Abstract
We investigated the predictive biomarkers for graft rejection in pig-to-non-human primate (NHP) full-thickness corneal xenotransplantation (n = 34). The graft score (0-12) was calculated based on opacity, edema, and vascularization. Scores ≥ 6 were defined as rejection. NHPs were divided into two groups: 1) graft rejection within 6 months; and 2) graft survival until 6 months. In the evaluation of 2-week biomarkers, none of the NHPs showed rejection within 2 weeks and the 34 NHPs were divided into two groups: 1) entire rejection group (n = 16); and 2) survival group (n = 18). In the evaluation of 4-week biomarkers, four NHPs showing rejection within 4 weeks were excluded and the remaining 30 NHPs were divided into two groups: 1) late rejection group (n = 12); and 2) survival group (n = 18). Analysis of biomarker candidates included T/B cell subsets, levels of anti-αGal IgG/M, donor-specific IgG/M from blood, and C3a from plasma and aqueous humor (AH). CD8+IFNγ+ cells at week 2 and AH C3a at week 4 were significantly elevated in the rejection group. Receiver operating characteristic areas under the curve was highest for AH C3a (0.847) followed by CD8+IFNγ+ cells (both the concentration and percentage: 0.715), indicating excellent or acceptable discrimination ability, which suggests that CD8+IFNγ+ cells at week 2 and AH C3a at week 4 are reliable biomarkers for predicting rejection in pig-to-NHP corneal xenotransplantation.
Keywords: biomarker, C3a, CD8+IFNγ+ cells, cornea, rejection, transplantation, xenotransplantation
1. Introduction
Corneal xenotransplantation using pig donors has been investigated as a substitute for human donor corneas.1–4 One of the challenges involving xenotransplantation relates to hyperacute rejection mediated by the natural antibody against Galalpha1,3Galbeta1,4GlcNAc-R (αGal) synthesized by α1,3-galactosyltransferase because non-human primates (NHPs) and humans carry natural anti-αGal IgM.5 Corneas express αGal, although at lower levels than vascular endothelial cells do.6 Further, the cornea is immune-privileged.7–9 Pig-to-NHP corneal xenotransplantation has been reported to result in longer graft survival compared with other solid organ xenotransplantation.10
Although the success rate of low-risk corneal allotransplantation is greater than 90%,8 the rejection rate can be up to 70% in high-risk cases.11,12 In allograft rejection, both innate and adaptive immunities are involved.13 Th1 cells play an important role.14,15 In corneal xenotransplantation, the rejection process may evoke higher innate response in addition to T-cell responses.5
Several biomarkers are associated with corneal allograft rejection. Cytokines such as interleukin (IL)-6, IL-10, interferon (IFN)-γ, monocyte chemoattractant protein-1, and inflammatory monocytes in aqueous humor (AH), and impaired regulatory T cells have been reported in corneal allograft rejection.15–18 In pig-to-NHP corneal xenotransplantation, the increase in IFN, tumor necrosis factor, IL-4, IL-5, IL-6, IL-10, and C3a levels was related to graft rejection.19–21 However, these biomarkers were based on a small population of subjects and no study analyzed the predictive ability of biomarkers in corneal xenotransplantation.
In this study, a retrospective analysis of various parameters was conducted to investigate the role of predictive biomarkers in graft rejection within 6 months in NHPs who underwent corneal xenotransplantation, using unpublished data or results of previous studies.2,3,22
2. Materials and Methods
This study adhered to ARVO Statement regarding the Use of Animals in Ophthalmic and Vision Research. This study was approved by Seoul National University (SNU) (IACUC: SNU-151102–3) and SNU Hospital (IACUC: 09–0156, 11–0152, 12–0207, 13–0221, 15–0171).
2.1. Recipient characteristics and study design
Between 2010 and 2018, 38 rhesus macaques which had undergone full-thickness porcine corneal xenotransplantation were included.2,3,22 Among them, four NHPs dying within 3 months without graft rejection were excluded. Donor pig characteristics, immunosuppressants used, and graft survival in all NHP recipients (n = 34) are summarized in Table 1. Briefly, 29 NHPs grafted with SNU wild-type (WT) miniature pig corneas reported previously,2,3,22 and five NHPs grafted with α1,3-galactosyltransferase gene-knockout (GTKO, n = 4) or GTKO/human CD39 knockin (n = 1) miniature pig corneas in a new experiment were included.
Table 1.
Systemic immunosuppressive regimen and corneal graft survival in NHPs (n = 34)
| Group No. | Group | Systemic Immunosuppressive regimen |
Subject number |
Donor pig | Graft survival | Reported year |
|---|---|---|---|---|---|---|
| 1 | Conventional steroid | Methylprednisolone | 3 | WT | 21, 28, 29 | 20152 |
| 2 | CD154 | Anti-CD154 Ab IVIG Methylprednisolone |
4 | WT | >192, >243, 318, 933 | 20152 |
| 3 | CD40 | Anti-CD40 Ab IVIG Methylprednisolone |
6 | WT | 41, >196, >203, >273, >422, >511 | 20183 |
| 4 | CD20 (Full dose) | Anti-CD20
Aba Tacrolimusc IVIG Basiliximab Methylprednisolone |
6 | WT | 134,>184,>210,>260,297,>470 | 20183 |
| 5 | CD20 (Low dose) | Anti-CD20
Abb Tacrolimusd IVIG Basiliximab Methylprednisolone |
7 | WT | 56, 92, 162, >181, >182, >182,>198 | 201822 |
| 6 | Tacrolimus only | Tacrolimuse IVIG Basiliximab Methylprednisolone |
5 | GTKO (n = 4), GTKO/hCD 39KI (n = 1) | 37, 55, 72, 91, 165 | Unpublished |
| 3 | WT | 29, 149, 161 | 201822 |
WT, wild type; Ab, antibody; IVIG, intravenous immunoglobulin; GTKO, alpha1,3-galactosyltransferase gene-knockout; hCD39KI, human CD39 knockin
Groups 1-6 with topical immunosuppressants: All NHPs received topical prednisolone acetate 1% (Pred forte®; Allergan, Irvine, CA, USA) daily for 3 months and injected subconjunctivally with dexamethasone 1.5 mg/0.3 mL (JW Pharmaceutical, Seoul, Republic of Korea) every week for 6 months.
