The aim of this study was to investigate liver microvascular adaptation following the intraportal infusion of pancreatic islets (pancreatic islet transplantation [islet-tx]) in diabetic patients using dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI). DCE-MRI was performed before and 7 days after islet-tx in six diabetic patients. Initial area under curve (AUC60) and volume transfer coefficient (Ktrans) were assessed as markers of liver perfusion. Clinical and metabolic monthly follow-up was performed in all patients, considering fasting C-peptide and β-score as main indices of graft function. High variability in the response of liver microvasculature to islet infusion was observed: two patients showed a significant reduction in liver perfusion after transplantation (pt.2: AUC60 = −23.4%, Ktrans = −31.7%; pt.4: AUC60 = −23.7%, Ktrans = −27.9%); three patients did not show any significant variation of liver perfusion and one patient showed a significant increase (pt.3: AUC60 = +31%, Ktrans = +42.8%). Interestingly, a correlation between DCE-MRI parameters and indices of graft function was observed and, in particular, both patients with DCE-MRI evidence of posttransplantation liver perfusion reduction experienced premature graft failure. Our preliminary study demonstrated that DCE-MRI may identify different adaptive responses of liver microvasculature in patients submitted to islet-tx. These different responses could have an impact on islet engraftment, although reported findings need confirmation from larger studies.
initial area under curve
dynamic contrast-enhanced magnetic resonance imaging
exogenous insulin requirement
instant blood-mediated inflammatory reaction
pancreatic islet transplantation
volume transfer coefficient
magnetic resonance imaging
type 1 diabetes mellitus
regions of interest
Pancreatic islet transplantation (islet-tx) is an appealing therapeutic approach for the treatment of Type 1 diabetes mellitus (T1DM), based on percutaneous intraportal infusion of pancreatic islets obtained from organ donors . Islet-tx demonstrated an improvement in glycometabolic control [2-6] and to allow insulin independence with rates that reach 44% at 3 years after transplantation . A major limitation of this procedure, which hampers the widespread clinical application of islet-tx, is the need for two to four donor pancreases per recipient to compensate for the loss of islet viability occurring during isolation and engraftment [8-10]. Multiple factors such as inflammatory, hypoxic and immune damage lead to the loss of at least 50% of infused islets in the earliest phases after transplantation [11-13], and a previous study demonstrated that the improvement of engraftment may allow for a successful single-donor, marginal-dose, islet transplantation . Among the factors responsible for this important early islet loss, hypoxia-induced cell death surely plays a critical role. Islets have a rich self-regulated microvascular network that guarantees they will receive 20–25% of the total pancreatic blood supply, even if they represent only 1–2% of the pancreatic mass [15, 16]. This microvascular network is disrupted during isolation and the proportion of islets that restore their vasculature after transplantation is limited and variable [17, 18]. Moreover, the microvascular damage induced during isolation and preparation leads islets to express the tissue factor that activates the pro-coagulant pathways determining the so-called instant blood-mediated inflammatory reaction (IBMIR) immediately after intraportal infusion [19, 20]. Therefore, a key factor influencing the success of the engraftment could be the balance between the neoangiogenetic process and the IBMIR occurring in the recipient's liver.
In the setting of islet-tx, magnetic resonance imaging (MRI) was used to identify periportal steatosis that may occur in some patients as a result of high local insulin concentration  and, above all, has been applied for in vivo detection and monitoring of iron-labeled transplanted islets [22, 23]. Despite the high value of the MRI in the functional assessment of the microvasculature, it has never been applied to patients submitted to islet-tx for investigating the critical microvascular phenomena characterizing islet engraftment. Dynamic contrast-enhanced MRI (DCE-MRI) is an accepted tool for the quantitative assessment of microvascular structure and function in oncology based on the analysis of the dynamic distribution of contrast media within tissues, which is influenced by different features of the microvascular bed . Chan et al  proposed DCE-MRI as a tool for the noninvasive assessment of islet revascularization in a mouse model of intrahepatic islet-tx for the first time.
Our hypothesis is that DCE-MRI may provide novel information about microvascular changes characterizing the engraftment of islet-tx. In this study, we describe the changes in liver perfusion observed by DCE-MRI in T1DM patients submitted to islet-tx, looking for a potential association between perfusion parameters and biochemical and clinical outcome.
