F8-IL10: A New Potential Antirheumatic Drug Evaluated by a PET-Guided Translational Approach.

Antibody fragment F8-mediated interleukin 10 (IL10) delivery is a novel treatment for rheumatoid arthritis (RA). F8 binds to the extra-domain-A of fibronectin (ED-A). In this study, in vivo biodistribution and arthritis targeting of radiolabeled F8-IL10 were investigated in RA patients, followed by further animal studies. Therefore, three RA patients (DAS28 > 3.2) received 0.4 mg of 30-74 megabecquerel [124I]I-F8-IL10 for PET-CT and blood sampling. In visually identified PET-positive joints, target-to-background was calculated. Healthy mice, rats, and arthritic rats were injected with iodinated F8-IL10 or KSF-IL10 control antibody. Various organs were excised, weighed, and counted for radioactivity. Tissue sections were stained for fibronectin ED-A. In RA patients, [124I]I-F8-IL10 was cleared rapidly from the circulation with less than 1% present in blood after 5 min. PET-CT showed targeting in 38 joints (11-15 per patient) and high uptake in the liver and spleen. Mean target-to-background ratios of PET-positive joints were 2.5 ± 1.2, 1.5 times higher for clinically active than clinically silent joints. Biodistribution of radioiodinated F8-IL10 in healthy mice showed no effect of the radioiodination method. [124I]I-F8-IL10 joint uptake was also demonstrated in arthritic rats, ∼14-fold higher than that of the control antibody [124I]I-KSF-IL10 ( p < 0.001). Interestingly, liver and spleen uptake were twice as high in arthritic than in healthy rats and were related to increased (∼7×) fibronectin ED-A expression in these tissues. In conclusion, [124I]I-F8-IL10 uptake was observed in arthritic joints in RA patients holding promise for visualization of inflamed joints by PET-CT imaging and therapeutic targeting. Patient observations and, subsequently, arthritic animal studies pointed to awareness of increased [124I]I-F8-IL10 uptake in the liver and spleen associated with moderate systemic inflammation. This translational study demonstrated the value of in vivo biodistribution and PET-CT-guided imaging in development of new and potential antirheumatic drugs'.

Introduction

Rheumatoid arthritis (RA) is a chronic inflammatory and destructive joint disease[1] for which treatment options currently include chemical and biological disease modifying antirheumatic drugs (DMARDs).[2,3] Nevertheless, 30–40% of patients are refractory to these treatments, leaving an unmet need for alternative therapies.[4] Targeted delivery of cytokines at the site of disease is an emerging novel approach to treat chronic inflammation, especially RA,[5] and has also been successfully investigated in the cancer setting over the past few years.[6−9] Immunocytokines (cytokine-antibody fusion proteins) represent a class of therapeutic agents that have previously shown their usefulness for targeting antigens at the site of inflammation followed by local activity of the cytokine.[5] Along these lines, target specificity and dose-mediated therapy response can be achieved, especially at synovium in RA.

Interleukin 10 (IL10) is an anti-inflammatory cytokine, produced by activated monocytes and T cells, and is involved in the regulation of inflammatory responses and immune reactions.[10,11] In the clinic, combination therapy of IL10 and methotrexate in a multicenter, placebo-controlled study in RA patients revealed clinical American College of Rheumatology 20 responses in 50–60% of patients compared with 10% for placebo.[12] Despite the anti-inflammatory properties of IL10, systemic administration in RA patients generated insufficient responses to pursue further development. However, when IL10 is fused with the single-chain antibody variable domain (Fv) fragment of antibody F8, targeted delivery of IL10 can be achieved locally, especially at RA synovium. Antibody fragment F8 binds to the extra-domain A (ED-A) of fibronectin, selectively expressed at sites of inflammation in RA[13] and in tumors[14] in humans as well as in animals. Preclinical studies show that F8 binds with comparable affinity to ED-A of murine, monkey, and human origin.[15] Previously, high levels of fibronectin ED-A have been demonstrated in the target tissues of arthritic as well as tumor-bearing animal models.[15,16] Besides, F8-IL10 displayed clear localization at sites of arthritis in a collagen-induced arthritis mouse model.[16] Recently, encouraging early signals of the therapeutic benefit of F8-IL10 were demonstrated in RA patients in a phase I study, underscoring the therapeutic potential of this new approach.[17] Positron emission tomography (PET) can provide unique whole-body molecular information about targeting and in vivo biodistribution of new therapeutic agents.[18,19] For imaging of antibodies, isotopes with long half-lives as iodine-124 (124I) and zirconium-89 (89Zr) allow for imaging of targeting and in vivo biodistribution over several days.[20−22] The purpose of the present translational study was to investigate the in vivo distribution and arthritis targeting of [124I]I–F8–IL10 in RA patients, followed by animal studies.

