Acute and Chronic Vascular Rejection in Nonhuman Primate Kidney Transplantation


* Corresponding author: G. Wieczorek,


A nonhuman primate (NHP) study was designed to evaluate in nonlife-supporting kidney allografts the progression from acute rejection with transplant endarteritis (TXA) to chronic rejection (CR) with sclerosing vasculopathy. Group G1 (n = 6) received high cyclosporine A (CsA) immunosuppression and showed neither TXA nor CR during 90 days post-transplantation. Group G2 (n = 6) received suboptimal CsA immunosuppression and showed severe TXA with graft loss within 46 days (median). Arterial intimal changes included infiltration of macrophages and T lymphocytes (CD3, CD4, CD8) with few myofibroblasts, abundant fibronectin/collagen IV, scant collagens I/III, high rate of cellular proliferation and no C4d accumulation along peritubular capillaries. Group G3 (n = 12) received suboptimal CsA and anti-rejection therapy (rabbit ATG + methylprednisolone + CsA) of TXA. Animals developed CR and lost grafts within 65 days (median). As compared to G2, the arterial intimal changes showed less macrophages and T lymphocytes, an increased number of myofibroblasts, abundant fibronectin/collagen IV and scar collagens I/III, C4d deposition along capillaries in 60% of animals and transplant glomerulopathy in 80% of animals.

In conclusion, CR is an immune stimulated process initiated during TXA with the accumulation and proliferation of myofibroblasts, and progressive deposition of collagens in the intima. Our experimental design appears well suited to study events leading to CR.


alpha smooth muscle actin


anti-thymocyte globulin


chronic rejection


cyclosporine A




major histocompatibility complex


mixed lymphocyte reaction


nonhuman primates


per os






transplant endarteritis


The late loss of kidney allografts is a major and largely unsolved problem in transplantation (Tx). Chronic renal allograft dysfunction can have different etiologies. Besides immunological processes of rejection, nonimmunologic factors, including donor age, brain death, hypertension, prolonged warm and cold ischemia times, hyperfiltration, recurrent renal diseases or calcineurin inhibitor toxicity, are known to be associated with fibrosis and can influence long-term graft survival (1,2). However, only little is known about the specific chain of events resulting in allograft fibrosis.

Chronic rejection (CR), characterized histologically by transplant arteriopathy, transplant glomerulopathy and fibrotic changes of the tubulointerstitial space, is the leading cause of late graft loss following solid organ transplantation (3,4). The most important injury to a kidney transplant is caused by cellular and/or humoral allo-responses best characterized in the setting of acute rejection. In particular, acute rejection episodes with transplant endarteritis (TXA, Banff type II acute rejection) seem to play a central role in the development of chronic rejection and graft loss (3,5,6). TXA often only incompletely responds to anti-rejection therapy including anti-lymphocytic preparations (5,7,8). Endarteritis in its early stage shows infiltration of mononuclear inflammatory cell elements. The recruitment of myofibroblasts into the inflamed intimal layers results in the accumulation of extracellular matrix components, i.e. sclerosing vasculopathy with concentric intimal fibrosis. Arterial intimal sclerosis can progress (chronic progressive rejection) in the setting of persistent or repetitive intimal inflammation, e.g. due to recurrent or smoldering rejection, which ultimately leads to significant luminal narrowing and impaired blood flow (3,4,9).

In humans, most data on CR is limited as histological studies are typically performed on biopsy cores taken for diagnostic purposes in complex clinical situations. Often, protocol biopsies do not meet the Banff criteria of specimen adequacy, i.e. >7 glomeruli and at least one artery (10,11). Thus, TXA and CR with sclerosing vasculopathy may be underestimated (3,7). In a retrospective review of a large series of kidney transplants in rhesus monkeys, Schuurman et al. (12) concluded that CR could develop in a high percentage of animals (up to 85%) within 60 days post-Tx under suboptimal baseline immunosuppression. Based on these observations, we designed a unique study to specifically induce CR in young, healthy cynomolgus monkey renal allograft recipients. We wanted to test the feasibility of our CR study design, to compare observations with human CR, and to describe key morphological aspects of arterial intimal thickening associated with sclerosis.

Material and Methods


All experiments were performed in accordance with the Swiss Animal Welfare Act dated March 9, 1978, and the accompanying Animal Welfare Regulation of May 28, 1981, under licences BS 1556 and 1782.

The 36 cynomolgus monkeys (Macaca fascicularis, 12 females and 24 males) used in this study were captive-bred and obtained from SICONBREC Inc. (Makati City, Philippines). All monkeys showed normal hematology and serum/urine chemistry. They were negative for tuberculosis, Salmonella/Shigella, viruses (herpes B, simian T-cell lymphotropic virus, simian immunodeficiency virus, simian retrovirus type D and hepatitis B) and ecto-/endoparasites. All female animals were used as donors and all male animals as recipients (one donor for two recipients). Donor–recipient combinations were made according to ABO blood type (13), and age matching, as well as major histocompatibility complex (MHC class II) mismatching [as indicated by one way mixed lymphocyte reaction (MLR) stimulation indexes >8 and <45 (14); for details see Table 1].

