Enzymatic conversion of blood group B kidney prevents hyperacute antibody-mediated injuries in ABO-incompatible transplantation

Abstract

Matching ABO blood group antigens between donors and recipients is critical to prevent hyperacute rejection in kidney transplantation. Enzymatic conversion of blood group antigens to the universal O type presents a promising strategy to overcome barriers in ABO-incompatible kidney transplantation. In this study, we employ α-galactosidase from Bacteroides fragilis to convert type B kidneys to type O during hypothermic machine perfusion. After 3 hours of perfusion with enzyme, more than 95% of blood group B antigens in the kidney endothelium are effectively removed. Subsequently, enzyme-treated kidneys are protected from antibody-mediated injuries in an ex vivo simulation of ABO-incompatible kidney transplantation. Encouraged by these results, a discarded type B kidney, following enzymatic conversion, is transplanted into a type O brain-dead recipient with high titer of anti-B antibody. The allograft survives for 63 hours without hyperacute rejection. Blood group B antigens re-express within 48 hours, with histopathological analyses indicating no evidence of antibody-mediated rejection. This enzymatic conversion approach holds the potential to broaden the practice of ABO-incompatible kidney transplantation, decrease waiting times and facilitate equitable organ allocation.

Introduction

Kidney transplantation (KT) is the optimal treatment for patients with end-stage renal disease, offering substantial improvements in survival and life quality compared to dialysis¹. Successful KT necessitates a precise match of the ABO blood group system and human leukocyte antigens (HLA) between donors and recipients. However, the presence of pre-existing antibodies in recipients targeting ABO blood type or HLA antigens in the donated kidney often poses immunological challenges, leading to rapid and severe graft destruction^2,3. Give the unmet gap between organ availability and the growing KT waiting lists⁴, these immunological barriers prolong the waiting time for these vulnerable individuals, resulting in significant mortality⁵. This situation is particularly critical for type O transplant candidates, who constitute more than half of the waiting list and can only accept type O kidneys due to antibodies against both blood group A and B antigens⁶. Living donor KT provides an alternative to finding a suitable deceased donor kidney on the waiting list⁷. However, approximately one-third of transplant candidates may be incompatible with their intended living donors due to mismatch of human blood group type⁸. Thus, any methods aimed at overcoming the ABO blood type barrier could substantially broaden the donor pool, improve the chances of successful donor-recipient matches, and ultimately save lives. This advancement would also greatly enhance equity and access to organ allocation, particularly for highly sensitized type O patients, who frequently face challenges in finding compatible donors.

Based on findings by Alexandre et al.⁹ in 1987, a recipient-centric desensitization strategy has been developed to overcome the ABO blood type barrier, incorporating immunosuppressants, anti-CD20 monoclonal antibody, and blood type antibody removing techniques such as plasmapheresis or immunoadsorption therapy^3,10. However, aggressive preconditioning has been associated with higher mortality rates due to increased severe infections, and heightened risks of early graft loss from surgical bleeding complications, and inferior patient and graft survival in ABOi KT compared to ABO-compatible KT^10,11,12. A reduction in donor blood group antigens represents a novel approach that may potentially mitigate the necessity for aggressive recipient preconditioning, thereby facilitating safer expansion of ABOi KT and addressing challenges posed by organ scarcity.

The ABO system is governed by the expression of A, B, and H antigens, encoded by the ABO, FUT1, and FUT2 loci³. The H antigen (H-Ag), which determines blood group O (often referred to as the universal blood type), is the enzymatic product of α-1,2-fucosyltransferase and serves as a substrate for A and B antigens^13,14. Therefore, the enzymatic modification of a single sugar in the oligosaccharide can convert blood group A, B, or AB to type O (ECO) using enzymes like α-N-acetyl-galactosaminidase, α-galactosidase, or a combination thereof¹⁵. Recently, several studies demonstrated the conversion of human blood group A lungs and kidneys to type O using A enzymes (FpGalNAc deacetylase and FpGalactosaminidase) during ex vivo machine perfusion, with subsequent ABOi transplant simulations showing no significant antibody-mediated injury^16,17. In addition, a recent short report¹⁸ demonstrated the conversion of blood group B kidneys to O kidneys under normothermic machine perfusion (NMP). Of note, the efficacy of B antigen (B-Ag) removal is not satisfactory and varies from 67% to 93%. Moreover, no simulations for ABOi KT were conducted.

In this study, we use recombinant α-galactosidase from Bacteroides fragilis (B-zyme)¹⁹ to convert type B kidneys into type O kidneys under hypothermic machine perfusion (HMP) and assessed its efficacy and safety in subsequent ABOi KT models. We find that B-zyme can convert B-Ag to H-Ag in type B RBCs and kidneys under hypothermic conditions, and it also provides protection against antibody-mediated injury in vitro ABOi KT simulations. Furthermore, we have successfully transplanted a B-zyme-converted kidney into a type O brain-dead recipient with high titer of anti-B antibody, with no hyperacute rejection observed. We propose that B-zyme has the potential to serve as a biotechnological and therapeutic tool to address ABO incompatibility in KT.

Results

B-zyme converts type B RBCs to universal O in kidney perfusate solution

The concept of synthesis and enzymatic removal of B-Ag is illustrated in Fig. 1a. The B allele encodes α-1,3-galactosaminyltransferase (GTB), which catalyzes the addition of a galactose residue to the H antigen (H-Ag) at the third position of the galactose residue, forming the B-Ag²⁰. To convert type B RBCs to universal type O (ECO), we employed α-galactosidase (B-zyme) to hydrolyze the α-1,3-galactose. Initial tests in a conversion buffer (250 mM glycine) demonstrated the efficacy of B-zyme (Supplementary Fig. 1–4).

Fig. 1: B-zyme enzymatically removed B-Ag from type B RBCs in kidney perfusate solutions.

a A schematic representation illustrating the enzymatic conversion of type B RBCs to type O RBCs (ECO) using B-zyme (α-galactosidase). The B allele encodes an α-1,3-galactosaminyltransferase (GTB), which transfers a galactose residue to the third position of the galactose in H-antigen, resulting in the formation of B-antigen. Accordingly, B-zyme hydrolyzes the α-1,3-galactose in the structure of B-Ag will result in exposure of the H-Ag. b Histograms illustrating B-Ag expression on RBCs treated with different doses of B-zyme in kidney perfusate solutions (UW, HTK, Celsior, and KPS-1) with 250 mM glycine at RT or 4 °C for 60 min. Positive control represents untreated type B RBCs, and negative control represents type O RBCs. c Heat map summary comparing the rate of B-Ag removal in different kidney perfusate solutions at RT or 4 °C for B-zyme-treated RBCs (n = 3). d Histograms showing B-Ag expression on RBCs treated with different doses of B-zyme in HD-40 solution (mixture of HTK, Glycine, and Dextran-40) at RT or 4 °C for 60 min. Positive control is untreated type B RBCs, and negative control is type O RBCs. e Heat map summary comparing the rate of B-Ag removal in HD-40 solution at RT or 4 °C for B-zyme-treated RBCs (n = 3). Data are presented as mean ± SEM, analyzed using two-way analysis of variance (ANOVA) followed by Bonferroni’s multiple comparisons test. ns, no significance. Source data are provided as a Source Data file. H-Ag H antigen, B-Ag B antigen, RBCs red blood cells, ECO enzyme converted O, UW University of Wisconsin, HTK histidine-tryptophan-ketoglutarate, KPS-1 kidney perfusion solution, Ctrl control, RT room temperature.