Methods

Patients

Three RA patients (aged ≥18 years) who fulfilled the American College of Rheumatology 1987 classification criteria for rheumatoid arthritis[23] were included during 2013 and 2014 (Table ). Patients eligible for this study required a high disease activity score across 28 joints (DAS28) of ≥3.2 and ≥2 clinically inflamed joints in hands. Exclusion criteria consisted of any hematologic, liver, or renal function test abnormalities; concurrent active infections/autoimmune disease or severe (malignant) diseases (other than RA); immunodeficiency or human immunodeficiency virus positivity; (history of) heart problems or insufficiency; uncontrolled hypertension; pregnancy or breastfeeding; major surgery within the previous 4 weeks; or any other treatment with investigational drugs within the previous three months. Stable doses of DMARDs and/or nonsteroidal anti-inflammatory drugs (NSAIDs) were continued if used at inclusion. All clinical data were obtained by an experienced research nurse blinded to the imaging data.

Table 1

Tracers and Dosages Used in Healthy/Arthritic Rat Studies

	tracer	dosage (mg/kg)
group A: healthy rats (F8-IL10low; n = 4)	[124I]I–F8–IL10	0.35
group B: arthritic rats control (KSF-IL10; n = 6)	[124I]I-KSF-IL10	0.35
group C: arthritic rats (F8-IL10low; n = 6)	[124I]I–F8–IL10	0.35
group D: arthritic rats (F8-IL10hgh; n = 6)	[124I]I–F8–IL10	12.9

Animals

Wistar rats (male, 150–200 g, Charles River International Inc., Sulzfeld, Germany) and healthy mice (6–8 weeks female, Envigo, Horst, NL) were provided with standard food (16% protein rodent diet, Harlan Laboratories Inc., Madison, WI, USA) and water ad libitum. Rats were housed in conventional cages and mice in filter top cages (in groups of 3–6 per cage) and kept in a room with a 12-h light/dark cycle and constant room temperature (21 °C) and humidity level (55 ± 10%).

For arthritis induction, Wistar rats were immunized systemically and additionally received 4× intra-articular methylated bovine serum albumin injections, 4 or 5 days apart, in the arthritic (right) knee, with the contralateral (left) knee serving as a control, as described previously.[24,25]

Tracer Synthesis

The full description of the tracer synthesis for administration to RA patients, mice, and rats is presented in Appendix A in the Supporting Information. Essentially, F8-IL10 was obtained from Philogen (Zurich, Switzerland) and was iodinated with iodine-124 for humans, rats, and mice, and iodine-131 for mice with a radiochemical purity of >95%, according to instant thin layer chromatography (iTLC) and high-performance liquid chromatography (HPLC). The immune reactive fraction was determined by measuring binding to a serial dilution of ED-A-coated beads, essentially, as described by Lindmo et al.[26] For the study in healthy mice, apart from labeling of F8-IL10 with 131I according to the iodogen method (as was done with labeling for humans and rats), F8-IL10 was also labeled with 124I according to the chloramine-T method used earlier by Philogen.[27] Radiolabeling of F8-IL10 and control antibody KSF-IL10 with 124I was applied in all experiments with rats, essentially, according to Tijink et al.[28] KSF is an antibody recognizing hen egg lysozyme and has no binding specificity in rats.[29]

[124I]I–F8–IL10 PET-CT of RA Patients

Before PET-CT scanning, patients received prophylaxis to block the thyroid by administration of 400 mg of potassium perchlorate 48, 24, and 1 h(s) before infusion of [124I]I–F8–IL10 and 1 h before PET-CT scanning at 24 and 72 h. Next, a total of 3 whole body PET-CT scans per patient and detailed images of wrists/hands were obtained, and blood samples were withdrawn at 5, 30, 90 min, 3, 24, and 72 h post injection (p.i.).

PET-CT scans were performed using both Gemini-64 and Ingenuity-128 PET-CT scanners (Philips Healthcare, Cleveland, USA). After intravenous (i.v.) injection of a single dose of [124I]I–F8–IL10 (0.4 mg (∼6 μg/kg), 59 ± 25 MBq), patients were scanned at 1, 20–22.5, and 69–76 h p.i.. The administration system was flushed with 20 mL of 0.9% NaCl, and after administration of the tracer, residual activity in the injection device was measured to determine the net injected dose. Subsequently, whole body scans from skull to symphysis pubis were acquired with 5 min/field of view. In addition, detailed images of wrists/hands were obtained (1 field of view, 6 frames of 5 min each). PET scans were preceded by a 30 mAs low-dose CT for attenuation correction and anatomical localization of the PET signal.

