Mimicking Clinical Rejection Patterns in a Rat Osteomyocutaneous Flap Model of Vascularized Composite Allotransplantation

Abstract

Background:

Graft loss in vascularized composite allotransplantation (VCA) is more often associated with vasculopathy and chronic rejection (CR) than acute cellular rejection (ACR). We present a rat osteomyocutaneous flap model using titrated tacrolimus administration that mimics the graft rejection patterns in our clinical hand transplant program. Comparison of outcomes in these models support a role for ischemia reperfusion injury (IRI) and microvascular changes in chronic rejection of skin and large vessel vasculopathy. The potential of the surgical models for investigating mechanisms of rejection and vasculopathy in VCA and treatment interventions is presented.

Materials and Methods

Four rodent groups were evaluated: syngeneic controls (Group 1), allogeneic transient immunosuppression (Group 2), allogeneic suboptimal immunosuppression (Group 3), and allogeneic standard immunosuppression (Group 4). Animals were monitored for ACR, vasculopathy, and CR of the skin.

Results

Transient immunosuppression resulted in severe ACR within two weeks of tacrolimus discontinuation. Standard immunosuppression resulted in minimal rejection but subclinical microvascular changes, including capillary thrombosis and luminal narrowing in arterioles in the donor skin. Further reduction in tacrolimus dose led to femoral vasculopathy and CR of the skin. Surprisingly, femoral vasculopathy was also observed in the syngeneic control group.

Conclusions

Titration of tacrolimus in the allogeneic VCA model resulted in presentations of rejection and vasculopathy similar to those in patients and suggests vasculopathy starts at the microvascular level. This adjustable experimental model will allow the study of variables and interventions, such as external trauma or complement blockade, that may initiate or mitigate vasculopathy and CR in VCA.

Introduction

Clinical experience with vascularized composite allotransplantation (VCA) of the upper extremity is now in its third decade; results to date have demonstrated good outcomes with complications similar to those in solid organ transplantation (SOT(1)). In contrast to the fairly homogenous tissue composition of SOT, upper extremity and some face VCA are composed of skin, muscle, bone, and other tissues, each with its own immunogenicity characteristics. More than 80% of VCA patients exhibit acute rejection events within the first year following transplant(2, 3). This high rate of rejection has been attributed to the immunogenicity of the skin(4–6), but may also be related to the ease of diagnosing rejection in an allograft that can be visualized directly(5). Lung transplants that are directly exposed to the environment also have high rates of rejection(7), albeit not as high as those observed in VCA recipients.

Historically, graft rejection was broadly defined as acute rejection, either of a cellular or humoral nature, or chronic rejection, an indolent process that can progress to obliterative vasculopathy and parenchymal fibrosis(8, 9). The risk of acute rejection is high in the first few months post-transplant, while chronic rejection occurs months to years post-transplant. Exposure of the recipient to allogeneic MHC on donor antigen presentation cells (APC) or direct allorecognition may result in a stronger T cell activation and response early post-transplant(10), while recipient exposure to donor MHC presented on the host APC (indirect presentation) leads to activation of a more limited T cell repertoire and graft targets, resulting in chronic rather than acute rejection processes(11). The definition of the different types of rejection is an ongoing and revolving process and is graft specific. Recently, the European Society of Organ Transplantation published an update on the current six different histological phenotypes of acute and chronic rejection in kidney transplantation(12). The definition of chronic rejection in VCA recipients is also being defined(13), and donor vasculopathy resulting in ischemic graft loss and fibrotic changes in the skin have been described(4, 14, 15). Vasculopathy in VCA is similar to that observed in cardiac transplantation, defined as diffuse stenosis caused by concentric intimal hyperplasia and inadequate compensatory outward modeling(16). Like VCA grafts(14, 17), all vessels may be involved, from the microcirculation to the large epicardial arteries and the coronary veins(18). Presentation occurs months to years after transplantation, indicating a chronic vs. acute immune response. While lesions are fairly uniform, the progression to vessel narrowing and eventually allograft ischemia(19–21) is quite focal(18). In addition to chronic rejection in the form of vasculopathy, VCA allografts have multiple CR targets, including skin and adnexal structures (hair follicles, sebaceous glands, nailbeds) as well as vessels(13).

While acute vs. chronic rejection can be somewhat defined by the temporal relation to time of transplant, acute cellular or humoral rejection can occur any time post-transplant. Acute rejection is differentiated from chronic rejection by a more intense mononuclear cell infiltrate in parenchymal tissue and can result in significant and even necrotic tissue damage. Acute humoral rejection is also referred to as acute vascular rejection; this form of rejection is not associated with cellular infiltrates and is characterized by focal ischemia, endothelial swelling and intravascular coagulation(22). Chronic rejection of the skin is associated with minimal cellular infiltrate, with extensive fibrosis and loss of adnexal structures (i.e. hair follicles and sebaceous glands) and normal skin architecture. Multiple presentations of chronic rejection have been documented in VCA patients, including ischemic vasculopathy in the absence of ACR or skin involvement, CR of the skin, and loss of fingernails with minimal vasculopathy as well as severe ACR, vasculopathy, and chronic skin changes(13).