To assess the conversion capability of B-zyme, type B RBCs were incubated with different concentrations (0.5, 1, 5, 10, and 50 μg/mL) in kidney preservation solutions containing 250 mM glycine for 60 min at room temperature (RT) or 4 °C. Flow cytometry analysis indicated optimal performance of B-zyme in HTK solution (Fig. 1b). Specifically, in HTK solution, treatment with 50 μg/mL B-zyme removed 98.1 ± 1.9% of B-Ag at RT and 94.1 ± 3.2% at 4 °C (Fig. 1c), corroborated by agglutination assays (Supplementary Figs. 5–6).

To enhance B-zyme efficiency at lower temperatures, we added Dextran-40 as a molecular crowder²¹ to the HTK solution. This modification reduced the effective enzyme dose from 50 μg/mL to 5 μg/mL at 4 °C, with no observed agglutination (Supplementary Fig. 7), maintaining the osmolality of the new perfusate solution (HTK mixed with 25 mM glycine and 60 mg/mL Dextran-40, termed HD-40) within physiological range (Supplementary Table 1). After 60 min of incubation in HD-40 solution, 5 μg/mL B-zyme demonstrated comparable B-Ag removal rates at RT and 4 °C (97.7% ± 1.3% vs. 96.3 ± 1.6%, p > 0.99) (Fig. 1d, e). Importantly, viability and integrity of ECO RBCs were preserved (Supplementary Fig. 8).

B-Ag was removed from kidneys by B-zyme treatment during HMP

Human kidneys from blood group B donors, initially deemed unsuitable for clinical transplantation, were treated with B-zyme. The donor characteristics were summarized in Supplementary Table 2. Given that HMP remains the prevailing technique for organ preservation, our experiment marked the inaugural use of the LifePort Kidney Transporter combined with B-zyme perfusate over a 4-h period to treat kidneys (Fig. 2a). To achieve more thorough B-Ag removal and reduce the time required for antigen elimination, kidneys from each donor were randomly allocated to receive either 5 μg/mL or higher concentration of B-zyme (10 μg/mL), with biopsies collected hourly for analysis by immunofluorescence microscopy (Fig. 2b).

Fig. 2: B-zyme enzymatically removed B-Ag from type B kidneys during HMP.

a Schematic illustration of the enzymatic conversion of blood group B kidneys using B-zyme during 4 h of HMP. Kidneys from the same donor were randomly assigned to different groups treated with varying doses of B-zyme (5 μg/mL and 10 μg/mL). b Immunofluorescence images showing CD31 (endothelial cell marker), B-Ag, and H-Ag expression in type B kidney glomeruli before and after B-zyme treatment during HMP at different time points. Scale bars, 20 μm. c Relative MFI of CD31 before and after treatment. d Quantification of B-Ag removal assessed by percent changes in MFI compared to pretreatment levels. e Relative MFI of H-Ag following B-zyme treatment in kidneys. Six fields of view were selected per kidney (n = 3) for 18 images total per group. Data in (c, d, e) are presented as mean ± SEM. All data were analyzed using two-way analysis of variance (ANOVA) followed by Bonferroni’s multiple comparisons test. ns, no significance. Source data are provided as a Source Data file. B-Ag B antigen, H-Ag H antigen, HMP hypothermic machine perfusion, h hour, MFI mean fluorescence intensity.

No significant difference in CD31 expression was observed following 4 h of B-zyme perfusion between the 5 μg/mL and 10 μg/mL treatment groups (Fig. 2c, p > 0.99). However, the 10 μg/mL B-zyme treatment group demonstrated a higher rate of B-Ag removal compared to the 5 μg/mL group at all time points in glomeruli (Fig. 2d; 52.2 ± 2.1% vs. 23.5 ± 1.6% at 1 h, 75.1 ± 1.5% vs. 38.1 ± 2.4% at 2 h, 94.4 ± 0.6% vs. 55.2 ± 2.1% at 3 h, 95.2 ± 0.4% vs. 73.3 ± 1.9% at 4 h; p < 0.0001), alongside increased H-Ag expression in glomeruli (Fig. 2e; 5.6 ± 0.1-fold vs. 1.3 ± 0.1-fold at 1 h, 7.2 ± 0.1-fold vs. 3.2 ± 0.1-fold at 2 h, 7.8 ± 0.1-fold vs. 4.0 ± 0.1-fold at 3 h, 7.7 ± 0.1-fold vs. 4.4 ± 0.2-fold at 4 h; p < 0.0001). Notably, no significant change in B-Ag removal was found after 4 h perfusion compared to 3 h treatment with 10 μg/mL B-zyme (p > 0.99). Similarly, the H-Ag expression remained unchanged between 3-h and 4-h perfusion (p > 0.99).

The efficacy of both treatment concentrations in removing intrarenal vascular blood group antigen was assessed (Supplementary Fig. 9). After 4 h of perfusion, both the 5 μg/mL and 10 μg/mL treatment groups effectively removed ~90% of intrarenal vessel B-Ag (89.6 ± 1.8% vs. 92.6 ± 0.4%, p > 0.99), with a comparable increase in H-Ag expression (8.1 ± 0.7-fold vs. 8.2 ± 0.8-fold, p > 0.99).

Furthermore, the 10 μg/mL B-zyme treatment group exhibited a significant reduction in B-Ag expression in peritubular capillaries compared to the 5 μg/mL group at all time points (4 h of perfusion, 90.1 ± 0.9% vs. 82.6 ± 1.1%, p = 0.02) (Supplementary Fig. 10). Similarly, the higher concentration group demonstrated a greater H-Ag conversion rate compared to the lower concentration group (6.6 ± 0.2-fold vs. 5.4 ± 0.1-fold, p < 0.0001). There were no significant differences in B-Ag removal (p = 0.98) or H-Ag expression (p > 0.99) between 3-h and 4-h of perfusion in the 10 μg/mL conversion group.

Effect of B-zyme treatment on kidney during HMP

During a 4-h HMP, functional parameters were assessed at half-hour intervals. No significant differences were observed in arterial pressure (p > 0.99), renal blood flow (p > 0.99), and arterial resistance (p > 0.99) between the groups treated with 5 μg/mL and 10 μg/mL B-zyme at all time points (Fig. 3a). Histological evaluation using hematoxylin and eosin (H&E) staining revealed no increase in acute renal injury scores before and after treatment with 5 μg/mL B-zyme during HMP (p = 0.83); similar findings were observed in 10 μg/mL B-zyme treated kidneys (Fig. 3b, c, p = 0.32). Following 4 h of treatment, no significant differences were found in terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick end labeling (TUNEL)-positive cells as a measure of kidney tissue apoptosis (Fig. 3d, e, p > 0.05). Furthermore, there were no discernible variations in renal injury cytokines in the perfusate, including NGAL (1.7 ± 0.1 vs. 1.5 ± 0.1 ng/mL, p = 0.56), KIM-1 (641.5 ± 65.6 vs. 626.3 ± 46.4 pg/mL, p = 0.59), CXCL9 (196.3 ± 21.8 vs. 202.3 ± 15.8 pg/mL, p = 0.77), and IL-18 (81.5 ± 13.0 vs. 86.5 ± 12.0 pg/mL, p = 0.20) between the 5 μg/mL and 10 μg/mL treatment groups at the end of perfusion (Fig. 3f).

