Blood was drawn from (i) healthy volunteers (Volunteers, n=20; ABO blood types A n=6; B n=6; AB n=3; O n=5), (ii) patients with ESRD (n=20), pre-renal transplantation (Pre) and on Day 1 (POD 1) and Day 14 (POD 14) after renal allotransplantation, (iii) brain-dead organ donors (DBD, n=20) and (iv) patients with renal allografts who were currently experiencing episodes of acute T cell-mediated rejection (Allo-TCMR, n=20) ( Table 1 ). Sera were obtained from de-identified remnant/discarded clinical laboratory samples.

Sera from Volunteers were obtained from the Second Affiliated Hospital of Hainan Medical University, and all experimental protocols were approved by the ethics committee of the Second Affiliated Hospital of Hainan Medical University. All procedures involving humans were performed in accordance with the relevant guidelines and regulations, and had no adverse effects on the subjects.

Blood was obtained from wild-type (WT, i.e., genetically-unmodified) pigs (n=4) and from different genetically-modified pigs (n=4) ( Table 1 ) (Chengdu Clonorgan Biotechnology, Chengdu, Sichuan, China).

Pig RBCs and PBMCs were stained for expression of Gal (by isolectin BSI-B4), Sda (Dolichos biflorus agglutinin,DBA), Neu5Gc (chicken anti-Neu5Gc mAb), and SLA (anti-human β2-microglobulin antibody, β2M).

Binding of human antibodies to pig cells was measured by flow cytometry using the relative geometric mean (rGM), as previously described (11). Briefly, pRBCs were separated from whole blood, washed x3 with phosphate-buffered saline (PBS), and centrifuged at 700g for 5min at 4°C. The washed RBCs were suspended in fluorescence-activated cell sorting (FACS) buffer (PBS containing 1% bovine serum albumin). pPBMCs were isolated using Ficoll (HaoYang, Tianjin, China) and suspended in FACS buffer for IgM/IgG binding assays. The isolated pRBCs (5x10⁵/tube) and pPBMCs (5x10⁵/tube) were incubated with heat-inactivated human serum at 4°C for 30min, respectively, and the final serum concentration was 20%. After incubation, cells were washed with PBS to remove unbound antibodies and were blocked with 10% goat serum for 15min at 4°C. After further washing with PBS, anti-human IgM or anti-human IgG (Jackson ImmunoResearch Laboratories, West Grove, PA, USA) (IgG: concentration 1:1000 for pRBCs and pPBMCs; IgM: concentration 1:1600 for pRBCs and pPBMCs) was added, and the cells were incubated for 30min at 4°C. After washing with PBS, 100μL PBS buffer was added. Flow cytometry was carried out using BD FACSCelesta (Becton Dickinson, San Jose, CA, USA).

Briefly, PBMCs (5×10⁵ cells in 250µL FACS buffer) were incubated with 50µL heat-inactivated human serum at 4°C for 1h. After washing with PBS, FACS buffer (200µL) and rabbit complement (50µL, Cedarlane, Hornby, CA, USA) were added (final concentration 20%), and incubation was carried out at 37°C for 30min. After washing with PBS, the cells were incubated in the dark at 4°C for 15min with propidium iodide, and finally 200µL FACS buffer was added. Flow cytometry was carried out using BD FACSCelesta.

Cytotoxicity was calculated, as follows (11):

where A represented the percentage of dead cells, B was the maximal percentage of dead cells (PBMCs fixed with 70% ethanol), and C was the minimal percentage of dead cells (PBMCs incubated with medium only).

Significance of the difference between two groups was determined by student t-test or Wilcoxon test. Continuous variables were expressed as mean ± SD. Comparisons among multiple groups were performed using a One-way ANOVA test (Tukey test) or nonparametric test (Dunn’s test). A p value of <0.05 was considered statistically significant. All statistical analyses were performed using social sciences software GraphPad Prism 8 (GraphPad Software, San Diego, CA, USA).

The aim was to compare the binding of different human sera to various pig PBMCs.

The aim was to compare the binding of different human sera to various pig RBCs.