There remains much speculation regarding the pathogenesis of vasculopathy; factors such as inadequate immunosuppression, extensive physical activity of the hand, vibrational injury or other traumatic mechanical damage, infection, ischemia-reperfusion injury, and cell- and antibody-mediated rejection have been proposed(4, 13–15, 17, 20). The common mechanisms and presentation of acute and severe vasculopathy and CR in the skin vs. non-progressive vasculopathy between VCA recipients have yet to be elucidated.

These unanswered questions regarding the targets and presentation of CR in hand transplant and other VCA patients necessitate the development of animal VCA models that will allow for the intervention or exacerbation of ACR, vasculopathy, and skin CR during a clinically relevant course of immunosuppression. Such models will offer insights into rejection and vasculopathy in the face of interventions such as complement blockade, infectious challenge, and vibrational or other traumatic injuries. In this study, we utilized a modified rat heterotopic osteomyocutaneous flap hind limb allograft(23) to develop an adaptable model of VCA with carefully titrated immunosuppression which allows for the study of ACR and CR of the skin, as well as vasculopathy, as seen in a clinical setting.

Materials and Methods

Animals

Complete major histocompatibility complex (MHC)-mismatched male Brown Norway (BN, RT1.Aⁿ) and Lewis (Lew, RT1.A^I) rats aged 12–16 weeks and weighing 275–350 grams were used as VCA donors and recipients, respectively. A cohort of Lew rats was used as syngeneic donors. Animals were obtained from Charles River Laboratories (Wilmington, MA, USA) and acclimated to the animal facility for at least one week prior to surgery. The procedures were performed in accordance with the University of Louisville Institutional Animal Care and Use Committee (IACUC protocol #18198) and NIH Guide for the Care and Use of Laboratory Animals(24).

Donor Surgery

Donor osteomyocutaneous flap grafts were recovered for implantation as previously described(23). Briefly, the femoral vessels of donor BN (allograft) or Lew (syngeneic) animals were mobilized by ligating and cutting all vessels other than those providing arterial flow and venous return from the graft, which consisted of the tibia, fibula, surrounding muscles, nerves, and skin. The knee and ankle joints were disarticulated and the artery and vein pedicles were cut proximally near the inguinal ligament. One milliliter of heparinized saline (30 Units/ml) was injected into the artery to flush the graft. This volume was sufficient to flush all blood from the graft with clear saline effluent post-flush. Recent experience of graft failure with increased cold ischemia times have resulted in our increasing the amount of flush used to 4 ml. For the experiments presented here, 1 ml of heparinized saline was used. After graft recovery, the donor rat was euthanized by pneumothorax.

Recipient Surgery

Recipient animals were anesthetized using 2.5% isoflurane via an endotracheal tube and then prepared for graft transplantation as previously described(23). Donor vessels were anastomosed venous end-to-side, arterial end-to-end using 10–0 nylon interrupted sutures. Visual assessment of the anastomoses confirmed effective graft reperfusion. The donor and recipient surgeries were performed concurrently and the total ischemic time for the donor limb flap was 60–90 minutes during which the graft was at room temperature while anastomoses were performed. The graft was inserted into the inguinal pocket and oriented upside down, with the ankle joint superior and the knee joint inferior. Upon removal from anesthesia, the rats were placed on a heating pad for thermal support, administered meloxicam (4 mg/kg body weight) subcutaneously for pain suppression, and monitored closely until recovery. Additional meloxicam doses were administered daily until postoperative day (POD) 3.

Immunosuppressive Regimens

A schematic diagram of the four groups and dosing regimens is shown in Fig 1. Group 1 syngeneic Lewis recipients (n=3) did not receive immunosuppression throughout the experiment, emulating clinical replant recipients without alloreactivity, and were monitored daily and explanted on POD60-POD66. Allogeneic recipients in Groups 2–4 were weighed twice weekly and administered daily subcutaneous injections of 2 mg/kg tacrolimus until POD11. In the Group 2 transient immunosuppression model, which was designed to produce early aggressive ACR, immunosuppression was discontinued after POD11; animals were monitored daily until POD41-POD44, when rejection necessitated explants (n=5). The Group 3 suboptimal Immunosuppression group was designed to produce a common presentation of chronic skin changes with large vessel vasculopathy, achieved via a stepwise decrease in immunosuppression after POD11, first to 0.5 mg/kg tacrolimus (POD12-POD26) and then to 0.25 mg/kg tacrolimus from POD27 until termination at around three months or in the case of graft failure. Termination in this group occurred between POD68 and POD97 (n=5). Finally, the Group 4 standard immunosuppression group was modeled on a clinical dose of tacrolimus known to maintain grafts with minimal ACR; tacrolimus was reduced after POD11 to 0.5 mg/kg until termination on POD60-POD64 (n=4).

Figure 1.

Schematic diagram of the tacrolimus dosing regimens utilized to in control (Group 1) and experimental groups (Groups 2–4).

VCA Recipient Monitoring

Recipient rats were housed singly and monitored daily for clinical signs of pain, dehydration, weight loss, and decreased activity, in addition to surgical failure (first 48–72h) or rejection. The initial clinical signs of rejection, including edema, maculopapular rash, erythema, and skin desquamation(25, 26) were recorded. Macroscopic images of the grafted area were captured daily and scored for rejection by a double-blinded observer.

