Detection of Insulitis by Pancreatic Scintigraphy With ^99mTc-Labeled IL-2 and MRI in Patients With LADA (Action LADA 10)

Type 1 diabetes (T1D) is a T-cell–mediated autoimmune disease derived from the selective destruction of insulin-secreting β-cells. Several studies have shown that lymphomononuclear cell infiltration of the endocrine pancreas (insulitis) is present for months or even years before the clinical diagnosis and is associated with residual β-cell mass (1–3).

With the exception of pancreatic biopsy, which is an invasive and risky procedure, to date no diagnostic markers are available to detect insulitis and identify responders to immunotherapy, which is currently being evaluated in trials of patients with recent-onset T1D (4). Lymphocytes and monocytes are responsible for the destruction of β-cells, and the “in vivo” visualization of islets infiltrated by mononuclear cells by means of a noninvasive technique could also be an important tool for the identification of prediabetic patients in whom the islet cell destructive process is potentially treatable and may be halted.

Therefore, the possibility of evaluating in vivo changes in insulitis could be useful for monitoring the disease process and the efficacy of an immune intervention. For this purpose, we developed an interleukin-2 (IL-2) radiolabeled with ^99mTc (^99mTc-IL-2), a cytokine that binds to a specific receptor expressed on activated lymphocytes (5). This radioactive probe has optimal characteristics for in vivo detection of activated T cells in infiltrated tissues. In previous studies, ^99mTc-IL-2 allowed the successful demonstration of insulitis in two animal models of autoimmune diabetes (6–8). Indeed, one of the most relevant characteristics of islet-infiltrating lymphocytes is their activation status, with surface expression of several receptors. Among these, the IL-2 receptor, an heterotrimeric transmembrane G protein–coupled receptor (α-chain CD25, β-chain CD122, γ-chain CD132). The α−β (CD25–CD122) heterodimer is responsible for IL-2 binding with high affinity. Large numbers of these receptors are expressed on the surface of activated T cells; a lesser amount is expressed on monocytes and other cells. Early attempts to image activated T cells with ¹²³I-IL-2 followed by ^99mTc-IL-2 have been made.

In humans, scintigraphy with either ¹²³I-IL-2 or ^99mTc-IL-2 has been used to image chronic inflammation-mediated disorders and lymphocyte homing in cancer. In particular, in patients with celiac disease it was possible to detect jejunal infiltration and to monitor the effect of diet on the immune process (9). Specific binding of radiolabeled IL-2 to activated lymphocytes was also clearly demonstrated by ex vivo microautoradiography and immunostaining; tissue specimens revealed that jejunal accumulation of ^99mTc-IL-2, as measured in vivo by single photon emission computerized tomography (SPECT), correlates with the number of infiltrating activated mononuclear cells per millimeter of mucosa (9). Similar results were achieved in patients with active Crohn’s disease in whom ¹²³I-IL-2 scintigraphy was used to monitor the efficacy of therapy with corticosteroids (10). ^99mTc-IL-2 scintigraphy was able to predict time to disease relapse in patients with inactive Crohn’s disease (11). Radiolabeled IL-2 is also able to evaluate the degree of lymphocytic infiltration of carotid atherosclerotic plaques, thus providing useful information for the selection of infiltrated vulnerable plaques at risk for rupture (12). In patients affected by cutaneous melanoma the uptake of radiolabeled IL-2 correlates with IL-2 receptor expression on activated T-lymphocytes in tumor infiltration. Thus, a positive response to ^99mTc-IL-2 scintigraphy could represent a prognostic marker to predict the efficacy of IL-2 immunotherapy in patients with melanoma (13). Our first study of ^99mTc-IL-2 scintigraphy in human patients with T1D, published in 2008, reported that 65% of patients >15 years of age with newly diagnosed T1D showed a pancreatic accumulation of ^99mTc-IL-2 (14). Most important, patients with positive scintigraphy at diagnosis showed better long-term metabolic control if treated with nicotinamide compared with patients with negative scintigraphy, suggesting that positive patients had persistent pancreatic inflammation at diagnosis with residual β-cell function and that nicotinamide was effective in protecting β-cells.