Materials and Methods
This study was approved by the Institutional Review Board (7/09/2009 EFSD/MSD ISLET MRI), and written informed consent was obtained from all patients. Six diabetic patients on the waiting list for islet-tx alone were enrolled. The patients enrolled (3 male; 3 female; mean age 46 ± 12.4 years) suffered from T1DM for 27 ± 8 years; all had inadequate glycometabolic control and frequent episodes of hypoglycemic unawareness. Exclusion criteria for the study were pregnancy, hypersensitivity to contrast agents, GFR ≤30 mL/min/1.73 m2, severe claustrophobia, the presence of pacemakers or implantable cardioverter-defibrillators and liver disease.
All patients underwent transplantation through intraportal vein infusion. Features of infused islet preparations are reported in Table 1. All patients underwent DCE-MRI immediately before islet infusion and 7 days after transplantation.
|Pt.||Age||Diabetes duration (years)||Retinopathy||Peripheral neuropathy||Nephropathy||Tx n.||IEQ||NA||Time culture (h)||Purification degree (%)||Infused volume (mL)||ΔPortal pressure (mmHg)|
|1||68||33||No||Yes||No||1||390 505||382 500||24||48||3||n.a.|
|2||248 425||207 000||22||50||2||−5|
|2||52||39||Yes||Yes||No||1||327 400||251 500||22||55||1.5||0|
|3||43||23||Yes||Yes||No||1||894 135||1 017 583||52||79.4||6||1|
|2||404 647||528 250||0||45||6||2|
|4||39||31||Yes||No||No||1||555 491||497 841||23||70||4||2|
|5||33||23||Yes||Yes||No||1||307 508||415 000||21||30||4||7|
|2||295 575||620 750||18||40||2.5||0|
|6||41||15||Yes||No||No||1||329 300||631 750||19||30||4.5||3|
Islet isolation, infusion and immunosuppression
Pancreatic islet preparations were obtained according to the Ricordi method . Intraportal islet infusion was performed in an angiographic suite using combined ultrasound (US) and fluoroscopic guidance: a US-guided puncture of the right portal vein branch allowed the placement of a straight-end 4-French catheter in the main portal vein, over a 0.018-in. guidewire. Through this catheter, islets were infused after a portal vein tree angiographic study. The suspension of purified islets was slowly infused (20 min) together with 2000 IU of heparin. Portography and portal vein pressure recording were performed before and after the infusion. Liver US was repeated at the end of the procedure and 24 h later to assess for potential complications. Prophylactic anticoagulation started at the time of infusion and was continued for 1 week with subcutaneous enoxaparin (2850 UI) twice daily. With the aim of preventing graft rejection, immunosuppression was based on Basiliximab 20 mg on days 0 and 4, plus Sirolimus (trough levels: 12–15 ng/mL during 1–3 months, and 10–12 ng/mL thereafter) and Tacrolimus (trough levels 4–6 ng/mL).
MRI protocol and analysis
Images were acquired on a 1.5 T MR scanner (Achieva-Nova, Philips Medical Systems, Best, the Netherlands) using a 4-channel phased-array coil positioned on the superior abdomen. DCE-MRI studies were performed immediately before (within 24 h) and 7 days after islet-tx, with a TRIVE-3D-FFE sequence acquired on the axial-plane encompassing the entire liver (45 dynamics; temporal resolution: 4 s; TFE-factor: 60, TR: shortest, TE: shortest, flip angle:10°, voxel size: 2.05 × 2.47 × 4 mm) during automatic injection of a contrast bolus at a flow rate of 3.5 mL/s. The contrast media used were gadofosveset-trisodium (Vasovist®, Bayer-Schering Pharma, Berlin, Germany) at a dosage of 0.1 mL/kg in pt.1, pt.2, pt.3 and gadoxetic-acid-disodium (Primovist®, Bayer-Schering Pharma, Berlin, Germany) at a dosage of 0.1 mL/kg in pt.4, pt.5 and pt.6. All patients were asked to breathe slowly and smoothly during the dynamic scan.
DCE-MRI analyses were performed using NordicICE software (NordicImagingLab, Bergen, Norway): initial area under curve taken up to 60 s (AUC60) and volume transfer coefficient (Ktrans) pixel-by-pixel maps were obtained for the entire liver. Mean AUC60 and Ktrans were measured in each liver segment drawing regions of interest, accurately avoiding the inclusion of large intrahepatic vessel branches. Mean values referred to the entire liver were achieved from the DCE-MRI parameters of each segment. Relative changes of AUC60 and Ktrans from pretransplantation to 7 days after islet-tx were calculated as follows:
Graft function assessment
Biochemical evaluation including blood glucose, C-peptide and glycated hemoglobin (HbA1c) was obtained pretransplantation, during the first week after transplantation, and every month thereafter. The exogenous insulin requirement (EIR) was monitored monthly after the first month. Based on these data, β-score was calculated at 1, 3 and 6 months after islet-tx . Fasting C-peptide levels and β-score measured at 3 and 6 months were considered as markers of mid-term global graft function, since after the third month, β-score value is independent from pretransplantation glycemic control. In pts.1 and 2, only 3-month graft function parameters were considered because pt.1 received a second infusion and pt.2 interrupted the clinical follow-up because of complete graft failure.
