The First Allogeneic Transplantation of Human Umbilical Cord-derived Mesenchymal Stem/stromal Cells for Bronchopulmonary Dysplasia: Preliminary Outcomes in Four Vietnamese Infants

Background: Bronchopulmonary dysplasia (BPD) is a severe condition in premature infants that compromises theirlung function and necessitatesoxygen support. Despite major improvements in perinatal care minimizing the devastating effects, BPD remains the most frequent complication of extreme preterm birth. Our study reports the safety ofthe allogeneic administration of umbilical cord-derived mesenchymal stem/stromal cells (allo-UC-MSCs) and the preliminary ecacy of the treatment in four infants with established BPD. Methods: UC tissue was collected from a healthy donor, followed by propagation at the Stem Cell Core Facility at Vinmec Research Institute of Stem Cell and Gene Technology. UC-MSC culture was conducted under xeno-free and serum-free conditions. Four patients with established BPD were enrolled in this study between May 25, 2018, and December 31, 2018. All four patients received two intravenous doses of allo-UC-MSCs (1 million cells/kg patient body weight (PBW) per dose) with an intervening interval of 7 days. Safety and ecacy were evaluated during hospitalization and at 7 days and 1, 6 and 12 postdischargemonths. Results: No transplantation-associated severe adverse events or prespecied adverse events were observed in the four patients throughout the study period. At the time of this report, all patients had recovered from BPD and been weaned off of oxygen support. Chest X-rays and CT scans conrmed the dramatic reduction inbrosis. Conclusions: Allo-UC-MSC transplantation is safe and might improve respiratory function anddecrease lung brosis in preterm infants with established BPD. Trial This study


Background
First discovered in 1967, bronchopulmonary dysplasia (BPD) has since emerged as the most prevalent chronic lung disorder in premature infants, resulting in reductions in alveolarization, vascular growth and overalllung function (1). The pathologic hallmarks of BPD are hyperoxia-induced pulmonary in ammation and tissue degeneration, resulting in signi cant cell death and culminating in impaired and immature alveolarization and dysregulated vascularization of the lung, nally leading to the formation of a brotic lung (2). These patients require prolonged mechanical ventilation and oxygen support (3).
Infants with gestational ages younger than30 weeks are at particularly high risk of immature development of the respiratory system and suffer from detrimental long-term outcomes includinghigh morbidity and mortality rates. In the last 50 years, advances in neonatal medicine, including the discovery of neonatal steroid treatments (4,5), surfactants (6)(7)(8), gentle ventilation treatments (9,10), and effective noninvasive ventilation devices, have signi cantly improvedthe clinical outcomes inpremature newborns with BPD. However, the rates of complications and mortality are still high among infants with BPD (11).
Recently, stem cell transplantation was used to treatBPD in an animal model. Proof-of-concept experiments in neonatal BPD rodent models demonstrated that the injection of bone marrow mesenchymal stem cells (BM-MSCs) via either the intravenous or intratracheal route exertedlungprotective functions, including reductions inf lung in ammation and pulmonary hypertension and reformation of the alveolar structure, leading toa subsequent improvement in the survival rate (12)(13)(14)(15). Furthermore, a single dose of human UC-MSCs administeredintratracheally prevented and rescued neonatal rats from hyperoxia-induced lung damage (16). In humans, Ahn and colleagues conducted the rst phaseI clinical trial using umbilical cord blood-derived MSC (UCB-MSC) transplantation to prevent the manifestation of BPD in premature infants in 2014. Their results con rmed that UCB-MSC transplantation was safe and could reduce the risk of BPD in premature infants (17). In 2017, our group reported the rst patient with established BPD treated successfully withautologous bone marrow mononuclear cells (18). However, obtaining bone marrow from established BPD newborns is a challenging task and carries a major risk of complications, especially infection. Therefore Donor screening criteria for UC tissue Healthy women with an uncomplicated, at term pregnancy underwent serological testing, including tests for HIV, cytomegalovirus (CMV), Epstein-Barr virus (EBV), hepatitis A virus (HAV), hepatitis B virus (HBV), hepatitis C virus (HCV), syphilis, and chlamydia, at 38 weeks of pregnancy. The umbilical cord tissues were collected at delivery and transferred to the laboratory for further processing.