Fig. 3: Effect of B-zyme treatment on kidney during HMP.

a Perfusion parameters of kidney function with B-zyme treatment during HMP (n = 3 per group). b H&E staining of kidneys before and after treatment with 5 μg/mL or 10 μg/mL B-zyme during HMP. Scale bar, 100 μm. c Quantification of kidney injury scores before and after 4 h of B-zyme treatment during HMP (n = 3). d Representative images of TUNEL-stained sections from kidneys treated with various doses of B-zyme during HMP. TUNEL positive cells merged with DAPI. Scale bars, 50 μm. e Quantification of TUNEL positive cells in kidneys treated with B-zyme. Six fields of view were selected per kidney (n = 3) for 18 images total per sample. f Concentration of kidney injury markers in the perfusate following 4 h of B-zyme treatment (n = 3). All data are presented as mean ± SEM. Data in (a, c, e) were analyzed by two-way ANOVA and Bonferroni’s multiple comparisons test. Data in (f) were analyzed using two-tailed paired t test. ns, no significance. Source data are provided as a Source Data file. HMP hypothermic machine perfusion, h hour, H&E hematoxylin and eosin, TUNEL terminal deoxynucleotidyl transferase dUTP nick end labeling, DAPI 4′,6-diamidino-2-phenylindole, NGAL neutrophil gelatinase-associated lipocalin, KIM-1 Kidney Injury Molecule-1.

Ex-vivo model of ABO-incompatible (ABOi) kidney transplantation

The efficacy of 10 μg/mL B-zyme in removing blood group antigens from kidneys has been previously demonstrated. To assess whether enzyme-treated kidneys are protected from antibody-mediated injury due to anti-B antibodies, we utilized an ex vivo NMP reperfusion model to simulate ABOi KT (Fig. 4a). Kidneys from the same donor were randomly assigned to either a control group (without B-zyme treatment) or an ECO group (treated with 10 μg/mL B-zyme). During Phase 1, all kidneys underwent 3 h of HMP with indicated doses of B-zyme. Subsequently, during Phase 2, kidneys were subjected to NMP with reperfusion using type O plasma with high-titer anti-B antibody (IgG ≥ 1:128; IgM ≥ 1:32) and type O RBCs for 4 h. Antibody-mediated injuries were evaluated through analysis of re-perfusate samples and renal biopsies.

Fig. 4: B-zyme-treated kidney mitigates antibody-mediated injury in an ex vivo model of ABOi KT.

a Schematic overview the ex vivo reperfusion model for ABOi KT. Kidneys from the same donor were randomly assigned to either the control group (HMP without B-zyme) or the ECO group (HMP with 10 μg/mL B-zyme). Following 3 h of HMP, kidneys were reperfused with high-titer anti-B antibody type O plasma and type O RBCs using a NMP device to simulate ABOi KT. b Representative H&E staining after reperfusion in the control group, showing RBC congestion, microthrombosis in glomeruli and arteries, and interstitial hemorrhage (arrows). Scale bar, 50 μm. c Representative H&E images after reperfusion in the ECO group, showing absence of RBC congestion and microthrombosis after 4 h of reperfusion. Scale bar, 50 μm. d Quantification of kidney injury scores before and after 4 h of reperfusion (n = 3). e Representative images of TUNEL positive cells in control and ECO kidneys after 4 h of reperfusion. Scale bars, 50 μm. f Quantification of TUNEL positive cells in control and ECO kidneys. Six fields of view were selected per kidney (n = 3) for 18 images total per sample. g Concentration of kidney injury markers in the urine after 4 h of reperfusion (n = 3). h Urine output from control and ECO kidneys during 4 h of reperfusion (n = 3). All data are presented as mean ± SEM. Data in (d and f) were analyzed by two-way ANOVA and Bonferroni’s multiple comparisons test. Data in g were analyzed using two-tailed paired t test. Data in h were analyzed using two-way ANOVA and Bonferroni’s multiple comparisons test (for urine output per hour) and two-tailed paired t test (for total urine output). ns, no significance. Source data are provided as a Source Data file. ABOi ABO incompatible, KT kidney transplantation, HMP hypothermic machine perfusion, NMP normothermic machine perfusion, ECO enzyme converted O, RBCs red blood cells, H&E hematoxylin and eosin, h hour, TUNEL terminal deoxynucleotidyl transferase dUTP nick end labeling, DAPI 4′,6-diamidino-2-phenylindole, NGAL neutrophil gelatinase-associated lipocalin, KIM-1 Kidney Injury Molecule-1.

Based on reperfusion hemodynamic parameters, the ECO group appears to exhibit more unimpeded perfusion, as indicated by the differences in parameter values (Supplementary Fig. 11). Following 4 h of reperfusion, there was no significant discrepancy in renal arterial pressure between the control and ECO groups (63.7 ± 2.4 vs. 70.1 ± 1.5 mmHg, p = 0.07) at the end of perfusion. However, the control kidneys showed lower renal blood flow (190.0 ± 9.7 vs. 243.7 ± 16.8 mL/min, p = 0.03) and higher arterial resistance (408.4 ± 12.5 vs. 322.1 ± 7.7 mmHg/(L/min), p = 0.04) compared to ECO kidneys at 4 h reperfusion.

Histological examination of H&E-stained images from the control group revealed substantial accumulation of RBCs in peritubular capillaries, glomeruli, and intrarenal arteries after 4 h of reperfusion (Fig. 4b). Additionally, microthrombus was observed in glomerular capillary loops. In contrast, no marked differences were observed in the ECO kidneys after 4 h of reperfusion compared to pre-reperfusion conditions (Fig. 4c and Supplementary Fig. 12). Renal injury grading indicated significantly higher injury scores in control kidneys after 4 h of reperfusion compared to pre-reperfusion (p < 0.0001), whereas no significant change was observed in the ECO kidneys (p = 0.19) (Fig. 4d). These changes correlated with increased TUNEL-positive cells in control kidneys (p < 0.0001) and no significant increase in apoptosis observed in the ECO group (Fig. 4e, f, p = 0.44).

Measurement of kidney injury markers in the urine after 4 h reperfusion showed a notable increase in KIM-1 (1293.5 ± 84.5 vs. 1088.8 ± 101.9 pg/mL, p < 0.01) and IL-18 (247.8 ± 25.4 vs. 200.0 ± 26.7 pg/mL, p = 0.01) in the control group compared to the ECO group, while NGAL (3.9 ± 0.4 vs. 2.8 ± 0.2 ng/mL, p = 0.05) and CXCL9 (479.4 ± 56.5 vs. 374.7 ± 46.7 pg/mL, p = 0.18) concentration were numerically higher in the control group but did not reach statistical significance (Fig. 4g). ECO kidneys produced more urine for 4 h reperfusion compared to control kidneys (Fig. 4h, total urine output, 100.7 ± 11.8 vs. 44.3 ± 10.5 mL, p < 0.01).