Laser Doppler Imaging

A Moor LDI2-IR laser Doppler scanner (Moor Instruments, Wilmington, DE, USA) with an infrared 785 nm laser diode was used to assess blood perfusion through the graft skin on POD14 and the day of explant. Perfusion through the skin was measured on a colorimetric scale from no flow (blue) to high flow (red) (Fig. 2).

Figure 2.

Laser Doppler imaging of a representative animal from each experimental group, with the border of the skin window marked by white dashed lines. Using transient immunosuppression (Group 2), robust perfusion through the graft as well as recipient collateral superficial blood supply was observed at POD14 (A) followed by complete loss of blood flow at POD43 (B). Exposure of the anastomosis (C, white arrow) at time of explant revealed severe vascular inflammation and edema distal to the anastomosis (C, black arrow). With suboptimal immunosuppression (Group 3), blood flow to the graft was normal at POD14 (D); by explant at POD90, nearly half of the graft exhibited loss of blood flow (E). The femoral artery appeared normal on both sides of the anastomosis (F, white arrow). With standard immunosuppression (Group 4), blood perfusion was normal throughout the entirety of the graft both at POD14 (G) and at explant on POD64 (H). The femoral artery exhibited normal morphology near the anastomosis site (I, white arrow).

Tissue Biopsy and Explant Surgery

On POD14 and biweekly thereafter, recipient rats were anesthetized using 2.5% inhaled isoflurane, 2 mm skin biopsies were obtained, and the wound was closed with absorbable 4–0 suture. For explant, animals were deeply anesthetized with 5% inhaled isoflurane; larger biopsies of the donor-recipient skin interface were procured, as well as donor muscle biopsy. Arterial samples (1–2 mm in length) were obtained proximally (recipient side) and distally (donor side) to the anastomosis (Fig. 4A). All tissue samples were fixed in 10% formalin before paraffin embedding.

Figure 4.

Femoral artery histopathology. Samples of the recipient femoral artery and donor femoral artery were obtained proximal and distal to the anastomosis site, respectively (A). In groups 2–4, recipient samples exhibited a thin intimal layer (yellow arrows) and comparable medial (red brackets) and adventitial layers (blue brackets)(B-D). Acute Rejection animals showed massive mononuclear cell infiltration in both the media (red bracket) and adventitia (blue bracket), as well as microvascular thrombosis of the vaso vasorum (black arrows show several examples), but no significant intimal hyperplasia (E). Suboptimal immunosuppression animals exhibited marked intimal hyperplasia (yellow bracket) with a layer of apoptotic endothelial cells lining the lumen, but relatively unchanged media (red bracket) and adventitia (blue bracket)(F). Group 4 standard immunosuppression animals exhibited signs of apoptosis of the endothelial lining and intima (yellow bracket), with slight outward remodeling of the media (red bracket) and unchanged adventitia (blue bracket)(G). Scale bar = 100 μm for all panels; images were captured at 20x magnification for all panels except E, which was captured at 10x magnification.

Histologic Evaluation

For morphometry, 6 μm sections of skin, muscle, and vessel samples were stained with hematoxylin and eosin (H&E) using standard staining methods. Skin biopsies were examined for changes in skin morphology and architecture as well as loss of adnexal structures (i.e., hair follicles and sebaceous glands). Skin and muscle biopsies were graded by a blinded observer, using the criteria listed in Table 1.

Table 1.

Rejection scoring for skin and muscle biopsies. All samples taken from Allogeneic transplants (Groups 2-4) at POD14 exhibited mild cell infiltration. Surprisingly Group 1 syngeneic recipients showed mild infiltrate at two weeks, which resolved by the time of explant. At explant, severe ACR with major epidermal necrosis was seen in Group 2, moderate infiltrate with some epidermal loss in Group 3 suboptimal immunosuppression animals, and mild to moderate infiltrate in Group 4 standard immunosuppression animals. Muscle biopsies at explant showed similar patterns, with mild cell infiltrate in Group1 syngeneic animals, heavy mononuclear cell infiltration with myocyte necrosis in Group 2 transient Immunosuppression animals, moderate cell infiltrate in Group 3 animals, and mild/diffuse cell infiltrate with minimal tissue damage in Group 4 standard immunosuppression animals. No significant differences were found between groups. However, analysis was affected by an N of 3 for most groups, and loss of data (some chronic rejection skin is scar and has no infiltrate). The biopsy data are presented as means as distribution of biopsy grades were shown to be normal using Shapiro-Wilk normality tests.