Allo-UC-MSC preparation
The UC sample was processed at the Stem Cell Core Facility at the Vinmec Research Institute of Stem Cell and Gene Technology under ISO 14644-1 (certi cation number: CR61119-1). Culture reagents were purchased from Thermo Fisher Scienti c (https://www.thermo sher.com/) and Pan Biotech (PowerStem MSC1 culture media, P04-77355K, hereinafter MSC culture media), unless stated otherwise. hUC-MSC cultures were conducted under xeno-free and serum-free conditions at 37°C in a humidi ed incubator containing 5% CO 2 . The medium was changed every 3 days until the culture reached 80% con uence, followed by passaging using CTS TM TrypLE TM Select (A1285901). Aliquots of hUC-MSCs at P3 were thawed inCTS™ CELLstart™ substrate-coated asks and cultured using TryPLE passaging in xeno-free and serum-free MSC culture media; under these conditions, hUC-MSCs were routinely passaged by incubation with 1X CTS TM TrypLE TM Select for 4 minutes at 37°C to liberate single cells or, preferably, small clumps of cells and subcultured for further expansion at a seeding density of 5000 cells/cm 2 . Following the absence of the detection of bacteria, fungi, mycoplasma, and endotoxin, the cells were suspended in 10 ml of NaCl 0.9% (Braun, USA) at a nal dose of 1x10 6 cells/kg PBW.

Patient enrollment
Four premature infants who required oxygen support for more than 28 days with ndings of lung brosis on chest CT were enrolled in the study. Prior to intervention, all patients underwent a thorough clinical examination with blood gas analysis, total blood count and chest CT. BPD diagnosis was determined as previously described, including preterm birth (at less than 37 weeks), evidence of respiratory distress syndrome, and the need for oxygen support higher than 21% for up to 56 days in preterm infants (> 32 weeks) and more than 36 weeks of age in preterm infants less than 32 weeks of age (19).

Mode of cell administration
All four patients received two transplantations of allo-UC-MSCs at a dose of 1 million cells/kg patient's bodyweight (PBW) with a 7-dayintervening interval via the intravenous (IV) route. On the day of infusion, harvested cells at the targeted dose were prepared in 10 mL of 0.9% NaCl (Braun, USA) and delivered to the transplantation ward for IV infusion at a rate of 20 mL/hour.

Outcome measures
To assess safety, any major or minor adverse events during the stem cell infusion (72 h) and during the 7 days after transplantation were monitored. Body temperature, blood pressure, respiratory rate, heart rate, and SpO2 were recorded regularly. All four patients were requested to attend re-examination at the hospital at 7 days, 1 month, 6 months and 12 months after discharge. At each visit, a full clinical assessment was performed, including measuring the infant's height and body weight. All medication, home oxygen therapy, and rehospitalization since the last visit were documented. SpO2 and arterial blood gas analysis (ABG) were examined at baseline and at each visit. Chest X-rays and CT scans were performed prior to transplantation, at the 6-month visit (CT scan), and at the 12-month visit (chest X-ray).

Statistical analysis
The data were analyzed using one-way ANOVA with Prism GraphPad software unless otherwise stated. ANOVA was performed to compare the means of the four patients as indicated in the test. Statistical signi cance was de ned as P<0.05 unless otherwise indicated.

hUC-MSC characterization
Our data showed that the UC-MSC line exhibited plastic adherent properties and a spindle-and broblastlike morphology ( Figure 1A), with a population doubling time of 24 ± 0.6 hours (n=3, mean ± SEM).