ECO kidneys resist blood group antibody deposition and protect against complement-mediated attacks

To evaluate the efficacy of ECO kidneys in mitigating blood group antibody deposition during ABOi KT, renal biopsies were subjected to immunofluorescence staining with anti-human IgM and IgG (Fig. 5a). Control kidneys exhibited substantial deposits of IgM (p < 0.0001) and IgG (p < 0.0001) in glomeruli (Fig. 5b) and peritubular capillaries (Supplementary Fig. 13) after 4 h of reperfusion, whereas ECO kidneys showed only minimal IgM binding. Additionally, the change in circulating IgM and IgG in the re-perfusate was assessed over time using flow cytometry with type B RBCs incubated with the perfusate (Supplementary Fig. 14). Upon reperfusion, circulating anti-B antibodies were rapidly absorbed in the control group but remained stable in the ECO group during ABOi KT simulation (Fig. 5c, d, p < 0.001 at all-time points). The changes in circulating anti-B IgG and IgM were summarized in Supplementary Fig. 15, suggesting that enzymatic conversion could prevent antibody binding. Following 4 h of reperfusion, the ECO group showed significantly lower concentrations of the inflammatory cytokines IL-6 (p = 0.02) in re-perfusate. Conversely, IL-1β, IL-2, IL-8, tumor necrosis factor-α (TNF-α), interferon-γ (IFN-γ), CX3CL1, CCL5, and CXCL9 were numerically lower than that in the control group, yet these differences did not reach a statistical significance (Fig. 5e).

Fig. 5: Antibody binding and inflammation markers in an ABOi KT model.

a Immunofluorescence images showing antibody deposition in glomeruli before and after reperfusion. Scale bars, 20 μm. Following 4 h of reperfusion, substantial amounts of IgM and IgG were deposited in the control kidneys (without B-zyme treatment), whereas only minimal IgM deposition was observed in the ECO group (10 μg/mL B-zyme treatment). b Relative MFI of IgM and IgG deposition in control vs. ECO kidneys. Six fields of view were selected per kidney (n = 3) for 18 images total per sample. c Histograms depicting anti-B IgM and IgG levels in the perfusate collected before and after 4 h of reperfusion in the control and ECO groups. d Changes in MFI values at all time points following reperfusion based on reperfusion 5 min (n = 3). e Changes in inflammation markers in the perfusate after 4 h of reperfusion in control vs. ECO groups (n = 3). All data are presented as mean ± SEM. Data in b and d were analyzed by two-way ANOVA and Bonferroni’s multiple comparisons test. Data in e were analyzed using two-tailed paired t test. ns, no significance. Source data are provided as a Source Data file. ABOi ABO incompatible, KT kidney transplantation, ECO enzyme converted O, h hour, MFI mean fluorescence intensity.

Having demonstrated that ECO kidneys can mitigate antibody deposition, we proceeded to assess activation of the complement pathway. Immunofluorescence of renal biopsies in Fig. 6a evaluated the presence of complement components deposition. Following ABOi KT simulation, untreated kidneys (control group) activated the classical complement cascade, depositing significant amounts of C1q (19.0 ± 0.7 vs. 7.2 ± 0.4-fold, p < 0.0001), C3 (20.1 ± 0.6 vs. 1.2 ± 0.1-fold, p < 0.0001), C4d (25.5 ± 0.8 vs. 1.1 ± 0.1-fold, p < 0.0001), and C5b-9 (9.9 ± 0.8 vs. 1.1-fold, p < 0.0001) in glomeruli (Fig. 6b) and capillaries compared B-zyme-treated kidneys (Supplementary Fig. 16), whereas ECO kidneys remained stable except for C1q deposition after 4 h reperfusion. At the end of reperfusion, circulating complement factors C1q (p = 0.02), C3 (p = 0.02), C3b (p = 0.10), and C4 (p = 0.23) were found to be numerically lower in the untreated group, while levels of C5b-9 (p = 0.02) were significantly higher compared to the ECO group (Supplementary Fig. 17). These biological changes contributed to differences in kidney physiology and histology between the two groups, which indicated that B-zyme-treated (ECO) kidneys alleviate injury mediated by blood group antibodies.

Fig. 6: Complement deposition in kidneys during ABOi KT model.

a Representative immunofluorescence images of glomeruli in control (without B-zyme treatment) and ECO kidneys (10 μg/mL B-zyme treatment) stained for C1q, C3, C4d, and C5b-9. Scale bars, 20 μm. b Relative MFI of complement deposition based on prereperfusion level between control and ECO groups. Six fields of view were selected per kidney (n = 3) for 18 images total per sample. All data are presented as mean ± SEM. Data in b were analyzed using two-tailed paired t test. Source data are provided as a Source Data file. ABOi ABO incompatible, KT kidney transplantation, ECO enzyme converted O, h hour, MFI mean fluorescence intensity.

B-zyme converted kidney was successfully transplanted into a human decedent

To investigate the foundational principles of the ECO kidney in human subjects, we conducted a pre-clinical study and mainly focused on safety and feasibility. A type B kidney, initially unsuitable for transplantation, was converted to a type O (ECO) kidney using 10 μg/mL B-zyme after 3 h HMP. Subsequently, we transplanted this ECO kidney into a type O brain-dead recipient who had high titers of anti-B antibodies (IgG, 1:128; IgM, 1:32), following the experimental design outlined in Fig. 7a. The recipient had a panel-reactive antibody level of 1% and five HLA mismatches with the ECO kidney donor. Baseline demographic characteristics of both donor and recipient were detailed in Supplementary Table 3. Prior to transplantation, the recipient received anti-thymocyte globulin, rituximab, and methylprednisolone, followed by maintenance immunosuppression based on Cyclosporine A and Mycophenolate Mofetil throughout the study (Supplementary Table 4).

Fig. 7: ECO kidney transplants in human decedent.

a Schematic overview of ECO kidney transplantation in human decedent. Following a 3 h treatment with 10 μg/mL B-zyme during HMP, a type B kidney was enzymatically converted to type O and subsequently transplanted into a type O brain-dead recipient. b Gross view of the ECO kidney during enzymatic conversion and operation. The ECO kidney turned pink after revascularization. c Photographs of the ECO graft after transplantation. The allograft survived for 63 h. d Anti-B antibody titers in recipients (n = 1). e Urine output from native kidneys and ECO grafts after transplantation (n = 1). The ECO kidney produced 125 mL of urine within the first 24 h. f Concentration of urinary kidney injury markers in ECO allograft (n = 1). Source data are provided as a Source Data file. ECO enzyme converted O, IS immunosuppressant, h hour, HMP hypothermic machine perfusion, NGAL neutrophil gelatinase-associated lipocalin, KIM-1 Kidney Injury Molecule-1.

The ECO kidney exhibited a pink color after revascularization and remained in the following days (Fig. 7b, c). However, the recipient suffered from severe anemia and thrombocytopenia (Supplementary Fig. 18), and the experiment terminated at 63 h post-transplantation due to systemic circulatory failure. The recipient’s anti-B antibody titer remained relatively stable without significant rebound after transplantation (Fig. 7d). The ECO allograft, owing to its poor quality and the recipient’s compromised systemic circulation, produced ~125 mL of urine on the first day, decreasing over time (Fig. 7e). Therefore, continuous renal replacement therapy was necessary throughout the study period. Urinary injury markers NGAL, KIM-1, CXCL9, and IL-18 (Fig. 7f), as well as inflammatory cytokines in peripheral blood (Supplementary Fig. 19), showed a slight elevation in subsequent days.