	Group 1 Syngeneic		Group 2 (Allo)		Group 3 (Allo)		Group 4 (Allo)
	No Immunosuppression		Transient Immunosuppression		Suboptimal Immunosuppression		Standard Immunosuppression
	Week 2	Terminal	Week 2	Terminal	Week 2	Terminal	Week 2	Terminal
Skin*	1.33±0.17	0.33±0.17	1.75±1.25	4.00±0.00	2.00±0.00	2.63±0.90	1.50±0.50	1.50±0.87
Muscle #		1.83+0.60		3.60±0.40		2.80±0.49		2.00±0.87
* Skin Biopsy Grade
0	No signs of rejection
1	Mild perivascular dermal cell infiltrate
2	Diffuse dermal cell infiltrate, interface reaction, sporadic cell infiltration of epidermis
3	Moderate to severe cell infiltration of epidermis, epidermal cell necrosis
4	Severe cell infiltration, major epidermal necrosis, loss of epidermis
# Muscle Biopsy Grade
0	Normal muscle tissue
1	Mild perivascular cell infiltrate
2	Diffuse mononuclear cell infiltrate
3	Moderate to heavy infiltrate, some disruption of tissue
4	Heavy infiltrate, myocyte necrosis, tissue damage, fibrotic scarring

Evaluation of Graft Vasculopathy

All measurements were performed by an observer blinded to the sample location and immunosuppression group. The vessel lumen area as well as the internal and external elastic laminae (EEL) were traced using ImageJ analysis software (NIH, Bethesda, MD), allowing for the calculation of intimal and medial thickness and area. The total number of cell nuclei within the medial layer of each vessel was determined. Percent intimal proliferation was calculated as intimal thickness/total laminae thickness, and the occlusion ratio was calculated as lumen diameter/total vessel diameter. For skin microvessel analysis, the total number of microvessels (defined as vessels with a lumen diameter of less than 200 μm) was determined for the donor skin area as well as the recipient skin area. The percentage of microvessels occluded by thrombosis, lumen narrowing, and/or microvessel wall thickening was calculated separately for donor and recipient skin areas.

Statistical Analyses

Statistical analyses were performed using SigmaPlot 14.0 (Systat, San Jose, California, USA) with p<.05 significance level. Continuous data were evaluated for normality using the Shapiro-Wilk test. All data sets were analyzed by Shapiro-Wilk in SigmaPlot, and required to demonstrate normality prior to testing for significance by ANOVA or t-test. Within group and within location (donor, recipient, and contralateral artery) one-way ANOVA was performed on all histological vessel measurements, followed by Holm-Sidak post hoc analyses where appropriate. To ensure rigor, histological images were coded and analyzed in a blind fashion with the code being broken only after all scoring quantifications were acquired.

Results

Operative results

The donor operative time was approximately 45 min, while the recipient operative time was approximately 1.5 to 2 h. On rare occasions, minor postoperative complications, such as focal skin necrosis, usually on the caudal edge of the skin paddle, and iatrogenic skin necrosis following subcutaneous tacrolimus injections, were observed. Upon explant analysis, one animal in the standard immunosuppression group was found to have numerous masses in the inguinal cavity, peritoneal cavity, and lungs; this animal was excluded from histological analyses.

Graft Perfusion

Laser Doppler imaging of syngeneic animals revealed proper blood flow through the skin of the graft throughout the study. Macroscopic examination of the femoral artery revealed no signs of trauma or hypertrophy on either side of the anastomotic site (data not shown).

Group 2 transient immunosuppression animals exhibited sufficient perfusion to the graft on POD14, including robust recipient collateral feeding vessels (Fig. 2A). Blood perfusion to the skin was completely ablated on the day of explant (POD43)(Fig. 2B). Surgical exposure of the anastomoses during the explant revealed a healthy recipient artery, whereas the donor artery distal to the anastomosis was severely inflamed and hypertrophied (Fig. 2C).

On POD14, Group 3 suboptimal immunosuppression animals presented with proper blood perfusion throughout the graft and abundant recipient collateral vessels (Fig. 2D), and terminal assessment revealed some areas of diminished blood flow within the VCA (Fig. 2E). Despite some loss of flow, macroscopic examination of the anastomoses did not indicate any obvious signs of vascular injury distal to the anastomosis (Fig. 2F).

In Group 4 standard immunosuppression animals, blood perfusion to the skin remained unchanged from POD14 (Fig. 2G) to POD64 (Fig. 2H), indicating sufficient blood flow throughout the course of the study. Exposure of the anastomoses at the explant revealed no outward signs of vascular trauma or thrombosis (Fig. 2I).

Clinical Signs of Rejection and Skin Histopathology

The Group 1 syngeneic control recipients exhibited normal healing, healthy skin components, and proper donor hair growth throughout the study. No clinical signs of rejection were noted. Biopsies from POD14 as well as at the time of explant (POD60-POD66) showed normal skin adnexal components. Interestingly, all the skin biopsies from Group 1 animals at two weeks showed grade1–2 cellular infiltrate (average histology score of 1.33, Table 1), with resolution of infiltrate by the time of explant (average skin histology score of 0.33, Table 1), possibly in response to IRI and surgical trauma.

In Group 2 transient immunosuppression animals, the skin showed normal wound healing and donor hair growth during the first 3 weeks post-surgery, indicating proper vascularization of the flap. The identification of donor skin vs. recipient skin was readily apparent because of the color and growth direction of the hair. The first signs of acute rejection, including skin redness, swelling, and partial loss of hair, were observed on POD19.4 ± 2.87 days (Table 2); by the time of explant, severe acute rejection of donor skin and graft failure was evident in all animals (Fig. 3A). Skin histopathology from POD14 revealed normal morphology with healthy adnexa, hair follicles, and sebaceous glands; mild mononuclear cell infiltration was apparent (Grade 1–2 ACR score). In the explant, the donor skin lacked adnexal structures, and the epidermis was necrotic and atrophic (Fig. 3B). Microvascular thrombosis (Fig. 3C, white arrow) and extensive mononuclear cell infiltration were observed in the donor skin.