Propagation of UC-MSCs up to passage 6 did not introduce any karyotypical abnormality, and the cells maintained normal 46XY as indicated by the G-banding technique ( Figure 1B). These cells were also able to form 519 ± 80 CFU/1000 cells (mean ± SEM, n=3) ( Figure 1C). Further analysis of the differentiation potential con rmed that the UC-MSC line could undergo adipogenic, chondrogenic, and osteogenic differentiation processes, illustrated by positive staining with Oil O Red, Alcian Blue, and Alizarin Red, respectively ( Figure 1D). Analysis of the expression patterns of positive markers, including CD73, CD90, and CD105, showed that more than 99% of the cells expressed all these markers and less than 2% expressed negative markers, including CD11b, CD19, CD34, CD45, and HLR-DR ( Figure 1E). These results ful lled the minimum criteria for mesenchymal stem cells proposed by the ISCT (Table 2).

Patient 1
An extremely premature girl (24 weeks and 5 days, rst born of twins) was born by C-section due to premature rupture of the placental membrane with a bodyweight (BW) of 720 grams. Soon after birth, the patient developed signs of respiratory distress syndrome with retraction followed by apnea and cyanosis (SpO 2 ranged from 60% to 70%) and bradycardia with a heart rate below 100 bpm. She was immediately intubated and placed on mechanical ventilation, with a peak inspiratory pressure (PIP) of 18 cmH 2 O and a positive end-expiratory pressure (PEEP) of 5 cmH 2 O. Chest X-ray showed a stage 2 hyaline membrane requiring one dose of surfactant (Curosurf) at 200 mg/kg BW. Heart ultrasound detected patent ductus arteriosus, which closed after one course of paracetamol (15 mg/kg/6 h) for seven days, with no evidence of pulmonary artery hypertension (PAH) on echocardiogram after treatment. In addition, the patient suffered from septicemia caused by Staphylococcus epidermidis,resulting in necrosis at the distal phalanx of the left little and ring ngers and requiring antibiotic treatment. In the rst 2 months, the patient was supported with synchronized intermittent mandatory ventilation (SIMV) and switched to continuous positive airway pressure therapy (CPAP) at7 cmH 2 O and 50% FiO 2 for the following 1.5 months. At 3.5months postnatal age, the patient was diagnosed with BPD and continued to receiveoxygen support via anasal cannula at 0.5 -1 L/min.Nebulized corticosteroids at 100 mcg/kg 4 times/day for a 1-month period wereadministered. A combination of diuretics (furosemide at 1 mg/kg/12 h), spironolactone (2 mg/kg/12 h) and bronchodilators (inhaled b2-agonists) together with nutrient enhancement (high-calorie nutrition and vitamins E, A, K supplementation) were initiated for 2 months.
However, at 4.5 months postnatal age, the patient's BPD was not improved, with the SpO 2 offoxygen support dropping to 90%. The chest CT scan and X-ray at 3.5months postnatal age con rmed the formation of diffuse brosis, atelectasis in the upper lobes of both lungs and signi cant air trapping in both lower lobes (Figure 2A and 3A).
Before transplantation, chronic hypercapnia was con rmed by ABG analysis with the following measured values: pH of 7.31, PaCO 2 of 68 mmHg, HCO 3 -of 41.3 mmol/L, and PaO 2 of 73 mmHg. PAH was determined based ona maximum pulmonary artery pressure (PAP) of 40 mmHg and illustrated onechocardiogram, and the pro-BNP level was high (1942 ng/mL). Oral sidena l (1.5 mg/kg/6 h) and bosentan (1 mg/kg/8 h) were administered when the patient was4 months old. The allo-UC-MSC transplantation was performed at 144 days postnatal age (47 weeks gestational age). No signs of serious adverse events were observed during the two transplantations. Three days after the second transplantation, the patient could breathe spontaneously withan SpO 2 of 96% without oxygen support.
The patient was discharged at 161 days postnatal age (17 days posttransplantation).