Since then, we continued investigating the use of radiolabeled IL-2 in several diseases, including T1D, and we developed a simple kit for easy and reproducible production of this radiopharmaceutical (15,16).

The main aim of the current study was to evaluate the ability of ^99mTc-IL-2 scintigraphy to detect the presence of activated lymphocytes in the pancreas of different types of diabetic patients, in particular patients with T1D with clinical onset at an adult age (latent autoimmune diabetes in adults [LADA]) compared with children with T1D and patients with long-standing type 2 diabetes (T2D). The IL-2 scan was evaluated further in a control group of patients with pancreatic carcinoma undergoing pancreatectomy because this type of cancer is usually infiltrated by activated T cells.

Sixteen T1D patients (8 males and 8 females; mean age 16.6 ± 8.4 years; median age 13 years; age range 8–33 years; all but one positive for anti-glutamic acid decarboxylase autoantibodies [GAD+] or anti–insulinoma-associated protein [IA]) within 4 weeks from clinical diagnosis and nine patients with LADA (6 receiving insulin therapy, 3 still receiving oral antidiabetic therapy; 4 males and 5 females; mean age 44 ± 9.1 years; all positive for anti-GAD or anti-IA2 or insulin autoantibodies [IAA]) (Table 1) were enrolled in this study. We selected patients with LADA according to the criteria established by the Action LADA project (17): basically a diagnosis of diabetes when more than 30 years old, no insulin required for at least 6 months, and positivity for at least one autoantibody islet cell–related autoantibody, including GAD, IA2, or IAA. These autoantibodies were measured centrally in a Diabetes Autoantibody Standardization Program–recognized laboratory. Patients with LADA were investigated within 12 months of diagnosis of LADA and recent introduction of insulin therapy.

As negative controls we studied six patients with long-standing T2D (four men and two women; mean age 54 ± 6.8 years; negative for islet cell autoantibodies and GAD; diabetes duration 5 ± 4.1 years) and, as positive controls, seven patients with pancreatic adenocarcinoma (ADK) (four men and three women; mean age 59 ± 8.7 years) in whom the neoplastic lesion was infiltrated by activated (CD25+) T cells. Patients were recruited at the Endocrinology and Surgery Departments of Universities “Campus Bio-Medico” and “Sapienza” in Rome, Italy.

All patients underwent IL-2 scintigraphy and contrast-enhanced MRI of the pancreas. A tissue sample was collected from patients with ADK during surgery for histological and immunohistochemical investigations. This type of cancer is usually infiltrated by activated T cells (18) and was used as a positive control and to correlate IL-2 uptake in the lesion with the histological findings.

This study was carried out in accordance with the principles of the Declaration of Helsinki and was approved by the Ethics Committee at Sapienza University. All patients or parents gave written informed consent before entering the study.

Radiolabeled IL-2 was prepared, as previously described (5), according to the principles of Good Manufacturing Practice. Subjects were intravenously injected with 74–185 MBq of ^99mTc-IL-2 (20–50 µg IL-2) and orally treated with KClO₄ (400 mg) 20 minutes before the study to prevent thyroid and stomach uptake of free ^99mTc originating from the metabolism of ^99mTc-IL-2. The dose in children was calculated by weight, according to international guidelines of the Pediatric Group of the European Association of Nuclear Medicine.

Patients were positioned under a computerized gamma camera, and SPECT images of the abdomen were acquired 45 to 60 min after injection using a 64 × 64 pixel matrix. The obtained transaxial abdominal section (3 cm thick), including the body/tail of the pancreas (or the pancreatic tumor in the head or body of the pancreas), was selected by image fusion with the corresponding section obtained by MRI (Figs. 1 and 2).