Pathological analysis and B-Ag re-expression in the ECO allograft

Protocol biopsy specimens underwent H&E staining for histological analysis and were stained for multi-immunophenotyping (Fig. 8a). Pathological scoring of the allograft was performed by two kidney transplant pathologists following the Banff 2019 criteria²². At 63 h post-transplantation, H&E staining of the ECO allograft revealed the absence of microvascular inflammation (MVI0) with glomerular capillaries (g0), peritubular capillaries (ptc0) and tubulitis (t0), along with moderate interstitial inflammation (i1). In addition, no arteritis, interstitial hemorrhage, and thrombotic microangiopathy were observed (Fig. 8b, c and Supplementary Table 5). Hematoxylin and eosin (H&E) stained biopsies were assessed pre-transplant and post-transplant 63 h showed no significant kidney injury grading in the ECO allograft before and after transplantation (Fig. 8d). Notably, no signs of hyperacute rejection or acute T cell-mediated rejection were detected in the ECO allograft at 63 h post-reperfusion.

Fig. 8: Histopathological and immunological assessments of biopsy samples at various timepoints post-transplantation.

a Biopsy from ECO allograft protocol. b H&E staining of ECO kidney pre-transplantation. Scale bars, 50 μm. c H&E staining of ECO allograft at 63 h post-transplantation. Scale bars, 50 μm. Black arrows show interstitial leukocytes infiltration. d Quantification of kidney injury scores before and after 63 h post-transplantation. Scores were derived from six randomly selected visual fields per sample (n = 1). e Immunofluorescence staining showing expression of B-Ag and H-Ag in ECO glomeruli before and after transplantation. Scale bars, 20 μm. B-Ag re-expression observed at 48 h post-transplantation. f Relative changes in MFI of B-Ag and H-Ag compared to pre-transplantation levels. Six fields of view were selected per kidney (n = 1) for 6 images total per timepoint. g Immunofluorescence staining of immunological markers in ECO allograft before and after transplantation. Scale bars, 20 μm. IgM, IgG, C1q, and C3 deposition observed in ECO graft at 63 h post-transplantation, with no detectable activation of C4d and C5b-9. h MFI values normalized to pre-transplantation levels at 63 h post-transplantation. Six fields of view were selected per kidney (n = 1) for 6 images total per group. All data are presented as mean ± SEM. Data in d and h were analyzed using two-tailed paired t test. Data in f were analyzed using one-way analysis of variance (ANOVA) and Bonferroni’s multiple comparisons test. ns, no significance. Source data are provided as a Source Data file. ECO enzyme converted O, H&E hematoxylin and eosin, h hour, B-Ag B antigen, H-Ag H antigen, MFI mean fluorescence intensity.

After 3 h of HMP with 10 μg/mL B-zyme treatment, 95.8 ± 0.9% and 97.0 ± 5.3% of B-Ag was removed from glomerular capillaries and peritubular capillaries compared to pre-treatment, respectively, accompanied by a corresponding increase of 6.0 ± 0.2-fold and 6.6 ± 0.4-fold of H-Ag in glomerular capillaries and peritubular capillaries (Supplementary Fig. 20). After 24 h transplantation, the level of B-Ag and H-Ag expression from glomerular and peritubular capillaries were in consistency with pre-transplantation (after B-zyme treatment). At 48 h post-kidney transplantation, re-expression of B-Ag was observed in the ECO kidney (Fig. 8e), with B-Ag levels increasing by 11.7 ± 0.6-fold in glomerular capillaries (p < 0.0001) and 7.9 ± 0.9-fold (p < 0.0001) in peritubular capillaries compared to pre-transplantation. In contrast, H-Ag expression in glomerular and peritubular capillaries dropped to 49.7 ± 4.3% (p < 0.001) and 37.5 ± 4.4% (p < 0.01) of pre-transplantation levels, respectively. By 63 h post-transplantation, B-Ag expression in ECO allograft glomeruli and peritubular capillaries increased to 19.0 ± 1.0-fold (p < 0.0001) and 14.8 ± 1.0-fold (p < 0.0001), respectively, whereas H-Ag expression decreased to 26.3 ± 2.0% (p < 0.0001) and 14.2 ± 1.5% (p < 0.0001) compared to pre-transplantation levels (Fig. 8f and Supplementary Fig. 21).

With re-expression of the B-Ag, we observed a 6.3 ± 0.6-fold increase in IgM (p < 0.001) and a 5.0 ± 0.7-fold increase in IgG (p < 0.01) deposition on the glomeruli at 63 h post-transplantation compared to pre-transplantation (Fig. 8g, h). Similar findings were evident in the peritubular capillaries (Supplementary Fig. 22). Initial activation of the classical complement pathway occurred through the binding of IgM and IgG to the ECO allograft. At 63 h post-transplantation, there was a 7.9 ± 0.6-fold increase in C1q and a 6.7 ± 0.4-fold increase in C3 binding to the glomerular capillaries in the ECO allograft, with no detectable C4d and C5b-9 (Fig. 8g, h). Comparable trends were observed in the peritubular capillaries (Supplementary Fig. 22). There was no significant increase observed in the concentration of soluble activation factors in recipient’s peripheral blood (Supplementary Fig. 23). Analysis of apoptosis in the ECO allograft showed no notable differences (Supplementary Fig. 24).

Discussion

This is the first study of B-zyme conversion in solid organs under HMP, demonstrating that B-zyme successfully converted ABO-B kidneys into type O (ECO) kidneys under hypothermic conditions, preventing antibody-mediated injury during ex-vivo ABOi KT simulation. In addition, there were no discernible changes in renal injury cytokines in the perfusate, and biopsy analysis confirmed that B-zyme treatment was safe during kidney preservation with HMP. Notably, a B-zyme-treated type B kidney was successfully transplanted into a type O brain-dead recipient, showing good tolerance without hyperacute rejection or early-phase acute antibody-mediated rejection (AMR). In this first reported pre-clinical decedent recipient, B-Ag gradually re-expressed within 48 h, providing essential evidence for future studies to optimize this desensitization approach, potentially offering an alternative strategy for ABOi KT and increasing accessibility for sensitized type O candidates.

ABOi KT has expanded the donor organ pool and reduced waiting times, though it has been limited to living donors due to the need for recipient preconditioning. However, the recipient-centric strategy has been associated with higher mortality rates, severe infections, and increased risks of early graft loss from surgical bleeding and acute rejection^10,11,12. Although paired kidney exchange and living kidney-sharing schemes have successfully lowered waiting times for more restrictive blood types, such programs are unavailable in some countries due to ethical and legal issues^23,24.

As a promising desensitization approach, limited studies have demonstrated the successful enzymatic conversion of human solid organ blood groups. Two type A conversions were performed on human lungs under ex vivo lung perfusion¹⁶ and on human kidneys under HMP and NMP¹⁷. The A-Ag removal rate exceeded 97% in lungs within 4 h and was ~80% in kidneys within 2 h. MacMillan et al. reported B-Ag removal rates ranging from 67% to 93% using an acellular perfusate supplemented with 2.5 µg/mL GH110B for 5 h during NMP¹⁸. The limited removal of these blood group antigens restricted their further clinical applications. Although NMP has shown promise and safety for human KT²⁵, HMP remains the predominant method for kidney preservation²⁶. In this study, we modified the widely used HTK perfusate, a commercial perfusion solution for kidney preservation, and safely increased the B-Ag removal rate in kidneys to 95% for 3 h using a commercial HMP system, the LifePort Kidney Transporter from Organ Recovery Systems. These findings enhance the potential clinical applications of B-zyme for blood group conversion.