At the rst follow-up visit, the patient was alert, hada BW of 4 kg and was spontaneously breathing, with an SpO 2 of 96% without oxygen support. Blood gas analysis revealed a signi cant reduction in thesaturated CO 2 in the blood as follows: pH of 7.5, PaCO 2 of 33.6 mmHg, HCO 3 -of 26.9 mmol/L, BE of 4 mmol/l and PaO 2 of 46 mmHg. These results suggested that the hypercapnic condition was ameliorated after stem cell transplantation. The pro-BNP level had dropped to 351.9 ng/ml, leading to the termination of PAH treatment at4 months posttransplantation. At the 1-month follow-up examination, the patient was cognizant and active, and her BW had increased to 4.3 kg with air fully entering both lungs. She was spontaneously breathing and had an SpO 2 of 97% onroom air without oxygen support. The laboratory tests revealed that herpH (7.37), PaCO 2 (46.3 mmHg), HCO 3 -(27 mmol/L), and PaO 2 (42 mmHg) remained stable posttransplantation. Hematological analysis also con rmedthe absence of in ammation andsepsis, as indicated by the Hgb level (129 G/L), white blood cell count (WBC, 6.1 G/L), and neutrophillevel(6.4%). At the 6-month visit, the patient no longer required oxygen support, with her SpO 2 reaching100%, good air entry into the lungs, no sign of dyspnea and ABG results in the normal ranges (pH: 7.37, PaCO 2 : 38 mmHg, HCO 3 -: 21.9 mmol/L, PaO 2 : 41 mmHg, and SpO 2 : 100%); the pro-BNP level was 283.1 ng/ml. A signi cant reduction in lung brosis was observed on chest CT without PAH ( Figure  2B). At 12months postdischarge, the patient's condition was improved. Her BW had increased to 7.5 kg. ABG results werein the normal ranges (pH: 7.34, PaCO2: 35.5 mmHg, PaO2: 87 mmHg, BE: -6; HCO3-: 19.3 mmol/l), and her pro-BNP level was 154.2 without PAH treatment. A signi cant reduction in lung brosis was observed on chest X-ray at the 12-month visit ( Figures 3B). The detailed progression of the patient's condition is described in Supplement Table 1.

Patient 2
The second premature neonate was patient 1's twin, who was enrolled in this study with a BW of 650 grams. Similar to her twin sister, the patient suffered from respiratory distress syndrome, including gasping, followed byapnea, bradycardia, and cyanosis with an SpO 2 between 50% and 60%. The patient was positive pressure ventilated by bag-mask and then intubated and placed on aventilator inSIMV mode (with ventilator parameters similar to those in the rst case). Chest X-ray revealed a stage 3 hyaline membrane, and a surfactant was given at a dose of 200 mg/kg BW on the 1 st and 3 rd days after birth. A large patent ductus arteriosus (PDA) was detected by cardiac echography, requiring one course of indomethacin (0.2 mg/kg/12 hours) within the rstpostnatalweek as previously described (20). After 2 months onSIMV, ventilation support was switched to CPAP with a PEEP of6 cmH 2 O and 40% FiO 2 . After treatment, the PDA size was reduced(1 mm), andthe shunt size was small; however, the size increased gradually and reached 3.6 mm at 3 months. The PDAwas maintained at a large size and required surgical closure at 3.5 months of age. Although the PDA was closed without complications, PAH wasobserved (38 mmHg), and the pro-BNP level was 2223 ng/ml, leading to treatment with 1 mg/kg/6 h sildena l and 2 mg/kg/8 h bosentan. After the operation, the patient was on CPAP at 6 cmH 2 O at a FiO 2 of 30% before switching to nasal cannula oxygen after 1 month at a rate of 1 L/min to maintain a stable SpO 2 between 93% and 97%. X-rays and chest CT scans at 4 months of age indicated diffuse brosis in the lung structures with atelectasis in the upper lobes of both lungs and signi cant air trapping in both lower lobes (Figure 2A and 3A). The patient was con rmed as havingBPD and treated with nebulized corticosteroids (100 mcg/kg 4 times/day), diuretics (furosemide 1 mg/kg/12 hours), spironolactone (2 mg/kg/12 hours), bronchodilators (inhaled β2-agonists) in combination with ipratropium bromide, and other supportive measures (high-calorie nutrition, vitamins E and A, etc.) for 6 weeks. However, the patient's condition didnot improve, and she remained dependent on oxygen support, leading to the indication forallo-UC-MSC transplantation at 151 days postnatal age. Prior to transplantation, ABG analysis revealed a pH of 7.6, PaCO 2 of 37.9 mmHg, PaO 2 of 35 mmHg, increased HCO 3 -of 29.1 mmol/L and BE of 8 mmol/L. Three days after transplantation, the patient was discharged with oxygen support via nasal cannula at 1 L/minute, a respiration rate of 64-67 times/minute, and an SpO 2 of 83% (FiO 2 : 21%).
