Using an IRB-approved protocol (Mass General Brigham IRB#2022P000210, Baylor College of Medicine IRB#H-42256), mini-bronchoalveolar lavage (mini-BAL) samples were obtained from mechanically ventilated patients within 72 hours post-lung transplantation. PGD was adjudicated using established ISHLT guidelines (5). The demographic characteristics of the human study cohort are summarized in Tables 1 and 2. Briefly, mini-BAL was performed by administering 30 ml of sterile saline via the endotracheal tube, followed by immediate suctioning using an 8-Fr suction cannula. Mini-BAL was then centrifuged at 3,500 revolutions/min for 15 min, and the supernatant was stored at −80°C until analysis using published methodologies (9).

Under an IACUC-approved protocol, murine lung transplant surgery was performed (MGB IACUC No. 2022N000156). A murine model of primary graft dysfunction through single orthotopic lung transplantation after prolonged cold ischemia of 18 hours (OLT-PCI) was performed using established techniques (14, 15). There were four groups in the study: A) Control group: a seven- to nine-week-old male BALB/c donor left lung was implanted into a C57BL/6 male (The Jackson Laboratory) mouse immediately after procurement without any cold ischemia, B) PGD group: seven- to nine-week-old male BALB/c donor left lung was implanted into a C57BL/6 male (The Jackson Laboratory) mouse after 18 hours of cold ischemia (donor lungs wrapped in perfadex solution and kept in 4 degree Celsius. C) Vehicle PGD group: seven- to nine-week-old male BALB/c donor left lung was implanted into a C57BL/6 male (The Jackson Laboratory) mouse after 18 hours of cold ischemia (donor lungs wrapped in perfadex solution and kept in 4 degree Celsius. Mouse received 50 μL of phosphate-buffered saline (PBS)intravenously, via penile vein injection, prior to reperfusion, D) RTR PGD group: seven- to nine-week-old male BALB/c donor left lung was implanted into a C57BL/6 male (The Jackson Laboratory) mouse after 18 hours of cold ischemia (donor lungs wrapped in perfadex solution and kept in 4 degree Celsius. Mouse received RTR, 250 μg in 50 μL PBS intravenously, via penile vein injection, prior to reperfusion. All mice were sacrificed after 4 hours of reperfusion using humane techniques and bronchoalveolar lavage (BAL), blood and tissue specimens collected. Likewise, we conducted a similar PGD model using syngeneic (B6→B6) transplant with similar ischemic parameters and RTR/vehicle to evaluate the role of adaptive versus innate immunity in RTR mediated PGD attenuation.

During the assessment of oxygenation and lung function, the right pulmonary hilum, including the accessory lobe, was occluded with a vascular clamp under median sternotomy, and the left lung was ventilated with a tidal volume of 5 mL/kg and a respiratory rate of 120 breaths/min. Mice were mechanically ventilated using a volume-controlled ventilator (Rovent, Harvard Apparatus, Holliston, MA, USA). Dynamic compliance and inspiratory capacity were measured using the built-in lung mechanics module of the Rovent ventilator. The fraction of inspired oxygen was maintained at 1.0. Five minutes after occlusion of the right hilum, arterial blood gas parameters were analyzed using an i-STAT handheld analyzer (Abbott Laboratories, Abbott Park, IL, USA) according to the manufacturer’s instructions.

Briefly, after anesthesia, the native right lung bronchus was clamped first, and then BAL was performed via a tracheal cannula using 0.7 mL of sterile PBS, which was instilled three times and suctioned back sequentially. BAL fluid was centrifuged at 1500 rpm for 10 minutes at 4°C. the supernatant was separated into aliquots and frozen at -80°C until analysis. The supernatant was then analyzed for PGP peptide levels. The remaining cell pellets from all the lavages were used for cell counts. Differential cell counting was performed on air-drive cytospin preparations stained by the Protocol HEMA3 stain set. At least 200 cells were counted and the absolute number of neutrophils was calculated.