Table 1.  Donor and recipient cynomolgus monkeys used in each study group, onset of transplant endarteritis (TXA) and survival of individual animals
GroupDonor no.Age (month)Weight (kg)Recipient no.Age (month)Weight (kg)MLR stimulation indexFirst signs of TXA (post-Tx day)Graft survival (days)
  1. Tx = transplantation.

  2. aAnimals terminated for nonrejection related reasons.

  3. bSevere rejection with necrosis.

  4. cAnimal not treated with anti-rejection therapy.

  5. dEndpoint of the experiment.


Several weeks prior to Tx, recipients were implanted with a transducer/emitter system (Data Sciences Inc, MN) for the telemetric monitoring of arterial blood pressure, heart rate and motor activity.

Kidney transplantation and post-operative monitoring

All transplantations were performed by the same Tx surgeon (MA). Two kidneys from one donor animal were removed and stored in University of Wisconsin (UW) solution at 4°C for 24–26 h. On the day of Tx, the donor kidney was implanted into a recipient that had undergone unilateral nephrectomy of one native kidney.

Following Tx, blood and urine samples were collected in the morning 1–2 times per week, just before administration of medication. Twenty-four-hour trough levels of CsA were measured in whole blood (EDTA anti-coagulated) 4 times per week during the first 2 weeks post-Tx and twice weekly thereafter, using the radioimmunoassay CYCLO-Trac® SP-Whole Blood (DiaSorin Inc, Stillwater, MN).

Since one native, life-supporting kidney remained in the recipient, serum creatinine levels could not be used for monitoring graft function. Instead, ultrasound examinations, at least twice weekly, and protocol biopsies every 2–3 weeks were used in order to assess the condition of the transplanted kidneys (15).

Study design

Animals were distributed into three treatment groups and received CsA (Neoral® or Sandimmun®, Novartis Pharma AG, Basel, Switzerland) as immunosuppressive treatments either orally (p.o.) or subcutaneously (s.c.), starting 1 day before Tx.

  • • Group G1 animals (n = 6) received high dose CsA treatment p.o.: 150 mg/kg/d for 15 days and subsequently 100 mg/kg/d. The administered dose of CsA immunosuppression was expected to result in long-term graft survival (16). The end of the observation period for this group was 90 days post-Tx.
  • • Group G2 animals (n = 6) received suboptimal CsA treatment p.o.: 150 mg/kg/d for 15–20 days followed by the step-wise reduction of the immunosuppression by 50 mg/kg/d every 15–20 days until ultrasound examination revealed severely impaired graft perfusion and TXA was diagnosed histologically. TXA was not treated with any specific anti-rejection therapy. Thus, animals in group G2 represent an unaltered course of acute rejection/TXA.
  • • Group G3 animals (n = 12) received suboptimal CsA treatment s.c.: 10 mg/kg/d for 21 days followed by 7.5 mg/kg/d up to 90 days post-Tx. Animals that developed TXA (Banff type II rejection) within 90 days post-Tx, received anti-rejection therapy with a polyclonal anti-thymocyte globulin (rabbit ATG, Fresenius HemoCare Immune Therapy GmbH, Germany, 20 mg/kg i.v.) together with methylprednisolone (Solu-Medrol®, Pfizer AG, Switzerland, 10 mg/kg i.v.) given five times every second day. At the same time, the s.c. administration of CsA was replaced with high dose of CsA p.o. (150 mg/kg/d for 15 days and subsequently 100 mg/kg/d). The end of the observation period for all animals treated with anti-rejection therapy was either graft loss or alternatively graft survival day 147 post-Tx.


Transcutaneous ultrasound-guided graft biopsies were performed every 2–3 weeks, starting 14–22 days post-Tx (protocol biopsies) or whenever ultrasound examination suggested impaired graft perfusion (according to previous reports 15, 17). All biopsies were obtained using a Bard® Magnum™ (Covington, Georgia) biopsy instrument with an 18G tru-cut biopsy needle. Usually 2 biopsy cores were collected and fixed in 4% buffered formalin.

At time of euthanasia, a full necropsy was performed. All organs were fixed in 4% buffered formalin. In addition, samples of kidney allograft, spleen and lymph nodes were frozen in Tissue-Tek® O.C.T™ Compound (Sakura Finetek Europe B.V., The Netherlands). Tissue processing and staining followed general recommendations (11).

All biopsy and necropsy kidney samples were examined blindly by two investigators (G.W., V.N.) and scored using the criteria recommended in the Banff 97 working classification of renal allograft pathology (11). Immunohistochemical staining results were semiquantitatively scored (see below). In case of a disagreement of lesion scoring, a consensus was reached by discussion at a double-headed microscope.