At the rst follow-up visit, the patient's body weight had increased to 4.3 kg, her heart rate was145 bpm, and she still required oxygen support at 0.5 l/min to maintain an SpO2 over 92% (83% without oxygen support). The ABG results showed a PaCO 2 of 67 mmHg and an HCO 3 -of 32.3 mmol/L, a PaO 2 of 36 mmHg, and a BE of 6 mmol/l. The total hemoglobin level, WBC level, and neutrophil percentage were 129 G/L, 6.1 G/L, and 6.4%, respectively. The patient's condition had improved by her 1-month follow-up visit, with both reductions in both herPaCO 2 and HCO 3 -, while her SpO 2 was maintained at 94-98% onoxygen via a cannula a 0.5 l/min. Two months after the rst hUC-MSC transplantation, home oxygen monitoring results con rmed that the patient could breathe normally, and her SpO 2 had reached 95%.
From the 6-month follow-up onwards, the patient's health had stabilized under normal conditions, with her BW reaching 8 kg at the 12-month visit. All ABG tests were within normal parameters at the 6-month visit, further con rming the recovery of the patient from BPD. Her SpO 2 was maintained at 95% at the 6month visit and reached 100% at the 12-month follow-up. Blood gas analysis at 12 months showed all parameters were within the normal limits without oxygen support. The chest CT scan at the 6-month visit revealed a signi cant reduction in lung brosis ( Figure 2B). A normal chest X-ray was observed at the 12month follow-up ( Figure 3B). It is important to note that the maximal PA recorded at the 6-month visit was 46 mmHg,with a pro-BNP level of 511 ng/ml, leading to the administration of sildena l (1 mg/kg/12 h).
At the 12-month visit, the maximum PA was 37 mmHg, and her pro-BNP levelwas reduced to 202 ng/mL; therefore, a lower dose of sildena l (0.5 mg/kg/12 h) was given.

Patient 3
A 34-week-old male infant was prematurely born due to premature rupture of the placental membrane and had a BW of 2.4 kg at birth. The patient was diagnosed with hyaline membrane disease and required ventilator support. After 3 consecutive treatments with a surfactant, he was successfully weaned off of mechanical ventilation at 3months postnatal age. However, he still depended on oxygen support at a rate of 1 L/min via a sponge cannula. The diagnosis of BPD with vocal cord cirrhosis and laryngomalacia combined with periventricular leukomalacia was con rmed using nasopharyngoscopy, CT and MRI.
Upon admission toVinmec International Hospital, the patient was supported with oxygen at a rate of 1 L/min via nasal cannula to maintain the target SpO 2 above 92%. The SpO 2 dropped dramatically to 60% without oxygen support or crying. The patient suffered from severe chronic hypercapnia with pH, BE, PaCO 2 and HCO 3 -levels maintained at 7.35, 12 mmol/L, 63.6 mmHg and 67.2 mmol/L, respectively, whereas his SpO 2 and PaO 2 were relatively low (60% at a FiO 2 of 21%, and 44 mmHg, respectively). No cardiovascular malfunction or PAH was detected on echocardiogram, with a pro-BNP level of 176.5 ng/ml. The patient was diagnosed with CMV infection, with a viral load of 1.44 x10 5 copies/ml in the endotracheal uid. When CMV treatment with 3 weeks with valganciclovir was complete, the chest CT scan and radiograph revealed lung brosis with signi cant air trapping in both lungs and lung in ammation (Figure 2A and 3A), and the patient could not be weaned off ofoxygen. He was dependent on oxygen at a rate of 1 L/min via nasal cannula to maintain an SpO 2 between 94-96%.
