Research ArticleTransplantation

A single localized dose of enzyme-responsive hydrogel improves long-term survival of a vascularized composite allograft

See allHide authors and affiliations

Science Translational Medicine  13 Aug 2014:
Vol. 6, Issue 249, pp. 249ra110
DOI: 10.1126/scitranslmed.3008778


Currently, systemic immunosuppression is used in vascularized composite allotransplantation (VCA). This treatment has considerable side effects and reduces the quality of life of VCA recipients. We loaded the immunosuppressive drug tacrolimus into a self-assembled hydrogel, which releases the drug in response to proteolytic enzymes that are overexpressed during inflammation. A one-time local injection of the tacrolimus-laden hydrogel significantly prolonged graft survival in a Brown Norway–to–Lewis rat hindlimb transplantation model, leading to a median graft survival of >100 days compared to 33.5 days in tacrolimus only–treated recipients. Control groups with no treatment or hydrogel only showed a graft survival of 11 days. Histopathological evaluation, including anti-graft antibodies and complement C3, revealed significantly reduced immune responses in the tacrolimus-hydrogel group compared with tacrolimus only. In conclusion, a single-dose local injection of an enzyme-responsive tacrolimus-hydrogel is capable of preventing VCA rejection for >100 days in a rat model and may offer a new approach for immunosuppression in VCA.


Vascularized composite allotransplantation (VCA), in particular hand and face transplantation, has the promise to circumvent hurdles associated with current clinical practices such as surgical reconstruction procedures using autologous tissues or prostheses. However, widespread clinical use of VCA is still hindered by side effects of immunosuppressive drugs. As with solid organ transplantation, systemic intake of immunosuppressive drugs increases the risk of opportunistic infections, end-stage renal disease, and malignancies such as lymphomas in VCA (1). In addition, metabolic complications including hyperglycemia, hyperlipidemia, impaired renal function, arterial hypertension, and aseptic hip necrosis have also been reported in VCA patients (2, 3).

Tacrolimus (FK506), a calcineurin inhibitor, is commonly used as a single drug or in combination with mycophenolate mofetil (MMF) or steroids as a maintenance therapy in VCA (46). However, systemic delivery of tacrolimus has been linked to nephrotoxicity (7, 8), diabetogenicity (9, 10), and malignancies (11) in patients. Trough levels of tacrolimus in VCA are set relatively high as compared to solid organ transplantation (12). In preclinical rat hindlimb transplantation experiments in the Brown Norway–to–F344 combination, administration of tacrolimus (10 mg/kg) on the day of transplantation, followed by intermittent injections of 3 mg/kg once a week, resulted in graft survival for up to 200 days. However, recipients with long-term surviving grafts suffered from drug-associated complications, in particular pneumocystis pneumonia, in most cases (13).

It has been reported that topical application of tacrolimus and clobetasol helped to reverse episodes of acute graft rejection in hand transplantation (14). Locally applied tacrolimus resulted in a several-fold increase in local drug concentration while maintaining a low systemic concentration (15), suggesting a potential benefit of local drug administration in VCA. Moreover, daily topical tacrolimus therapy combined with the standard immunosuppressive induction protocol prevented skin rejection in a rat hindlimb allotransplantation model by inhibiting immune cell functions in the skin (16). While maintaining blood levels of 0.5 ng/ml of tacrolimus, topical-only administration resulted in reduced rejection of skin allografts (17).

Tacrolimus-encapsulated biodegradable glycolide-co-clatide-co-caprolactone polymer has been used to prolong corneal allograft survival without any adverse effects, suggesting a potential for sustained-release systems in transplantation (18). Self-assembling poly(ε-caprolactone)–poly(ethylene glycol)–poly(ε-caprolactone) micelles, loaded with tacrolimus, were shown to release the drug much slower than standard formulations, which would help to avoid toxic peak blood levels (19). Encouraging data were also obtained in a rat liver transplantation model using tacrolimus-loaded poly(ethylene glycol)-poly(d,l-lactide) nanoparticles (20).

Self-assembled hydrogels can encapsulate hydrophobic drugs and can be formed from low–molecular weight amphiphilic gelators with enzyme-responsive bonds that may provide persistent drug delivery in response to an inflammatory rejection episode. Here, we report an injectable self-assembled hydrogel that can encapsulate and release the immunosuppressive drug tacrolimus in response to proteolytic enzymes and prolong graft survival to more than 100 days.


Design and gelation studies of triglycerol monostearate

To identify suitable hydrogel compositions, we initially scanned a pool of hundreds of agents from the GRAS (generally recognized as safe) agent list by the U.S. Food and Drug Administration (FDA) to identify amphiphilic agents with enzyme-cleavable bonds. We rationalized that GMP (good manufacturing practice)–grade GRAS agents are available in large scale typically at low cost, which could make the proposed approach a more attractive translatable solution. We identified and confirmed that triglycerol monostearate (TGMS) could self-assemble into hydrogels and disassemble in an enzyme-responsive manner. TGMS is an amphiphile that comprises a polyhydroxyl sugar head group for the formation of a hydrogen-bonding network, making self-assembled TGMS nanofibers soluble in water, and a polymethylene hydrocarbon chain for efficient encapsulation of hydrophobic molecules via van der Waals forces. In addition, TGMS has an ester linkage that enables cleavage by esterases and matrix metalloproteinases (MMPs) that are present during inflammatory conditions (Fig. 1A). TGMS forms robust gels in a wide range of solvents from 5 to 9% (w/v). TGMS hydrogels could encapsulate therapeutically relevant doses of the immunosuppressive drug tacrolimus. Tacrolimus-encapsulated TGMS hydrogels (TGMS-TAC) showed transformation of gel phase to liquid phase at temperatures between 55° and 60°C (table S1), suggesting that the TGMS-TAC hydrogel will be stable and remain in gel phase at physiological temperatures. Transmission electron microscopy showed a nanofibrous morphology of TGMS-TAC hydrogel (Fig. 1B).