The data presented in relation to differences in binding of various human sera to pig cells ( Figures 2A – E ) were re-presented to more clearly illustrate the effect of different pig genotypes ( Figures 2F – J ).

Patients with ESRD were divided into 2 groups based on whether they were undergoing hemodialysis (n=15) or peritoneal dialysis (n=5).

Serum samples from renal allotransplant recipients were collected pretransplant (Pre), and on POD1 and POD14, and the levels of anti-pig IgM and IgG antibodies were measured ( Figure 5 ). In the majority of patients (n=18), there was no change in IgM/IgG binding to PBMCs or RBCs between Pre, POD 1, and POD 14, although in some patients anti-pig IgM/IgG decreased transiently on POD1, but recovered to Pre levels by POD14. (This may possibly be related to hemodilution by perioperative fluid infusion.) However, in 2 of the recipients (red dot), serum IgG binding to GTKO PBMCs increased by POD14, compared with Pre and POD1. There was no change in IgG binding to GTKO RBCs (that do not express SLA), nor in binding to PBMCs and RBCs from the other pig genotypes.

Healthy volunteers were divided into 4 groups based on ABO blood type.

The first important observation in this study was that our data showed that healthy volunteer serum IgM binding to WT, GTKO, GTKO/β4GalNT2KO PBMCs was significantly greater than to CMAHKO and TKO/hCD55 PBMCs, strongly suggesting that CMAHKO will be important for clinical renal xenotransplantation, as others have also reported (24, 25). However, deletion of expression of Neu5Gc appeared to play a more important role in reducing human antibody binding to the pig cells than deletion of expression of Gal, which is in contrast to some previous studies in which GTKO had a much greater impact than CMAHKO on IgM binding (14, 26). The reduction in IgM and IgG binding to GTKO pig cells when compared to WT cells was less than reported in most previous studies (25), but, in contrast, the reduction in binding and CDC was significantly greater after deletion of expression of Neu5Gc. This is most likely explained by differences in the antibody profiles of the Chinese participants in this study compared with those of some other ethnic groups.

Neu5Gc is expressed in pigs, apes and OWMs, but not in humans (27–29), and therefore only humans develop anti-Neu5Gc antibodies (30). Gao and her colleagues drew attention to the fact that natural antibodies are largely associated with exposure to glycans expressed on flora in the gastrointestinal tract (26), as suggested by others previously (31). In humans, anti-Neu5Gc antibodies develop during the first 6 months of life, and reach adult levels by the end of the first year (32). Both anti-Neu5Gc IgM and IgG increase soon after the infant is exposed to cow’s milk and baby foods containing red meat, which express Neu5Gc (32).

All of the subjects from whom blood was drawn in the present study were Chinese and had been resident in China throughout most of their lives, whereas those in other reported studies were from a variety of ethnic and geographic backgrounds. This suggests that maybe in Chinese patients (or in patients who have been exposed to Chinese environmental factors, e.g., diet, for a prolonged period of time), an absence of expression of Neu5Gc in the pig organ may be as important, if not more important, than absence of expression of Gal.

We therefore suggest that the Chinese subjects included in the present study expressed different gastrointestinal flora (perhaps based on differences in diet), than other groups that have been studied in other geographic regions (26, 33), thus rendering Neu5Gc a more important stimulus to natural antibody production. However, of interest, no correlation between diet and anti-pig antibody levels was found in a previous study in Taipei (34). There may other ‘ethnic’ or ‘environmental’ differences in other population groups that have not yet been investigated.

There are increasing in vitro data indicating that TKO pig organs will prove to be a major advance over GTKO organs for transplantation into humans (5, 7, 35) which is consistent with our conclusions from the present study.

A second major observation made in this study was that (i) patients with ESRD, brain-dead donors, and immunosuppressed patients with kidney allografts generally had significantly lower levels of anti-pig antibodies than healthy human volunteers, except in regard to WT pig cells. The trends in CDC were similar to those of IgM and IgG. In addition, the ABO blood type of the donor of the serum appeared to play no part in influencing the results.