For the evaluation of the inflammatory status and early liver response to islets infusion aspartate transaminase (AST), alanine transaminase (ALT), lactate dehydrogenase (LDH) and C-reactive protein (CRP) were assessed every day during the first week, then at 15, 21 and 30 days after infusion. Platelet count, fibrinogen and XDP (D-dimer) were evaluated during the first 30 days after transplantation.
Auto- and alloimmune profiles were assessed as previously described .
Data were expressed as mean ± standard deviation. Statistical analysis was performed using IBM-SPSS Software 20 (Chicago, IL). Correlations were performed using Spearman and Pearson tests. Student's t-test was applied to compare means.
DCE-MRI assessment of liver perfusion changes associated with islet infusion
Changes in liver perfusion, revealed by the comparison between the DCE-MRI values before and 7 days after islet-tx, demonstrated high variability in the response of liver microvasculature to islet infusion and engraftment. In particular, a significant decrease of both AUC60 (pt.2: −23.4%; pt.4: −23.7%) and Ktrans (pt.2: −31.7%; pt.4: −27.9%) was observed after transplantation in two patients (Fig. 1; Table 2). Nonsignificant changes were observed in three patients (pt.1, pt.5 and pt.6; Fig. 2; Table 2). Finally, a single patient (pt.3) reached a markedly significant increase in both DCE-MRI parameters of liver perfusion (AUC60: +31%, Ktrans: +42.8%; Fig. 3; Table 2).
|Patients||DCE-MRIpretx||DCE-MRIday7||ΔDCE-MRI: [(7d-pretx)/pretx] × 100 (%)||p-Value|
|AUC60||41.67 ± 4.34||44.38 ± 4.816||6.5||n.s.|
|Ktrans||0.053 ± 0.0072||0.052 ± 0.0091||−2.0||n.s.|
|AUC60||22.56 ± 1.621||17.96 ± 1.466||−23.4||<0.01|
|Ktrans||0.059 ± 0.0074||0.041 ± 0.0064||−31.7||<0.01|
|AUC60||24.96 ± 3.432||32.69 ± 4.609||31||<0.01|
|Ktrans||0.027 ± 0.0158||0.038 ± 0.0083||42.8||<0.01|
|AUC60||27.62 ± 4.610||21.08 ± 2.431||−23.7||<0.01|
|Ktrans||0.047 ± 0.0119||0.034 ± 0.0077||−27.9||<0.01|
|AUC60||22.66 ± 3.236||23.53 ± 2.453||3.84||n.s.|
|Ktrans||0.025 ± 0.0034||0.025 ± 0.0097||−1.3||n.s.|
|AUC60||23.65 ± 2.854||23.64 ± 2.214||−0.1||n.s.|
|Ktrans||0.030 ± 0.0042||0.027 ± 0.0034||−10.6||n.s.|
Biochemical and clinical outcome of islet-tx
No major complications related to intraportal vein infusion were documented. In all patients, as expected, CRP increase was observed between 2 and 4 days after the infusion, followed by a peak of AST and ALT at 7 days in five patients (pts.2–6), whereas in one patient (pt.1) a slight increase of transaminases at 21 days was observed; finally, a progressive increase of LDH until 30 days after transplantation was documented in all patients (data not shown). Moreover, all patients had a slight platelet consumption with the nadir between 5 and 7 days after transplantation (mean maximum reduction −30% from pre-transplantation) without significant modifications of coagulation factors. Laboratory indices (CRP, AST, ALT, LDH, platelets) showed a similar trend in all patients and no significant correlation was found between these indices and the DCE-MRI parameters.