Groups 1-6: Methylprednisolone was used with the same protocol in all groups. It was intramuscularly administered at an initial dose of 2 mg/kg/day and tapered over 5 weeks.
Groups 2-5: IVIG was used with the same protocol in groups 2-5. It was intravenously administered on preoperative day 1 and postoperative day 14 at a dose of 1 g/kg.
Group 2: Recombinant anti-CD154 Ab (V-regions from mouse 5C8 clone; C-regions human IgG1k) was intravenously administered 15 to 19 times at a dose of 20 mg/kg. (Am J Transplant. 2015;15:628-641)
Group 3: A mouse-rhesus chimeric monoclonal anti-CD40 Ab (2C10R4, NIH Nonhuman Primate Reagent Resource) was intravenously administered 15 times at a dose of 30-50 mg/kg. (Am J Transplant. 2018;18:2330-2341.)
Groups 4 and 5: Anti-CD20 Ab (Rituximab; MabThera®, Hoffmann-La Roche, Basel, Switzerland) was intravenously administered at a dose of 20 mg/kg on postoperative days 0 and 7, and every 2a or 3b months. (Am J Transplant. 2018;18:2330-2341.; Xenotransplantation. 2018;25:e12442)
Groups 4-6: Tacrolimus (Prograf®; Astellas Pharma US, Deerfield, IL, USA) was intramuscularly administered twice daily at a dose of 0.05c or 0.035e mg/kg or at a dose of 0.05 mg/kg for 4 weeks followed by 0.035 mg/kgd.
Groups 4-6: Basiliximab was intravenously administered at a dose of 0.3 mg/kg on postoperative days 0 and 4.
Penetrating keratoplasty procedures were described previously.2,3 NHPs were administered systemic and topical immunosuppressants listed in Table 1. Systemic immunosuppression was scheduled for six months. All NHPs received topical prednisolone acetate 1% (Pred forte®; Allergan, Irvine, CA, USA) daily for 3 months and injected subconjunctivally with dexamethasone 1.5 mg/0.3 mL (JW Pharmaceutical, Seoul, Republic of Korea) every week for 6 months.
Postoperative 2 or 4-week biomarkers for predicting graft rejection within 6 months were evaluated. To investigate 2 or 4-week biomarker candidates, 34 NHPs were divided into two groups: rejection group (entire or late) and survival group (Table 2). The entire rejection group included all NHPs whose graft was rejected within a 6-month period, while late rejection group included NHPs whose graft was rejected after more than 4 weeks up to 6 months. In the evaluation of the 2-week biomarkers, none of the NHPs showed rejection and the 34 NHPs were divided into two groups: entire rejection group (n = 16) and survival group (n = 18). In the evaluation of 4-week biomarkers, four NHPs showing rejection within 4 weeks were excluded and the remaining 30 NHPs were divided into two groups: late rejection group (n = 12) and survival group (n = 18).
Table 2.
Schematic of the study design for the 2 or 4-week biomarker candidates and group characteristics at weeks 2 and 4.
| Analysis time |
Biomarker candidates | Rejection group number |
Average graft survival of each rejection group, mean ± SD (ranges) |
Survival group number |
Total number |
|---|---|---|---|---|---|
| Week 2 | Blood: C3a, DS Abs, anti-αGal Abs, T and B cell subsets | 16* | 82.63 ± 54.38 (21~161) | 18 | 34 |
| Week 4 | Blood: C3a, DS Abs, anti-αGal Abs, T and B cell subsets AH: C3a |
12† | 101.25 ± 50.15 (41~161) | 18 | 30 |
DS, donor pig-specific; Abs, antibodies; αGal, galactose-alpha-1,3-galactose; AH, aqueous humor
named as “Entire rejection group”. Entire rejection group includes NHPs whose grafts were rejected within 6 months.
named as “Late rejection group”. Late rejection group includes NHPs whose grafts were rejected at > week 4 up to month 6.
To analyze 2 or 4-week biomarker candidates, blood or AH was collected to obtain the T and B cells, Abs, and C3a. Biomarker candidates were evaluated by comparing the rejection and survival groups at baseline, week 2, and week 4. AH C3a assay was performed only at week 4. In addition, we performed subgroup analysis to evaluate the effect of GTKO on predictive biomarkers. The subgroup analysis involved NHPs carrying rejected WT grafts and those carrying rejected GTKO grafts. Similar analysis was conducted to compare NHPs carrying surviving WT grafts and NHPs with rejected WT grafts after excluding those with GTKO xenografts. Receiver operating characteristic (ROC) curve analysis was performed to determine the predictive ability of the biomarkers, and areas under the curves (AUCs) were calculated to determine the level of discrimination.23
2.2. Graft score and definition of rejection
The corneal graft score (0–12) was calculated based on opacity, edema, and vascularization as described previously.24 Scores ≥ 6 were defined as graft rejection. Success criteria for corneal xenograft are based on 6-month graft survival.24 Therefore, data were analyzed up to 6 months.
2.3. T and B cell assays
Sub-populations of T cells (CD28+CD95+ central memory, CD28−CD95+ effector memory, and CD4+CD25+Foxp3+ regulatory cells) and activated B cells (CD3−CD20+CD28+) in blood were evaluated.1,2 For extracellular surface staining, cell suspensions were incubated for 30 minutes at 4°C with fluorescein-conjugated mouse anti-human Abs as follows: CD3-FITC (1:40), CD4-FITC (1:200), CD8-PerCp-Cy5.5 (1:200), CD25-APC, CD28-APC (1:200), CD95-PE (1:200), and CD20-PE (1:200). For intracellular Ab staining, cell suspensions were incubated at 4°C with fluorescein-conjugated mouse anti-human Abs as follows: IFN-γ-PE (1:200, 30 minutes) and Foxp3-PE (1:200, 1 hour). Intracellular IFN-γ staining was performed after stimulation overnight with anti-CD3 Ab (2.5 μg/mL) and anti-CD28 Ab (0.25 μg/mL) in the presence of GolgiPlug (brefeldin A; 1 μL/1 mL). All Abs were purchased from eBioscience (San Diego, CA, USA), except CD3-FITC (from BD PharMingen, San Diego, CA, USA) and anti-CD3 Ab (U-CyTech, Utrecht, The Netherlands). Data were acquired using a FACSCanto flow cytometer (Becton-Dickinson, Mountain View, CA, USA) and analyzed using FlowJo software (Tree Star, Ashland, OR, USA) (Supplementary Figure S1). Data were presented as the absolute number of cells per unit volume or percentage of peripheral blood mononuclear cells.