Fig. 1. Encapsulation of tacrolimus in TGMS hydrogel and enzyme-responsive drug release.

(A) Schematic of TGMS self-assembly and encapsulation of tacrolimus. (B) Transmission electron micrograph of TGMS-TAC hydrogel. (C and D) Proteolytic enzyme-responsive tacrolimus release. Hydrogels incubated in PBS remained hydrolytically stable and did not release the drug for at least 3 months, and then addition of proteolytic enzymes (lipase, MMP-2, and MMP-9) induced the drug release. *P < 0.002 and **P < 0.01, PBS versus enzyme-treated groups. (E) Schematic of LPS activation of RAW 264.7 macrophages to mimic inflammatory conditions. (F) Cell culture supernatant from activated macrophages induced drug release when added to TGMS-TAC (gray symbols, supernatant added on days 0, 3, and 6), whereas supernatant from nonactivated macrophages or PBS did not induce drug release (***P < 0.03). Raw data and details on the statistical tests are given in tables S2 to S4.

Drug release from TGMS-TAC in response to enzymes and inflammation-like conditions in vitro

Previously, we have demonstrated the enzyme-triggered controlled delivery of hydrophobic drugs using nanofibrous hydrogels (21). Proteolytic enzymes are significantly up-regulated under inflammatory conditions like acute rejection in VCA (22), and their expression and concentration correlate with the degree of inflammation (23). We therefore evaluated the ability of self-assembled gel to release the encapsulated drug in response to proteolytic enzymes. TGMS-TAC hydrogels were immersed in phosphate-buffered saline (PBS) and incubated at 37°C with lipase (esterase), MMP-2, or MMP-9 enzyme (100 ng/ml). At regular intervals, aliquots of samples were collected, and release of tacrolimus was quantified using high-performance liquid chromatography (HPLC). Plotting cumulative release of tacrolimus (%) versus time (Fig. 1C and table S2) revealed that lipase and MMPs trigger hydrogel degradation to release the encapsulated tacrolimus, whereas gels in PBS controls remained stable and did not release significant amounts of the drug. Gels in PBS remained stable for at least 3 months, indicating that the presence of enzymes is required for gel disassembly and the release of encapsulated agents (Fig. 1D). In the present system, we did not observe burst release, which is consistent with self-assembled prodrug-based gels that we have previously synthesized (21). To investigate the potential for enzyme-responsive disassembly, MMP-9 was added after a preincubation of TGMS-TAC in PBS for 90 days and triggered gel degradation and release of the drug (Fig. 1D and table S3).

To investigate whether an inflammation-like condition would trigger hydrogel degradation, leading to the release of the encapsulated drug, murine macrophages (RAW 264.7) were activated by treatment with lipopolysaccharide (LPS; Fig. 1E). After 24 hours, cell culture supernatant from activated macrophages was added to TGMS-TAC hydrogel and incubated at 37°C, and the release of tacrolimus was quantified. Plotting cumulative tacrolimus release in percent versus time (Fig. 1F and table S4) revealed that supernatant from LPS-activated macrophages triggered gel degradation, leading to the release of the drug, whereas supernatant of nonactivated macrophage cultures did not. In addition, repeated addition of 100 μl of activated macrophage supernatant to TGMS-TAC in 3-day intervals resulted in repeated surges of released drug as shown in Fig. 1F. Control experiments with supernatant from nonactivated macrophages and PBS showed no significant drug release.

Vascular composite allograft survival and macroscopic assessment

Hind limb transplantations from Brown Norway–to–Lewis rats were performed, and TGMS-TAC was injected subcutaneously into the graft at day 1 after transplantation. Graft rejection was evaluated macroscopically and graded as 0 = no rejection, 1 = erythema and edema, 2 = epidermolysis and exudation, and 3 = desquamation, necrosis, and mummification. The median survival time (MST) of groups I (no treatment) and II (TGMS hydrogel vehicle) was 11 days (n = 4 each; Fig. 2, A and B, and table S5). All recipients in these two groups showed acute rejection of the graft with edema formation and necrosis (Fig. 2C). Recipients in group III (local injection tacrolimus only) showed an MST of 33.5 days (n = 6). Macroscopic signs of rejection in group III were slightly different to the ones seen in the acute rejection groups I and II. Early signs of rejection included loss of hair, mild edema, and dry skin. Later, rejection signs continued to develop, characterized by epidermolysis, exudation, and necrosis (Fig. 2C). MST of group IV (TGMS-TAC, ipsilateral injection) was >100 days and thus significantly higher than in all other groups (n = 6). Contralateral injection of TGMS-TAC, group V, resulted in an MST of 75 days, which is significantly lower than in group IV (n = 5; P = 0.0068, Mantel-Cox test).