In addition to achieving a satisfactory B-Ag removal after 3 h of perfusion, we also evaluated the safety of B-zyme. The modified HD-40 solution preserved osmolality within the physiological range, with no detectable alterations observed in renal injury cytokines in the perfusate following B-zyme treatment. Histological analysis revealed no signs of additional vascular damage, suggesting that B-zyme treatment under HMP is safe. Previous ex-vivo ABOi KT simulations used type O RBCs, off-the-clot type AB human serum, and mouse monoclonal anti-A and anti-B IgM to provide antibodies, soluble complement factors, and metabolites for complement pathway activation. However, this approach fails to mimic clinical practice, which involves polyclonal antibodies and activation of the coagulation cascade, a critical process in hyperacute rejection¹⁷. In this study, we used type O RBCs and type O fresh frozen plasma containing high-titer anti-B antibodies (IgG ≥ 1:128; IgM ≥ 1:32) as a reperfusate, which is more relevant to the clinical situation. Notably, substantial accumulation of RBCs in peritubular capillaries, glomeruli, and intrarenal arteries, as well as microthrombi in glomerular capillary loops, was observed in the control group (without B-zyme treatment) after 4 h of reperfusion. In contrast, no marked differences were observed in the ECO kidneys (10 μg/mL B-zyme treatment). The renal injury score, apoptosis rate, and kidney injury markers in the re-perfusate and urine reflected these findings. These ex-vivo results confirmed that B-zyme conversion protects against antibody-mediated injury.

Last but not least, ex-vivo experiments cannot fully replicate the true immunological response of the human body to an ECO kidney. Additionally, the mechanisms behind blood group antigen re-expression and its subsequent interaction with the recipient’s immune system post-transplant remain unclear. In this pre-clinical study, we demonstrated that a B-zyme-treated kidney was successfully transplanted into a type O brain-dead recipient, showing good tolerance without hyperacute rejection or early-phase AMR within 63 h. These results are the first to confirm the safety of the B-zyme conversion strategy during the early phase in an in-vivo decedent model. The study also revealed that B-Ag began to re-generate after 48 h, providing critical evidence for subsequent immunosuppression regimens modification and additional antigen removal to accommodate the allograft. At 63 h, there was increased deposition of anti-B IgM and IgG in the ECO allograft, along with activation of C1q and C3. However, no C4d or C5b-9 depositions were detected, and no histological signs of AMR were observed. For the first time, these results offer important data on the kinetics and dynamics of B-Ag re-expression in an intact human kidney after transplantation.

There are several areas for improvement in this paper that require further research. Firstly, due to the limited availability of donated organs, we were only able to use discarded kidneys and brain-dead patients unsuitable for organ donation as recipients. The recipient in this pre-clinical study experienced anuria after transplantation due to severe anemia and circulatory disorders, leading to the termination of the experiment 63 h post-transplantation after unsuccessful rescue efforts. The poor function of the ECO kidney and the poor condition of the brain-dead recipient restricted long-term follow-up and limited our ability to gather more information about the immunological reaction to the re-expressed B-Ag. Secondly, B-Ag began to re-generate 48 h after transplantation, with anti-B IgM and IgG deposition detected in the ECO allograft at 63 h. Although no histological signs of hyperacute rejection or AMR were observed within this period, it remains unknown whether rejection or accommodation will occur over a longer timeframe. As a supplement to traditional recipient-centered desensitization therapy, this method may allow for more flexible preparation time prior to kidney transplantation and potentially reduce the use of immunosuppressants. In addition, B-zyme could potentially be administered as a drug infusion during periods of B-Ag re-expression, to maintain low B-Ag levels over the long term and support ECO allograft survival during critical periods. Further studies are required to investigate long-term changes in the ECO kidney and to develop optimized immunosuppressive regimens.

In conclusion, this study demonstrates that enzymatic removal of blood group antigens can mitigate hyperacute antibody-mediated injuries in simulated ABOi KT and in a pre-clinical human model. Furthermore, re-expression of blood group antigens within 48 h post-enzymatic conversion did not elicit hyperacute rejection. The enzymatic treatment described holds promise for expanding ABOi transplantation and enhancing equity and accessibility in organ allocation.

Methods

Study approval

Ethical approval was granted by the Ethics Committee on Biomedical Research, West China Hospital of Sichuan University (2023SHEN2088), and the Ethics Committee of the Second Affiliated Hospital of Chongqing Medical University (2023SHEN116). Written informed consent was given by the donor and recipient families. This study aimed to evaluate the feasibility and safety of employing α-galactosidase (B-zyme) to remove the B antigen (B-Ag) from human kidneys during hypothermic machine perfusion (HMP). After enzymatic conversion, a pre-clinical study explored the safety of the enzyme-converted O (ECO) kidney transplantation into a brain-dead recipient, mainly focusing on hyperacute rejection.

Materials

The recombinant α-galactosidase¹⁹ (B-zyme) was provided by the Institute of Transfusion Medicine, Academy of Military Medical Sciences, Beijing, China. Briefly, α-galactosidase from Bacteroides fragilis was expressed in E. coli BL21(DE3), purified using cation and anion exchange column chromatography, and confirmed proteins were formulated in Disodium hydrogen phosphate-citrate buffer (pH 6.8) at a final concentration of 1.2 mg/mL, stored at 4 °C until use. Type B RBCs and discarded kidneys were obtained from deceased donors with informed consent. Four widely used commercial organ preservation solutions were employed: the University of Wisconsin (UW), histidine-tryptophan-ketoglutarate (HTK), Celsior, and kidney perfusion solution (KPS-1). The HMP system used was the Lifeport Kidney Transporter (Organ Recovery Systems Inc., IL, USA), and the normothermic machine perfusion (NMP) system was the Life-X100 (Singularity Medical Technology Co., Ltd, Guangzhou, China).

Enzymatic conversion of RBCs

Type B and O whole blood from deceased donors, collected in citrate anticoagulant tubes with consent, was centrifuged at 1000 × g for 5 min at room temperature (RT), and RBCs were isolated and washed three times with PBS. A 250 mM glycine solution served as an initial buffer to assess the efficacy of B-zymes. The appropriate dose of B-zyme was then added to 1 mL of different kidney perfusion solution (UW, HTK, Celsior, and KPS-1) containing 250 mM glycine and 1% hematocrit RBCs to achieve final concentrations of 0.5, 1, 5, 10, and 50 µg/mL of B-zymes. Additionally, Dextran-40, known for enhancing enzyme efficiency²¹, was added in the most effective conversion buffer at specified concentrations to optimize performance. The conversion reaction was incubated for 60 min with gentle agitation at RT and 4 °C, followed by three rounds of washing with 1 mL PBS. Finally, each sample was diluted to 200 µL with PBS per sample.