Table 2.

Gross clinical observation of graft. A t-test on time to first rejection event revealed significantly delayed rejection events in Group 3 suboptimal Immunosuppression animals compared to Group 2 transient immunosuppression animals (*p=.002).

	First External Rejection Event
	Group 1 Syngeneic	Group 2 (Allo)	Group 3 (Allo)	Group 4 (Allo)
	No Immunosuppression	Transient Immunosuppression	Suboptimal Immunosuppression	Standard Immunosuppression
Days Post Surgery	No events	19.4±2.87	39.8±3.57*	No events

Figure 3.

Skin histopathology. At time of explant, Group 2 transient immunosuppression animals exhibited necrotic skin over the graft (A). Skin samples taken at the donor-recipient skin border (A, black box) showed massive mononuclear cell infiltration on both sides of the border, and necrotic epidermis as well as loss of skin adnexal components on the donor side (B); further magnification of donor skin (B, black box) showed perivascular infiltration as well as microvascular thrombosis (C, white arrow). Group 3 suboptimal immunosuppression animals exhibited pervasive hair loss and contraction of donor skin consistent with chronic rejection of the skin (D). A skin sample traversing donor-recipient-donor skin border revealed moderate mononuclear cell infiltrate localized to the donor tissue (E). A healthy new layer of epidermis was observed over the donor skin component of the graft; further magnification of donor skin (E, black box) revealed perivascular infiltration and narrowing of the arteriole lumen (F, black arrow). Animals in the standard immunosuppression group exhibited normal skin and hair growth throughout the entirety of the study (G). A skin sample taken at the donor-recipient skin border (G, black box) revealed normal skin on both sides of the graft, including healthy epidermis and normal skin adnexal components (H). Further magnification (H, black box) indicated vessel wall thickening of a microvessel (I, black arrow) and microvascular thrombosis (I, white arrow). Scale bar for B, E, and H = 500 μm; scale bar for C, F, and I = 100 μm.

In Group 3 animals receiving suboptimal immunosuppression, clinical follow-up showed a normal VCA skin appearance with healthy hair growth on POD14. The identification of donor vs. recipient skin was readily apparent. Skin biopsies on POD14 showed normal morphology with diffuse mononuclear cell infiltration (Grade 2 ACR score, Table 1). When immunosuppression therapy was decreased to 0.25 mg/kg, rejection events including maculopapular lesions, gradual hair loss, and atrophy of donor muscle and skin were noted; the first rejection event was observed at POD39.8 ± 3.57 days (Table 2). The donor skin over the flap gradually reduced in size over the course of the experiment and was eventually covered with a new epidermis (Fig. 3D). In explants, mononuclear cell infiltration was apparent in the donor skin (Fig. 3E), but to a lesser extent than that observed in ACR animals. Histopathology revealed a lack of adnexal structures in the donor skin, with severely contracted fibrotic donor skin with loss of the rete ridges and epidermal thinning. New skin likely of recipient origin (hair growth in this area was white in color suggesting Lewis strain origin) covered the rest of the muscle and bone graft. The recipient skin exhibited healthy adnexal structure, epidermis, and vessel components. At greater magnification, arteriolar wall thickening with perivascular infiltration (Fig. 3F, black arrow) was observed in the donor area of the skin.

In the Group 4 standard immunosuppression group, clinical follow-up revealed normal VCA skin and hair throughout the study (Fig. 3G). When immunosuppression was maintained at 0.5 mg/kg, the recipient animals did not exhibit gross signs of rejection. One animal in this group showed redness and swelling on POD44. However, this animal developed tumors throughout the graft and in the peritoneal cavity and lungs, necessitating its removal from the study. Skin biopsies on POD14 showed no signs of adnexal structure loss with minimal mononuclear cell infiltration (primarily ACR Grade 1), similar to both rejection groups (Table 1). Skin biopsy of the recipient/donor graft border at the explant revealed that the donor skin resembled the recipient skin. The skin exhibited normal epidermis and adnexa without morphological change, and mild mononuclear infiltration (Fig. 3H). At a higher magnification, occasional microvascular thrombosis (white arrow) and vessel wall thickening (black arrow) were observed in the donor skin (Fig. 3I).

Vessel Histopathology and Quantitative Measurement

Samples of the femoral artery proximal (recipient) and distal (donor) to the anastomosis were obtained (Fig. 4A). Histological analyses of the Group 1 syngeneic donor, recipient, and contralateral artery samples revealed no differences in either the total number of cells (Fig. 6A) or cells/μm² (Fig. 6B) in the intimal and medial laminae of the vessel. The donor lumen area was slightly, but not significantly, constricted (Fig. 6C). A layer of apoptotic cells lined the lumen along the intimal interface on both the recipient and donor sides of the anastomosis. No differences were observed in intimal thickness (Fig. 6D), medial thickness (Fig. 6E), total lamina thickness (Fig. 6F), intimal proliferation (Fig. 7A), occlusion ratio (Fig. 7B), percentage of occluded skin microvessels (Fig. 7C), or skin microvessels/mm² (Fig. 7D).