2.4. Antibody assay
Plasma anti-αGal IgM/G Abs were measured by ELISA as previously described.25 Concentrations of anti-αGal Abs were expressed as artificial units (AU)/mL. Plasma levels of donor pig-specific (DS) IgM/G Abs were measured using flow cytometric cross-match technique using donor PBMCs as targeting cells.4 Concentrations of DS Abs were semi-quantitatively expressed as mean fluorescence intensity (MFI).
2.5. Complement (C3a) assay
Levels of C3a in the plasma or in the AH were measured using the OptEIA™ Human C3a ELISA Kit (BD Biosciences, San Diego, CA, USA) according to the manufacturer’s protocols. The upper detection limit of C3a concentration of AH was 25 ng/mL.
2.6. Statistical analysis
Normality was assessed by Shapiro-Wilk test. Independent continuous variables were compared using the Mann-Whitney U test or independent t-test. To determine the predictive ability of biomarkers, we performed ROC curve analysis. AUCs over 0.7, 0.8, or 0.9 were considered as acceptable, excellent, or outstanding discrimination, respectively.23 The value with the maximum Youden index (J = sensitivity + specificity − 1) was regarded as the optimal cut-off.26 Mann-Whitney U test/independent t-test or ROC curve analysis was performed using SPSS v20.0 (SPSS Inc., Chicago, IL, USA) or the pROC package in R (V.3.5.0; R Foundation, Vienna, Austria), respectively. A P value of < 0.05 was considered statically significant.
3. Results
3.1. Comparison of biomarker candidates in the rejection and survival groups
Baseline levels of biomarker candidates in the rejection and survival groups are shown in Table 3. Biomarker candidate levels at baseline did not significantly differ between the two groups.
Table 3.
Baseline levels of biomarker candidates in rejection and survival groups.
| Biomarkers | Entire rejection (n = 16) |
Late rejection (n = 12) |
Survival (n = 18) |
P (Entire rejection vs. Survival) |
P (Late rejection vs. Survival) |
|---|---|---|---|---|---|
| AH C3a (ng/mL) | 3.98 ± 1.74 | 3.87 ± 1.69 | 4.10 ± 1.09 | 0.810* | 0.659* |
| Plasma C3a (ng/mL) | 178.11 ± 130.47 | 125.80 ± 31.54 | 162.93 ± 128.36 | 0.308† | 0.783† |
| DS IgG (MFI) | 175.52 ± 229.95 | 99.28 ± 123.88 | 123.64 ± 208.17 | 0.905† | 0.514† |
| DS IgM (MFI) | 217.42 ± 318.36 | 130.79 ± 137.07 | 214.22 ± 127.15 | 0.190† | 0.071† |
| Anti-αGal IgG (AU/mL) | 102.33 ± 207.52 | 106.08 ± 236.04 | 61.84 ± 159.23 | 0.512† | 0.799† |
| Anti-αGal IgM(AU/mL) | 269.77 ± 243.15 | 281.14 ± 277.50 | 366.29 ± 202.75 | 0.053† | 0.075† |
| CD4+IFNγ+ (cells/mm3) | 12.14 ± 16.74 | 11.75 ± 17.24 | 17.02 ± 22.16 | 0.352† | 0.330† |
| CD8+IFNγ+ (cells/mm3) | 34.28 ± 30.39 | 35.25 ± 29.68 | 54.45 ± 51.71 | 0.317† | 0.498† |
| CD4+ CM (cells/mm3) | 329.44 ± 157.63 | 365.36 ± 154.82 | 287.15 ± 130.26 | 0.398* | 0.146* |
| CD4+ EM (cells/mm3) | 30.31 ± 32.82 | 35.68 ± 35.67 | 53.78 ± 63.30 | 0.370† | 0.735† |
| CD8+ CM (cells/mm3) | 130.52 ± 90.86 | 143.32 ± 95.86 | 110.71 ± 44.94 | 0.704† | 0.421† |
| CD8+ EM (cells/mm3) | 388.75 ± 239.71 | 373.99 ± 248.89 | 487.93 ± 319.19 | 0.408† | 0.352† |
| ActB (cells/mm3) | 8.21 ± 6.58 | 9.78 ± 6.65 | 8.00 ± 6.83 | 0.945† | 0.397† |
| Treg (cells/mm3) | 41.85 ± 29.41 | 49.66 ± 29.93 | 27.05 ± 24.07 | 0.112† | 0.057† |
| CD4+IFNγ+ (%) | 0.33 ± 0.47 | 0.30 ± 0.52 | 0.51 ± 0.70 | 0.334† | 0.204* |
| CD8+IFNγ+ (%) | 0.88 ± 0.79 | 0.82 ± 0.79 | 1.58 ± 1.62 | 0.157† | 0.138* |
| CD4+ CM (%) | 7.75 ± 3.40 | 8.19 ± 3.75 | 8.29 ± 4.08 | 0.679* | 0.945† |
| CD4+ EM (%) | 0.75 ± 0.90 | 0.84 ± 1.01 | 1.65 ± 2.24 | 0.301† | 0.472* |
| CD8+ CM (%) | 3.00 ± 1.70 | 3.18 ± 1.90 | 3.19 ± 1.39 | 0.719* | 0.988† |
| CD8+ EM (%) | 9.59 ± 5.47 | 8.30 ± 4.56 | 13.10 ± 7.96 | 0.138† | 0.054* |
| ActB (%) | 0.18 ± 0.16 | 0.22 ± 0.16 | 0.24 ± 0.22 | 0.469† | 0.832* |
| Treg (%) | 1.02 ± 0.59 | 1.10 ± 0.64 | 0.75 ± 0.60 | 0.195* | 0.140† |
Independent T-test (two-tailed),
Mann-Whitney U test (two-tailed)
Data are presented as mean ± SD.
vs, versus; AH, aqueous humor; DS, donor pig-specific; αGal, galactose-alpha-1,3-galactose; IFN, interferon; CM, central memory T cells; EM, effector memory T cells; ActB, activated B cells; Treg, regulatory T cells
Entire rejection group includes NHPs whose grafts were rejected within 6 months.