Flow cytometry

One microliter RBC was fixed for 10 min in 0.1% glutaraldehyde in 50 µL PBS to prevent agglutination. After three washes with 200 µL PBS, cells were suspended in 100 µL of cell staining buffer (420201, BioLegend, USA). Subsequently, cells were incubated with murine monoclonal Anti-B reagents (1:100, 70502, DIAGAST, France) for 10 min at RT. Following three washes with 200 µL PBS, cells were again suspended in 1 mL of cell staining buffer and incubated at RT for 15 min with Alexa Fluor 647 Goat anti-Mouse IgM (1:1000, A-21238, Invitrogen, USA). The cells were then washed twice with staining buffer and incubated with PE Mouse anti-Human CD235a (1:100, 561051, BD Biosciences, USA) for 15 min at RT. Finally, samples were washed once, resuspended in cell staining buffer, and acquired using a CytoFLEX cytometer (Beckman-Colter, USA). The complete gating strategy is provided in Supplementary Fig. 1. Flow cytometry data were analyzed using FlowJo™ v10.4. The percentage of B-Ag removal was calculated using the formula: % B-Ag removal = [MFI (ABO-B RBCs without B-zyme)—MFI (ABO-B RBCs with B-zyme)]/[MFI (ABO-B RBCs without B-zyme)—MFI (ABO-O RBCs)] *100%.

Agglutination assay

Type-B RBCs without B-zyme treatment (positive control), type-O RBCs (negative control), and B-zyme-treated RBCs were diluted to 0.3% hematocrit with PBS. Subsequently, 20 μL of RBC suspension was incubated with Anti-B regents (ABO2) (1:1, 70502, DIAGAST, France) at RT for 30 min. The absence of RBC aggregation indicates a negative/no-agglutination result, while evident RBC rouleaux formation indicates a positive result for B-Ag presence. Agglutination of the RBCs was observed using the EVOS™ XL Core imaging system (Invitrogen, USA). Positive and negative control agglutination images were provided in Supplementary Fig. 3.

Glucose-6-phosphate dehydrogenase (G6PDH) assay

The WST-8 kit (S0189, Beyotime Biotechnology, China) was used to assess G6PDH activity in RBCs. Absorbance at 450 nm was measured for each well using a SpectraMax 190 microplate reader (Molecular Devices, USA) in accordance with the manufacturer’s instructions to determine G6PDH activity.

Erythrocyte membrane osmotic fragility assessment

Various concentrations of NaCl were utilized to assess erythrocyte membrane osmotic fragility. NaCl concentrations ranged from 0.9% to 0% in decrements of 0.05%. 5 μL RBCs were added to each test tube containing NaCl solutions with varying concentrations (1% hematocrit). The tubes were gently mixed and incubated for 1 h at RT. After centrifugation at 12000 rpm for 3 min, 100 μL of supernatant was transferred to a 96-well plate, and absorbance at 415 nm (A415nm) was measured using a UV spectrophotometer. The hemolysis rate of RBCs in the 0.9% NaCl solution was set to 0%, while in ddH₂O (0% NaCl solution) it was set to 100%. The erythrocyte lysis rate was calculated as follows: Erythrocyte lysis rate = A415nm (test group) / A415nm (0.0% NaCl).

Scanning electron microscopy

The RBCs were suspended and initially fixed in 1% glutaraldehyde solution at RT for 2 h. Subsequently, the fixed RBCs were rinsed and further fixed in 1% osmium tetroxide working solution for 1 h. Following three washes with ultrapure water, the RBCs were dehydrated with ethanol and subsequently mounted onto glass slides for imaging using a scanning electron microscope (JSM-IT700HR, JEOL, Japan).

Anti-blood group B antibody assay

Five milliliters whole blood from type O recipient, collected in ethylene diamine tetra acetic acid (EDTA) anticoagulant tubes, was centrifuged at 1500 × g for 5 min at room temperature (RT), and the serum was separated into two separate batches for anti-B IgM and IgG testing.

For anti-B IgM titer assay, test tubes and Neutral Gel Cards (BX1010, Bioxun Biotech, China) were labeled NEAT, 1:2, 1:4, 1:8, up to 1:2048 for experimental serum. Each tube was filled with 200 μL 0.9% NaCl solution, and 200 μL serum from the NEAT tube was taken and transferred to the 1:2 tube; 200 μL was then taken from the 1:2 tube and transferred to 1:4 tube. Continuing serially up to the 1:2048 tube. Next, 0.8% type B RBCs (BX2001-5-3, Bioxun Biotech, China) were added to the corresponding Neutral Gel Cards, with 50 μL in each hole, followed by 50 μL of diluted type O serum from each test tubes. The mixtures were incubated for 10 min at RT and then centrifuged at 1000 × g for 10 min using a BIO-RAD ID-centrifuge. The results were then assed.

For anti-B IgG titer detection, the serum was pretreated by 2-mercaptoethanol (2-ME) (1:1, 31350010, Invitrogen, USA) in a test tube, followed by incubation at 37 °C for 30 min to remove IgM activity. The preconditioned serum was then serially diluted as previously described. Next, 0.8% type B RBCs were added to Gel Coombs Test Cards (BX1016, Bioxun Biotech, China) at 50 μL per well, followed by 50 μL of diluted serum. The mixtures were incubated for 15 min at RT, centrifuged with a BIO-RAD ID-centrifuge at 1000 × g for 10 min, and then the results were evaluated following the manufacturer’s instructions.

Enzymatic conversion of kidneys during HMP

Type B donor kidneys, deemed unsuitable for transplantation, were perfused using a LifePort Kidney Transporter following the manufacturer’s instructions. Each pair of kidneys was randomly assigned to receive the kidney perfusate solution (determined as the most effective through RBCs conversion) containing either 5 μg/mL or 10 μg/mL B-zymes. The kidneys underwent HMP for 4 h. Biopsies were taken before and after treatment at hourly intervals, fixed with formalin, and embedded in paraffin (FFPE). Renal arterial pressure, blood flow, and resistance were monitored every 30 min. Perfusate samples were collected pre- and post-perfusion, flash-frozen in liquid nitrogen, and stored long-term at −80 °C. Donor information is detailed in Supplementary Table 2. Immunofluorescence was used to analyze the expression of blood group antigens following the described methods.

Ex vivo ABO-incompatible (ABOi) kidney transplant model

An ex vivo ABOi transplant model was established using the Life-X100 system based on a previously published protocol²⁷. In phase 1, kidneys from the same donor were randomly assigned to either the control group (HMP without B-zyme treatment) or the ECO group (HMP with 10 μg/mL B-zyme treatment). The kidneys underwent initial perfusion with kidney conversion solution for 3 h during HMP. Subsequently, they were flushed with 500 ml of Ringer’s solution upon removal from the LifePort Kidney Transporter, then placed on ice pending initiation of ABOi simulation.

In phase 2, following 3 h of HMP, the kidneys were re-perfused under NMP with 2 units of type O RBCs and 600 mL of fresh frozen type O plasma containing anti-B antibodies (IgG titer ≥ 1:128, IgM titer ≥ 1:32). Specifically, during phase 2, kidneys were transferred to the organ reservoir for ex vivo NMP and perfused for 4 h at a targeted mean arterial pressure of 65–75 mmHg, with oxygen supplied at 20 ml/min and maintained at 37 °C. Biopsies were obtained before and after treatment, fixed in formalin, and embedded in paraffin (FFPE). Throughout the procedure, renal arterial pressure, blood flow, and resistance were monitored at 30-min intervals. Samples of re-perfusate and urine (if any) were collected before and after perfusion, flash-frozen in liquid nitrogen, and stored long-term at −80 °C.