PET data were normalized and corrected for attenuation, decay, and scatter. All scans were reconstructed according to international guidelines.[30]

Image Analysis

An experienced nuclear medicine physician (O.S. Hoekstra) qualitatively interpreted PET-CT data for whole body biodistribution and joints with visually enhanced tracer uptake (PET-positive joints). For quantitative comparison of tracer uptake, volumes of interest (VOIs) were drawn using software developed in-house,[31] with the corresponding low dose CT as an anatomical reference. For biodistribution, VOIs were drawn over lungs, liver, spleen, kidneys, heart (blood pool), and vertebrae. This was performed using in-house standard operating procedures by drawing VOIs over the entire organ (CT as reference), and results were presented as percentage of the injected dose (%ID).

For analysis of visually defined PET-positive joints, VOIs were drawn on top of joints with clear focal tracer enhancement using thresholds exceeding local background uptake (e.g., PET negative joints). In addition, for PET-negative hand joints, standardized spherical VOIs were drawn on wrists (∼60 mL), metacarpophalangeal joints 1–5 (8 mL), and (proximal) interphalangeal joints 1–5 (2.0 mL) centered in the middle of the joints (n = 22 joints per patient). Standardized uptake values were calculated by dividing the PET tracer tissue concentration by the activity injected (MBq) per body weight (kg). Tracer uptake in targets is presented as SUVpeak which is defined as the highest average uptake within a sphere of 1.2 mL within the VOI.[32,33] In addition, spherical VOIs of 0.5 mL in the second metacarpal bone were used as background to calculate target-to-background ratios.

Animal Studies

Ex Vivo Biodistribution of [124I/131I]I–F8–IL10 in Healthy Mice Comparing Radio-Iodination Methods

Healthy mice (n = 14) were injected with either [124I]I–F8–IL10 (chloramine-T method) or [131I]I–F8–IL10 (iodogen method) at 10 min (n = 3) and 24 h (n = 4). In both cases, the F8-IL10 protein dose was 3 mg/kg. Each mouse received ∼3.5 MBq [124I]I–F8–IL10 or [131I]I–F8–IL10.

At the end of the study, all mice were sacrificed, and blood and various tissues were excised, rinsed, dipped dry, and weighed. The amount of tissue radioactivity was determined using an LKB 1282 Compugamma CS gamma counter (LKB, Wallac, Turku, Finland). Results were expressed as a percentage of the injected dose per gram tissue (%ID/g).[24]

Ex Vivo Biodistribution of [124I]I–F8–IL10 in Healthy and Arthritic Rats

Iodine-124 was used in the rat experiments to allow for a better and direct comparison of findings in rat experiments and human studies. Six days after the last intra-articular injection of methylated bovine serum albumin, all rats received ∼3.5 MBq tracer injection (summarized in Table ). The control and low dose groups receiver was 3.5 MBq/100 μg, and the high dose group was 3.5 MBq/3.75 mg. 24 h post tracer injection; all rats were sacrificed, and knee joints, as well as various tissues, were removed and analyzed as described above in the mice experiment.

Immunofluorescence of Liver and Spleen of Arthritic versus Healthy Rats

Tissues (liver and spleen) of arthritic rats (n = 4) and healthy rats (n = 4) were snap frozen in liquid nitrogen and stored at −80 °C. Sections of 8 μm were cut and stained with hematoxylin and eosin (HE), whereas staining of splice isoforms of fibronectin ED-A was performed by F8-small immunoprotein (SIP) (Philogen, Siena, Italy) or isotype control antibody. Tissues were embedded in a suitable medium (OCT; SKU4583, Tissue-Tek, Netherlands), cut using a cryotome cryostat (−20 °C) (Leica, Netherlands), and placed on Superfrost (4951PLUS4, ThermoFisher, Netherlands) glass slides for immunofluorescence staining.