For data analysis, pancreatic radioactivity (or tumor radioactivity) was calculated by drawing an irregular region of interest over the pancreatic region (body/tail of the pancreas in diabetic patients or the pancreatic tumor mass in patients with ADK) in the most representative transaxial section. On the same section, a circular region of interest was drawn over the spine; this was considered background radioactivity. Pancreatic or tumor uptake of ^99mTc-IL-2 was calculated as the pancreas (target)-to-background radioactivity ratio (T:B), as described in a previous study (14). The effective dose equivalent (EDE) of a diagnostic dose of ^99mTc-IL-2 (3 mCi, 110 MBq) is approximately 0.9 mSv, which is significantly lower than the EDE of most radiological investigations (the EDE of a computed tomography [CT] scan of the abdomen is 4–5 mSv) and falls into category IIa of the International Committee of Radiation Protection. The amount of IL-2 injected (20–50 µg) is nontoxic and not immunogenic.

Pancreatic scintigraphy was considered positive if the T:B was higher than the average of that in T2D patients +2 SD. Scintigraphies were evaluated by two independent nuclear medicine physicians (A.S. and G.C.).

Contrast-enhanced MRI of the pancreas was performed using 1.5T Siemens Vision Plus (gradient power 25 mT/m; gradient rise time 600 µsec) using fat-suppressed sequences. Slices (3 mm) were acquired for better anatomic detail to determine signal and enhancement differences within the pancreas and between patient groups. Volumetric slices (3 cm thick) were acquired for image fusion with SPECT images. MRI images were evaluated by two independent radiologists (C.C. and A.M.Q.).

Specimens were collected from the edges of the tumor mass in patients with pancreatic ADK. Tissue samples were fixed in formalin and embedded in paraffin for further histological examination. Paraffin sections were stained for CD25 antigen expression by an initial incubation with a normal goat serum blocking reagent followed by a primary monoclonal antibody and were revealed with a peroxidase-conjugated secondary antibody using diaminobenzidine as a chromogen. Cells were counted using a Quantimet 920 Image Analyzer with a monochrome charge-coupled device camera (Hamamatsu C3077). Using an off-peak intensity of 546 nm with a Reichert Polyvar microscope connected to the Hamamatsu camera, the diaminobenzidine immunostaining was evaluated at a final magnification of 1,700× to quantify the number of positive cells in at least five randomly selected fields.

Data are presented as mean ± SD. To evaluate the differences of T:B between studied populations we used the Student t test for unpaired data.

Regression analysis between ^99mTc-IL-2 pancreatic uptake and other parameters was performed. Pearson correlation for normally distributed data was used to calculate statistical significance. We used Excel version 14.4.5 and AcaStat version 8.3.10 to calculate means and SDs, t tests, and correlations and to create graphs. SPSS software version 22.0 was used for Pearson correlations.

No adverse effects were observed following the administration of ^99mTc-IL-2. The procedure to radiolabel IL-2 was carried out under sterile conditions, and all labelings satisfied the criteria for biological and radiochemical quality controls (radiochemical purity >95% by instant thin-layer chromatography). All patients completed the study.

MRIs clearly showed the pancreas in all patients, with no significant differences between patient groups. Both T2-weighted and fat-suppressed images showed homogenous intensity of the body and tail of the pancreas in patients with T1D, LADA, and T2D. No significant differences between contrast enhancement of MRI within the pancreas and between diabetic patients were detected. We found differences in focal intensity only in patients with ADK. There was always agreement between the reports from the two radiologists.

Good correlation also was observed between the two nuclear medicine readers (r = 0.91; P < 0.001); the mean results of reading of the two operators were used for statistical analysis.

Pancreatic ^99mTc-IL-2 uptake in T2D patients (negative controls) was 2.81 ± 0.63. Thus, using a positive threshold of 4.06 (mean + 2 SD), 8 of 16 T1D patients (50%) and 6 of 9 patients with LADA (66.6%) showed a significant accumulation of IL-2 in the pancreas.

In T1D patients, the T:B did not correlate with age or body weight (T:B vs. age, r = 0.16; T:B vs. body weight, r = 0.31) but was higher than in T2D patients, although it did not reach statistical significance (4.45 ± 1.99 vs. 2.81 ± 0.63; P = 0.06).