Late rejection group includes NHPs whose grafts were rejected at > week 4 up to month 6.
At week 2, the graft score did not significantly differ between the groups (0.94 ± 1.73 and 0.33 ± 1.41, respectively; P = 0.145; Table 4). Both the concentration and percentage of CD8+IFNγ+ cells at week 2 were significantly higher in the entire rejection group (52.32 ± 51.69 cells/mm3 and 1.13 ± 1.16%, respectively) than in the survival group (17.68 ± 16.26 cells/mm3 and 0.48 ± 0.56%, respectively; all P = 0.032), and those at week 4 and last examination showed no group-wise significant differences (Tables 4 and 5). The other biomarker candidates revealed no significant group-wise differences at week 2. At the last follow-up, the AH and plasma levels of C3a, DS IgG, and anti-αGal IgG were significantly higher in the entire rejection group than in the survival group.
Table 4.
Values of biomarker candidates at postoperative week 2 in the entire rejection and survival groups.
| Biomarkers | Week 2 |
Last FU |
||||
|---|---|---|---|---|---|---|
| Entire rejection (n = 16) |
Survival (n = 18) | P | Entire rejection (n = 16) |
Survival (n = 18) | P | |
| Graft score | 0.94 ± 1.73 | 0.33 ± 1.41 | 0.145* | 6.56 ± 0.51 | 0.11 ± 0.47 | < 0.001* |
| AH C3a (ng/mL)a | NA | NA | NA | 22.01 ± 5.94 | 6.50 ± 5.69 | < 0.001* |
| Plasma C3a (ng/mL) | 167.83 ± 131.33 | 137.67 ± 62.22 | 0.863* | 183.38 ± 97.42 | 124.06 ± 52.63 | 0.028* |
| DS IgG (MFI) | 680.30 ± 804.89 | 395.23 ± 349.09 | 0.748* | 1124.00 ± 1269.31 | 240.83 ± 439.75 | < 0.001* |
| DS IgM (MFI) | 299.66 ± 534.32 | 224.14 ± 105.64 | 0.343* | 252.13 ± 278.58 | 245.18 ± 144.93 | 0.485* |
| Anti-αGal IgG (AU/mL) | 234.64 ± 248.13 | 107.31 ± 75.46 | 0.255* | 385.48 ± 677.19 | 51.99 ± 111.92 | 0.007* |
| Anti-αGal IgM (AU/mL) | 416.53 ± 337.88 | 396.11 ± 209.65 | 0.512* | 353.34 ± 220.04 | 349.73 ± 171.86 | 0.730* |
| CD4+IFNγ+ (cells/mm3) | 19.41 ± 31.94 | 7.33 ± 10.52 | 0.098* | 8.19 ± 11.83 | 15.93 ± 24.74 | 0.350* |
| CD8+FNγ+ (cells/mm3) | 52.32 ± 51.69 | 17.68 ± 16.26 | 0.032* | 33.88 ± 28.99 | 58.08 ± 52.49 | 0.142* |
| CD4+ CM (cells/mm3) | 397.38 ± 194.67 | 394.60 ± 377.43 | 0.317* | 352.13 ± 156.95 | 268.24 ± 125.61 | 0.093† |
| CD4+ EM (cells/mm3) | 40.38 ± 58.57 | 25.04 ± 32.95 | 0.558* | 31.44 ± 35.26 | 71.96 ± 107.06 | 0.617* |
| CD8+ CM (cells/mm3) | 133.02 ± 70.14 | 138.69 ± 130.00 | 0.605* | 138.69 ± 94.06 | 109.26 ± 56.17 | 0.270† |
| CD8+ EM (cells/mm3) | 307.89 ± 159.73 | 261.28 ± 253.64 | 0.121* | 498.63 ± 521.24 | 515.40 ± 430.83 | 0.641* |
| ActB (cells/mm3) | 6.40 ± 5.76 | 3.77 ± 6.02 | 0.073* | 6.88 ± 12.41 | 4.32 ± 5.68 | 0.794* |
| Treg (cells/mm3) | 33.08 ± 32.11 | 45.21 ± 72.34 | 0.490* | 27.75 ± 18.14 | 21.80 ± 16.84 | 0.233* |
| CD4+IFNγ+ (%) | 0.48 ± 0.85 | 0.18 ± 0.25 | 0.112* | 0.32 ± 0.62 | 0.29 ± 0.33 | 0.388* |
| CD8+ IFNγ+ (%) | 1.13 ± 1.16 | 0.48 ± 0.56 | 0.032* | 0.98 ± 1.18 | 1.41 ± 1.32 | 0.227* |
| CD4+ CM (%) | 8.30 ± 2.87 | 9.69 ± 4.54 | 0.448* | 8.39 ± 2.93 | 6.80 ± 3.45 | 0.162† |
| CD4+ EM (%) | 0.79 ± 1.16 | 0.66 ± 0.66 | 0.863* | 0.70 ± 0.55 | 1.29 ± 1.47 | 0.370* |
| CD8+ CM (%) | 2.79 ± 1.20 | 3.26 ± 1.56 | 0.878† | 3.26 ± 1.97 | 2.77 ± 1.50 | 0.617* |
| CD8+ EM (%) | 7.14 ± 4.80 | 6.35 ± 4.14 | 0.334* | 10.52 ± 6.96 | 12.51 ± 8.66 | 0.605* |
| ActB (%) | 0.12 ± 0.10 | 0.08 ± 0.09 | 0.138* | 0.13 ± 0.22 | 0.12 ± 0.17 | 0.822* |
| Treg (%) | 0.81 ± 0.69 | 0.92 ± 0.97 | 0.918* | 0.80 ± 0.48 | 0.67 ± 0.54 | 0.546* |
Mann-Whitney U test (two-tailed),
Independent T-test (two-tailed).
AH, aqueous humor; DS, donor pig-specific; αGal, galactose-alpha-1,3-galactose; IFN, interferon; CM, central memory T cells; EM, effector memory T cells; ActB, activated B cells; Treg, regulatory T cells
Entire rejection group includes NHPs whose grafts were rejected within 6 months.
Last FU; last follow-up; examination performed during the rejection period before sacrifice in the entire rejection group and at month 6 in the survival group.