For immunostaining, frozen tissue sections were first brought to room temperature for 30 min. The sections were fixed in acetone (439126, Sigma-Aldrich, Netherlands) for 10 min at −20 °C and air-dried for 10 min at room temperature. Sections were marked with a DAKO pen (S2002, DAKO, Santa Clara, CA, USA). After being washed 3× with PBS on a shaker, sections were treated with 100% fetal bovine serum (FBS) for 30 min at room temperature to avoid nonspecific binding and then washed again with PBS (3 × 5 min). Thereafter, sections were incubated with F8-small immunoprotein (SIP) (2 μg/mL) in 10% FBS/PBS for 24 h at 4 °C or with 10% FBS/PBS. After a washing step (3 × 5 min in PBS on a shaker), sections were incubated with rabbit anti-human IgE 1:1000 in 10% FBS/PBS (DAKO, A0094, Santa Clara, United States) for 60 min at room temperature. After a final washing step, sections were incubated with goat-anti-rabbit Alexa 488 (A-11008, ThermoFisher Scientific, Netherlands) in 10% FBS/PBS), air-dried, and mounted (2 μL of MOWIOL mounting medium (81381, Merck, Zwijndrecht, The Netherlands).

Microscopy and Image Analysis of Immunofluorescent Tissue Sections

The slides were imaged with a Zeiss Axiovert 200 M Marianas inverted microscope (40× oil-immersion lens), equipped with a motorized stage (stepper-motor z-axis increments: 0.1 μm), and a turret of four epifluorescence cubes (fluorescein isothiocyanate, Cy-5, Cy-3, aminomethylcoumarin, as well as a differential interference contrast bright field cube). A cooled charge-coupled device camera (Cooke Sensicam SVGA [Cooke Co., Tonawanda, NY, USA], 1280 × 1024 pixels) recorded images with true 16-bit capability. The camera was linear over its full dynamic range (up to intensities of over 4000) with dark/background currents (estimated by the intensity outside the cells) of typically <100. Exposure, objective, montage, and pixel binning were automatically recorded and stored with each image (Dell Dimension workstation: Quad-core processor, 16GB RAM). The microscope, camera, and data processing were controlled by SlideBook software (SlideBook version 6, Intelligent Imaging Innovations, Denver, CO, USA).[34]

All images obtained in 2D were first deconvolved (no neighbor) in order to improve the signal-to-noise ratio. The arthritic and healthy liver/spleen images were segment masked (using background from negative liver/spleen sections). Thereupon, mask statistics were performed on the entire mask, and the sum intensities were recorded. Sum intensities of all multiple representative images from arthritic and healthy rat liver/spleen were averaged (±SD).

Histology images were captured using a Leica 4000B microscope and a Leica digital camera DC500 (Microsystems B.V. Rijswijk, The Netherlands).

Statistical Analysis

PET data of RA patients were reported in a descriptive manner because of the small sample size. Quantitative human and RA-rat model data are presented as the mean ± standard deviation (SD) or as median and interquartile range (IQR), in the case of skewed distribution. Mann–Whitney U (exact) tests were performed to analyze differences in tissue distribution and fibronectin ED-A staining between different rat groups (A–D). A p-value < 0.05 was considered as statistically significant. Statistical analyses were performed using SPSS version 22.0 for Windows (SPSS, Chicago, IL, USA).

Results

Clinical Data of RA Patients

Clinical characteristics of the three patients included are summarized in Appendix B, Supporting Information.

All patients had high disease activity despite stable doses of DMARDs, with a DAS28 > 4.0. Out of 66 evaluated hand/wrist joints of RA patients, 25 were clinically tender and/or swollen (38%), i.e., 5, 15, and 5 in patients 1, 2, and 3, respectively. Clinical symptoms of arthritis (tenderness and/or swelling) were present in wrists (5/66; 8%), metacarpophalangeal joints (15/66;22%), and proximal phalangeal joints (5/66; 8%). [124I]I–F8–IL10 was injected without any side effects occurring during the observation period.

Joint Targeting of [124I]I–F8–IL10 in RA Patients

Detailed images of wrists/hands showed evident targeting in joints (Figure A). The highest uptake and target-to-background ratios were determined at 24 h p.i. Within the one field of view of the hands, a total of 57 hand joints across the 3 patients could be imaged using [124I]I–F8–IL10, and a cumulative number of 38 PET-positive joints were observed (11–5 joints per patient). Out of 38 [124I]I–F8–IL10 PET-positive joints, 19 were also clinically swollen and/or tender (50%). In addition, there was enhanced uptake in 19 hand joints that lacked clinical signs of arthritis. The remaining 19/57 joints were PET-negative, of which 17 (89%) joints were also clinically negative.

Figure 1

(A) Example of a [124I]I–F8–IL10 PET-CT scan of the hands of RA patient 3 with clinically active disease. (B) [124I]I–F8–IL10 PET scan of RA patient 1 showing clear tracer uptake in liver and spleen. %ID = percentage of the injected dose.