The clinical and metabolic follow-up revealed premature dysfunction of the graft in two patients: pt.2 experienced complete graft failure already 1 month after infusion, despite a transient C-peptide secretion between day 7 (C-peptide 0.59 ng/mL) and 14 (C-peptide 0.51 ng/mL); pt.4 showed transient moderate graft function (1 month: 0.11 ng/mL; 3 month: 0.71 ng/mL) followed by a progressive functional exhaustion after the third month and complete graft failure at 6 months (C-peptide <0.01 ng/mL; Table 3).
|Patients||Indices of graft function||Time-points after islet-tx|
|Month 1||Month 3||Month 6|
|Fasting C-peptide (ng/mL)||1.72||1.13|
|Fasting C-peptide (ng/mL)||<0.01||<0.01|
|Fasting C-peptide (ng/mL)||0.91||1.14||0.88|
|Fasting C-peptide (ng/mL)||0.11||0.71||<0.01|
|Fasting C-peptide (ng/mL)||1.7||1.51||0.77|
|Fasting C-peptide (ng/mL)||0.68||1.5||1.37|
The remaining four patients (pts.1, 3, 5 and 6) showed a good mid-term graft function demonstrated by the fasting C-peptide and β-score values recorded during the monthly follow-up, which are reported in Table 3.
Shortly after transplantation, pt.3 had a positivization of Ab anti-GAD titer with an early peak on day 16 after transplantation (>threefold increase).
In the long term, pt.1 and pt.5 became insulin-free after a second infusion (pt.1 had the second infusion at 3 months and persist insulin-free for 6 months before auto-suspended immunosuppression; pt.5 received a second infusion after 9 months and she is still insulin-free 2 years after transplantation); Pt.3 received a second infusion 6 months after the first, maintaining a partial function with progressive decrease of EIR (from 0.79 to 0.21 InsU/die/kg). He showed a marked increase in Ab anti-GAD at 372 days after islet-tx, followed by a progressive exhaustion of graft function until 14 months posttransplantation. Pt.6 never received a second infusion and maintained good function until 15 months after transplantation, when immunosuppression was auto-suspended.
None of the remaining patients showed a rise in autoantibody titers after transplantation; moreover, no other immunological marker associated with early graft failure was documented in pt.2 or pt.4.
Comparison between perfusion changes and clinical-biochemical indices
Both patients, pt.2 and pt.4, who experienced a complete graft failure at 1 and 6 months after transplantation, respectively, showed a strong decrease of liver perfusion parameters measured on day 7. In the remaining four patients, DCE-MRI parameters increased or did not significantly change after 7 days from transplantation; no significant change was observed in pts.1, 5 or 6 (Table 2), while a significant increase of both DCE-MRI parameters was observed only in pt.3, who is also the only patient who showed a significant autoantibody increase, with a peak at 16 days after transplantation.
Statistical analysis revealed the following correlations:
- - ΔAUC600–7days − fasting C-peptide: R = 0.721; p = 0.05
- - ΔAUC600–7days − β-score: R = 0.754; p < 0.05
- - ΔKtrans0–7days − fasting C-peptide: R = 0.567; p = n.s.
- - ΔKtrans0–7days − β-score: R = 0.899, p < 0.01.
As previously described, pt.3 had a strong liver perfusion increase with concomitant autoantibody titer increase. This immunological phenomena may independently impact on liver perfusion; therefore, statistical analysis were re-performed excluding pt.3 and they revealed stronger correlations:
- - ΔAUC600–7days − fasting C-peptide: R = 0.984; p < 0.01
- - ΔAUC600–7days − β-score: R = 0.932; p < 0.05
- - ΔKtrans0–7days − fasting C-peptide: R = 0.982; p < 0.01
- - ΔKtrans0–7days − β-score: R = 0.980; p < 0.01.
Islet-tx has progressively gained importance during the last years, although widespread clinical application is still hindered by the need for a large amount of pancreatic islets and by the limited long-term clinical success [1, 7]. Both critical issues seem to be mainly related to the large loss of islets that occurs early after intraportal infusion, with only 30–50% of infused islets that survive and contribute to long-term graft function [8-12]. Preclinical data suggest that microvascular phenomena play a crucial role in the islet-tx engraftment [16, 18-20, 24]; however, the same phenomena are largely unknown in the clinical setting. Chan et al  demonstrated a strong correlation between liver perfusion modifications assessed by DCE-MRI in a mouse model of islet-tx and the amount of newly formed islet vessels (von Willebrand factor staining). We extended this observation to six T1DM patients, for the first time. The two parameters considered were the most commonly used AUC60 and Ktrans, which are characterized by high interstudy reproducibility . AUC60 is a semi-quantitative parameter of tissue perfusion largely used in oncological imaging to assess the effects of anti-angiogenic treatment on tumor vasculature, where posttreatment AUC60 decrease indicates a reduction in tumor perfusion. Ktrans is the most commonly used quantitative parameter obtained through a fitting process of the DCE-MRI data set and may have different physiological interpretations depending on the tissue of interest: Ktrans changes may represent a modification either of permeability or blood flow according to the limiting factor .