AH C3a assay was not performed at week 2 to avoid possible damage to the graft in the early postoperative period.
Table 5.
Values of biomarker candidates at postoperative week 4 in the late rejection and survival groups.
| Biomarkers | Week 4 |
Last FU |
||||
|---|---|---|---|---|---|---|
| Late rejection (n = 12) |
Survival (n = 18) | P | Late rejection (n = 12) |
Survival (n = 18) | P | |
| Graft score | 0.92 ± 1.56 | 0.22 ± 0.73 | 0.122* | 6.25 ± 0.87 | 0.11 ± 0.47 | < 0.001* |
| AH C3a (ng/mL) | 16.56 ± 8.87 | 6.25 ± 2.82 | 0.001* | 22.84 ± 5.87 | 6.50 ± 5.69 | < 0.001* |
| Plasma C3a (ng/mL) | 136.58 ± 38.23 | 134.51 ± 90.86 | 0.122* | 159.75 ± 46.15 | 124.06 ± 52.63 | 0.067† |
| DS IgG (MFI) | 504.96 ± 684.17 | 337.04 ± 460.07 | 0.783* | 585.91 ± 519.57 | 240.83 ± 439.75 | 0.001* |
| DS IgM (MFI) | 182.66 ± 170.71 | 237.86 ± 105.88 | 0.477† | 188.64 ± 164.677 | 245.18 ± 144.93 | 0.290* |
| Anti-αGal IgG (AU/mL) | 112.82 ± 104.00 | 69.10 ± 81.70 | 0.057* | 92.39 ± 115.24 | 51.99 ± 111.92 | 0.065* |
| Anti-αGal IgM (AU/mL) | 337.70 ± 249.59 | 463.41 ± 320.68 | 0.138* | 331.62 ± 246.16 | 349.73 ± 171.86 | 0.374* |
| CD4+IFNγ+ (cells/mm3) | 18.43 ± 29.32 | 9.03 ± 10.49 | 0.374* | 7.58 ± 12.89 | 15.93 ± 24.74 | 0.268* |
| CD8+IFNγ+ (cells/mm3) | 60.23 ± 72.74 | 51.46 ± 68.79 | 0.352* | 32.75 ± 30.14 | 58.08 ± 52.49 | 0.144* |
| CD4+CM (cells/mm3) | 398.93 ± 187.66 | 305.11 ± 173.46 | 0.171† | 357.83 ± 150.53 | 268.24 ± 125.61 | 0.088† |
| CD4+EM (cells/mm3) | 79.95 ± 157.66 | 36.55 ± 45.12 | 0.498* | 38.67 ± 38.15 | 71.96 ± 107.06 | 0.816* |
| CD8+CM (cells/mm3) | 140.89 ± 66.14 | 105.92 ± 54.43 | 0.125† | 140.67 ± 92.68 | 109.26 ± 56.17 | 0.256† |
| CD8+EM (cells/mm3) | 557.01 ± 453.75 | 390.70 ± 305.89 | 0.253* | 504.17 ± 567.70 | 515.40 ± 430.83 | 0.719* |
| ActB (cells/mm3) | 3.74 ± 3.46 | 2.82 ± 4.04 | 0.175* | 4.42 ± 5.27 | 4.32 ± 5.68 | 0.966* |
| Treg (cells/mm3) | 41.51 ± 44.26 | 42.58 ± 56.44 | 0.949* | 29.25 ± 18.40 | 21.80 ± 16.84 | 0.182* |
| CD4+IFNγ+ (%) | 0.45 ± 0.93 | 0.25 ± 0.26 | 0.933* | 0.27 ± 0.66 | 0.29 ± 0.33 | 0.189* |
| CD8+IFNγ+ (%) | 1.23 ± 1.55 | 1.23 ± 1.50 | 0.657* | 0.81 ± 1.08 | 1.41 ± 1.32 | 0.138* |
| CD4+ CM (%) | 8.21 ± 2.70 | 8.82 ± 3.55 | 0.866* | 8.38 ± 3.05 | 6.80 ± 3.45 | 0.209† |
| CD4+ EM (%) | 1.43 ± 2.30 | 1.02 ± 0.95 | 0.672* | 0.83 ± 0.58 | 1.29 ± 1.47 | 0.703* |
| CD8+ CM (%) | 3.06 ± 1.53 | 3.23 ± 1.23 | 0.611* | 3.39 ± 2.22 | 2.77 ± 1.50 | 0.366† |
| CD8+ EM (%) | 11.18 ± 6.52 | 11.12 ± 5.55 | 0.882* | 10.57 ± 7.07 | 12.51 ± 8.66 | 0.582* |
| ActB (%) | 0.08 ± 0.08 | 0.10 ± 0.12 | 0.703* | 0.09 ± 0.09 | 0.12 ± 0.17 | 0.916* |
| Treg (%) | 0.89 ± 0.78 | 1.25 ± 1.16 | 0.421* | 0.82 ± 0.52 | 0.67 ± 0.54 | 0.512* |
Mann-Whitney U test (two-tailed),
Independent T-test (two-tailed).
Data are presented as mean ± SD.
AH, aqueous humor; DS, donor pig-specific; αGal, galactose-alpha-1,3-galactose; IFN, interferon; CM, central memory T cells; EM, effector memory T cells; ActB, activated B cells; Treg, regulatory T cells
Late rejection group includes NHPs whose grafts were rejected at > week 4 up to month 6.
Last FU; last follow-up; examination performed during the rejection period before sacrifice in the entire rejection group and at month 6 in the survival group.
The graft score at week 4 was not different between the groups (0.92 ± 1.56 and 0.22 ± 0.73, respectively; P = 0.122; Table 5). The AH C3a concentration at week 4 was significantly higher in the rejection group (16.56 ± 8.87 ng/mL) than in the survival group (6.25 ± 2.82 ng/mL; P = 0.001), and the other biomarker candidates did not differ between the groups. At the last follow-up, the AH C3a and DS IgG concentrations were significantly higher in the late rejection group than in the survival group.
In subgroup analysis, the level of DS IgM was higher in the WT xenografted NHPs than in the GTKO xenografted NHPs throughout the follow-up. However, the DS IgG level was significantly higher in the WT xenografted NHPs at week 2 than in the GTKO xenografted NHPs without baseline differences (Table 6 and Supplementary Table S1). Excluding GTKO xenografted NHPs, no significant differences in anti-αGal and DS Abs were found between the rejection and the survival groups (Supplementary Table S2).