Before transplantation, the patient still suffered from chronic hypercapnia with the following parameters: pH 7.51, PaCO 2 59 mmHg, HCO 3 -47.2 mmol/L, and PaO2 57 mmHg. He required oxygen support via nasal cannula at 1 L/min to maintain an SpO 2 between 92%-97%; without oxygen support, his SpO 2 was as lowas 70% (FiO 2 : 21%). The PCR results and hematological analysis (WBC: 23.9 G/L, neutrophils: 20.9%, and Hgb: 95 G/L) con rmed that the patient no longer carried CMV; he did not suffer from in ammation, nor did he have sepsis. Allo-UC-MSC transplantation was performed at 173 days postnatal age with no signs of severe adverse events. The patient was discharged 13 days after the rst transplantation with oxygen support via nasal cannula at 0.5 L/min with an SpO 2 ranging between 93% and 98%.
At the rst visit, the patient's general condition was fair, and he was cognizant, with his BW slightly increased to 5.3 kg. He was still receiving oxygen at 0.5 l/min via cannula to maintainan SpO 2 at92-98%.
The patient's hypercapnic condition was reduced, with the following ABG test results: pH of 7.46, PaO 2 of 45 mmHg, PaCO 2 of 52.6 mmHg and HCO 3 -of 38 mmol/L. His SpO 2 without oxygen support had increased to 85% on room air. The total blood count results remained in the normal ranges. The blood CRP level was 0.2 mg/L, con rming that the patient hadnot developed anin ammatory response. At the 1-month follow-up, the patient was still dependent on oxygen support at a rate of 0.5 L/min to maintainan SpO 2 level between 95% and 98. Two months posttransplantation, the patient was independent of active oxygen support, with an SpO 2 of 96-98%.
The clinical team observed improvements at the 6-month visit. The patient was cognizant and was able to crawl, laugh, and actively respond to his parent's voice. Due to the complication of periventricular leukomalacia, an additional Denver II test was conducted at the 6-month examination, and the results con rmed that the patient's gross motor function was similar to that expected at 3 months, his language ability was equivalent to that expected at 5-6 months, his ne motor adaptive skills were equivalent to those expected at 3 months, and his personal-social skills were equivalent to those expected at 5 months. Moreover, improved respiratory function was also documented, with better air ow in both lungs, no crackles or rales, and no signs of retraction ornasal aring atthe 6-month visit. All ABG results remained stable at the 12-month visit, with no sign of respiratory distress syndrome, an improved saturated oxygen level (SpO 2 : 100%) and a normal CO 2 level in the blood (pH of 7.4, PaO 2 of 72 mmHg, PaCO 2 of 34.8 mmHg, HCO 3 -of 21.5 mmol/L; BE of -3 mmol/l) (Supplemental Table 4). Investigation of the patient's lungswithCT at the 6-month visit indicated areduction in brosis and the recovery offunction. Chest X-rays at the 12-month visit further con rmed the signi cant improvement ( Figure 3B).

Patient 4
A premature female infant was born at another hospital at 28 weeks gestation with a birth weight of 1400 grams due to premature rupture of the placental membranes. She rapidly developed respiratory distress syndrome and required mechanical ventilation. A single dose of surfactant was given (100 -200 mg/kg) on the rst day. After that, the patient was placed on CPAP for a month, followed by oxygen support at 0.5 -1 L/min until she reached 36 weeks old. Dexamethasone treatment using the DART protocol was advised for one week to further improve the patient's condition. The patient was successfully weaned from oxygen supportand discharged at 37 weeks with an SpO 2 ranging between 93% and 95%. However, 2 days postdischarge, the patient developed dyspnea with acute respiratory distress and returned to the hospital, where she stayed for the next 2 months.
The patient was referred to VinmecHospital at 4 months old with malnutrition (BW of 3 kg). Although the oxygen support was maintained at 1 L/min via nasal cannula, herSpO 2 was relatively low (80%).
