All treatment protocols were approved by the Istituto Superiore di Sanità [ISS, Rome, authorization number no. 45253(06)-PRE.21-882 and no. 28689(13)321-1223] and by Agenzia Italiana del Farmaco (AIFA) on October 10, 2007 and September 30, 2013, respectively, and by the Institutional Review Board of the Ospedali Riuniti/Azienda Ospedaliera Papa Giovanni XXIII of Bergamo (authorization no. 352, March 18, 2008 and no. 110/13, November 6, 2013). The study is registered with ClinicalTrials.gov (NCT00752479 and NCT02012153). Written informed consent was obtained from all recipients and living donors in accordance with the Declaration of Helsinki.

Patients with ESRD were enrolled in a phase 1, single-center, open-label pilot study conducted at the Ospedale Papa Giovanni XXIII, Bergamo, Italy, that aimed primarily to characterize the safety and feasibility of peri-transplant infusion of ex vivo expanded, autologous bone marrow-derived MSC in living-related donor kidney transplant recipients. Patients #1 and #2 adopted the initial study protocol (Protocol 1, see below), with MSC given intravenously 7 days posttransplantation (Figure 1) (11). Since transient acute renal insufficiency due to engraftment syndrome occurred after cell infusion, the protocol was revised (Protocol 2, see below) in the next patients, #3 and #4, who received pretransplant (day −1) MSC infusion (Figure 1) (12). Moreover, patients #1 and #2 received induction therapy with basiliximab and low-dose rabbit anti-thymocyte globulin (RATG), whereas patients #3 and #4 were given RATG alone to avoid the possible negative impact of basiliximab on Treg development and function (11, 12). As controls for patients #1 and #2, six kidney transplant recipients from living-related (n = 3) and deceased (n = 3) donors given basiliximab/low-dose RATG as induction therapy were considered (11). For patients #3 and #4, an additional control group (n = 6) of historical deceased-donor kidney transplant recipients receiving the same low-dose RATG induction therapy alone was studied, since in our living-donor transplant program only dual induction therapy with basiliximab and low-dose RATG is adopted. In all MSC-treated and control patients, maintenance immunosuppressive therapy comprised low-dose ciclosporin (CsA) and low-dose mycophenolate mofetil (MMF) (Figure 1).

MSC-treated patients and the two groups of control kidney transplant recipients were comparable for gender, cause of renal failure, and donor age (Table 1).

Four months before transplantation patients underwent sterna bone marrow aspiration under local anesthesia. MSCs were isolated and ex vivo expanded according to Good-Manufacturing Practice procedures (18, 19). On day 7 after kidney transplant, autologous MSCs were administered intravenously (1.7 × 10⁶ and 2.0 × 10⁶ cells/kg body weight, respectively) after premedication with chlorphenamine and acetaminophen. Patients received induction regimen with basiliximab (20 mg intravenously pretransplant and on day 4 posttransplant) and low-dose RATG infusion (thymoglobulin, 0.5 mg/kg, daily from day 0 to day 6 posttransplant) as per center practice (20). Maintenance immunosuppression was with CsA (target trough blood levels of 300–400 ng/mL up to day 7 postsurgery, and 100–150 ng/mL at month 5 posttransplantation), MMF, and steroids. Five hundred milligrams of methylprednisolone were administered before the first RATG infusion to minimize the possible cytokine release reaction related to the antibodies, and continued for two more days posttransplant (250 and 125 mg, respectively). Subsequently, oral prednisone (75 mg) was administered, which was progressively tapered and discontinued after day 7 postsurgery.