Immunohistochemistry was performed on sections of formalin-fixed and paraffin-embedded or frozen tissue samples harvested at time of necropsy. Standard procedures were applied using monoclonal and polyclonal antibodies and the streptavidin-biotin-peroxidase complex technique (for details see Table 2). Spleen tissue served as positive control sample for the CD markers and Ki-67 reaction. Negative immunohistochemical staining controls were obtained by replacing the primary antibodies with antibody isotype controls (Zymed Laboratories, Inc., San Francisco, CA).

Table 2.  Antibodies used for immunochistochemical stainings
AntibodyCloneVendorDilution tissueIncubationAntigen retrieval
  1. P = paraffin section; F = frozen sections; ON = overnight; MW = microwave oven; RT = room temperature.

  2. aSerotec Ltd., Oxford, UK.

  3. bNovocastra Laboratories Ltd., Newcastle upon Tyne, UK.

  4. cBD Biosciences Pharmingen, San Diego, CA.

  5. dDAKO A/S, Glostrup, Denmark.

  6. eAbcam Limited, Cambridge, UK.

  7. fRockland Inc., Gilbertsville, PA.

  8. gBiogenex, San Ramon, CA.

  9. hBiomedica Gruppe, Vienna, Austria.

CD3CD3–12Seroteca1:500 (P)ON +4°C15 min MW, EDTA pH 8.0
CD41F6Novocastrab1:100 (P)ON +4°C15 min MW, EDTA pH 8.0
CD8RPA-T8Pharmingenc1:25 (F)60 min RTNo
CD20L26Dakod1:1000 (P)ON +4°C20 min MW, citrate pH 6.0
CD68KP-1Dakod1:500 (P)ON +4°C20 min MW, citrate pH 6.0
αSMA1A4Dakod1:100 (P)60 min RT10 min MW, citrate pH 6.0
FibronectinPolyclonalAbcame1:50 (P)ON +4°C15 min, 0.1% protease XIV
Collagen IPolyclonalRocklandf1:50 (P)ON +4°C2 h, 0.25% pepsin/HCl
Collagen IIIHWD1.1BioGenexg1:50 (P)ON +4°C2 h, 0.25% pepsin/HCl
Collagen IVCIV22Dakod1:50 (P)ON +4°C15 min, 0.2% trypsin
Ki-67MIB-5Dakod1:500 (P)ON +4°C20 min MW, citrate pH 6.0
C4dPolyclonalBiomedicah1:40 (P)60 min RT20 min autoclave

C4d immunohistochemical staining was performed on formalin-fixed and paraffin-embedded sections, following recommendations of the supplier. Briefly, after deparaffinization and rehydration, slides followed an antigen retrieval in an autoclave (20 min at 120°C). Subsequently, they were incubated in methanol containing 0.5% H2O2 for 20 min, PBS containing 4% of fat-free powdered milk for 60 min and with the primary polyclonal antibody rabbit anti-human C4d (diluted 1:40, overnight incubation at +4°C). The immunoreaction was visualized using the streptavidin-biotin-peroxidase complex technique (Vector Laboratories, Inc., Burlingame, CA) and AEC+ (DakoCytomation Corp., Denmark). Slides were mounted with an aqueous mounting medium.

The immunohistochemical staining analysis was focused on arterial intimal changes of arteries, in particular interlobar and arcuate vessels. The most inflamed and thickened intimal segments were scored. Cells expressing a distinct phenotype (CD markers, Ki-67 and alpha-smooth muscle actin, αSMA) were scored semi-quantitatively (percentage of cells in the intima). The accumulation of extracellular matrix components, i.e. fibronectin, collagens I, III and IV, was semi-quantitatively scored based on the stained intimal surface area [no intimal staining, minimal (<10%), mild (10–25%), moderate (26–50%), marked (>50%)]. The accumulation of the complement degradation product C4d along peritubular capillaries in the cortex and/or medulla was scored as absent or present according to previous descriptions (18,19).


There were no surgical failures in the series of 24 transplants performed in this study. A total of 105 transplant biopsies were obtained, all of which fulfilled the Banff 97 criteria for at least minimal adequacy (11). All monitored serum/urine chemistry parameters as well as blood pressure and heart rates remained within normal limits during the entire observation period. Transient lymphocyte depletion was observed during anti-rejection therapy in group G3 (data not shown). No sings of nephrotoxicity or other histopathological abnormalities were detected in native kidneys.

Group G1 (n = 6), high CsA p.o. immunosuppression: no vascular lesions

Three recipients (no. 212, no. 142 and no. 269) were monitored up to 90 days post-Tx (end of protocol). Three other recipients (no. 280, no. 265 and no. 281) were euthanized on days 57, 86 and 87 post-Tx, respectively, due to infectious complications (post-transplant lymphoproliferative disorder, broncho-pneumonia or pyelonephritis) (Table 1, Figure 1). During the entire study period, a total of 32 biopsies were performed. Only one animal (no. 142) showed interstitial cellular rejection at the end of the observation period (Banff type IB: i3, t3, v0, g0, ci0, ct0, cv0, cg0). The arteries and glomeruli in all grafts were without abnormality. Viral inclusion bodies were not present. Immunohistochemical stainings were not performed.