Fig. 2. Vascular composite allograft survival and histopathological features.

(A and B) Brown Norway–to–Lewis orthotopic hindlimb transplantation was performed. Control groups were left untreated (I) or were treated with TGMS as a vehicle control (II). Experimental groups were treated with a single injection of 7-mg tacrolimus subcutaneously (III) or 7-mg TGMS-TAC (group IV) into the transplanted or contralateral limb (TGMS-TAC/ConLat, group V), respectively, at postoperative day (POD) 1. Kaplan-Meier graft survival curves are shown with P values analyzed by log-rank (Mantel-Cox) test. Raw data are given in table S5. (C) Representative macroscopic images of hindlimb allografts. (i) Groups I and II showed an acute rejection with an MST of 11 days. (ii) Group III allografts, treated with a single injection of 7-mg tacrolimus, rejected with an MST of 33.5 days. (iii) No signs of rejection were seen in the long-term survival group IV at day 100. (D and E) Representative photomicrographs of the histology (hematoxylin and eosin staining) of skin (D) and gastrocnemic muscle (E) of normal rats, no treatment, TGMS-treated, tacrolimus only–treated, and TGMS-TAC–treated groups. Rejected grafts showed severe cell infiltrations, edema formation, and necrosis. Magnification, ×50; scale bars, 400 μm.


In groups I and II (no treatment and TGMS control), severe disruption of the tissue architecture of both skin and muscle was observed at the time of rejection, accompanied by mononuclear cell infiltration and massive edema formation, with no difference between the two groups (Fig. 2, D and E). In the dermis, particularly the perivascular regions were massively infiltrated by mononuclear cells. Although histopathological changes were less severe in group III (tacrolimus only), infiltrates of mononuclear cells and edema formation were also evident. No signs of rejection and minimal mononuclear cell infiltration were observed in grafts of group IV (TGMS-TAC), analyzed at POD 100.

Allospecific antibody

We compared the tissue deposition of allospecific antibodies between groups III (tacrolimus only) and IV (TGMS-TAC), and normal rats were used as control. Deposition of immunoglobulin G (IgG) was significantly increased in both skin and muscle tissues of group III compared with normal rats and with group IV [n = 4 to 6; P < 0.01, analysis of variance (ANOVA) with Bonferroni posttest; Fig. 3, A and B, and table S6]. Muscle and skin tissues from group IV rats showed no increased IgG deposition as compared with normal rats. Deposition of IgM in skin was significantly increased in group III compared with healthy rats (n = 4 to 6; P < 0.05, ANOVA with Bonferroni posttest; Fig. 3C and table S6). In contrast, no or minimal skin deposition of IgM was observed in group IV. In muscle, no significant deposition of IgM was found in groups III and IV as compared with normal rats (Fig. 3D and table S6). Detection of Brown Norway–specific IgG and IgM antibody levels in plasma showed that anti-donor antibodies slightly increased over time in both groups III and IV. In group III, IgG and IgM levels were increased by 16.1 and 14.5%, respectively, at the time of rejection compared with baseline (Fig. 3, E and F, and tables S7 and S8). In group IV, IgG and IgM levels were increased by 18.3 and 4.1%, respectively, at POD 100 compared with baseline. However, the levels of donor-specific IgG and IgM were not significantly different between these two groups.

Fig. 3. Humoral immune response.

Allografts were immunostained for deposition of IgG and IgM. (A and C) Deposition of IgG (A) and IgM (C) in skin tissue. (B and D) Deposition of IgG (B) and IgM (D) in muscle tissue. Quantitative analysis of fluorescence intensity was performed by ImageJ software (n = 4 to 6; *P < 0.05, **P < 0.01, ***P < 0.001, ANOVA/Bonferroni). (E and F) Plasma from TAC- and TGMS-TAC–treated groups was collected at the indicated time points. Splenocytes from Brown Norway rats were incubated with the plasma at 1:4 dilution and stained with fluorescein isothiocyanate (FITC)–conjugated anti-rat IgG and phycoerythrin (PE)–conjugated anti-rat IgM, respectively, and the fluorescence was measured in a 96-well plate reader. Fluorescence intensities of plasma IgG (E) and IgM (F) alloantibody levels are shown. Data are represented as means ± SD. Magnification, ×50; scale bars, 500 μm. Raw data are given in tables S6 to S8.

Complement deposition and cytokines

Complement C3c deposition was significantly increased in the skin samples of group III compared with healthy controls and group IV. Only minimal staining for C3c was observed in group IV (Fig. 4, A and B, and table S9). We could not detect C3c deposition in muscle tissue in any of the groups. In hand transplant recipients, the cytokines interleukin-2 (IL-2), tumor necrosis factor–α (TNF-α), and interferon-γ (IFN-γ) have been suggested as indicators of an anti-graft immune response (24). This prompted us to measure the levels of IL-2, TNF-α, IFN-γ, and IL-1β in this study. The latter cytokine was included because of its known importance in allograft rejection (25). Levels of IL-2, TNF-α, IFN-γ, and IL-1β were increased in all three analyzed groups (II, III, and IV) on POD 7 (Fig. 4C and table S10). During the later follow-up, levels decreased or remained constant both in group III, with an MST of 33.5 days, and in group IV, with an MST of >100 days.

Fig. 4. Cytokines and complement activation.