Four to six months before transplantation patients underwent right posterior superior iliac crest aspiration under local anesthesia. MSCs were isolated and ex vivo expanded according to Good-Manufacturing Practice procedures. The day before kidney transplantation (day −1) autologous MSCs were administered intravenously (2.0 × 10⁶ cells/kg body weight) after premedication with chlorphenamine and acetaminophen. Patients received induction therapy with low-dose RATG infusion (0.5 mg/kg, daily from day 0 to day 6 posttransplant). Maintenance immunosuppression was with CsA (target trough blood levels of 300–400 ng/mL up to day 7 postsurgery, and 100–150 ng/mL at month 5 posttransplantation), MMF and steroids. Five hundred milligrams of methylprednisolone were administered before the first RATG infusion to minimize the possible cytokine release reaction related to the antibodies, and continued for two more days posttransplant (250 and 125 mg, respectively). Subsequently, oral prednisone (75 mg) was administered, which was progressively tapered and discontinued after day 7 postsurgery.

In all MSC-treated patients clinical and hematochemical parameters, acute graft rejection episodes and incidence and severity of adverse events have been prospectively evaluated up to their last clinical follow-up, spanning from 5 to 7 years after transplantation. Similar parameters were retrospectively evaluated in control patients during the same follow-up period after kidney transplantation. Both MSC-treated and control transplant recipients had periodically monitoring of renal graft function as glomerular filtration rate (GFR), measured by the plasma clearance of iohexol (21, 22), every 6 months after kidney transplantation. At the same time points, 24 h urinary samples were collected to evaluate urinary protein excretion. Immunophenotyping of peripheral blood lymphocyte subpopulations, ex vivo functional immunological assays to assess donor-specific hyporesponsiveness and monitoring of donor-specific anti-human leukocyte antigen (HLA) antibodies were sequentially evaluated in MSC-treated patients and compared with those in controls. Donor cells for functional immunological studies were available for six control patients given basiliximab/low-dose RATG, but for only one receiving low-dose RATG alone.

Mesenchymal stromal cells were processed and cultured according to Good-Manufacturing-Practice procedures (Centre of Cellular Therapy “G. Lanzani” Azienda Socio Sanitaria Territoriale Papa Giovanni XXIII, Bergamo, Italy, authorization no. aM.-189/2008 AIFA) (18, 19). Cells were classified as MSC based on their ability to differentiate into bone, fat and cartilage, and by flow cytometric analysis (positive for CD44, CD29, CD73, HLA-ABC, CD90, and CD105, but negative for CD14, CD34, CD45, and HLA-DR) responding to defined criteria for MSC established by the International Society for Cellular Therapy (23). The final product was characterized with respect to viability, purity, and therapeutic potential.

Peripheral blood mononuclear cells (PBMCs) were collected pretransplant, at 12 months posttransplant and at yearly intervals thereafter. PBMCs were stained with fluorochrome-conjugated murine monoclonal antibodies against human CD3 (clone HIT3a), CD4 (clone RPA-T4), CD8 (clone RPA-T8), CD45RA (clone HIT100), CD45RO (clone UCHL1), CD25 (clone M-A251), CD127 (clone HIL-7R-M21), CD19 (clone HIB19), CD27 (clone M-T271), CD38 (clone HIT2), IgD (clone IA6-2), CD24 (clone ML5), and FoxP3 (clone 236A/E7) (all from BD Bioscience). Multicolor flow cytometry was used to identify T- and B-cell subsets with standard technique and equipment (FACSAria, BD Bioscience) (24, 25). At the same time points, alloimmune response against donor and third-party antigens was assessed by IFNγ ELISpot assay and cell-mediated lympholysis (26).

For NK and monocyte subset analysis, 10⁶ PBMC, resuspended in 50 µL Brilliant Buffer (BD Bioscience), were labeled with a mix of antibodies for 30 min. The mix antibodies contained BUV395 mouse anti-human CD45 (clone HI30), BV510 mouse anti-human CD3 (clone HIT3a), APC-R700 mouse anti-human CD27 (clone M-T271), BB700 mouse anti-human CD16 (clone B73.1), BV650 mouse anti-human CD56 (clone B159), BV711 mouse anti-human CD11b (clone D12), BUV737 mouse anti-human CD14 (clone M5E2), PE-CF594 mouse anti-human CD64 (clone 10.1), BV605 mouse anti-human HLA-DR (clone G46-6), and PE-Cy5 mouse anti-human CD33 (clone WM-53), all antibodies were purchased from BD Bioscience. At the end of incubation, PBMCs were washed with Stain Buffer (BD Bioscience). Cells were suspended in Stain buffer, and 10 min prior the analysis, Cell Viability Solution (BD) was added to exclude death cells. Cells were acquired by FACS Fortessa X-20 (BD Bioscience) and analyzed with FlowJo software.