Figure 1.

Kaplan-Meier survival curve of animals in groups G1—high immunosuppression, no TXA; G2—unaltered severe acute rejection with TXA and G3—severe rejection with sclerosing transplant vasculopathy.

Cyclosporine blood levels:  Animals exhibited a high degree of inter- and intra-individual variability of CsA levels (mean whole blood CsA 24-h trough level for the entire group: 478 ± 379 ng/mL, range: 79–1566 ng/mL, measured within the last 2 weeks of the observation period).

Group G2 (n = 6), low CsA p.o. immunosuppression: severe transplant endarteritis

All animals were terminated within 20–128 days post-Tx (median 46 days) due to severe untreated TXA (Table 3, Figure 1). TXA with interstitial cellular rejection was diagnosed in all grafts between days 20 and 85 post-Tx (median 40 days). In three animals, the diagnosis of TXA was made in allograft biopsies (Banff type IIA in animal no. 181 on day 85 post-Tx; Banff type IIB in animals no. 244 and no. 249 on day 43 and 29 post-Tx, respectively). In the other three grafts, TXA was first diagnosed at time of necropsy (Banff type IIB in animals no. 268, no. 274 and no. 278 on day 64, 20 and 37 post-TX, respectively). For technical reasons the graft of animal no. 181 could not be scored and was excluded from further analysis.

Table 3.  Sequence of events detected in protocol biopsies and at necropsy in group G2
Group G2
Animal IDDays post-TxNo. of glomeruli/arteriesBanff scoring
  1. B = biopsy; N = necropsy; Tx = transplantation; *First diagnosis of TXA; #Mild fibrosis in areas of interstitial edema.

No. 24414/B11/5i3, t2, v0, g0, ci0, ct0, cv0, cg0
29/B10/1i3, t3, v0, g0, ci2#, ct0, cv0, cg0
43/B*15/4i3, t3, v3, g1, ci0, ct0, cv0, cg0
55/NWhole grafti3, t3, v2, g2, ci1#, ct0, cv0, cg0
No. 24914/B14/4i2, t2, v0, g0, ci0, ct0, cv0, cg0
29/B*14/5i3, t3, v1, g1, ci2#, ct0, cv0, cg0
31/NWhole grafti3, t3, v2, g1, ci1#, ct0, cv0, cg0
No. 26815/B10/1i3, t2, v0, g0, ci0, ct0, cv0, cg0
29/B26/2i3, t2, v0, g0, ci0, ct0, cv0, cg0
43/B23/4i3, t2, v0, g0, ci1#, ct0, cv0, cg0
56/B14/1i3, t3, v0, g0, ci0, ct0, cv0, cg0
64/N*Whole grafti3, t3, v2, g1, ci2#, ct0, cv0, cg0
No. 27415/B11/1i3, t2, v0, g0, ci0, ct0, cv0, cg0
20/N*Whole grafti3, t2, v2, g1, ci2#, ct0, cv0, cg0
No. 27815/B33/4i0, t0, v0, g0, ci0, ct0, cv0, cg0
27/B22/5i3, t2, v0, g0, ci0, ct0, cv0, cg0
37/N*Whole grafti3, t3, v2, g1, ci1#, ct0, cv0, cg0
No. 18115/B28/3i1, t2, v0, g0, ci0, ct0, cv0, cg0
27/B21/3i2, t2, v0, g0, ci0, ct0, cv0, cg0
42/B25/3i1, t1, v0, g0, ci1#, ct0, cv0, cg0
57/B12/3i2, t2, v0, g0, ci2, ct0, cv0, cg0
71/B28/4i3, t2, v0, g0, ci2, ct0, cv0, cg0
85/B*12/5i3, t2, v1, g0, ci1, ct0, cv0, cg0

Histology and immunohistochemistry:  Although TXA involved the entire arterial tree to different degrees, the most prominent changes were seen in the arcuate and interlobar arteries. In the inflammed intimal layer, spindle and mononuclear cells were present (Figure 2A). Trichrome stains revealed focal minimal extracellular matrix accumulation (Figure 2C). The internal elastic lamina remained unaltered (Figure 2E). In 2/5 animals, severe intimal inflammation and hypercellularity resulted in focal vascular occlusion and parenchymal infarction. TXA was also noted in extraparenchymal arteries along the ureters. Transplant glomerulitis (i.e accumulation of inflammatory cells in dilated peripheral glomerular capillaries) was seen in all animals (Banff score g1–2) without evidence of transplant glomerulopathy (i.e. splitting and double contouration of peripheral capillary walls) (Table 3, Figure 2G). All animals showed diffuse mild interstitial fibrosis, predominately in areas of edema (i.e. so-called scleredema, Banff score ci1–2). Viral inclusion bodies were not present.