Mean pancreatic ^99mTc-IL-2 uptake in patients with LADA and pancreatic ADK was significantly higher than in patients with T2D (4.79 ± 1.1 and 4.54 ± 1.62 vs. 2.81 ± 0.63; P = 0.01 and P = 0.04, respectively) (Fig. 3). No significant differences between patients with T1D and LADA, T1D and pancreatic carcinoma, or LADA and pancreatic carcinoma were observed.

Pancreatic tissue specimens were obtained in five of seven patients studied. Two patients were nonoperable. Pancreatic accumulation of ^99mTc-IL-2 in patients with pancreatic ADK correlated with IL-2 receptor expression in cancer-infiltrating T-lymphocytes (r = 0.8; P = 0.01) (Fig. 4).

Pancreatic uptake of radiolabeled IL-2 did not correlate in all patients with age, body weight, or autoantibody titer (GAD, IA2, and IAA), but in patients with LADA it correlated with disease duration (r = 0.7; P = 0.03).

Autoantibodies to islet cell antigens and mononuclear cell infiltration in the islet, the pathological hallmark of T1D, appear long before the onset of the disease, persist for some years after clinical diagnosis, and progressively decrease as the β-cell mass declines (19,20). Although both genetic and autoantibody screening can help to predict autoimmune diabetes in some cases, the identification and evolution of insulitis can offer new clues for disease pathogenesis and follow-up therapy after an immune intervention.

Performing multiple pancreatic biopsies would be the ideal method to study the presence and the severity of insulitis; as an invasive technique, however, biopsies are routinely impracticable. The goal of imaging modalities is to integrate diagnostic information with anatomic and functional data to characterize the insulitis process. The relatively small inflammatory process of insulitis is difficult to study by CT and MRI. Furthermore, these radiological techniques are not able to supply functional information about β-cells.

Nuclear medicine provides a wide spectrum of radiopharmaceuticals to image chronic inflammation. Pancreatic insulitis in T1D patients was first detected 30 years ago by ¹¹¹In-labeled lymphocytes but without convincing and conclusive results (21). Other authors showed that ¹¹¹In-labeled lymphocytes do not image the infiltrated pancreas in biobreeding rats (22). Instead, radiolabeled IL-2 has optimal characteristics for in vivo detection of activated T cells in infiltrated tissues. It allowed detection of insulitis in two animal models of autoimmune diabetes, and a positive correlation between ¹²³I-IL-2 scintigraphic uptake and the degree of lymphocytic infiltration was observed by gamma camera imaging and proved histologically (9–13).

In the current study we used ^99mTc-IL-2 to attempt evaluation of ongoing inflammation in the endocrine pancreas. Our results showed that ^99mTc-IL-2 accumulates in the pancreas of approximately 50% of T1D patients at the time of clinical diagnosis and in 66.6% of patients with LADA (a sign of the presence of activated mononuclear cells), particularly if they are diagnosed within 1 year of the first episode of hyperglycemia. The differences in the percentage of positivity can be related to different stages of the disease. All T1D patients were insulin dependent, and in some of these there probably was no residual inflammation in pancreatic β-cells (23). Patients with LADA were all newly diagnosed and three of them were noninsulin dependent at the time of study; therefore they can be considered as being in a preclinical phase of the disease in which we expect to find insulitis. Different degrees of radiolabeled IL-2 uptake in the pancreas have already been reported in newly diagnosed T1D patients, concluding that it may reflect a different pattern of infiltration at the time of diagnosis (24). Thus, because it is able to identify the degree of lymphocytic infiltration, IL-2 could be used as a marker to highlight the relationship between the autoimmune process and progression of the disease in patients with LADA.

In this study we also demonstrated that ^99mTc-IL-2 uptake in patients with pancreatic carcinoma correlates with IL-2 receptor expression, as evaluated by immunostaining. Although we had specimens from only five patients, these results confirm those previously obtained in other diseases (9–13) and support the feasibility of imaging insulitis in autoimmune diabetes.