DCE-MRI demonstrated a significant reduction of both parameters of liver perfusion in two of the six transplanted patients. None of them demonstrated significant differences in the trend of laboratory indices (transaminases, LDH and platelets) in comparison with the other four patients and, moreover, nonrelevant immunological events were detected by autoantibody and alloantibody titers monitoring. Nevertheless, both these patients experienced an unfavorable outcome of the graft: pt.2 only had a transient C-peptide secretion between days 7 and 14 after transplantation and lost the graft function completely within the first month; pt.4 showed a modest function during the first 3 months, followed by premature exhaustion of islet secretion. The complete loss of graft function as soon as it was observed in pt.2, in the absence of positivization of allo- and autoantibodies, suggests a complete failure of the engraftment. In pt.4, the modest function observed in the first months, followed by premature exhaustion of graft function, even without a rise in antibodies, may also suggest an impaired engraftment with only few surviving islets. In the other four patients, the DCE-MRI study demonstrated the absence of significant modification or an increase in perfusion parameters after islet transplantation; all these patients experienced a good mid- and long-term graft function. In particular, three patients showed unchanged posttransplant perfusion parameters, while only pt.3 showed a marked increase in liver perfusion at 7 days after transplantation and, interestingly, he is also the only patient to show an increase in autoantibodies with a peak observed 16 days after transplantation. This association may suggest that the immune response may be at the origin of the increased liver perfusion, even if no definitive conclusion can be drawn considering that this behavior was only observed in a single patient.
It is of interest to note that we found a correlation between the changes in DCE-MRI parameters that were observed 7 days after islet infusion, and the indices of graft function measured between the third and sixth month after transplantation. These results, although requiring confirmation in larger populations, may suggest that the functional adaptation of liver microvasculature observable with DCE-MRI 7 days after islet infusion may impact on islet engraftment. The changes in liver perfusion observed may result from the balance between various phenomena that are known to coexist in a liver that receives islet transplantation, such as the neoangiogenetic processes driven by the need for islet revascularization and the microthrombotic events linked to the IBMIR reaction.
Interestingly, if pt.3 were excluded from our statistical analysis on the basis that there could be another explanation for the observed changes in the perfusion parameters, as suggested by the almost concomitant increase of autoantibodies, the observed correlations found between DCE-MRI parameters and indices of graft function become stronger and more significant.
We are aware of some major limitations of this study: first of all, the study was performed on a small number of patients and so the results need to be confirmed in larger studies. Another limitation is the use of two different gadolinium-based contrast agents for DCE-MRI caused by the sudden commercial discontinuation of Vasovist® during the study. To overcome this limitation, we ensured that the pre- and posttransplant DCE-MRI studies were performed with the same contrast agent in each patient, so that the changes in DCE-MRI parameters observed at different time points were consistent and reliable. Moreover, the contrast media applied in the last three patients is actively accumulated in the hepatocytes after infusion with a peak of liver enhancement 20 min after injection. However, this property of the used contrast media does not impact on the perfusion parameters measured in this study because these parameters are substantially influenced by the first part of the enhancement–time curve and, as demonstrated in previous studies, the liver perfusion indices measured using Primovist® are reliable . Finally, the anticoagulation prophylaxis used in our Institution may differ from that used by other groups, perhaps influencing the observed perfusion findings.
In conclusion, this is the first clinical study aiming to investigate in noninvasive way liver perfusion modifications occurring during islet engraftment and attempting to correlate them with clinical parameters in T1DM patients. DCE-MRI identifies various adaptive responses of liver microvasculature in patients submitted to islet transplantation; these responses could have an impact on islet engraftment, but further investigations on larger patient populations are required.
This research was supported by a grant from the European Foundation for the Study of Diabetes (EFSD) and Merck Sharp & Dohme (MSD; noninvasive in vivo MRI of Endorem®-labeled pancreatic islets following transplantation in humans). Principal Investigator Prof. Antonio Secchi; CoPI: MLM and ADM. The funders had no role in the study design, data collection and analysis, decision to publish or preparation of the manuscript. We thank Prof. Alessandro Ambrosi of Vita-Salute San Raffaele University for the assistance in the statistical analysis.
The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.