The clinical impact of ESRD on anti-pig immunity remains to some extent uncertain because immune dysfunction in ESRD includes both immunoactivation and immunosuppression. Heparin-induced antibodies (HIA) (36, 37) and anticardiolipin antibodies (IgG-ACA) (38) are elevated in ESRD and the complement system can be activated (39, 40). However, It is well-recognized that ESRD patients are to some extent immunocompromised (41, 42), e.g., reduced number of NK cells, reduced phagocytic activity of neutrophils  (43). Anti-pig antibodies are low in infants (44) and in patients with ESRD (14), the last of which observations is consistent with our present results.

Patients with kidney allografts are receiving chronic immunosuppressive therapy, and presumably this has prevented an increase in antibody levels even though T cell activation had taken place in the patients we studied. The explanation in brain-dead donors is not so obvious, but may be associated with the infusion of fluids to maintain an adequate hemodynamic state, resulting in hemodilution.

However, there was no difference in serum IgM/IgG binding to GTKO/CMAHKO and TKO/hCD55 PBMCs or RBCs in the 4 groups, indicating that the effect of deletion of Neu5Gc expression on the pig cell was sufficient to reduce antibody binding and CDC to negligible levels whatever the source of the serum. Expression of CD55 on TKO pig PBMCs further reduced CDC of the cells.

Our data showed that there was no significant difference in serum IgM/IgG binding to any of the pig cell types between hemodialysis and peritoneal dialysis patients, with the exception of lower IgM binding to GTKO/CMAHKO cells in patients on hemodialysis. This suggests that hemodialysis might remove anti-β4GalNT2 antibodies. However, it is unlikely that hemodialysis directly removes antibodies since standard hemodialysis filters typically have a cut-off size of between 10–20 kDa, whereas IgM/IgG molecules are >150 kDa, and are thus not removed by hemodialysis (45).

As is well-known, the most widely used xenotransplantation model is the genetically-modified pig-to-OWM. A major difference between humans and OWMs is that OWMs express Neu5Gc on the vascular endothelium, whereas humans do not. OWMs, therefore, are far from ideal models for xenotransplantation, as there is markedly increased OWM serum antibody binding and CDC to TKO pig cells. Although New World monkeys have some advantage in this respect (5), their small size and the ineffectiveness of some immunosuppressive drugs in them negates their suitability as a surrogate recipient for pig organ transplantation (46–50). Is there an alternative recipient as a surrogate for living humans?

Could DBD subjects be used as recipients in preclinical studies of xenotransplantation? Our data indicate that their levels of anti-pig antibodies are significantly lower than healthy volunteers (possibly associated with hemodilution through the need for fluid administration to maintain hemodynamic stability), and comparable to those in patients with ESRD. However, activation of innate immunity and inflammation can occur in brain-dead subjects (21). This observation, and because of their hemodynamic instability that may limit follow-up to days rather than months, reduces their suitability as potential surrogates for living recipients (51, 52).

Our observation that 2 of 20 patients undergoing kidney allotransplantation developed increased anti-pig antibody binding on POD14 may suggest that, at least in some patients, there may be cross-reactivity between anti-HLA antibodies and SLA epitopes, but the data are too few to draw any definite conclusions. (Unfortunately, anti-HLA antibodies were not investigated.)

In summary, on the basis of the present study, (i) CMAHKO in the pig may be critical to the success of clinical pig kidney xenotransplantation, and may be the most important after α1,3-galactosyltransferase gene be knockout, at least in Chinese patients; (ii) subjects with ESRD, or who are immunosuppressed after kidney allotransplantation, and brain-dead organ donors, all have lower levels of antibody binding and CDC to genetically-engineered pig cells than do healthy human volunteers; (iii) brain-dead subjects may mimic ESRD patients in that they both have low levels of anti-pig antibody levels, but experimental pig organ transplants in this group are unlikely to provide significant information of real value; (iv) TKO pigs with selected human ‘protective’ transgenes, e.g., CD55, are likely to prove to be optimal sources of kidneys for clinical xenotransplantation. (v) The role of SLAKO or SLA knockdown remains uncertain.