Auscultation showed poor air entry into the lungs with crackles andrales. Her heart rate was high (200 -220 bpm), with evident cyanosis and an SpO 2 of 80% on 24%oxygen. The patient was intubated immediately and placed ona ventilator inSIMV mode (PIP at 23 cm H 2 O, PEEP at 5.5 cm H 2 O, and FiO 2 at 50%). Five days after the treatment, ventilation support was switched to sponge cannula with oxygen owing at 1 L/min. The ABG examinations revealed the following: pH of 7.49, PaCO 2 of 38.6 mmHg, HCO 3 -of 29.5 mmol/l, and PaO 2 of 60 mmHg with FiO2: 40%. Furthermore, a complete blood count showed a low platelet count (53 G/L), while the WBC, neutrophil, and Hgb results were 5.8 G/L, 1.3 G/L and 112 G/L, respectively. Anechocardiogram was performed when the patient was stable and showed a pressure gradient through the tricuspid valve at 28 mmHg. The pro-BNP level was 8065 pg/ml. Hence, the patient was treated with 0.5 mg/kg/8 h sildena l. The viral tests con rmed a CMVinfection (460 copies/ml), which was treated with valganciclovir for 21 days. The results of a chest X-ray and CT scan indicated severe lung brosis and substantialairtrapping in both lungs (Figure 2A and 3A).
Two UC-MSC transplantations were carried out without adverse events when the patient was160 days old. Four days after the rst transplantation, the patient could breathe spontaneously at 55 -62 breaths/minute. On the day of discharge (a week after the second transplantation), the patient breathed spontaneously with an SpO2 of 95% without oxygen support.
At the 7-day examination, the patient still suffered from dyspnea, with a respirationrate of 53 breaths/minutes. An increase in the SpO 2 level to 95% was also recorded. The pro-BNP level was reduced signi cantly to 136.7 ng/ml. The hematological analysis con rmed that no sepsis or in ammatory reaction had occurred after stem cell transplantation, with a WBC count of 9.8 G/L, neutrophil percentage of 12.1%, Hgb level of 112 G/L and platelet count of 61 G/L. Onemonth postdischarge, the patient was cognizant and active, with a BW of 4kg. There was an increase in the SpO 2 to 98% without oxygen support, suggesting that the patient's respiratory function had recovered.
At the 6-month visit, respiratory distress was assessed as mild. The SpO 2 had stabilized at 97%. The 12month follow-up corroborated the conclusion that the patient had recovered from BPD, witha normal SpO2 of 97%, pH of 7.34, PaCO 2 of 39.8 mmHg, HCO 3 -of 20.8 mmol/l, BE of -4 mmol/l, and PaO2 of 73 mmHg. At 12 months after transplantation, it is worth mentioning that the patient hadrecovered well with regard to both her SpO 2 and PaO 2 levels, which were 100% and 72 mmHg, respectively. Evaluation of the lung structure onCT scans demonstrated that the brotic area was reduced ( Figure 2B), while alveolation and maturation of the lung had becomeobvious. A further assessment of the lung structure using chest X-rays at the 12-month follow-up showed no signs of atelectasis or hyperexpansion in either lung ( Figure  3B).

Discussion
All four patients tolerated the allo-UC-MSC infusion well, and no prespeci ed infusion-related adverse events were recorded after either the rst or second transplantation. Speci cally, no signi cant changes in heart rate, mean arterial pressure, oxygen saturation or body temperature were observed in any of the four infants (Supplementary Figure 1). These results, together with the detailed hematological analysis reported in each case, con rmed that allo-UC-MSC transplantation does not trigger an in ammatory response during or 72 h after infusion. Previously, a study reported the safety outcomes of the allogeneic transplantation of human umbilical cord blood-derived MSCs, in which nine preterm infants received either a single dose of 1x10 7 cells/kg or 2x10 7 cells/kg (21). In another single-center, open-label phase 1 trial, Lim's group administered 1x10 6 human amnion epithelial cells tosix preterm infants with established BPD and reported the safety pro le 2 years posttransplantation. In these trials, as in ours, no infusion toxicity or allogeneic UC-MSC transplantation-associated adverse events were recorded. These results were also supported by a preclinical study evaluating the long-term safety and e cacy of the allogeneic transplantation of MSCs in a rodent model of BPD, which reported no adverse lung effects posttransplantation together with persistent improvement in respiratory function and lung condition (22).