Figure 6.

Quantification of total nuclei in the intimal+medial laminae (A), nuclei/μm² in intimal+medial laminae (B), lumen area (C), intimal thickness (D), medial thickness (E), and total laminae thickness (F) from femoral artery histological samples. All graphs display group means + standard error. *p<.05 when compared to contralateral artery sample from the same group. Transient immunosuppression was characterized by cell infiltration to the donor vessel, increased medial thickening and increased total laminae area due to outward remodeling, whereas suboptimal immunosuppression resulted in cell infiltration to both donor and recipient vessel as well as intimal hyperplasia. Standard immunosuppression resulted in no large vessel differences compared to the animals’ unoperated control vessel or syngeneic surgical controls. Note that box and whisker charts show distribution of data into quartiles (indicated by horizontal line), highlighting the mean (shown by an X) and outliers. The boxes have lines extending vertically called “whiskers” indicating variability outside the upper and lower quartiles. Note that no points fall outside these whiskers.

Figure 7.

Quantification of intimal proliferation (A), occlusion ratio (B), and percentage of skin microvessels occluded (C). All graphs display group means + standard error. *p<.05 when compared to contralateral vessels from the same group, #p<.05 when compared to syngeneic samples, ^p<.05 when compared to recipient vessel samples from the same group. All groups exhibited significantly higher percentage of occluded microvessels in donor skin compared to syngeneic controls. The same markers of the box and whisker plots used for Figure 6 are used in Figure 7.

In the Group 2 transient immunosuppression animals, designed to develop severe acute rejection of the skin, the femoral artery exhibited changes consistent with necrotic rejection rather than vasculopathy. A robust mononuclear cell invasion in the medial layer of the donor artery (Fig. 4E) was detected compared to that in the recipient artery (Fig. 4B). An apoptotic layer of cells lining the lumen of the distal artery sample was observed (Fig. 4E), with pronounced microvascular thrombosis in the adventitial layer (Fig. 4E; examples shown with black arrows). Severe vasculitis of the donor artery in this group required imaging of the vessel sample at a lower magnification (10x); all other vessel images in Fig. 4 were acquired at 20x magnification. Quantitative vessel measurements did not show significant differences in total cells (Fig. 6A) and cells/ μm² (Fig. 6B) in the intimal+medial laminae. The lumen area (Fig. 6C) and intimal thickness (Fig. 6D) were unchanged, with increased medial thickness in the donor artery distal to the anastomosis compared to contralateral vessel samples (one-way ANOVA,DF=2, F=8.081, p=.006, Holm-Sidak post hoc p=.005)(Fig. 5E). Similarly, the total laminae thickness (intimal+medial layers) exhibited significant outward remodeling in the donor artery compared to the contralateral control vessel (one-way ANOVA, DF=2, F=6.090, p=.015, Holm-Sidak post hoc p=.016. Intimal proliferation and occlusion ratios were unchanged between donor, recipient, and contralateral vessel samples (Fig. 7A and 7B, respectively). Significantly more donor skin capillaries were occluded compared to syngeneic donor skin (one-way ANOVA, DF=3, F=9.785, p=.001, Holm-Sidak post hoc p=.001)(Fig. 7C).

Figure 5.

Femoral artery histopathology of Group 1 syngeneic control animals. Samples of the recipient femoral artery and donor femoral artery were obtained proximal and distal to the anastomosis site, respectively (A). Unlike Groups 2–4, two of the three animals in Group 1 showed some level of intimal hyperplasia on the recipient side of the anastomosis. (C,D). In addition, microvascular thrombosis of the vaso vasorum (black arrows show several examples) were seen in the adventitia of the recipient side sample in the third animal (D). When the donor side of the femoral artery was examined, again, two of the three animals showed significant intimal proliferation, even more than seen on the recipient side of the anastomosis (E,F). The third animal showed no intimal proliferation and normal histology of the donor femoral artery (G). Scale bar = 100 μm for all panels; images were captured at 20x magnification for all panels.