Table 6.
Summary of subgroup analysis indicating differences in DS IgG and IgM levels in the rejection group according to donor pig type. The DS IgM level was higher in the WT xenografted NHPs than in the GTKO xenografted NHPs throughout the follow-up, which was not clinically relevant. However, DS IgG was significantly higher in the WT xenografted NHPs at week 2 than in GTKO xenografted NHPs without baseline differences, suggesting a possible association between the DS IgG level and rejection in WT xenografted NHPs.
| Biomarkers | Baseline |
Week 2 |
Week 4 |
Last FU |
||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| WT (n = 11) | GTKO (n = 5) | P* | WT (n = 11) | GTKO (n = 5) | P* | WT (n = 7) | GTKO (n = 5) | P* | WT (n = 11) | GTKO (n = 5) | P* | |
| DS IgG (MFI) | 234.73 ± 250.48 | 27.50 ± 4.36 | 0.157 | 915.14 ± 824.28 | 34.50 ± 14.73 | 0.004 | 694.51 ± 792.01 | 239.60 ± 442.05 | 0.062 | 1219.91 ± 1436.50 | 860.25 ± 716.81 | 0.896 |
| DS IgM (MFI) | 298.49 ± 347.61 | 14.75 ± 5.19 | 0.005 | 400.26 ± 598.23 | 23.00 ± 16.04 | 0.004 | 275.11 ± 145.20 | 20.88 ± 11.56 | 0.008 | 334.27 ± 284.22 | 26.25 ± 11.15 | 0.004 |
| Anti-αGal IgG (AU/mL) | 59.11 ± 79.90 | 197.40 ± 359.30 | 0.610 | 274.93 ± 257.86 | 146.00 ± 224.40 | 0.157 | 87.62 ± 68.43 | 148.10 ± 141.60 | 0.223 | 539.51 ± 776.69 | 46.60 ± 52.78 | 0.079 |
| Anti-αGal IgM (AU/mL) | 290.84 ± 286.04 | 223.40 ± 115.11 | 0.955 | 485.14 ± 386.37 | 265.60 ± 116.10 | 0.336 | 382.91 ± 318.97 | 274.40 ± 100.58 | 0.935 | 380.68 ± 242.14 | 293.20 ± 168.56 | 0.533 |
Mann-Whitney U test (two-tailed)
DS, donor pig-specific; WT, wild type; GTKO, α-1,3-galactosyltransferase gene knockout; NHPs, non-human primates; αGal, galactose-alpha-1,3-galactose; IFN, interferon
Last FU; last follow-up; examination performed during the rejection period before sacrifice in the entire rejection group and at month 6 in the survival group.
3.2. Predictability of presumptive biomarkers for graft rejection within 6 months
The CD8+IFNγ+ cells at week 2 and AH C3a at week 4 were presumptive biomarkers, which showed significant differences between the rejection and survival groups. The predictive abilities of these biomarkers were assessed (Figure 1).
Figure 1.
Receiver operator characteristic curve analysis of CD8+IFNγ+ at week 2 (A) and aqueous humor C3a at week 4 (B) for predicting graft rejection. (A) The area under the curve (AUC) of the CD8+IFNγ+ cells at week 2 was 0.715 (both concentration and percentage), indicating acceptable discrimination ability. The CD8+IFNγ+ cell concentration of 47.15 cells/mm3 (sensitivity, 44%; and specificity, 94%) and percentage of 0.56% (sensitivity, 69%; and specificity, 78%) were the best cut-off values. (B) The AUC of the AH C3a at week 4 was 0.847, indicating excellent discrimination ability. The AH C3a level of 14.785 ng/mL (sensitivity, 0.58%; and specificity, 100%) represented the best cut-off value. Positive and negative predictive values of AH C3a of 14.785 ng/mL were 1.00 and 0.78, respectively. Round dot denotes the optimal cut-off value. AUC, area under the curve; PPV, positive predictive value; NPV, negative predictive value
The AUC of CD8+IFNγ+ cells at week 2 (both the concentration and percentage: 0.715; P = 0.032) showed acceptable discrimination ability for predicting rejection. The concentration of CD8+IFNγ+ cells estimated at 47.15 cells/mm3 (sensitivity, 44%; and specificity, 94%) and the percentage of 0.56% (sensitivity, 69%; and specificity, 78%) represented the optimal cut-off values. In addition, the AUC of the AH C3a at week 4 (0.847; P = 0.001) showed excellent discrimination ability. AH C3a of 14.785 ng/mL (sensitivity, 58%; and specificity, 100%) was the best cut-off value. The positive and negative predictive values of AH C3a level of 14.785 ng/mL were 1.00 and 0.78, respectively, which indicates that AH C3a concentration > 14.785 ng/mL at postoperative week 4 predicted rejection with a probability of 100%. Sensitivity, specificity, and positive and negative predictive values for each predictive biomarker are described in Supplementary Table S3.
4. Discussion
Corneal xenograft rejection is mediated by both innate and adaptive immune systems. The innate immune response is immediate, while the adapted immune response occurs within several days or weeks.27 As shown in Tables 1 and 2, rejection in corneal xenotransplantation occurred frequently between months 1 and 3. The graft scores were similar between the two groups at weeks 2 and 4, which indicates that changes in predictive biomarker levels precede corneal morphological changes during the rejection process. This finding suggests the clinical relevance of predictive biomarkers for the detection of rejection earlier than slit-lamp microscopy. Therefore, our study showed that the 2 or 4-week predictive biomarker profiles may facilitate early intervention against rejection. In this study, the levels of CD8+IFNγ+ cells at week 2 and AH C3a at week 4 were significantly higher in the rejection group than in the survival group and showed acceptable or excellent discrimination abilities for predicting rejection within 6 months.