(A and B) Allografts were immunostained for deposition of the complement activation product C3c. Deposition of C3c in skin and muscle is shown. C3c deposition was significantly increased as compared with normal tissue only in the skin of the tacrolimus only–treated grafts, which were rejected with an MST of 33.5 days (n = 4 to 6; **P < 0.01 and ***P < 0.001, ANOVA/Bonferroni). Raw data are given in table S9. Quantitative analysis of fluorescence intensity was performed by ImageJ software. (C) Levels of IL-2, TNF-α, IFN-γ, and IL-1β were analyzed at the indicated time points using the Bio-Plex suspension array system. Data are means ± SD. Magnification, ×50; scale bars, 500 μm. Raw data are given in table S10.

Plasma and tissue levels of tacrolimus

Plasma levels of tacrolimus (Fig. 5A and table S11) were compared between groups III (tacrolimus only), IV (TGMS-TAC), and V (TGMS-TAC, contralateral). Administration of 7 mg of tacrolimus subcutaneously on POD 1 resulted in a plasma level of 210.1 ± 17.7 ng/ml on POD 3 in group III. In group IV, the peak tacrolimus levels were significantly lower (127.2 ± 54.98 ng/ml) than in group III (n = 5; P < 0.05, ANOVA with Bonferroni posttest). Plasma levels of tacrolimus rapidly declined over time in group III to reach undetectable levels at the time of graft rejection (PODs 31 to 62). In contrast, tacrolimus levels decreased much slower in group IV. At PODs 11 and 21, they were significantly higher than in group III (n = 4 to 6; P < 0.05 and P < 0.001, respectively, ANOVA with Bonferroni posttest) and then gradually fell to reach 0.38 ± 0.08 ng/ml at POD 100. Administration of TGMS-TAC to the contralateral leg (group V) also resulted in a rapid decrease of tacrolimus plasma levels from 236.2 ± 199.5 ng/ml on POD 3 to 3.66 ± 1.48 ng/ml at the time of graft rejection or termination of the experiment, that is, PODs 35 to 100. There was no significant difference in total tacrolimus release between groups as estimated by area under the curve using ANOVA (n = 3 to 6).

Skin tacrolimus concentrations (Fig. 5B and table S11) in transplanted limbs were compared between groups III, IV, and V. Similar to plasma levels, tissue levels were highest at POD 7 and then rapidly decreased in group III, but not in group IV (TGMS-TAC). Injection of TGMS-TAC in the contralateral leg (group V) led to significantly lower initial skin levels of tacrolimus (n = 4; P < 0.05, group V versus groups III and IV, POD 7, ANOVA with Bonferroni posttest). At rejection, skin concentrations of tacrolimus were less than 20 ng/g of protein in group III (5.9 ± 2.8 ng/g; MST, 33.5 days) and group V (10.5 ± 5.4 ng/g; MST, 75 days). In contrast, the long-term surviving grafts, group IV, had skin tacrolimus levels of 87.0 ± 91.5 ng/g at POD 100, at which time the experiments were terminated.

Fig. 5. Plasma and skin levels of tacrolimus.

(A) Plasma levels of tacrolimus were measured for group III (TAC), group IV (TGMS-TAC), and group V (TGMS-TAC, contralateral). Tacrolimus levels were significantly higher on POD 3 and then significantly lower on PODs 11 and 21 in group III compared with group IV (n = 5; *P < 0.05, ANOVA/Bonferroni). PODs 3, 7, and 11 are depicted as column graphs in the insert. In group IV, tacrolimus levels were detectable up to POD 100. No tacrolimus could be detected at rejection (Rej) in groups III (#) and V (§). Plasma tacrolimus levels declined rapidly in groups III and V. Data are means ± SD. (B) Tissue levels of tacrolimus in the transplanted limb were elevated in group III (TAC) on POD 3 compared to group IV. Subsequently, these levels were reduced at POD 21 and at rejection. In group IV, tacrolimus levels were maintained constant at PODs 3, 21, and 100. Tissue levels of tacrolimus were lower in group V from POD 3 to graft rejection. Data are mean values. Raw data are given in table S11.

Assessment of plasma markers for kidney and liver damage

We measured creatinine and blood urea nitrogen levels as indicators of kidney damage due to the use of tacrolimus. For liver damage, the markers aspartate aminotransferase and alanine aminotransferase were measured. Plasma samples were taken on PODs 11 and 35 from groups III (TAC) and IV (TGMS-TAC) rats. All observed values were within the normal range expected for rats (26), and no statistically significant differences were found between groups or within the groups for the two different time points (fig. S1).

Resorption of TGMS-TAC hydrogel

To assess the persistence of the hydrogel, we palpated the hydrogel depots throughout the lifetime of the grafts. Postsurgical edema of the allograft disappeared after 10 to 14 days. After this period, the hydrogel depot was evident and palpable (Fig. 6A, data from POD 21). In groups IV and V, disappearance of the hydrogel depots were 60 ± 5 days and 40 ± 4 days, respectively, as assessed by palpation (Fig. 6B and table S12). Resorption of TGMS-TAC hydrogel injected in the transplanted limb was thus significantly slower compared with injection into the contralateral limb (n = 4 to 6; P = 0.0266, t test).