Pre-clinical ABO-incompatible kidney transplantation in a human decedent

A discarded type B kidney underwent perfusion using the HMP system with 10 μg/mL B-zyme for 3 h to produce an ECO kidney. Subsequently, the ECO kidney was flushed with 200 mL of the Celsior solution to eliminate residual enzymes. The ECO kidney was then transplanted into a type O brain-dead recipient at the Second Affiliated Hospital of Chongqing Medical University. The recipient is blood group type O with high titer anti-B antibodies (IgG, 1:128; IgM, 1:32). Detailed characteristics of the type B donor and type O recipient are provided in Supplementary Table 3. This study received ethics approval from the Ethics Committee of the Second Affiliated Hospital of Chongqing Medical University (2023SHEN116). Written consent for research was given by the donor families and was obtained.

During transplantation, the renal artery and vein were anastomosed to the external iliac vessels in an end-to-side manner, and the ureter was directly connected to a urinary collection system. The recipient received induction therapy with antithymocyte globulin (ATG) and Rituximab, followed by triad immunosuppression for maintenance, as outlined in Supplementary Table 4. After transplantation, the kidney was placed on the body surface and covered with sterile gauze to facilitate observation and biopsy. A multi-point biopsy (16 G, two stitches daily) of the ECO allograft was conducted. Biopsy tissues were divided: one part was fixed in formalin and embedded in paraffin for subsequent histological and immunological analyses, while another part was snap-frozen and stored at −80 °C. Peripheral blood, urine from the ECO allograft, and native kidneys were collected daily post-transplantation and processed by centrifugation at 4000 rpm for 15 min at 4 °C. The supernatants were split: one portion was preserved at −80 °C for cytokine analysis, and the other was used for anti-B antibody detection.

Histological analysis

Four microliter sections of FFPE kidney tissues were subjected to standard hematoxylin and eosin (H&E) staining. Renal injury was scored according to the following system: 0 = no injury, 1 = 1%–20% of area, 2 = 21%–50% of area, 3 = 51%–75% of area, and 4 ≥ 75%. Tubular injury was defined as tubular sloughing, cast formation, dilatation, degeneration, atrophy, or tubulitis. All allograft biopsies were assessed by two expert nephropathologists who were masked in this study, and the following lesion scores (0–3) of allograft was scored according to Banff 2019 criteria²²: interstitial inflammation in non-scarred cortex (i), tubulitis (t), arteritis (v), glomerulitis (g), peritubular capillaritis (ptc), microvascular inflammation (MVI, g+tc); mesangial matrix expansion (mm); arteriolar hyalinosis (ah); chronic glomerulopathy (cg), interstitial fibrosis (ci), tubular atrophy (ct), arterial intimal fibrosis (cv), total inflammation (ti), interstitial fibrosis and tubular atrophy (IFTA), interstitial inflammation in scarred cortex (i-IFTA), and tubulitis in area of IFTA (t-IFTA). Six randomly selected fields per sample were captured using a Zeiss AX10 Imager (ZEISS, Oberkochen, Germany).

Multiplex immunofluorescence staining

Four microliter FFPE sections were deparaffinized with xylene and dehydrated with an ethanol gradient before heat-induced epitope retrieval using a Tris-EDTA buffer (pH 9). Sections were stained with the following primary antibodies: anti-CD31 (1:1, MAB-0720, MXB Biotechnologies, China); anti-blood group B antigen (1:50, ab2524, abcam, UK); anti-blood group H2 (1:50, sc-59467, Santa Cruz Biotechnology, USA); anti-human-IgM (1:1000, ab134159, abcam, UK); anti-human-IgG (1:500, ab109489, abcam, UK); anti- C1qA (1:200, ab189922, Abcam, UK); anti-C3 (1:200, ab200999, abcam, UK); anti-C4d (1:100, ab167093, abcam, UK); anti-C5b-9 (1:500, ab55811, abcam, UK). For multiplex immunofluorescence staining, Opal Polaris 7-Color Manual IHC Kit (NEL861001KT, Akoya Biosciences, USA) was used according to the manufacturer’s instructions. After each staining cycle, slides were microwaved at room temperature (45 s at 100% power and 15 min at 20% power in Opal Slide Processing Jar) to strip the primary-secondary-HRP complex. Sections were cooled naturally after antigen retrieval, washed with PBS, stained with DAPI for 5 min in the dark, and imaged using Vectra Polaris Quantitative Pathology Imaging Systems (Akoya Biosciences, USA). Six randomly chosen fields per sample were acquired using QUPATH (v0.5.0) for semi-quantitative analysis with Image-Pro Plus 6.0 software. Relative MFI = MFI (Post-treatment)/ MFI (Pre-treatment). % B-Ag removal = 1-(Relative MFI of B-Ag) * 100%.

Circulating antibody assay

The levels of circulating anti-B IgG and IgM in re-perfusates were quantified as described previously¹⁶. Briefly, re-perfusate supernatants were incubated with ABO-B type RBCs (1:1, FRBIOMEABO, Brother Biotech, China) for 30 min at RT. After two washes with cell staining buffer, the RBCs were stained with Alexa Fluor 488-conjugated goat anti-human IgM (1:100, A-21215, Invitrogen, USA) and Alexa Fluor 647-conjugated goat anti-human IgG (1:100, A-21445, Invitrogen, USA) at RT for 30 min. Subsequently, the RBCs were washed twice with staining buffer, resuspended in staining buffer, and analyzed using a CytoFLEX cytometer (Beckman-Colter, USA). The complete gating strategy is provided in Supplementary Fig. 14.

Enzyme-linked immunosorbent assay

The concentrations of IL-1β, IL-2, IL-6, IL-8, TNF-α, INF-γ, CCL5, CX3CL1, CXCL9, C1q, C3, C3b, C4, and C5b-9 in the organ perfusate and peripheral blood, as well as the concentrations of CXCL9, IL-18, Kim-1, and NGAL in perfusate and urine, were quantified using commercial ELISA kits (Bio-Swamp Bioengineering Institute, Wuhan, China) following the manufacturer’s instructions. Average of two technical replicates for different biological replicates.

TUNEL analysis

Four microliter FFPE sections were assessed using the DeadEndTM Fluorometric TUNEL system (G3250, Promega, USA) following the manufacturer’s instructions. The stained sections were visualized and scanned using Vectra Polaris Quantitative Pathology Imaging Systems (Akoya Biosciences, USA). The extent of apoptosis was quantified and normalized based on the percentage of TUNEL-positive cells relative to the total number of cells. Six randomly selected fields of view were measured.

Statistical analysis

Statistical analyses were conducted using GraphPad Prism 9.0 (GraphPad Software, USA), and all data are presented as mean ± standard error of the mean (SEM) unless stated otherwise. Two-tailed paired t-test was used to compare two paired experimental conditions or one-way and two-way analysis of variance (ANOVA) with Bonferroni’s multiple comparisons test to compare more than 2 experimental conditions involving 1 or 2 independent variables. Statistical significance was considered at p ≤ 0.05 (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001); p > 0.05 was deemed not significant (ns).

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.