Serum samples collected before surgery and at yearly intervals after transplantation from MSC recipients and control patients given basiliximab/low-dose RATG induction therapy were tested for class I and class II anti-HLA antibodies using LABScreen Mixed kit (One Lambda, Canoga Park, CA, USA) (27). Positive tests were quantified by Single-Antigen Beads (One Lambda). A mean fluorescence intensity (MFI) threshold of 2,000 was used to identify a donor-specific anti-HLA antibody (DSA). The tests were carried out according to the manufacturer’s instructions and the analysis was performed with One Lambda software (HLA Visual Version 2.2).

Individual GFR slopes (GFR changes over time) were calculated by linear regression analysis. Difference in GFR slopes between MSC recipients and control groups was assessed by Mann–Whitney U test. Variations over time in counts and percentages of T- and B-cell subsets, frequency of IFNγ-producing T cells and cell-mediated cytolysis in either control group not given MSCs were assessed by repeated-measures ANOVA, whereas differences between the two control groups were evaluated by ANOVA. All calculations were made using MedCalc 12.3.0.0 (MedCalc Software, Mariakerke, Belgium). P value below 0.05 was considered as statistically significant.

Table 1 shows demographics and follow-up of MSC-treated patients and controls. Both patients who received MSC infusion at day 7 after transplantation (#1 and #2) developed transient renal dysfunction shortly after cell administration. A graft biopsy few days after MSC infusion in patient #2 excluded acute graft rejection, but showed focal inflammatory infiltrate, mostly granulocytes (11). Following corticosteroid administration, graft function improved and the serum creatinine level stabilized throughout the entire follow-up (7 years) in both patients (patient #1, serum creatinine: 1.83–2.27 mg/dL; patient #2, serum creatinine: 2.38–2.21 mg/dL) (Table 1; Figure S1 in Supplementary Material). At month 77 posttransplantation, proteinuria became detectable in patient #2 and a renal biopsy revealed recurrence of immunoglobulin A nephropathy (Figure S1 in Supplementary Material). In patient #3, in whom pretransplant infusion of MSC proceeded without complications, renal graft function remained normal over the entire 6-year follow-up (serum creatinine: 1.35–0.97 mg/dL) (Table 1; Figure S1 in Supplementary Material). At variance, patient #4, who also received MSC-pretransplant, experienced an acute cellular rejection episode 2 weeks after kidney transplantation. After methylprednisolone pulses, renal function recovered to normal values within 10 days and remained stable until the last available evaluation at 5 years posttransplantation (serum creatinine: 1.37–1.49 mg/dL) (Table 1; Figure S1 in Supplementary Material).

In the control group given induction therapy with basiliximab/low-dose RATG, serum creatinine levels averaged 1.77 ± 0.22 mg/dL 7 years after transplantation (Table 1). Similarly, in the other control group receiving induction therapy with low-dose RATG alone, mean serum creatinine concentrations were 1.94 ± 0.40 mg/dL at year 5 after transplantation, the average time of follow-up for patients #3 and #4 (Table 1). During the observation periods, one patient in both control groups developed acute cellular rejection that resolved with intravenous methylprednisolone pulses. Interestingly, the median rate of measured GFR decline tended to be numerically slower in the first four MSC recipients throughout their respective follow-up than in control patients given basiliximab/low-dose RATG or low-dose RATG alone (−0.278 vs −0.938 mL/min/1.73 m²/year, respectively, P = NS) (Figures 2A,B). Similarly, measured GFR decreased to a significantly lower rate in the first four MSC-treated patients compared with living-related donor kidney transplant recipients in the basiliximab/low-dose RATG control group (n = 3) (−0.278 vs −1.834 mL/min/1.73 m²/year, respectively, P < 0.001).