Using ultrasonography, we also measured the size of the hydrogel depots at POD 21. In the transplanted limb, the average volume of the hydrogel was significantly larger compared to the contralateral side (group IV versus group V; n = 5; P = 0.0058, t test). The average volumes of the hydrogel in the transplanted and the contralateral limbs were 0.050 ± 0.027 cm3 and 0.004 ± 0.003 cm3, respectively (Fig. 6, C and D, and table S12).

Fig. 6. Assessment of hydrogel drug depot.

(A) Representative pictures of formation of hydrogel drug depots at POD 21 in the contralateral and transplanted limbs. Arrows indicate the subcutaneous tacrolimus drug depots. (B) TGMS-TAC depots remained palpable for a mean time of 60 ± 13 days in the transplanted limb, compared with 40 ± 8 days in the contralateral limb (n = 4 to 6; P = 0.0266, t test). (C and D) The volume of the drug depots as assessed by ultrasonography was significantly larger in the transplanted as compared with the contralateral limb at POD 21 (n = 5; P = 0.0058, t test). Raw data are given in table S12.


Currently, adverse side effects of immunosuppressive drugs are a major concern in transplantation medicine. In VCA, as a non–life-saving transplantation, the use of immunosuppressive drugs is highly controversial and a crucial factor currently impeding the widespread use of VCA. Therefore, many strategies such as steroid-free maintenance immunosuppression with tacrolimus and MMF have been tested. However, because acute rejection episodes reached 85% during the first 2 years (2), treatment with steroids has been reintroduced in these patients (4). Tolerance induction protocols for use in VCA may be a promising option for the future, but they are still at an early stage of development. Meanwhile, current immunosuppressive protocols need to be improved, aiming at a better therapeutic window and minimizing side effects to encourage better patient compliance.

It has also been demonstrated that long-term effective immunosuppression can be achieved with biodegradable tacrolimus pellets in spinal trauma–injured Sprague-Dawley rats injected with human neural precursor cells (27). Self-assembling hydrogels are particularly interesting as drug delivery system because they have features such as encapsulation of large amounts of hydrophobic drugs, biodegradability, sustained release, and flexibility to engineer stimuli-responsive release (28). This prompted us to evaluate an enzyme-responsive injectable drug delivery system based on self-assembling hydrogels in VCA, and also because we envisioned that the microenvironment within the allograft might favorably influence the release of encapsulated drugs.

Our data demonstrate that injection of hydrogel-encapsulated tacrolimus 1 day after transplantation prolongs allograft survival to more than 100 days. Previously, it has been shown that Lewis recipients transplanted with a Brown Norway limb receiving a 14- or 90-day course of tacrolimus at 6 mg/kg per day had mean graft survival times of 28 and >90 days, respectively (29). In the latter study, it was presumed that an average-sized recipient of 300 g would have received about 25 mg of tacrolimus at the end of a 14-day period or 160 mg at the end of a 90-day treatment period. In our protocol, 7 mg of tacrolimus encapsulated in TGMS as a one-time injection prevented allograft rejection for >100 days, suggesting extended release of the drug in a concentration able to prevent immune activation. Macroscopic appearance of the allografts that survived >100 days showed no evidence of rejection. Histological evaluation of rejected grafts showed severe tissue damage and cell infiltration in both muscles and skin, whereas we observed only mild cell infiltration with otherwise normal histology in both muscle and skin of the >100-day survival group. This is in sharp contrast to earlier reports, showing that cessation of tacrolimus after POD 50 in a Brown Norway–to–Lewis transplantation model resulted in rejection at POD 60 with a dense infiltration of cells and necrosis in both epidermis and muscle (30).

We confirm data previously reported by Unadkat et al. that allospecific antibodies are elevated in a Brown Norway–to–Lewis rat transplantation model (31). In our experiments, we found tissue deposition of allospecific IgM, IgG, and complement C3c. The role of the humoral immune response is so far poorly understood in VCA because detection of complement products and donor-specific antibodies has rarely been reported (1). However, our data show activation of the humoral immune system in the single-dose, nonencapsulated tacrolimus–treated group (group III). The observed binding pattern of antibodies and complement is in line with the fact that skin is the most antigenic part of the allograft (30). With respect to cytokine profiles, we could not observe the changes that were reported earlier in hand transplant recipients, namely, that IL-2, TNF-α, and IFN-γ levels were decreased initially and then returned to pretransplant levels, followed by a further drop (24). On the contrary, we observed elevated levels of IL-2, TNF-α, and IFN-γ at POD 7, which then started to decrease. No significant correlation was found between the analyzed cytokines and graft rejection. It has been proposed that intragraft cytokine and chemokine expression profiles, rather than the systemic cytokine profiles, should be used as a diagnostic tool to predict graft dysfunction (32), suggesting a necessity to investigate intragraft cytokine profile in VCA. It was indeed shown that intragraft cytokine expression patterns were significantly influenced by topical immunosuppression in a rat face transplantation model (33).

In the ipsilateral TGMS-TAC group, tissue levels of the drug were increased 10-fold over systemic concentrations and 12 times higher at POD 100 compared with the groups treated with tacrolimus alone or TGMS-TAC given on the contralateral side. In an earlier study, topical application of tacrolimus resulted in prolonged allograft survival with a higher concentration in skin compared with peripheral blood in a rat hindlimb transplantation model (16). Hence, locally high concentrations of tacrolimus may play a role in prolonged allograft survival when delivered by the hydrogel system. Hydrogels have been used in many studies to deliver drugs efficiently, particularly locally, to reduce the side effects of systemic administration (34, 35). It has been reported that doxorubicin-loaded hydrogel can be used as postsurgical therapy for solid tumors (36) and that anastomotic intimal hyperplasia due to both autologous vein and synthetic vascular grafting can be treated with a cyclosporine-loaded poly(ethylene glycol) hydrogel system (37). However, the use of hydrogels as a potent drug delivery system in transplantation has been poorly investigated.