Group 3 was designed to replicate chronic rejection patterns seen in clinical VCA recipients, including chronic skin rejection with fibrosis and loss of adnexal structures and rete ridges as well as intimal hyperplasia and medial hypertrophy or graft vasculopathy in the donor vessels. In Group 3 suboptimal immunosuppressed animals, vessel samples from the recipient side of the femoral artery anastomosis exhibited healthy intimal and medial layers (Fig. 4C). Marked intimal hyperplasia was apparent in the donor artery (Fig. 4F, yellow bracket) and apoptotic cells lined the innermost layer of the intima along the lumen. Significantly more infiltrating cells/μm² of the intimal+medial area was found (one-way ANOVA, DF=2, F=6.092, p=.015) both in the recipient (p=.023) and donor arteries (p=.031) than in the contralateral arteries (Fig. 6B). The lumen area was unchanged (Fig. 6C), while the intima was significantly thicker in the donor vessel than in the contralateral control artery (one-way ANOVA, DF=2, F=5.747, p=.018, Holm-Sidak post hoc p=.016)(Fig. 6D). The medial thickness and total lamina thickness remained unchanged (Fig. 6E and 6F, respectively). Intimal proliferation was significantly higher in donor vessels (one-way ANOVA, DF=2, F=13.310, p<.001) than in both recipient (p=.005) and contralateral arteries (p=.001)(Fig. 7A). The occlusion ratio was unchanged (Fig. 7B), while the percentage of occluded skin microvessels was significantly higher in the donor skin than in Group 1 syngeneic donor skin (one-way ANOVA, DF=3, F=9.785, p=.001, Holm-Sidak post hoc p=.004)(Fig. 7C). Skin microvessels/mm² remained unchanged (Fig. 7D).

Groups 2–4 all exhibited normal vessel components on the recipient side of the anastomosis (Fig. 4B,4C and 4D). Unexpectedly, 2/3 animals in Group 1 showed mild intimal hyperplasia on the recipient side of the anastomosis (Fig. 5C, 5D). In addition, the syngeneic control animals showed moderate to severe intimal hyperplasia in 2/3 animals on the donor side of the anastomosis (Fig. 5E, 5F). Evidence of intimal hyperplasia was observed in the donor vessel of Group 4 (Fig. 4G), with a layer of apoptotic cells lining the innermost layer of the intima, but not to the extent seen in Group 3 animals who showed overt intimal hyperplasia in the donor vessel (Fig 4F). Group 2 animals who received transient immunosuppression that resulted in severe necrotic acute rejection had donor arteries with extensive lymphocytic infiltration and capillary thrombosis of the adventitial layer, and acentric intimal and medial proliferation consistent with severely inflamed tissue (Fig. 4E).

In addition to skin histology, we also scored biopsies of muscle taken at time of explant. Muscle showed cellular infiltrate and evidence of tissue damage in a similar pattern as seen for skin biopsies taken at explant. Biopsies were scored as defined in Table 1, and showed mild to diffuse cellular infiltrate in Group 1 syngeneic animals, and Group 4 standard Immunosuppression. As expected, muscle at explant from animals with severe ACR (Group 2) revealed robust cellular infiltration, lack of cross striation and hypochromagenicity of the muscle, and arteriole hyperplasia (black arrow, Fig. 8A) (Table1). Muscle biopsy from Group 3 Suboptimal Immunosuppression animals revealed moderate cellular infiltration and decreased lumen diameter in the muscle arterioles (Fig. 8B, black arrows) (Table 1). In contrast, Histological analysis of the explant muscle samples from Group 4 standard immunosuppression animals revealed mild cell infiltration and healthy microvessels (Fig. 8C, white arrow) (Table 1).

Figure 8.

Representative images of muscle biopsy at the time of explant collection for each group. Animals in the transient immunosuppression (Group 2) exhibited robust mononuclear cell infiltration as well as hyperplasia of capillary vascular structures (A, black arrow). Group 3 suboptimal immunosuppressed animals showed moderate cell infiltration, combined with compromised lumen diameter in muscle capillaries (B, black arrows). Group 4 standard immunosuppression resulted in minimal cell infiltration, healthy muscles, and normal capillaries (C, white arrow). Scale bar = 100 μm.

Discussion

Multiple rodent experimental models of VCA have been described, each with its own advantages and drawbacks(27). The model presented here was developed at our center(23) and offers key components that mimic various presentations of rejection observed in the clinic. The groups presented were successful in modeling patterns seen in clinical replant patients (Group 1 – Syngeneic transplants in the absence of immunosuppression), and in clinical VCA transplant recipients including ACR without vasculopathy (Group 2 – Transient immunosuppression), evidence of both chronic skin changes and vasculopathy (Group 3 – Suboptimal immunosuppression), and minimal ACR with some microvessel thickening in the absence of significant large vessel involvement or significant acute chronic skin rejection (Group 4 – Standard immunosuppression).

The rejection spectrum demonstrated by the groups in this study is influenced by four variables: IRI/surgical trauma, allogeneic donor MHC, the level of immunosuppression and duration of immunosuppression. Group 1 animals showed an unexpected amount of intimal proliferation on both the recipient and donor side of the anastomosis. Of relevance to the observation in this report of vasculopathy in the donor vessel of our syngeneic control groups, vasculopathy can also occur when digits and upper extremities are replanted following amputation(28, 29). This vasculopathy is not related to alloantigen and likely originates in the ischemia reperfusion injury (IRI) to the replanted tissue as a result of the trauma and surgical procedure.