Very few serious adverse events were reported in MSC-treated patients, and none was deemed to be related to cell infusion. Table 2 focuses on infectious and neoplastic complications. In the cohort of MSC recipients, a total of seven infectious complications were reported during follow-up. Hospitalization was required only for patient #1 at month 23 after transplantation due to an Escherichia coli septic episode originating from the urinary tract (Figure S1 in Supplementary Material). Overall, 20 infectious episodes were recorded in the control group during the long-term observation period, but hospital admission was required for only two patients for cytomegalovirus infection and labyrinthitis at months 2 and 62 posttransplant, respectively (Table 2). One other patient’s initial postsurgery hospitalization was prolonged due to bronchopneumonia (Table 2).

With regard to tumor development, dual benign renal hemangiomas were identified in the right native kidney of patient #2 2 months after transplantation, and dual nephrectomy of the native kidneys was performed (11). A control patient developed squamous papillomas of the uterine cervix 71 months after transplantation and was successfully treated with loop electrosurgical excision procedure (Table 2).

The percentage of circulating memory CD45RO⁺RA⁻CD8⁺ T cells in patient #1 was reduced by 50% at 1-year posttransplant compared with basal value and did not change further until 5 years posttransplant, then returned to pretransplant levels (Figure 3A). In patient #2, conversely, the percentage of memory CD8⁺ T cells did not undergo relevant variations during follow-up. In control patients given basiliximab/low-dose RATG induction therapy, the percentage of memory CD8⁺ T cells remained comparable to pretransplant values during the 7-year observation period (Figure 3A). In MSC recipients #3 and #4, too, the percentage of memory CD8⁺ T cells was reduced by 50% at 1-year posttransplant, compared with pretransplant levels (Figure 3B). While in patient #4 memory CD8⁺ T cell percentage remained lower than basal values until 5-year posttransplant, in patient # 3 the percentage of these cells in the circulation was very low during the entire follow-up period (Figure 3B). At variance, in control patients given induction therapy with low-dose RATG alone, the percentage of memory CD8⁺ T cells significantly increased at 1-year posttransplant compared with pretransplant values, returning to basal levels thereafter (Figure 3B). FACS gating strategy and representative dot plots of memory CD45RO⁺RA⁻CD8⁺ T cells are shown in Figure S2 in Supplementary Material. A very similar trend was found when considering absolute counts of circulating memory CD8⁺ T cells in MSC-treated and control patients (Figure S3 in Supplementary Material).

In the peripheral blood, the percentage of Treg (defined as percentage of Foxp3⁺CD127⁻ on CD4⁺CD25^high T cells, see Figure S2 in Supplementary Material) in patients #1 and #2 decreased slightly at 1-year posttransplant compared with the pretransplant value, then further declined progressively in patient #1, while it remained stable in patient #2 (Figure 3C) up to 7 years after transplantation. No significant changes in Treg percentage were documented in control patients given basiliximab/low-dose RATG during follow-up (Figure 3C). Interestingly, in patients #3 and #4, the Treg percentage remained similar to the pretransplant value during the first 3 years posttransplant, with a trend to further increase thereafter, particularly in patient #3 (Figure 3D). This was at variance with control patients given induction therapy with low-dose RATG alone, where the Treg percentage approximated a mean value of 34–45% throughout the entire follow-up (Figure 3D).