In vitro, TGMS-TAC hydrogels released the drug predominantly in response to proteolytic enzymes and cell culture supernatants from activated macrophages. The observed in vivo drug release profile suggests that TGMS-TAC hydrogel released the drug at high concentrations up to POD 11, akin to injection of nonencapsulated tacrolimus, but then maintained a significantly higher concentration than seen in the tacrolimus group. The most obvious reason for this observation is the inflammatory reaction during the early PODs, caused by surgical trauma and ischemia/reperfusion injury. Analogous to the in vitro situation with supernatant of activated macrophages, this leads to drug release in high concentrations, whereas later on, with diminishing inflammation, tacrolimus is released at a much lower concentration, and TGMS-TAC acts like a sustained drug release system, prolonging VCA graft survival to more than 100 days. That graft survival was shorter when TGMS-TAC was injected in the contralateral leg as compared to the transplanted leg fits with the lower tissue concentrations of tacrolimus that were found in the contralateral injection group. The reason for the quicker degradation of the TGMS-TAC depots in the contralateral legs remains unclear at the moment. Swelling in the transplanted limb and impairment of lymphatic circulation, as recently described in patients after hand transplantation (4), could possibly favor preservation of the drug depots.

The development of a single-dose sustained-release drug delivery system is also important for better patient compliance. It has been reported that noncompliance with immunosuppressive protocols in hand transplantation (2) accounts for a significant percentage of failure in transplantation, suggesting that this is a critical factor influencing long-term graft survival (38). Our study showed that a one-time injection of tacrolimus-encapsulated TGMS hydrogel was able to maintain effective immunosuppression for >100 days. The use of TGMS-TAC, injected subcutaneously in the transplanted limb, may thus also represent a therapeutic strategy to increase patient compliance during maintenance treatment.

In our rat model, tacrolimus monotherapy is known to prevent rejection of the transplanted limb. This is not the standard treatment in the clinical situation with humans, who usually receive triple immunosuppression therapy with tacrolimus or sirolimus, MMF, and prednisone (4). The rats are also hosted in individually ventilated cages with the aim to prevent infection by microorganisms. Because it is known that infections can trigger rejection reactions, it remains to be proven that the local, enzyme-responsive drug delivery system will indeed be able to abrogate rejection episodes that are, for example, triggered by viral infections.

In conclusion, our study demonstrates the successful use of self-assembled hydrogel-based immunosuppressive drug delivery in a transplantation model. This approach suggests that local delivery of the drug to the allograft may improve therapeutic outcomes and lead to a paradigm shift in clinical immunosuppressive therapy in VCA.


Study design

In vitro. We initially examined the chemical structures of hundreds of agents from the GRAS list by the U.S. FDA to identify an amphiphilic molecule with an enzyme-cleavable bond. After its identification, TGMS was tested for its ability to self-assemble into enzyme-responsive hydrogels and encapsulate tacrolimus.

In vivo. A series of hindlimb transplantation experiments with rats were designed to test the hypothesis that local immunosuppression by one-time injection of TGMS-TAC will be able to prevent rejection of the transplanted limb for more than 100 days.

Preparation of hydrogels

To examine encapsulation, typically, solvents [800 μl of water and 200 μl of dimethyl sulfoxide (DMSO)] were added to a glass scintillation vial with TGMS (10%, w/v) and tacrolimus (7 mg) and sealed with a screw cap. The vial was heated to 60° to 80°C until TGMS and tacrolimus were dissolved. The vial was placed on a stable surface and allowed to cool to room temperature. After 15 to 45 min, the solution transitioned into a viscous gel. Gelation was considered to have occurred when no gravitational flow was observed upon inversion of the glass vial. The resulting hydrogels were easily injectable through 20-gauge needles.

Enzyme-responsive drug release

TGMS-TAC hydrogel (200 μl) (tacrolimus, 7 mg/ml) samples were placed in dialysis tubing (10-kD molecular weight cutoff, Spectrum Labs) and suspended in PBS (800 μl). To these samples, PBS (100 μl, control), cell culture supernatants (100 μl), lipase (10,000 U in 100 μl), and MMP-2/9 (100 ng/ml, 100 μl) were added. The tubes were then closed, placed in a PBS reservoir (30 ml), and incubated at 37°C with a shaking speed of 150 rpm. At each time point, an aliquot (1 ml) from the PBS reservoir was removed and lyophilized, the solid was dissolved in 100 μl of DMSO, and the concentration of tacrolimus was analyzed by HPLC. After withdrawing each aliquot, the incubation medium was replenished with PBS (1 ml). The concentration of tacrolimus was determined in HPLC and a μBondapak C18 column (3.9 mm in internal diameter and 15 cm in length) and maintained at 60°C, and the eluents were monitored at 214 nm. The mobile phase (80% methanol and 20% water, acidified to pH 6 with HCl) was used at a flow rate of 1.5 ml/min, and the retention time for tacrolimus was 4.8 min.