Figure 2.

Graft histology: A, C, E, G — group G2, animal 249 — unaltered severe acute rejection with TXA (A, C, E) and glomerulitis (G) 31 days post-Tx; B, D, F, H — group G3, animal 277 — severe rejection with sclerosing transplant vasculopathy (B, D, F) and transplant glomerulopathy (H) 65 days post-Tx. A and B — hematoxylin and eosin stained sections (original magnification ×200). Note marked intimal inflammation and hyperplasia. C and D — trichrome stained sections (original magnification ×200). Arrows in D indicate collagen deposition (in blue) in the thickend intima. E and F — Verhoeff stain for elastic fibers (original magnification ×200). Note that intimal inflammation and sclerosis is not associated with ruptures of the internal elastic lamina or elastosis. G and H — periodic acid Schiff reaction (original magnification ×630). The arrow in G indicates a dilated glomerular capillary filled with inflammatory cells, typical of transplant glomerulitis. The arrow in H indicates double contoured peripheral glomerular basement membranes, typical of transplant glomerulopathy. M indicates the unchanged arterial medial smooth muscle layer.

Transplant endarteritis and arterial remodeling were further studied by immunohistochemistry (Table 4, Figure 3 A,C,E,G and Figure 4 A,C,E,G). The arterial intimal cell infiltrates consisted of T cells (5–25% CD3-, 10–25% CD4-, up to 5% CD8-positive cells), macrophages (20–50% CD68-positive cells) and exceptionally rare B cells (CD20-positive). αSMA, a myofibroblast marker, was expressed by 10–50% of intimal cells, predominantly spindle cell elements. Individual αSMA positive cells were located throughout the entire intimal thickness. They formed a single cell layer adjacent to the internal elastic lamina. The hypercellular intima revealed a high proliferative activity (20–50% of cells expressed Ki-67). Among the tested extracellular matrix components, fibronectin was most prominent with marked immunoreactivity throughout the entire thickness of the inflamed intimal layers in all animals. Fibronectin surrounded individual cells in a lacelike pattern. Collagen IV showed a similar distribution pattern, however only mild to occasionally moderate immunoreactivity. The intimal distribution pattern of collagen I and III deposits was very restricted. Only 3/5 animals demonstrated collagens I and III in some arteries with minimal to mild immunoreactivity adjacent to the internal elastic lamina in areas of myofibroblast clustering. No animal (0/5) demonstrated C4d deposits along peritubular capillaries (Figure 5A).

Table 4.  Immunohistochemical staining patterns observed in thickened arterial intimal layers in groups G2 and G3 at time of necropsy (percentages of immunopositive cells and extent of staining)
GroupsAnimal no.CD3CD4CD8CD20CD68Ki67SMAFNCol IVCol IIICol IC4d
Figure 3.

Immuhistochemical staining pattern of different cell populations in the intima of arteries. A, C, E, G — group G2, animal 249 — acute TXA 31 days post-Tx; B, D, F, H — group G3, animal 277 — sclerosing transplant vasculopathy 65 days post-Tx. A and B — CD3 staining of T cells. C and D — CD68 staining of macrophages. E and F — Ki-67 staining of proliferating cells. G and H — alpha smooth muscle actin staining of myofibroblats and myocytes. The arrowhead in G indicates a single αSMA-positive cell. Note reduced percentage of CD3- and CD68-positive cells as well as abundant myofibroblats in arteries with sclerosing vasculopathy. Intimal cell proliferation (Ki-67-positive cells) remains a conspicuous finding during intimal sclerosis. M indicates the unchanged arterial medial smooth muscle layer. Arrows mark the internal elastic lamina (all images with the original magnification ×400).

Figure 4.

Immuhistochemical staining pattern of different extracellular matrix components in the intima of arteries A, C, E, G — group G2, animal 249 — acute TXA 31 days post-Tx; B, D, F, H — group G3, animal 277 — sclerosing transplant vasculopathy 65 days post-Tx. A and B — fibronectin, C and D — collagen I, E and F — collagen III, G and H — collagen IV. While fibronectin is a conspicuous matrix component in acute arteritis and sclerosing vasculopathy, collagen I, III and IV deposits are markedly increased over time. M indicates the unchanged arterial medial smooth muscle layer. Arrows mark the internal elastic lamina (all images with the original magnification ×400).

Figure 5.

Immuhistochemical staining pattern of the complement degradation product C4d along peritubular capillaries. (A) No C4d deposits in acute TXA in animal 249, 31 days post-Tx. (B) Strong C4d staining in animal 264, 57 days post-Tx.

Cyclosporine blood levels:  Animals showed a high degree of inter- and intra-individual variability of CsA levels (mean whole blood CsA 24-h trough level for the entire group: 111 ± 86 ng/mL, range <10–311 ng/mL, measured within 2 weeks before termination).