In a previously published study, ¹²³I-IL-2 accumulation also was observed in subjects at risk of developing autoimmune diabetes. Patients with a positive scan developed diabetes within 1 year (25), suggesting that this technique may be useful for predicting disease. In this context it is of interest that our patients with LADA were investigated in the early phases of the autoimmune process (i.e., within 1 year from diagnosis) and showed abnormal ^99mTc-IL-2 uptake in the pancreas.

These results open new avenues in the management of diabetic patients. Thus, it will be possible to identify in the preclinical phase the presence of insulitis so that patients could be treated with immunotherapies at an early stage and followed up. IL-2 scintigraphy can be used at time of diagnosis to identify a subgroup of patients with persistent insulitis (and therefore residual β-cell mass) in whom the efficacy of treatment can modulate the autoimmune process.

Other groups used different radiopharmaceuticals to image β-cells, but the results of SPECT were not promising. Better results have been achieved using positron emission tomography (PET) radiopharmaceuticals to evaluate residual β-cell mass (26).

After clinical diagnosis during the first year of disease, a small proportion of T1D patients undergo spontaneous clinical remission. This finding has encouraged the beginning of clinical trials with several adjuvant therapies to improve the rate of clinical remission at the time of diabetes onset. Patients with active, ongoing pancreatic inflammation may benefit from immunotherapy. Therefore, using a receptor ligand such as radiolabeled IL-2 for mononuclear cells is the ideal method to quantify in vivo the severity of the inflammatory process. One of the principal difficulties in studying the insulitis process in vivo by scintigraphic techniques, however, is related to the low resolution and the poor anatomical definition of system detection.

In our study MRI was performed in each patient, and scintigraphic images of the pancreatic region were fused with MRIs to anatomically localize the pancreas. Although this fusion method is efficacious, young patients had a poor compliance with MRI, and this technique may not be used routinely, particularly in children. This problem could be overcome by the use of hybrid SPECT–CT or PET–CT cameras that simultaneously provide a multislice CT scan and a tomographic scintigraphic scan of the target organ with image fusion and high anatomic and functional definition of the target. Indeed, ¹⁸F-labeled IL-2 has recently been synthesized, thus allowing the use of PET–CT camera's for imaging insulitis (27,28).

In conclusion, our results show that ^99mTc-IL-2 scintigraphy is able to identify CD25+ lymphocytic infiltration in the pancreas of patients with pancreatic ADK, and the high pancreatic uptake of ^99mTc-IL-2 in patients with T1D and LADA suggests that we were able to image insulitis. In recent years research has focused on the identification of therapeutic agents interfering with the islet destruction process. A major problem related to these aspects is the poor availability of markers to evaluate the efficacy of treatment and to select responders to different therapies. Our data demonstrate that ^99mTc-IL-2 scintigraphy, as a noninvasive technique, allows the in vivo detection of CD25+ cells in tissues and may be a suitable tool for monitoring immune intervention in patients with autoimmune diabetes, although large intervention studies are needed to confirm the clinical role of this technique.

Acknowledgments. The authors acknowledge the Association of Nuclear Medicine Discovery (Nu.Me.D.) for help and support during this study and Dr. E. Maddaloni, Department of Endocrinology and Diabetes, University Campus Bio-Medico, Rome, Italy, for statistical analysis.

Funding. This research was supported by the Italian Ministry of Education, University and Research (MIUR) and the Action LADA Group from the European Union (contract no. QLG1-CT-2002-01886).

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. A.S. received funding for the study, performed scans, analyzed data, and wrote the manuscript. G.C. performed scans, analyzed data, and wrote the manuscript. M.C. performed scans and analyzed data. E.B. performed histological analysis. F.G. performed all labeling procedures. C.C. and A.M.Q. performed nuclear magnetic resonance scans and analyzed data. G.D.T. selected patients with pancreatic adenocarcinoma and performed surgery. S.M. selected diabetic patients and those with LADA. P.P. recruited patients, analyzed data, and wrote the manuscript. A.S. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.