Cell culture

RAW 264.7 murine macrophages (American Type Culture Collection) were maintained on plastic petri dishes at 37°C and 5% CO2 in culture medium consisting of RPMI 1640 supplemented with 5% heat-inactivated fetal bovine serum, 1 mM glutamine, penicillin/streptomycin/fungizone, 10 mM Hepes buffer, 100 μM nonessential amino acids, and 2.5 × 10–5 M 2-mercaptoethanol (Invitrogen Corp.). Cells were passaged every 3 days. For preparation of culture supernatant, cells at passages 3 to 5 were seeded, medium was changed on day 2, and LPS (from Escherichia coli 0111:B4, Sigma-Aldrich) was added at 100 ng/ml. After 24 hours, culture supernatant was collected and frozen at –20°C.

Rat hindlimb transplantation

Inbred male Brown Norway (RT1n) and Lewis (RT1l) rats, 8 to 12 weeks of age, were purchased from Charles River. Rat hindlimb transplantation was performed as described previously with several modifications (39). Briefly, buprenorphine at 50 μg/kg was given as a preemptive analgesic followed by 5% isoflurane in pure oxygen inhalation anesthesia for the induction and 1.5% for maintenance. Donors (Brown Norway) received 300 IU of heparin intravenously followed by hindlimb amputation at the midfemur level. Recipient (Lewis) rats were prepared by amputating the corresponding hindlimb. Osteosynthesis was performed with a blunted 18-gauge needle. The femoral artery was anastomosed in an end-to-end technique with 10/0 interrupted sutures, and the vein anastomosis was performed by a cuff technique using a polyimide tube (RiverTech Medical) as described previously (40). After ensuring adequate vascularization of the transplanted limb, the femoral and sciatic nerves were anastomosed with interrupted 10/0 sutures followed by muscle and skin suturing with 4/0 resorbable sutures. All experiments in this study were performed according to current Swiss Laws on Animal Protection.

Treatment groups and drug administration

Animals were randomly divided into five groups (Fig. 1B). In group I, recipients were left untreated. In group II, 1 ml of TGMS as a vehicle control was injected into the transplanted limb. Group III was injected with 7 mg of tacrolimus, dissolved in 1 ml of sterile 20% DMSO in saline, into the transplanted limb on POD 1. Groups IV and V were injected with tacrolimus-encapsulated TGMS hydrogels into the transplanted and the contralateral limbs, respectively. In groups II to V, a total volume of 1 ml was divided into four aliquots of 0.25 ml and injected subcutaneously at four sites around the limb at POD 1.

Histology and immunofluorescence staining

Tissue samples from the grafts, retrieved at the end of the experiments, were preserved in 4% buffered formaldehyde and stained with hematoxylin and eosin, whereas acetone-fixed cryosections were used for immunofluorescence staining. Anti-C3b/c (Dako) primary antibody was used to detect C3 deposition in the tissue. Goat anti-rat IgG FITC (Southern Biotechnology Associated) and goat anti-rat IgM Alexa 674 (Jackson ImmunoResearch) were used to detect tissue deposition of IgG and IgM, respectively. Images were acquired with a fluorescence microscope (Leica DMI4000B), and all images were captured with identical exposure times and settings in each experiment. ImageJ software ( was used to quantify fluorescence intensities.

Detection of plasma levels of alloantibody

To analyze donor-specific alloantibody responses, EDTA plasma was collected at PODs 3, 7, 11, 21, 35, 49, 63, 77, 94, and 100 from recipients. Aliquots containing 1 × 106 splenocytes from DA rats (donor strain) were incubated for 60 min on ice with diluted Lewis recipient plasma (1:4). The washed cells were then incubated for 30 min on ice with goat anti-rat IgG FITC diluted 1:100 (Southern Biotechnology Associated) and goat anti-rat IgM PE diluted 1:100 (Jackson ImmunoResearch). Washed cells were resuspended and transferred to a 96-well plate. Fluorescence levels of the cells were analyzed using a TECAN Infinite M1000 reader. Then, 20 μl of PrestoBlue Cell Viability Reagent (Invitrogen) was added to each well and incubated for an additional 20 min at 37°C before fluorescence at 586 nm was measured. Ratios of fluorescence readings for antibodies and PrestoBlue were used to assess the amount of IgG and IgM binding to donor cells.

Plasma cytokine levels

Levels of IL-2, IL-1β, TNF-α, and IFN-γ in the plasma collected at PODs 3, 7, 11, 21, 35, 49, 63, 77, 94, and 100 were analyzed using a rat cytokine multiplex panel according to the manufacturer’s instructions (Bio-Rad). Briefly, plasma samples were incubated with magnetic beads conjugated with the respective antibodies for 30 min. After a washing step, beads were incubated with detection antibodies for 30 min. After another washing step, magnetic beads were incubated with streptavidin-PE (Qiagen) for a further 10 min. After the final washing step, concentrations were measured using a Bio-Plex 100 system (Bio-Rad).