In contrast to solid organ transplantation, corneal graft rejection can be detected by slit lamp examination. However, at early stages of immune reaction, the cornea may retain transparency, which may contribute to detection failure of early rejection.28 Corneal edema can be reversed upon early detection of rejection and appropriate management before irreversible graft failure occurs.29 In this regard, the changes in CD8+IFNγ+ cells may represent a key 2-week biomarker for the early detection of rejection. At the last follow-up, no systemic differences in CD8+IFNγ+ cells were found, which may be explained by the localization of cells in the cornea, a finding supported by previous studies showing infiltration of CD8+ cells in rejected grafts.1,2,4,30
AH complement activation is related to both innate and adaptive immunity.31 Our previous studies indicated the presence of C3c deposits as well as high levels of AH C3a in NHPs with rejected grafts, but rarely in NHPs with surviving grafts.1–4 The combined data suggest that AH complement is a critical factor for rejection. The AH C3a assay was performed at postoperative week 4 to avoid possible graft damage in the early period. Therefore, further studies are needed to investigate the potential role of AH C3a as a 2-week biomarker.
We are planning a clinical trial of corneal xenotransplantation.32 The results obtained in this study will be used as a standard of reference to predict rejection in the clinical trial. In particular, our results indicated that AH C3a is a potentially critical biomarker with a positive predictive value of 1.0 at the optimal cut-off value. In our study, no complications occurred during AH collection,1–4 which is considered as a routine procedure for patients undergoing PCR testing for virus,33 and can be performed safely with adequate precaution.34
DS IgG and anti-αGal IgG were not significant predictors of rejection. In WT pig-to-NHPs corneal xenotransplantation, high levels of αGal epitope or IgG deposits are present in the rejected graft.1–3 Therefore, subgroup analysis was performed to determine whether the inclusion of GTKO porcine corneal grafts in NHPs affected the changes in DS or anti-αGal Abs as biomarkers (Table 6 and Supplementary Tables S1 and S2). The level of DS IgM was higher in WT xenografted NHPs than in GTKO xenografted NHPs during the follow-up, which was not clinically relevant. However, DS IgG was significantly higher in WT xenografted NHPs at week 2 than in GTKO xenografted NHPs without significant baseline differences, suggesting a possible association between the DS IgG level and rejection in WT xenografted NHPs. As shown in Supplementary Table S2, subgroup analysis was performed after excluding GTKO xenografted NHPs from the rejection groups (11 in entire rejection / 7 in late rejection), because the survival group did not include GTKO xenografted NHPs. Although no significant differences in anti-αGal and DS Abs were found between the rejection and survival groups, we observed changes in DS IgG at week 2 in the rejection group, suggesting that the inclusion of GTKO xenografted NHPs might alter the DS IgG biomarker levels. Therefore, our study limitation related to inclusion of both WT and GTKO donor grafts. Another limitation involved inclusion of NHPs under various immunosuppression regimens. Heterogeneous immunosuppressants exhibit varied effects on the immune response. Further biomarker studies including homogeneous optimal donors and immunosuppressant types are needed.
In conclusion, CD8+IFNγ+ cells at week 2 and AH C3a concentrations at week 4 showed potential as useful biomarkers for predicting graft rejection in pig-to-NHP corneal xenotransplantation. To the best of our knowledge, this study is the first to report predictive biomarkers for graft rejection in corneal xenotransplantation.
Supplementary Material
Supplementary Figure S1. Representative multi-color flow cytometry gating strategies for IFNγ+ CD4+ or CD8+ T cells (A), CD28+CD95+ central memory T cells, CD28−CD95+ effector memory T cells (B), CD3−CD20+CD28+ activated B cells (C), and CD4+CD25+Foxp3+ regulatory T cells (D).
Acknowledgement
Study was supported by a grant from the Korea Healthcare Technology R&D Project, Ministry for Health & Welfare (Project No. HI13C0954). Anti-CD154 and anti-CD40 antibodies had been provided by the Nonhuman Primate Reagent Resource supported by U.S. National Institutes of Health NIAID contract HHSN272200900037C and grants AI126683 and OD010976.
This study was consulted by a statistician at the Medical Research Collaboration Center of our hospital (http://mrcc.snuh.org; MRCC No. 2019–0006).
Abbreviations:
- Ab
antibody
- αGal
Galalpha1,3Galbeta1,4GlcNAc-R
- AH
aqueous humor
- AU
artificial unit
- AUC
areas under the curve
- C3a
complement activation fragment
- DS
donor pig-specific
- GTKO
α−1,3-galactosyltransferase gene knockout
- IL
interleukin
- IVIG
intravenous immunoglobulin
- IFN
interferon
- MFI
mean fluorescence intensity
- NHP
non-human primate
- ROC
receiver operating characteristic
- SNU
Seoul National University
- WT
wild type
Footnotes
Conflict of Interest Disclosures
The authors declare no competing financial interests.