Additionally, our preliminary data support the e cacy of allo-UC-MSC transplantation in neonates with established BPD. This is, to our knowledge, the rst report of the use ofallo-UC-MSC transplantation to treat premature infants with established BPD. During the follow-up period, all patients exhibited signi cant improvements in lung function. All patients stopped being dependent on mechanical oxygen support as early as 4 days after the rst transplantation and as late as 2 months posttransplantation. These results make our study different from previous studies, as our study focused on the treatment of established BPD in premature infants. Ahn's study suggested that the injection of allogeneic MSCs could reduce the risk of BPD (11,23,24). It was suggested that the earliest time at whichBPD can be predicted is day 4 of life, when the peak inspiratory ventilator pressure is recorded and assisted ventilation is required (25). To con rm the diagnosis of BPD, subsequent studies found that BPD patternstypical o ung diseasesemerge during the rst 14 days of life. Hence, during this period, several risk factors could affect the disease, including late surfactant de ciency, sepsis, in ammation, and PDA (26,27). Furthermore, it was reported that 50% of infants with pulmonary deterioration and nearly 70% of infants with early persistent pulmonary deterioration develop BPD (28). Therefore, intervention withstem cell therapy during the rst 14 postnatal days shouldbe considered as a preventive treatment ininfants at risk of BPD.
Although it was suggested that the direct delivery of stem cells to a patient's lung via intratracheal administration could allow the stem cells to reach their designated location, a systematic analysis of preclinical studies suggested that the intravenous administration of MSCs resulted insuperior effects than the intratracheal administration of MSCs (29). Moreover, it was previously reported that the intravenous route is more effective than local infusion via intratracheal delivery (30). Our results support the argument that using a less invasive administration method, such as the intravenous infusion of stem cells or drugs, in established BPD infants could further enhance the effectiveness of the therapy andreduce the risks of both respiratory and systemic in ammation (31,32). Moreover, the delivery of a high concentration of stem cells in a relatively small volume (<5 ml) might increase the risk of pulmonary embolism regardless of whether intravenous administration orintratracheal infusion was used, as was demonstrated previously in mouse models (33).
BPD is a clinical syndrome characterized by the disruption of the alveolarization process and the microvascular development of the lung. In ammation is the common pathway that leads to the phenotype of the disease and further compromises the structure and function of the patient's respiratory system (24). However, pro-in ammatory cascades are complex and redundant, involving numerous different cell types and molecular pathways triggered by either cell-to-cell interactions or paracrine signals from damaged tissues (34). Therefore, MSC transplantation emerges as a promising therapy investigated in a recently reported phase I clinical trial (21). Two potential mechanisms supporting the action of MSCs in BPD patients are the immunological regulatory functions and the secretion of a wide range of growth factors, cytokines, and exosomes that are involved in the initiation of tissue regeneration (35,36). Despite the rapid evolution of MSC therapy forthe treatment of BPD, it is important to emphasize that cell-based therapies for lung regeneration are still in their infancy, and major knowledge gaps regardinghow these MSCs may be involved in the repair of the damaged lung remain to be addressed. Hence, a better understanding of the biological activity of MSCs and technological advances in the manufacturing of these cells couldsupport the development of such therapies in the near future (36).

Conclusions
Our current study provides the rst evidence of the safety of the allogeneic transplantation of hMSCsin established BPD patients.The results also showed that the allogeneic transplantation of hMSCsmight decrease lung brosis and improve lung function. However, our major limitation is the small cohort size (only 4 preliminary cases) and the lack of a control group. A study with a larger sample size and a control group is needed to draw more accurate conclusions.