When we considered the ratio of Treg/memory CD8⁺ T cells, an increase was observed in patient #1 at 1-year posttransplant, and then this ratio remained relatively high at least until 4 years posttransplant (Figure S1 in Supplementary Material). In patient #2, the ratio did not change during follow-up, with a profile very similar to that observed in control patients given basiliximab/low-dose RATG induction therapy (Figure 3E). On the other hand, the Treg/memory CD8⁺ T cell ratio increased progressively in patient #3, starting from 1-year posttransplant, reaching very high levels in years 4–6 follow-up (Figure 3F; Figure S1 in Supplementary Material). Instead, in patient #4, the ratio did not undergo relevant variations, remaining at around the mean value of 2, similar to control patients given low-dose RATG induction therapy alone (Figure 3F).

Next, we analyzed the phenotype of peripheral B cells according to IgD/CD27 classification (28–30). We found that B cells with a naïve phenotype progressively expanded in the peripheral blood of patients #1 and #3 starting from the third and second year posttransplant, respectively, reaching very high cell counts at year 6 after transplantation (Figure 4A; Figure S1 in Supplementary Material). The expansion of naïve B cells was confirmed by the increase in naïve B cell percentages (Figure 4B). In these two patients, the increase in naïve B cell percentages was paralleled by a decrease in switched B cells (Table S1 in Supplementary Material). By contrast, in patients #2 and #4, naïve B cell counts (Figure 4A) and percentages (Figure 4B) did not show appreciable variations during follow-up, remaining at levels similar to those observed pretransplant and in control patients (Figures 4A,B). As shown in Figures 4C,D, the increase in naïve B cell count in patients #1 and #3 was also paralleled by a marked increase in circulating transitional B cells. Indeed, these B cell counts increased progressively starting from the second year posttransplant in patients #1 and #3, reaching, in the long term, 10-fold higher levels than those observed before transplantation (Figure 4C; Figure S1 in Supplementary Material) and in the other MSC-treated patients and in both control groups, in whom transitional B cells remained at low levels during the entire follow-up period (Figure 4C). The increased transitional B cell counts in patients #1 and #3 were also confirmed by the remarkable increase in the percentages of this B cell subpopulation (Figure 4D). FACS gating strategy and representative dot plots of naïve and transitional B cells are shown in Figure S4 in Supplementary Material.

To assess the pro-tolerogenic milieu, MSC-treated and control transplant recipients were monitored for alloimmune response against donor and third-party antigens. In all MSC-treated patients, the frequency of IFNγ-producing memory T cells in response to donor antigens was below the positive threshold of 25 spots/300,000 cells (31, 32) at basal evaluations, and remained below this threshold during the entire follow-up (Figure 5A). However, in control patients, the basal frequency of anti-donor IFNγ-producing memory T cells was below the positive threshold as well, but then tended to increase during long-term follow-up, reaching significantly higher values than pretransplant at years 2 and 3 posttransplant (Figure 5A). The frequency of IFNγ-producing memory T cells in response to third-party cells was more variable (Figure 5B; Figure S5 in Supplementary Material).

In MSC-treated patients, the CD8⁺ T cell-mediated lympholysis of donor cells sharply decreased after transplantation, remaining lower than pretransplant values throughout their respective follow-up, with the exception of patient #2, in whom anti-donor CD8⁺ T cell cytotoxicity resumed basal levels from year 5 posttransplant (Figure 6A; Figure S1 in Supplementary Material). Notably, in control patients, the anti-donor CD8⁺ T cell-mediated lympholysis did not significantly change and remained similar to pretransplant levels during follow-up (Figure 6A). Moreover, in response to third-party cells, the CD8⁺ T cell cytolytic activity increased during the first 3 years posttransplant in patients #1 and #3, and fluctuated in patients #2 and #4 (Figure 6B). In control patients, there were no relevant changes to the CD8⁺ T cell cytolytic response to third-party antigens during the 7 years of follow-up (Figure 6B).

The interactions between MSC and NK cells are very complex (33–35). Up to now, however, no experimental studies or clinical evidence are available on the possible effect of MSC on NK cell phenotype and vice versa in solid organ transplantation.

