CMV or EBV reactivation (defined as a CMV−/EBV− patients receiving a graft from a CMV+/EBV+ donor) were prophylactically treated with valganciclovir up until 6 months after transplantation. Informed consent in accordance with the Declaration of Helsinki was obtained from all the patients, and this study was approved by the medical ethical committee of the University Medical Center Utrecht (METC 06-144). All methods were carried out in accordance with the approved guidelines. Serum samples from 20 healthy controls (HC) who donated blood for research purposes were obtained, processed, and stored at −80°C until further usage.

To reduce the risk for thromboembolic complication shortly after LTx (21), the grafted lungs were flushed antegradely via the pulmonary artery with perfadex solution. During this procedure, the lungs were ventilated at tidal volume and topically cooled. After explantation, the lungs were flushed for a second time, but now via the pulmonary vein, until the fluid became clear and upon inspection was free of blood clots. This flush fluid was collected and centrifuged for 10 min at 1,800 rpm. The cell pellet was dissolved in phosphate-buffered saline (PBS) (Sigma-Aldrich, St. Louis, MO, USA) and subsequently loaded on Ficoll-Paque (GE Healthcare, Milwaukee, WI, USA) for cell separation. The cells were isolated from the interphase, put in aliquots of 1 ml in RPMI-1640 (Gibco, Waltham, MA, USA) supplemented with DMSO (Sigma-Aldrich) and 20% v/v fetal bovine serum (FBS) (Bodinco, Alkmaar, The Netherlands), and stored in liquid nitrogen until further processing.

Both whole perfusate and frozen cell aliquots were inspected for the presence of lung EC via FACS analysis using a FACSCanto and FACSDiva software (BD Biosciences, Durham, NC, USA) and FlowJo v.10 (FlowJo, LLC, Ashland, OR, USA). Either whole perfusate cell samples or thawed frozen cell aliquots were stained with a combination of CD45-PO (Thermo Fisher Scientific, Rockford, IL, USA), CD31-FITC (BD Biosciences), CD13-PE (BD Biosciences), CD90-BV421 (BioLegend, San Diego, CA, USA), VEGFR-2-AF647 (BD Pharmingen), or respective isotypes (Biolegend and R&D Systems, Minneapolis, MN, USA). Intracellular staining for Von Willebrand factor (vWF) was conducted using vWF-APC (R&D systems) and Intrastain kit (DAKO, Glostrup, Denmark).

For cell culturing, frozen cell samples were thawed at 37°C, washed with RPMI-1640 20% FBS, centrifuged for 10 min at 1,800 rpm, and the pellet was dissolved in EC-specific growth medium (EBM-2, Lonza, Basel, Switzerland) and additives for vascular EC (EGM-2 SingleQuots, Lonza). Cells were cultured in gelatin (Sigma-Aldrich) pre-coated nunclon T-25 culture flasks (Sigma-Aldrich) in EBM-2 supplemented with primocin (Invivogen, San Diego, CA, USA) until confluency. Confluent flasks were treated with accutase (BD Biosciences), and cells were subcultured (50,000 per flask) in nunclon T-75 culture flasks. Cell morphology was assessed via light microscopy (Leica DM IL, Leica Microsystems, Wetzlar, Germany), and cell numbers analyzed via Bürker–Türk cell counting chambers (Laboroptik Ltd., Lancing, UK).

Jurkat cells (ATCC, Manassas, VA, USA) were cultured in T-75 culture flasks in RPMI-1640 10% FCS supplemented with 1% penicillin/streptomycin (Thermo Fisher Scientific) and subcultured 1:6 twice a week.

IgG was purified from frozen serum samples using Magne Protein G Beads (Promega, Madison, WI, USA) according to manufacturer’s instructions. Briefly, 50 μl beads were incubated with 50 μl pre-centrifuged thawed serum samples for 30–60 min at room temperature while shaking. Supernatant was removed via magnetic separation of the beads, which were subsequently washed multiple times. Purified IgG was eluted from the beads, and the purified IgG quality and concentration was obtained using the NanoDrop^® ND-1000 system (Thermo Fisher Scientific). Purified IgG fractions were stored at 4°C until further usage.