Group G3 (n = 12), suboptimal CsA s.c. immunosuppression: development of CR

In 7 out of 12 animals (58%), TXA and interstitial cellular rejection were diagnosed in graft biopsies obtained between 20–35 days post-Tx (median: 21 days; Banff type IIA in animals no. 236, no. 251, no. 264 and no. 284; Banff type IIB in animals no. 240, no. 248 and no. 277; Tables 1 and 5). Six of these animals received anti-rejection therapy (animal no. 251 was not treated). Treatment resulted in graft survival ranging between 44 and 147 days post-Tx (median: 65 days; Table 5 and Figure 1), corresponding to 22–127 days of survival after the initial diagnosis of TXA (median: 45 days). It was noted that the grafts with an initial diagnosis of Banff type IIA rejection generally survived longer (57, 114, 147 days) than those with Banff type IIB rejection (44, 47, 65 days).

Table 5.  Sequence of events detected in protocol biopsies and at necropsy in group G3 in animals with transplant endarteritis (TXA) treated with anti-rejection therapy
Group G3
Animal IDDays post-TxNo. of glomeruli/ arteriesBanff scoring
  1. B = biopsy; N = necropsy; Tx = transplantation, *First diagnosis of TXA and beginning of anti-rejection treatment.

  2. **Post-transplant lymphoproliferative disorder.

27720/B*13/3i3, t3, v2, g1, ci0, ct0, cv0, cg0
43/B7/1i3, t2, v2, g1, ci1, ct1, cv1, cg1
56/B28/6i3, t3, v2, g1, ci1, ct1, cv1, cg1
65/NWhole grafti3, t3, v2, g1, ci2, ct3, cv1, cg3
28420/B*18/6i3, t2, v1, g1, ci0, ct0, cv0, cg0
43/B22/6i1, t1, v0, g0, ci1, ct0, cv1, cg0
56/B12/2i2, t2, v0, g0, ci1, ct0, cv0, cg1
71/B7/2i1, t2, v0, g0, ci1, ct0, cv1, cg0
85/B34/3i1, t1, v0, g0, ci1, ct0, cv1, cg0
99/B19/6i2, t2, v0, g0, ci1, ct0, cv1, cg0
113/B30/7i2, t2, v0, g0, ci2, ct1, cv0, cg0
127/B21/6i0, t1, v0, g0, ci0, ct0, cv0, cg0
141/B14/2i0, t0, v0, g0, ci0, ct0, cv0, cg0
147/NWhole grafti0, t1, v1, g0, ci1, ct0, cv1, cg0
24021/B*15/3i3, t3, v2, g1, ci0, ct0, cv0, cg0
44/B24/10i3, t3, v2, g2, ci1, ct1, cv1, cg1
47/NWhole grafti3, t3, v1, g2, ci1, ct1, cv1, cg1
24821/B*7/5i3, t3, v2, g1, ci0, ct0, cv0, cg0
44/B24/7i3, t3, v2, g1, ci1, ct1, cv0, cg0
44/NWhole grafti3, t3, v1, g2, ci0, ct0, cv1, cg1
26422/B15/5i2, t2, v0, g0, ci0, ct0, cv0, cg0
35/B*11/2i3, t3, v1, g1, ci0, ct0, cv0, cg0
55/B13/5i3, t3, v2, g1, ci1, ct1, cv1, cg2
57/NWhole grafti3, t3, v1, g1, ci0, ct0, cv1, cg1
23622/B12/4i3, t3, v0, g1, ci0, ct0, cv0, cg0
28/B*13/4i3, t3, v1, g2, ci0, ct0, cv0, cg0
49/B21/5i3, t2, v0, g0, ci2, ct1, cv1, cg0
63/B16/7i3, t3, v1, g1, ci1, ct1, cv1, cg1
82/B20/2i3, t2, v0, g1, ci1, ct1, cv0, cg2
91/B26/5i3, t2, v1, g1, ci1, ct1, cv1, cg1
105/B7/1i3, t2, v0, g0, ci2, ct1, cv1, cg1
114/NWhole graftExcluded**

Four of the remaining animals showed interstitial rejection (Banff type IA) without evidence of TXA up to day 90 post-Tx (end of enrollment period into anti-rejection treatment group). The last animal (no. 276) of group G3 succumbed to acute vascular rejection with thrombi and infarction on day 13 post-Tx.; in this animal anti-rejection therapy was not initiated.

Histology and immunohistochemistry:  Renal transplants from five animals treated with anti-rejection therapy were further analyzed at time of necropsy (one animal was excluded from detailed scoring due to the development of a post-transplant lymphoproliferative disorder). Vascular rejection with arterial remodeling affected the entire arterial tree; the most prominent lesions were seen in arcuate and interlobar arteries. The thickened and hypercellular arterial intimal layers demonstrated various cell elements including spindle cells (most prominent along the internal elastic lamina), mononuclear cells as well as occasional foam cells (Figure 2B). In trichrome-stained sections, mild to moderate extracellular matrix deposition was noted that was most pronounced adjacent to the internal elastic lamina, where numerous spindle cells were located (Banff score cv1, Table 5, Figure 2D). Elastic tissue stains did not reveal any evidence of elastosis in the thickened intima or ruptures of the internal elastic lamina (Figure 2F). Arterial remodeling was also noted in extraparenchymal vessels along the ureters.