In our MSC-treated patients, circulating NK cell counts fluctuated in the range 100–200 cells/μL, with no significant variation during the 5-year posttransplant follow-up. Similar NK cell counts were shown in control kidney transplant recipient groups (Figure 7A).

Human NK cells are classified into two different subsets based on their expression of CD56 and CD16 with different ability to mediate cytotoxicity and to release cytokines (36). The population of CD56^dimCD16⁺ NK cells is mostly responsible of cytotoxicity following specific recognition of their targets and is the predominant population in the peripheral blood (36). In our setting, CD56^dimCD16⁺ were positive for CD11b expression and negative for CD27 and were identified as cytotoxic NK cells (Figure 7B). The CD56^bright CD16^−/low NK cells, which are the predominant NK cell subset in peripheral lymphoid organs, are the major producer of cytokines (36). Among the CD56^bright NK cell subset, the expression of CD27 further divided NK with cytotoxic capacity (CD27⁻) and NK with higher capacity to produce cytokines (CD27⁺) (37) (Figure 7B).

The frequencies of CD56^dim NK cell, CD56^bright NK cells either positive or negative for CD27 expression were very similar among MSC-treated patients and kidney transplant recipients given basiliximab/low-dose RATG or only low-dose RATG as induction therapy, and did not undergo significant modifications during the follow-up (Figures 7C–E). Similarly, the frequency of NKT cells (Figures 7B,F) during the 5-year follow-up was comparable between MSC-treated patients and control kidney transplant recipients given RATG induction therapy. Percentages of NKT cells in patients given induction therapy with basiliximab/low-dose RATG were lower compared with MSC-treated patients and patients given induction therapy with low-dose RATG alone at each time point analyzed, reaching statistical significance 3 years after transplantation.

Overall, our analysis showed that MSC cell therapy did not have a major impact on NK cell counts and phenotype and on frequency of NKT cells on the long-term posttransplant.

We also examined the phenotypic profile of circulating monocytes, which are classified on the basis of relative expression of CD14 and CD16 surface markers in classical monocytes (CD14⁺⁺CD16⁻), intermediate monocytes (CD14⁺⁺CD16⁺), and non-classical monocytes (CD14⁺CD16⁺⁺) (38) with different functions. Classical monocytes are phagocytic cells and preferentially express genes involved in angiogenesis, wound healing, and coagulation, intermediate monocytes are pro-inflammatory and with the capability of antigen presentation and T cell activation, while non-classical monocytes display patrolling properties. These subsets expressed high levels of HLA-DR and CD33 and different levels of CD64 (39, 40) (Figure 8A). Very few studies are available on the possible interaction between MSC and monocytes (41, 42). The frequency of classical monocytes in our MSC-treated patients was around 70% during the follow-up and similar to that found in kidney transplant recipients in the two control groups (Figure 8B). The expression of HLA-DR on this monocyte subset ranged around value of MFI of 7,500 in MSC-treated and control transplant patients. HLA-DR expression on classical monocytes remained constant during the follow-up in the three groups of patients (Figure 8C). The frequencies of intermediate and non-classical monocyte were more variable among and between subjects than classical monocytes, with no significant differences during the 5-year follow-up in the three groups of kidney transplant recipients (Figure 8D).

Therefore, MSC infusion in kidney transplant patients did not associate with significant changes of the phenotype of circulating monocytes in the long-term.

All MSC recipients and control patients given basiliximab/low-dose RATG induction therapy were negative for DSA at the time of kidney transplantation. Patient #1 developed de novo class I DSA against B57 (MFI: 5,740) and B58 (MFI: 5,590) antigens 7 years after transplantation (Figure S1 in Supplementary Material). Notably, DSA were never detected in the remaining MSC recipients throughout their respective follow-up. By contrast, three of six control patients were transiently positive for de novo class II DSA reactive to DQ and/or DR antigens during the 7-year observation period. In particular, in a patient DSA were detected at years 4 (anti-DQ6, MFI: 6,800) and 5 (anti-DQ6, MFI: 5,600; anti-DR0404, MFI: 2,730; anti-DR16, MFI: 2,760) after transplantation. A second patient tested positive for DSA at years 2 (anti-DQ6, MFI: 3,000) and 5 (anti-DQ5, MFI: 2,690), while in a third patient DSA were identified in a single occasion 5 years after transplantation (anti-DQ0603, MFI: 3,260).