Upon euthanasia (CO₂ inhalation with cervical dislocation as a secondary method), the lungs were removed and immersed in fresh fixative for at least 24 hours. Following paraffin embedding, sections were cut and stained with hematoxylin and eosin for histologic and morphometric analysis. Histological scores of acute lung injury were graded on a scale based on morphological appearance: normal (0%), mild (<10%), moderate (10–50%), or severe (>50%) abnormalities, corresponding to scores of 0, 1, 2, and 3, respectively (16). Five randomly selected alveolar areas at high-power fields (HPFs) were evaluated from each lung section, with a total of four sections analyzed per group (20 HPFs in total). The number of macrophages, was quantified using ImageJ2 software (NIH, Bethesda, MD, USA). Likewise, for immunofluorescence studies, the tissue sections were deparaffinized, rehydrated, and subjected to antigen retrieval. The sections were then blocked and incubated overnight at 4°C with primary antibodies. After washing, secondary antibodies were applied, and the slides were examined under a microscope. The complete immunofluorescence protocol, including details on the antibodies and detection methods, is provided in the Supplementary Methods.

For murine lung lysate analysis, one section of lung was collected, homogenized, and resuspended in PBS. Total MMP-9 and prolyl endopeptidase (PE) concentrations in the lung tissue lysates were quantified using commercially available ELISA kits according to the manufacturers’ protocols (MMP-9: R&D Systems; PE: MyBioSource). For human BAL samples, MMP-9 and PE levels were measured using human-specific ELISA kits from Abcam based on a sandwich ELISA format. Values below detection limits were excluded. Murine lung tissue was analyzed for inflammatory cytokines using a multiplex cytokine assay. Details in Supplementary Methods.

Lung tissue homogenates were prepared in lysis buffer without protease inhibitors and centrifuged at 12,000 × g for 10 min at 4°C. Equal amounts of protein (54 μg) were mixed with non-reducing sample buffer and loaded onto 10% SDS–polyacrylamide gels containing 0.1% gelatin (Sigma-Aldrich). After electrophoresis, gels were washed twice in 2.5% Triton X-100 for 30 min to remove SDS and renature the gelatinases. Gels were then incubated in enzyme activation buffer (50 mM Tris-HCl, pH 7.5, 5 mM CaCl₂, 0.02% NaN₃) at 37°C for 18 h. Following incubation, gels were stained with 0.5% Coomassie Brilliant Blue R-250 and destained until clear lytic bands appeared against a blue background. Gel images were captured, and band intensities corresponding to MMP-9 activity were quantified using ImageJ software (NIH, Bethesda, MD,USA).

PGP peptides in murine and human BAL samples were quantified using an MDS Sciex (Applied Biosystems) API-4000 triple-quadrupole mass spectrometer coupled with a Shimadzu high-performance liquid chromatography system and a 2.0 x 150 mm Jupiter 4u Proteo column (Phenomenex) [7].

For statistical testing, normality for each data set was assessed by Shapiro-Wilk test followed by the appropriate statistical testing (t -test or non-parametric, 2-tailed, alpha cutoff for statistical significance was <0.05). data was presented as mean+/- SEM. Details of individual tests used are provided in the figure legends. GraphPad Prism version 10.0 was used to analyze and plot the data. Illustrations were generated using BioRender Illustrator.