References
- 1.Kim J, Kim DH, Choi HJ, et al. Anti-CD40 antibody-mediated costimulation blockade promotes long-term survival of deep-lamellar porcine corneal grafts in non-human primates. Xenotransplantation. 2017;24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Choi HJ, Lee JJ, Kim DH, et al. Blockade of CD40-CD154 costimulatory pathway promotes long-term survival of full-thickness porcine corneal grafts in nonhuman primates: clinically applicable xenocorneal transplantation. Am J Transplant. 2015;15:628–641. [DOI] [PubMed] [Google Scholar]
- 3.Kim J, Choi SH, Lee HJ, et al. Comparative efficacy of anti-CD40 antibody-mediated costimulation blockade on long-term survival of full-thickness porcine corneal grafts in nonhuman primates. Am J Transplant. 2018;18:2330–2341. [DOI] [PubMed] [Google Scholar]
- 4.Choi HJ, Kim MK, Lee HJ, et al. Efficacy of pig-to-rhesus lamellar corneal xenotransplantation. Invest Ophthalmol Vis Sci. 2011;52:6643–6650. [DOI] [PubMed] [Google Scholar]
- 5.Hara H, Cooper DK. The immunology of corneal xenotransplantation: a review of the literature. Xenotransplantation. 2010;17:338–349. [DOI] [PubMed] [Google Scholar]
- 6.Lee HI, Kim MK, Oh JY, et al. Gal alpha(1–3)Gal expression of the cornea in vitro, in vivo and in xenotransplantation. Xenotransplantation. 2007;14:612–618. [DOI] [PubMed] [Google Scholar]
- 7.Hara H, Koike N, Long C, et al. Initial in vitro investigation of the human immune response to corneal cells from genetically engineered pigs. Invest Ophthalmol Vis Sci. 2011;52:5278–5286. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Niederkorn JY, Larkin DF. Immune privilege of corneal allografts. Ocul Immunol Inflamm. 2010;18:162–171. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Hori J, Vega JL, Masli S. Review of ocular immune privilege in the year 2010: modifying the immune privilege of the eye. Ocul Immunol Inflamm. 2010;18:325–333. [DOI] [PubMed] [Google Scholar]
- 10.Kim MK, Hara H. Current status of corneal xenotransplantation. Int J Surg. 2015;23:255–260. [DOI] [PubMed] [Google Scholar]
- 11.Jabbehdari S, Rafii AB, Yazdanpanah G, Hamrah P, Holland EJ, Djalilian AR. Update on the Management of High-Risk Penetrating Keratoplasty. Curr Ophthalmol Rep. 2017;5:38–48. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Flynn TH, Ohbayashi M, Ikeda Y, Ono SJ, Larkin DF. Effect of allergic conjunctival inflammation on the allogeneic response to donor cornea. Invest Ophthalmol Vis Sci. 2007;48:4044–4049. [DOI] [PubMed] [Google Scholar]
- 13.Niederkorn JY. Immune mechanisms of corneal allograft rejection. Curr Eye Res. 2007;32:1005–1016. [DOI] [PubMed] [Google Scholar]
- 14.Amouzegar A, Chauhan SK, Dana R. Alloimmunity and Tolerance in Corneal Transplantation. J Immunol. 2016;196:3983–3991. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Lapp T, Zaher SS, Haas CT, et al. Identification of Therapeutic Targets of Inflammatory Monocyte Recruitment to Modulate the Allogeneic Injury to Donor Cornea. Invest Ophthalmol Vis Sci. 2015;56:7250–7259. [DOI] [PubMed] [Google Scholar]
- 16.Funding M, Hansen TK, Gjedsted J, Ehlers N. Simultaneous quantification of 17 immune mediators in aqueous humour from patients with corneal rejection. Acta Ophthalmol Scand. 2006;84:759–765. [DOI] [PubMed] [Google Scholar]
- 17.Maier P, Heizmann U, Bohringer D, Kern Y, Reinhard T. Distinct cytokine pattern in aqueous humor during immune reactions following penetrating keratoplasty. Mol Vis. 2010;16:53–60. [PMC free article] [PubMed] [Google Scholar]
- 18.Inomata T, Hua J, Di Zazzo A, Dana R. Impaired Function of Peripherally Induced Regulatory T Cells in Hosts at High Risk of Graft Rejection. Sci Rep. 2016;6:39924. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Dong X, Hara H, Wang Y, et al. Initial study of alpha1,3-galactosyltransferase gene-knockout/CD46 pig full-thickness corneal xenografts in rhesus monkeys. Xenotransplantation. 2017;24. [DOI] [PubMed] [Google Scholar]
- 20.Zhiqiang P, Cun S, Ying J, Ningli W, Li W. WZS-pig is a potential donor alternative in corneal xenotransplantation. Xenotransplantation. 2007;14:603–611. [DOI] [PubMed] [Google Scholar]
- 21.Jie Y, Liu L, Pan Z, Wang L. Survival of pig-to-rhesus corneal xenografts prolonged by prior donor bone marrow transplantation. Mol Med Rep. 2013;7:869–874. [DOI] [PubMed] [Google Scholar]
- 22.Choi SH, Yoon CH, Lee HJ, et al. Long-term safety outcome of systemic immunosuppression in pig-to-nonhuman primate corneal xenotransplantation. Xenotransplantation. 2018;25:e12442. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Hosmer DW, Lemeshow S. Applied Logistic Regression, 2nd Ed. New York: Wiley, 2000. [Google Scholar]
- 24.Kim MK, Choi HJ, Kwon I, et al. The International Xenotransplantation Association consensus statement on conditions for undertaking clinical trials of xenocorneal transplantation. Xenotransplantation. 2014;21:420–430. [DOI] [PubMed] [Google Scholar]
- 25.Kang HJ, Lee H, Park EM, et al. Increase in anti-Gal IgM level is associated with early graft failure in intraportal porcine islet xenotransplantation. Ann Lab Med. 2015;35:611–617. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Youden WJ. Index for rating diagnostic tests. Cancer. 1950;3:32–35. [DOI] [PubMed] [Google Scholar]
- 27.Platt JL. New directions for organ transplantation. Nature. 1998;392:11–17. [DOI] [PubMed] [Google Scholar]
- 28.Armitage WJ, Dick AD, Bourne WM. Predicting endothelial cell loss and long-term corneal graft survival. Invest Ophthalmol Vis Sci. 2003;44:3326–3331. [DOI] [PubMed] [Google Scholar]
- 29.Larkin DF. Corneal allograft rejection. Br J Ophthalmol. 1994;78:649–652. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Oh JY, Kim MK, Ko JH, et al. Acute cell-mediated rejection in orthotopic pig-to-mouse corneal xenotransplantation. Xenotransplantation. 2009;16:74–82. [DOI] [PubMed] [Google Scholar]
- 31.Asgari E, Zhou W, Sacks S. Complement in organ transplantation. Current opinion in organ transplantation. 2010;15:486–491. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Choi HJ, Yoon CH, Hyon JY, et al. Protocol for the first clinical trial to investigate safety and efficacy of corneal xenotransplantation in patients with corneal opacity, corneal perforation, or impending corneal perforation. Xenotransplantation. 2018:e12446. [DOI] [PubMed] [Google Scholar]
- 33.Koizumi N, Yamasaki K, Kawasaki S, et al. Cytomegalovirus in aqueous humor from an eye with corneal endotheliitis. Am J Ophthalmol. 2006;141:564–565. [DOI] [PubMed] [Google Scholar]
- 34.Kitazawa K, Sotozono C, Koizumi N, et al. Safety of anterior chamber paracentesis using a 30-gauge needle integrated with a specially designed disposable pipette. Br J Ophthalmol. 2017;101:548–550. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary Figure S1. Representative multi-color flow cytometry gating strategies for IFNγ+ CD4+ or CD8+ T cells (A), CD28+CD95+ central memory T cells, CD28−CD95+ effector memory T cells (B), CD3−CD20+CD28+ activated B cells (C), and CD4+CD25+Foxp3+ regulatory T cells (D).