Transplant glomerulitis and glomerulopathy, with global or segmental double contouration of peripheral basement membranes, were seen in 4 of 5 animals (Banff scores g1–2 and cg1–3, Figure 2H). Interstitial fibrosis and tubular atrophy were observed in 3 of 5 animals (Banff score ci1–2 and ct1–3). Viral inclusion bodies were not present.

Arteries undergoing remodeling were further studied by immunohistochemistry (Table 4, Figure 3 B, D, F, H and Figure 4 B, D, F, H). The markedly thickened and hypercellular arterial intimal layers showed a predominance of αSMA-positive cells (50–85% of cells). The myofibroblasts were located throughout the entire intimal thickness and formed dense multilayered clusters along the internal elastic lamina. Between 5 and 20% of intimal cells expressed either T cell markers (CD3, CD4 and CD8) or the macrophage marker CD68. B lymphocytes (CD20-positive cells) were exceptionally rare. Evidence of proliferation (Ki-67 nuclear positivity) varied between 0 and 40% of intimal cells. Among the tested extracellular matrix components, fibronectin was most prominent with marked accumulation throughout the entire intimal thickness. Collagens were noted with moderate to marked immunoreactivity, mainly in the outer intimal zones containing densely clustered myofibroblasts. Collagen III deposits were limited to these outer intimal zones, whereas some collagen I and IV deposits were also noted toward the arterial lumens in the inner intimal layers. The complement degradation product C4d was detected along peritubular capillaries in 3/5 animals (no. 240, no. 264 and no. 277, Figure 5B). Other features of antibody mediated rejection, such as peritubular capillary inflammatory cell accumulation, were not conspicuous.

Cyclosporine blood levels:  Subcutaneous application of CsA resulted in low inter- and intra-individual variability of CsA levels (mean whole blood CsA 24-h trough level: 540 ± 103 ng/mL, range: 337–677 ng/mL) for 7 animals with TXA measured within 2 weeks prior to the initial diagnosis.


In renal allograft recipients, CR represents a major problem resulting in progressive graft failure. The pathophysiological events leading to CR are poorly understood and adequate treatment strategies are currently not available. Triggering events promoting the development of CR include acute rejection episodes, in particular those with TXA (5–7). In general, the clinical diagnosis of CR is made years after Tx, however, it has occasionally been described as early as 2–3 months post-grafting, especially in poorly compliant patients (20). In humans, CR is morphologically characterized by fibrous intimal thickening of arteries and allograft glomerulopathy, as well as interstitial fibrosis and tubular atrophy (4,11). Arterial remodeling in CR is of particular clinical importance because it can lead to stenosis, impaired perfusion and allograft dysfunction/loss (9). To our knowledge, systematic studies of CR in nonhuman primate (NHP) models have not been described. NHP models are considered to be most relevant for Tx (21), and they could provide close insights into the development of CR in humans. Here, we report such a study.

Our unique experimental design in cynomolgus monkeys demonstrated that CR was an inducible, immune-mediated injury caused by smoldering TXA in animals treated with suboptimal immunosuppression. The observed arterial changes with intimal inflammation, proliferation of myofibroblasts, the deposition of collagens, leading to sclerosing vasculopathy as well as transplant glomerulitis and glomerulopathy, and peritubular capillary C4d deposits were very similar to those described in humans (3,4,22). Our study resembled in part clinical scenarios seen in noncompliant patients, patients with an incomplete response to anti-rejection therapy, or those with subclinical smoldering rejection episodes (5,7,8). Interstitial fibrosis and tubular atrophy were not very prominent features in our animals, presumably due to the relatively short duration of impaired blood flow and ischemia-induced atrophy/fibrosis secondary to arterial stenosis.

Our experimental protocol included kidney transplantation into healthy, normotensive recipients that had undergone unilateral nephrectomy in order to enable the full development of rejection under nonlife-supporting conditions, similar to a recently described rodent model (23). The presence of the contralateral kidney in situ, during allograft rejection episodes, allowed for long-term graft monitoring without compromising an animal welfare and gave an additional information about the absence of nephrotoxic effects of the immunosuppressive therapy on the native kidneys. Although normal renal function and absence of hypertension made our study situation dissimilar from the clinical settings, there was no evidence that the functioning native kidneys had affected the immunological response or altered the pathology of rejection in groups G2 and G3.