To determine whether the matrikine proline-glycine-proline (PGP) contributes to primary graft dysfunction (PGD), we examined specimens from our murine PGD model. Histology revealed extensive interstitial and perivascular neutrophilic inflammation, pleural thickening, and disruption of alveolar architecture in PGD allografts relative to controls (Figure 1A). To confirm the known pro-inflammatory phenotype of PGD, we conducted multiplex cytokine analysis of lung tissue lysates, which revealed a trend toward increased levels of inflammatory cytokines, including IL1α, IL1β, IP10, RANTES, and IL6, in PGD compared to controls (Supplementary Figure 1). Immunofluorescent staining for myeloperoxidase (MPO) confirmed a pronounced increase in neutrophil accumulation within PGD grafts (Figures 2A, B), underscoring neutrophil invasion as an early pathological feature. Because PGP arises during neutrophil-mediated extracellular-matrix degradation, we quantified its bioactive acetylated form (acPGP) in whole-lung lysates and found significantly higher levels in PGD allografts than in controls (Figure 1B). Our previous work showed that neutrophils are an inducible source of matrix metalloproteinase-9 (MMP-9) and prolyl endopeptidase (PE), enzymes essential for PGP generation. Consistent with this pathway, PE and MMP-9 were elevated in PGD lung lysates compared with controls (Figures 2D, E).

To confirm human clinical relevance, we measured PGP levels in bronchoalveolar lavage (BAL) fluid collected 72 hours after transplantation from recipients with grade 3 PGD (PGD3) and from non-PGD (PGD 0) controls. BAL PGP concentrations were markedly elevated in PGD samples (Figure 1C). To add rigor and to validate the pathway further, we utilized an independent human BAL cohort and found that BAL MMP9 and PE levels were significantly higher in the PGD3 specimens compared to PGD0 group (Figures 1D, E) paralleling the murine findings and implicating the PGP axis in human PGD.

To evaluate whether pharmacologic blockade of PGP can attenuate PGD, we treated allografts with the PGP-neutralizing tripeptide L-arginine-threonine-arginine (RTR). Gross inspection after prolonged ischemia revealed visibly less edema and erythematous inflammation in RTR-treated grafts compared to the vehicle-treated PGD group (Figure 3A, left panel). Histology confirmed a significantly reduced neutrophil infiltration, interstitial edema (Figure 3A, right panel), and overall PGD injury scores in RTR-treated grafts (Figure 3B). Total protein concentration in BAL fluid in RTR-treated PGD mice tended to be lower compared with vehicle-treated PGD mice (Supplementary Figure 2). Oxygenation, as assessed by arterial blood gas analysis, and allograft lung function were significantly improved in the RTR-treated group compared with the vehicle group (Figures 3C–E). Further, H&E staining showed reduced infiltration of macrophage-shaped cells in the allografts of the RTR-treated group (Figure 3F). Immunofluorescent staining for myeloperoxidase (MPO) demonstrated lower signal intensity compared with vehicle-treated PGD grafts (Figures 2A, B), underscoring the attenuation of neutrophilic inflammation. Consistent with these findings, bronchoalveolar lavage showed a marked decrease in total neutrophil cell counts in RTR-treated lungs compared to the vehicle-treated PGD group (Figure 2C). Next, to investigate whether adaptive immunity was associated with the PGP forward feedback lung injury observed in PGD, we used a syngeneic (B6→B6) transplant model of PGD. Our results showed a similar attenuation of PGD in RTR-treated syngenic transplant compared to vehicle control (Supplementary Figure 3), indicating that the observed benefit from RTR was not dependent on adaptive alloimmunity.

Because RTR attenuated both PGD severity and neutrophilia, we next interrogated its impact on the protease network that generates PGP. MMP-9 and PE are integral to generation of PGP. ELISA assays demonstrated that RTR treatment significantly reduced MMP-9 expression in PGD allografts compared to the vehicle-treated PGD group (Figure 2D). Likewise, Prolyl endopeptidase, required for PGP release, was down-regulated in RTR-treated mice (Figure 2E), suggesting that RTR intercepts the PGP-mediated feed-forward loop, thereby curbing further neutrophil recruitment and protease release. Importantly, RTR reduced both proteases to near-control levels. To add further rigor, we conducted MMP9 immunofloroscence which confirmed reduced MMP-9 expression with RTR treatment (Figures 4A, B). Next, we performed gelatin zymography to measure active MMP-9. Active MMP-9 levels were significantly lower in the RTR-treated group compared to vehicle controls (Figures 4C, D).