Detection of plasma and skin levels of tacrolimus

EDTA plasma samples were collected at PODs 3, 7, 11, 21, 35, 49, 63, 77, 94, and 100 and stored at –80°C until analysis. After shaving the fur, skin was cleaned with 70% alcohol, and about 2 × 2–mm skin biopsies were excised from the allograft under general anesthesia using surgical scissors. Skin biopsies were homogenized using a Qiagen TissueLyser II. Briefly, 100 μl of radioimmunoprecipitation assay buffer with a protease inhibitor cocktail (Sigma) was added to the tubes and homogenized for 2 min at 30 Hz. Tubes were centrifuged at 10,000 rpm for 10 min at 4°C. Supernatants were collected, and protein concentration was measured using a Bio-Rad DC Protein Assay kit. Consequently, 50 μl of EDTA plasma or 250 μg of protein extract from the skin biopsies was analyzed using a PRO-Trac II Tacrolimus ELISA kit (DiaSorin) according to the manufacturer’s instructions. Tacrolimus concentrations in skin were expressed as nanograms per gram of total protein.

Evaluation and monitoring of drug depot

In vivo formation of drug depot was detected using palpation started at POD 14 and then weekly. A negative palpation of the entire drug depot was defined as complete resorption of TGMS-TAC. In addition, the size of the hydrogel depot was measured using a real-time ultrasound scanner (ATL 3500, Philips Medical Systems) with a 3- to 12-MHz linear array transducer in the small parts mode.

Statistical analysis

Statistical analysis was performed using the GraphPad Prism version 5 program (GraphPad Software). The results are expressed as means ± SD. Survival of the allografts was examined using Kaplan-Meier analysis, and groups were compared using the log-rank test. Data were analyzed using Student’s t test for single pairwise comparisons or one-way ANOVA, and Bonferroni post hoc test for multiple comparisons. Significance was defined as P < 0.05.


Fig. S1. Determination of plasma markers for kidney and liver damage.

Table S1. Transformation of TGMS-encapsulated tacrolimus from gel to liquid phase.

Table S2. Statistical analysis of enzyme-responsive tacrolimus release (raw data of Fig. 1C).

Table S3. Enzyme-responsive tacrolimus release (raw data of Fig. 1D).

Table S4. Tacrolimus release in response to conditioned medium from activated macrophages (raw data of Fig. 1F).

Table S5. Vascular composite allograft survival (raw data of Fig. 2A).

Table S6. Fluorescence intensities measured for IgG and IgM staining (raw data of Fig. 3, A to D).

Table S7. Plasma levels of anti-donor IgG (raw data of Fig. 3E).

Table S8. Plasma levels of anti-donor IgM (raw data of Fig. 3F).

Table S9. Fluorescence intensities measured for C3c deposition in skin and muscle tissues (raw data of Fig. 4, A and B).

Table S10. Quantitative analysis of cytokines in plasma in pg/ml (raw data of Fig. 4C).

Table S11. Plasma (ng/ml) and tissue (ng/g) levels of tacrolimus (raw data of Fig. 5, A and B).

Table S12. Palpability and volume of hydrogel depots (raw data of Fig. 6, A and D).


  1. Acknowledgments: We thank V. Gorantla and J. Schnider (University of Pittsburgh Medical Center) for help with the establishment of the rat hindlimb transplantation model and for helpful discussions, and J. Denoyelle for technical support. The gift of polyamide cuffs by RiverTech Medical is acknowledged. Images were acquired on equipment supported by the Microscopy Imaging Center of the University of Bern. We thank B. Leuenberger and his team at the Institute of Pathology, University of Bern, for preparation of the histological slides and M. Fiedler and his team for analysis of kidney and liver damage markers. A.D. thanks the Indian University Grants Commission for Junior Research Fellowship. P.K.V. thanks the Department of Biotechnology, India, for Ramalingaswami Re-Entry Fellowship and the Central Imaging and Flow Cytometry Facility (CIFF) at National Centre for Biological Sciences. Funding: Supported by a grant from the American Society for Surgery of the Hand and the Olga Mayenfisch Foundation. Additional grant support was received from the Julia Bangerter-Rhyner Foundation and a Harvard Institute of Translational Immunology/Helmsley Trust Pilot grant to J.M.K. and an NIH grant DE023432 to J.M.K. Author contributions: T.G. and R.O. designed and performed animal experiments, analyzed the respective data, and wrote the manuscript; F.M.L. designed and performed limb transplantation experiments; A.D. performed in vitro experiments; Z.Y. designed the research and made the link between the groups in Bern and Boston/Bangalore; A.K.B. designed and performed analyses of rat plasma samples; Y.B. analyzed the histology; M.A.C. designed and supervised the transplantation experiments; J.M.K. designed and supervised the research related to TGMS-TAC hydrogels, analyzed the respective data, and wrote the manuscript; P.K.V. designed the research related to TGMS-TAC hydrogels, analyzed the respective data, and wrote the manuscript; R.R. carried the overall responsibility for the project, designed the research, analyzed the data, and wrote the manuscript; and E.V. designed the research, analyzed the data, and wrote the manuscript. Competing interests: J.M.K. is a founder of, and has a financial interest in, Skintifique, a company that is developing self-assembled hydrogels for skin applications. J.M.K.’s laboratory developed the hydrogel technology, and it was licensed by the Brigham and Women’s Hospital to Skintifique. J.M.K. holds patent no. 20110229565, entitled “Drug delivery composition comprising a self-assembled gelator,” filed 22 September 2011. J.M.K.’s interests were reviewed and are managed by the Brigham and Women’s Hospital and Partners HealthCare in accordance with their conflict of interest polices. The other co-authors declare no competing interests.
View Abstract

Stay Connected to Science Translational Medicine

Navigate This Article