Anti-AJC and anti-AEC antibodies were assessed via flow cytometry. Jurkat cells were co-incubated with 1 μl of 1 μM staurosporine (Sigma-Aldrich) O/N at 37°C to induce apoptosis. Cells were centrifuged for 5 min at 1,800 rpm, the supernatant discarded, and the cell pellet was dissolved in 5 ml culture medium. One-hundred microliters of cell suspension were added to a 96-well round bottom plate (Sigma-Aldrich), centrifuged (5 min, 1,500 rpm), and subsequently were incubated with 60 μl purified IgG (diluted 1:2 in PBS) for 30 min at 37°C. Each well was washed twice with 150 μl PBS and centrifuged (5 min, 1,500 rpm). The supernatant was discarded, and the cells were incubated with mouse anti human IgG PE (BioLegend), diluted 1:12.5 in PBS for 30 min at 4°C. Cells were washed once with 150 μl PBS and once with 150 μl annexin V binding buffer (1 l H₂O + 10 mM Hepes + 140 mM NaCL + 2.5 mM CaCl₂⋅2H₂O; pH 7.4). Cells were centrifuged, the supernatant discarded, and100 μl binding buffer supplemented with 4 μl 7AAD (BD Pharmingen, San Diego, CA, USA) and annexin V FITC (ITK Diagnostics, Uithoorn, The Netherlands), diluted 1:2, was added followed by 10 min incubation at room temperature and fluorescence measurement.

In order to induce apoptosis, lung EC were subjected to serum starvation for 48–72 h at 37°C, culturing with staurosporine O/N or exposed to UV light using an UV stratalinker (Stratagene, Santa Clara, CA, USA). Serum starvation yielded the largest amount of apoptotic EC (35–40%). After serum starvation, cells were washed with PBS and detached with 4 ml accutase for 5 min at 37°C. Cell detachment was determined via light microscopy. Six milliliters of culture medium was added to the culture flask and centrifuged for 5 min at 1,500 rpm. The supernatant was discarded and the cell pellet dissolved in PBS. One hundred microliters cell suspension/well (V bottom 96-well plate, Sigma-Aldrich) and incubated with 60 μl purified IgG (mean concentration 548 ± 154 μg/ml), diluted 1:2 in PBS for 30 min at 37°C. Subsequent antibody incubations were identical to the previous described Jurkat protocol. All measurements for either Jurkat or primary EC were conducted on the same day. To determine the intra- and inter-assay variation, cells were seeded in duplicate and identical sera measured on multiple days. The inter- and intra-assay variability for this assay were 18 and 4.1%, respectively.

The D’Agostino and Pearson omnibus normality test was used to test the data for Gaussian distribution. Differences in not-normally distributed data were analyzed via the Mann–Whitney test, whereas Student’s t-tests were used to asses differences in normally distributed data. Correlation analyses on non-normally distributed data were performed with the Spearman’s rank correlation coefficient. Statistical analyses were performed using GraphPad Prism version 6.02 (GraphPad Software Inc., San Diego, CA, USA) and SPSS version 20 (IBM Corp., Armonk, NY, USA). Information on statistical testing and graph display is depicted in the respective figure legend. A p-value <0.05 was considered to be statistically significant.

A total of 80 patients who underwent LTx within our center between September 2003 and November 2012, and of whom pretransplantation serum was available, were included in this study. Twenty-nine of these patients were diagnosed with BOS based upon ISHLT criteria (22), and in our total cohort 22 patients died during follow-up, of whom 9 were diagnosed with BOS. Seventeen patients developed one or more episodes of AR, diagnosed, since biopsies were unavailable, as a spontaneous decline of FEV₁ reversible via steroid-pulse treatment. One patient was positive for anti-HLA antibodies directed at DQ4, whereas none of the other patients presented with pretransplant anti-HLA antibodies, as described previously (7). Further clinical and demographic parameters are depicted in Table 1.

First, we analyzed perfusate samples from six patients for the presence of vascular EC as determined by FACS. Vascular lung EC were identified as being CD45−/VEGFR-2+/CD13+ and confirmed by intracellular staining of vWF. The percentage of EC in each perfusate was assessed by analyzing 1,000,000 cells. This percentage ranged from 0.0002 to 0.0256% (see Figure 1A). For further analysis of ECs, cells in the perfusate were first expanded in primary cell cultures. Individual ECs were observed within the first week after initial cell seeding. Cell numbers sufficient for subculturing or cell marker analysis were reached in 2–5 weeks (Figure 1B). ECs were cultured for a maximum of nine passages.