The development of acute and chronic rejection was monitored by protocol biopsies (15,17) because graft function could not be accurately assessed by noninvasive methods, i.e. blood and urine biochemistry. A 24-h cold ischemia time was included into the experimental design as a presumed nonimmunological costimulatory factor for the development of CR (24). Under these conditions, suboptimal CsA immunosuppression led to acute rejection with TXA in 100% of animals in group G2 (CsA p.o.) and 58% in group G3 (CsA s.c.) at 6 and 3 weeks (median) post-Tx, respectively. Compared to p.o. administration of CsA, the s.c. application in group G3 had the advantage of being more convenient for daily animal handling, and it resulted in more reproducible 24-h trough levels (lower intra- and inter-individual variability, as well as an earlier onset of TXA). The s.c. application, however, did not improve the predictability of the onset of acute rejection and TXA. Our approach to slow down the progression of TXA with ATG treatment in group G3 was based on clinical experience in humans (5,8) as well as in vivo data from NHP, where it was used to induce lymphocyte depletion and prevention of acute rejection (25,26). It represents, to the best of our knowledge, the first reported trial of anti-rejection therapy in rejecting cynomolgus kidney allografts. However, our study design did not aim at fully treating rejection. Rather, ‘suboptimal’ anti-rejection therapy resulted in smoldering TXA and the development of progressive CR within weeks.

Histologic changes of acute rejection with TXA (group G2) included marked arterial intimal hypercellularity with influx and proliferation of different cell types including CD68-positive macrophages, CD3-positive T cells and scattered αSMA-positive myofibroblasts. To our knowledge, the detection of myofibroblasts during the early phase of TXA lacking sclerosis has not been reported before. Myofibroblasts can produce extracellular matrix proteins including collagens I and III (27). Thus, their appearance seems to mark an initial key event in TXA that sets the stage for subsequent collagen deposition and the development of sclerosing transplant vasculopathy (28). Myofibroblasts are activated by various cytokines, growth factors and fibronectin, in particular so-called cellular fibronectin isoforms (29,30). This pathway appears to play an important role in our model because marked fibronectin accumulation was constantly found in all animals with arterial inflammation and remodeling. Within weeks, smoldering TXA evolved in the G3 group into chronic sclerosing transplant vasculopathy characterized by large numbers of myofibroblasts and pronounced accumulation of collagens I, III and IV. In CR, myofibroblasts constituted the dominant cell type, likely due to persistent activation and proliferation, whereas the percentage of inflammatory cell elements decreased over time, likely due to the anti-rejection therapy. These low numbers of T cells and macrophages, however, are most likely sufficient to maintain a ‘permissive’ milieu promoting myofibroblastic activation and the progressive synthesis of collagens. Interestingly, in CR, arterial intimal thickening showed zonal differentiation: ‘older’ zones, in the outer intimal layer adjacent to the internal elastic lamina, demonstrated myofibroblasts and dense deposition of fibrillar collagens I and III, whereas ‘younger’ inner zones, under the endothelium, predominately revealed inflammation lacking conspicuous collagen accumulations.

At this juncture, we can only speculate about the origin of myofibroblasts in the intima of arteries with sclerosing transplant vasculopathy. Some studies suggest that they are of stem or progenitor cell origin, i.e. recipient derived, and migrate from the circulation into sites of injury and inflammation (31–34). Alternatively or additionally, they may also originate from adventitial fibroblasts (35) or the medial smooth muscle layer (4), i.e. donor derived. It seems tempting to hypothesize that myofibroblasts in our model originated from the donor based on their ‘clustering’ adjacent to the medial layer/internal elastica.

The role of humoral mediators in the development and promotion of CR is controversially debated (18,19,36). Some authors have proposed the term ‘chronic humoral rejection’ for C4d positive CR episodes (22). Our finding that 60% of animals with CR demonstrated C4d deposits along peritubular capillaries is similar to Mauiyyedi's et al. report in humans (22). Interestingly, all C4d positive animals showed transplant glomerulopathy. It supports the hypothesis that not only cellular but also humoral mediators contribute to the development of sclerosis and CR. Future experiments have to be designed to specifically monitor for circulating donor-specific antibodies, and the accumulation of C4d and their specific role in promoting CR.

In conclusion, this paper describes a unique study design promoting CR in cynomolgus macaque kidney transplants. The observed changes are very similar to human CR, and the experimental design seems well suited to further examine specific pathophysiologic events leading to sclerosing transplant vasculopathy. Our study suggests that myofibroblasts play a central role in the development of CR and that acute and chronic rejection should not be regarded as ‘two’ disease entities, but rather as a continuum. We believe that further experiments using our study protocol could be implemented to assess the beneficial effects of new treatment strategies for CR, in particular, those targeting myofibroblasts.


The authors are grateful to R. Apolloni, J. M. Blum, E. Braun, C. Cannet, A. Cattini, M. Erard, A. Harnist, A. Kunkler, A. Marcantonio, C. Maurer, B. Puissant, N. Stuber, C. Vedrine and G. Vogt for their excellent technical assistance during pre- and post-Tx monitoring and to Prof. R. Morris and Drs. E. Perentes and Ch. Bruns for review of the manuscript.