A characteristic shared by Group 1 and clinical replant patients is that neither received tacrolimus immunosuppression. As we did not see significant hyperplasia on the recipient side of the anastomosis in Group 2–4, nor in the donor femoral artery in Group 2 animals, it may be that the tacrolimus interfered with the development of non-allogeneic related intimal hyperplasia. A future direction of ours is to repeat Group 1 and also test the ability of tacrolimus to interfere with the development of intimal hyperplasia due to IRI and surgical trauma. In the Group 2 transient immunosuppression group, when immunosuppression was halted 12 days postoperatively, severe acute rejection events occurred in less than two weeks, resulting in necrotic compromise of the skin and ischemic changes before vasculopathy could develop. This model mimics the rare non-compliance situation where patients stop taking immunosuppression. In the Group 3 suboptimal immunosuppression group, tacrolimus administration was extended but reduced in a stepwise fashion and maintained at lower doses until terminal time points, which allowed both chronic skin rejection with loss of adnexal structures and moderate to severe vasculopathy to develop. In the Group 4 standard immunosuppression animals, the tacrolimus dose was lowered on POD12 but maintained at a dose sufficient to keep the graft outwardly healthy. On the surface, the skin and hair appeared normal throughout the course of the study and blood perfusion was uniform throughout the flap at the explant. This group models patients on standard immunosuppression that are doing well clinically, but over time develop histologic changes in skin biopsies and mild intimal hyperplasia of donor vessels. In Group 4 animals, skin and vessel histology at explant showed scattered microvascular narrowing and thrombosis and mild medial hyperplasia of the donor femoral artery distal to the anastomosis. We expect mild intimal hyperplasia would also develop if animals were maintained longer. This presentation is most similar to the clinic, where the target is the lowest amount of immunosuppression that will maintain the graft, while minimizing immunosuppression complications. While Group 4 models most transplant recipients with good graft function but evidence of ongoing low-level rejection in the skin and vessels, it does not model the confluent aggressive ischemic vasculopathy and graft loss in the absence of skin rejection seen in our fourth transplant recipient(20). We hypothesize that this type of rejection is a combination of alloreactivity and external physical trauma to the graft. We have evidence that vibrational trauma can exacerbate vasculopathy in this rat model of VCA (unpublished data), and other groups have demonstrated the potential role of physical and thermal trauma in promoting graft rejection in VCA recipients(30, 31).

These models will be used in experiments to test potential interventions or define mechanisms that result in graft rejection. VCA patients that are doing well clinically but have histologic evidence of alloimmune-mediated damage are vulnerable to progression from a number of causes. Reduced doses of immunosuppression occur clinically because of related complications, neoplasia, or compliance issues. In addition, we expect accumulation of cellular- and antibody-mediated rejection events, infection, and/or physical trauma will promote progression to chronic rejection of the skin and vasculopathy in our animals.

In the Group 4 standard immunosuppression model, it appears the vascular injury in the graft begins in the small vessels. Microvessels in the donor skin exhibited thrombosis and lumen narrowing, with minimal signs of vasculopathy in the donor femoral artery. This observation may indicate a cascade of vascular involvement in rejection that begins in the smallest vessels. Strengthening this argument, Kanitakis et al. have observed capillary thrombosis prior to vasculopathy that ultimately led to the loss of two grafted fingers in a VCA patient(32). While multiple mechanisms have been proposed for developing vasculopathy, few studies have focused on the initiating events, and provide evidence that the microvessels are affected first, as seen in Group 4.

It is important to consider the limitations and opportunities for development of the current study. A significant limitation was the use of only male animals for this study. We used only males as the femoral artery and vein anastomosis is much easier if the donor and recipient are larger and approximately the same size. Female Brown Norway and Female Lewis rats tend to weigh 100 grams less than males, and do not usually achieve weights of 300 grams, which is the average weight of our recipient Lewis rats. This disparity can be addressed in future studies where only female rats are used, and results compared to the current study. The most important impact of sex bias on the present study is that female rats may metabolize tacrolimus at a different rate than male rats. This would definitely impact the timing and severity of graft rejection in this model. While the analysis of acute vs. chronic rejection of the skin was based on mononuclear cell infiltrate and patterns of damage and fibrosis by H&E, as is done clinically, we performed staining of CD3, CD45RA and several other lineage markers which produced expected staining patterns. However, the background staining was unexpectedly high, despite multiple attempts to optimize with different blocking agents. We suspect that fixation in 10% formalin for extended periods may have contributed to the high backgrounds and this data could not be included. Additionally, two other clinical presentations of rejection seen on our hand transplant recipients remain to be developed. First is a model of development of chronic skin changes in the absence of significant vasculopathy, which has also been observed in VCA patients(13). Additionally, we are working on a model that mimics aggressive confluent ischemic vasculopathy without ACR in the skin that resulted in early graft loss in our fourth hand transplant recipient(20).

Conclusions

Here, we present a reliable and reproducible tool for studying the underlying mechanisms of VCA graft vasculopathy and acute and chronic rejection of the skin. Careful titration of immunosuppression in this model can push a graft towards different levels of rejection, allowing for later administration of additional graft stressors (i.e. infection, mechanical insult, etc.) or interventions such as complement blockade to ascertain potential thresholds that could ultimately identify mechanisms and potential for prevention or treatment of skin CR and vasculopathy.

Abbreviations

ACR

acute cellular rejection

CR

chronic rejection

EEL

external elastic lamina

IEL

internal elastic lamina

IRI

ischemia reperfusion injury

MHC

major histocompatibility complex

POD

post-operative day

SOT

solid organ transplant

VCA

vascularized composite allotransplantation