Endothelial cells were characterized via cell morphology, cell surface, and intracellular markers. Cells in culture presented typical EC morphologic characteristics. Cell density per passage differed. Furthermore, monolayers were observed in later culture passages, and cell morphology remained constant during culturing (Figure 1C, passages 1–4). ECs were non-lymphoid (CD45−). Cells could be cultured without any morphological indications of cell death until passage nine. FACS analysis showed that cultured EC expressed both EC-specific markers CD13 and CD31 on their cell surface. Furthermore, intracellular staining revealed these cells to be positive for vWF (Figure 1D).

As these cells have been used frequently as a source of apoptotic cells, we first tested the presence of antibodies against apoptotic Jurkat cells. Apoptosis was induced as described in the Section “Materials and Methods.” Overall apoptosis rates in Jurkat cells were around 60–70% of total cell numbers. Higher levels of antibodies against apoptotic Jurkat cells were observed in LTx patients before transplantation as compared to HC (p < 0.001, Figure 2A).

Endothelial cells induced to apoptosis were analyzed with flow cytometry. Annexin V+/7AAD− cells were considered to be early apoptotic cells and Annexin V+/7AAD+ cells late apoptotic. Overall apoptosis rates in EC were around 20–30% of total cell numbers. Apoptosis induction via exposure to UV radiation did not result in sufficient apoptotic EC numbers (<5%, data not shown). Purified pre-LTx serum IgG of 80 end-stage lung disease patients was analyzed for binding to apoptotic EC, from 2 different lung donors, together with serum IgG of 20 HC. We observed a higher IgG reactivity to apoptotic ECs in pre-LTx serum samples compared to HC (p = 0.0018; Figure 2B).

The difference between apoptotic specific IgG levels between end-stage lung disease patients and HCs was higher on apoptotic Jurkat cells. We did not observe differences in IgG antibody reactivity to apoptotic targets when early and late apoptotic ECs or Jurkat cells were differentiated. Stratification of serum IgG titers per end-stage lung disease, clustered as COPD, CF, and ILD did not show any significant differences. Also, no significant differences were observed when age, autoimmune disease, or anti-HLA antibodies were taken into account. Furthermore, we observed no correlation between anti-AJC and anti-AEC (Figure 2C). A heat-map representation of the individually observed mean fluorescent intensity (MFI) values is depicted in Figure 2D.

We analyzed the relation of pre-LTx levels of anti-AJC and anti-AEC and outcome after LTx. First, we stratified patients by development of BOS during follow-up. Twenty-nine patients were diagnosed with BOS. No differences were observed between serum IgG levels of anti-AEC or anti-AJC in BOS+ and BOS− patients (Figures 3A,B, respectively). Also, these pre-LTx levels did not correlate to time to BOS diagnosis or time to death (data not shown).

Second, we investigated the predictive value of pre-LTx serum IgG levels on the incidence of AR post-LTx. In total, 17 patients were diagnosed with one or more episodes of AR. We observed no significant differences between serum levels of anti-AEC or anti-AJC in LTx patients with or without AR (Figures 3C,D, respectively).

Since we did not observe any predictive value concerning pretransplant levels of serum IgG reactivity to apoptotic cells, we investigated levels of these antibodies at the time of BOS diagnosis. We isolated serum IgG from serum samples obtained from 10 patients within a month that BOS was diagnosed. In parallel, we selected serum samples from 10 patients who did not develop BOS and were matched for gender, age, primary lung disease, and month post-LTx. IgG was purified from these samples as described in Section “Materials and Methods.” No increase in IgG reactivity was observed against either apoptotic lung EC (Figure 4A) or apoptotic Jurkat cells (Figure 4B).

In order to gain insight into the course of IgG antibodies against apoptotic cells, we divided the time from transplantation to BOS diagnosis in four even time frames (quartile-based selection). Subsequently, we isolated IgG from serum samples obtained during each quartile and analyzed these samples for IgG reactivity. This procedure was also conducted for the matched BOS− patients. Overall, we observed an increased level of IgG antibodies against apoptotic Jurkat cells in BOS+ patient compared to BOS− patients during follow-up. However, this difference is not significant and does not seem to be discriminative at time points prior to clinical manifestations of chronic rejection (Figure 4C).