Patient #3 had a favorable posttransplant clinical course with evidence of sustained pro-tolerogenic immune milieu. In particular, no features of subclinical rejection were observed in the 1-year protocol biopsy (11), and graft function remained stable for 2 years after transplantation (measured GFR: 67.07 mL/min/1.73 m²), without evidence of de novo DSA development. At this point, the patient exhibited a pro-tolerogenic profile characterized by an increased ratio of Treg/memory CD8⁺ T cell percentage and expansion of naïve and transitional B cells in the peripheral blood, along with suppression of donor-specific CD8⁺ T cell cytotoxicity. In light of these clinical and immunological findings, we chose to attempt to wean patient off immunosuppressive therapy while monitoring renal graft function closely. Accordingly, the CsA dose was gradually tapered starting from month 26 and was eventually discontinued at month 73 posttransplant (Figure S1 in Supplementary Material), leaving the patient on low-dose MMF monotherapy since then. No clinical or laboratory evidence of rejection was observed during or after CsA withdrawal. Graft function remained normal and actually improved over time, as demonstrated by measured GFR, which increased by 2.30 mL/min/1.73 m²/year over 6 years of follow-up. De novo DSA were never detected in serum samples, even after CsA was discontinued. For safety reasons, immunosuppressive therapy weaning was not attempted in the other MSC recipients due to either transient acute renal insufficiency or acute cellular rejection early after transplantation, even though signs of immunoregulation were observed in patient #1.

Since patient #4 developed acute cellular rejection 2 weeks after kidney transplantation, to limit the risk of early posttransplant acute rejection the study protocol was modified again for patient #5, who still received MSC before surgery (day −1) but was given both basiliximab and RATG as induction therapy. Patient #5 was a 42-year-old woman on hemodialysis due to ESRD secondary to tubular interstitial nephropathy. She received a renal transplant from her sister, mismatched for two HLA alleles (one mismatch on HLA-B and one on HLA-DR, while HLA-A alleles were coincidental). In this patient, pretransplant infusion of MSC was uneventful. Renal function gradually recovered and normalized within the first two post-operative weeks (serum creatinine: 1.32 mg/dL). Thereafter, graft function remained relatively stable up to 12 months posttransplantation (serum creatinine: 1.52 mg/dL, measured GFR: 52.59 mL/min/1.73 m²; proteinuria: 0.09 g/24 h). At this time, a protocol biopsy showed mild mesangial expansion and patent capillary lumina surrounded by well-preserved tubulointerstitial compartment (Figure 9A).

Immunophenotypic analysis of peripheral blood lymphocytes showed that the proportion of memory CD8⁺ T cells was reduced by almost 70% 1 year after transplantation compared with pretransplant value, whereas Treg percentage did not change over the same period (Figures 9B,C). Thus, the ratio of Treg/memory CD8⁺ T cells markedly increased 1 year following transplant compared with basal level (Figure 9D). Circulating naïve and transitional B cell counts and percentages did not change appreciably 1-year posttransplant compared with pretransplant values (Figures 9E,F). Regarding functional immunological studies, the basal frequency of IFNγ-producing memory T cells in response to donor antigens was lower than the positive threshold of 25 spots/300,000 cells, and tended to reduce 1 year after transplantation (Figure 9G). Instead, the frequency of IFNγ-producing memory T cells in response to third-party cells exceeded the positive threshold at baseline evaluation, but it fell below the above indicated threshold 1-year posttransplant (Figure 9G). At this time, CD8⁺ T cell-mediated lympholysis against donor and third-party antigens decreased compared with pretransplant values (Figure 9H).