The mean standard uptake value of PET-positive joints was 0.25 ± 0.15, and the mean target-to-background ratio was 2.5 ± 1.2. PET-positive joints that were also clinically active had 1.5-fold higher target-to-background ratios than those of clinically silent PET-positive joints (i.e., subclinical joints) (target-to-background ratios 2.6 ± 1.2 vs 1.6 ± 0.8, respectively).

Whole-Body Distribution of [124I]I–F8–IL10 in RA Patients

Blood sample analysis in RA patients indicated very fast clearance of [124I]I–F8–IL10 from the circulation after i.v. injection of the drug. In fact, after 5 min, less than ∼1% of the injected dose was detected in blood (Figure ). Another remarkable finding was the high uptake of [124I]I–F8–IL10 in liver (48.0 ± 3.5% of the injected dose, %ID) and, to a lower extent, spleen (5.3 ± 3.5%ID), as seen in the whole body PET-CT scans performed at 1 h p.i. (Table , Figure B). Thereafter, [124I]I–F8–IL10 liver uptake decreased from 48.0 (%ID) to 13.4 (%ID) at 72 h p.i. (Table ).

Figure 2

Pharmacokinetic analysis of [124I]I–F8–IL10 in (A) whole blood and (B) plasma of three RA patients.

Table 2

Organ Uptake of [124I]I–F8–IL10 Expressed as a %ID 1 and 72 h p.i.

	%ID at 1 h p.i.	%ID at 72 h p.i.
	mean ± SD	mean ± SD
heart (blood pool)	0.005 ± 0.002	0.001 ± 0.0007
kidneys	1.5 ± 0.4	0.5 ± 0.3
liver	48.0 ± 3.5	13.4 ± 2.7
spleen	5.3 ± 3.5	1.4 ± 1.5
bone marrow	0.2 ± 0.1	0.004 ± 0.001

To further understand the high uptake in liver and spleen, animal experiments were performed.

Effect of Radioiodination on the in Vivo Biodistribution of F8-IL10 in Healthy Mice

To investigate whether radioiodination affected F8-IL10, resulting in increased uptake in liver and spleen, healthy mice were analyzed for biodistribution after i.v. injection of F8-IL10 labeled according to the chloramine-T ([124I]I–F8–IL10) and iodogen method ([131I]I–F8–IL10), respectively (Figure ). At 10 min p.i., both tracers had comparable blood levels of 16–25%ID/g, decreasing to low values, <0.3%ID/g, at 24 h p.i. Biodistribution in various organ tissues was also similar over time, with no accumulation in liver or spleen at 24 h. All tissues showed clearance of both tracers within the time span of 24 h. Urine excretion was similar.

Figure 3

Ex vivo tissue distribution of healthy mice with [131I]I–F8–IL10 (iodogen) and [124I]-F8-IL10 (chloramine-T) at 10 min and 24 h. The results are expressed as a percentage of the injected dose per gram (%ID/g ± SD).

Ex Vivo Biodistribution of [124I]I–F8–IL10 in Arthritic Rats

After arthritis induction, all rats showed macroscopic thickening of the right, arthritic knee compared with the contralateral knee (data not shown). None of the rats showed any adverse effects, and no major changes in body weight were observed. Similar to observations in RA patients, uptake in arthritic knee joints was found in the F8-IL10low group. In this group, uptake in the arthritic knee joint of [124I]I–F8–IL10 was ∼2× higher than that (p < 0.01) in the (right) knee joint of a healthy (nonarthritic) rat (Figure ). Specificity of uptake was demonstrated by a ∼14-fold increased uptake of F8-IL10 in the arthritic knee joints as compared with uptake of the control antibody KSF-IL10 (p < 0.001). Interestingly, up to two times higher uptake of F8-IL10 was also observed in liver and spleen of arthritic rats as compared with that in healthy rats, while uptake in other organs was comparable between arthritic and healthy rats (Figure ). When the dose of F8-IL10 was increased significantly (F8-IL10high group), [124I]I–F8–IL10 uptake decreased in several tissues with a pronounced decrease in arthritic knee joints, liver, and spleen (p < 0.01), while blood levels were comparable at 24 h p.i.

Figure 4

Ex vivo tissue distribution at 24 h p.i. in healthy rats, [124I]I–F8–IL10low (white bars), and in arthritic rats, [124I]I-KSF-IL10-control (black bars), [124I]I–F8–IL10low (light gray bars), and [124I]I–F8–IL10high (dark gray bars). The results are expressed as percentage injected dose per gram (%ID/g ± SD).

Investigation of the Effect of the Disease RA on Liver and Spleen Expression of Fibronectin ED-A

Increased expression of fibronectin ED-A in liver and spleen was observed in arthritic rats compared with healthy rats (Figure ). In Figure , the white pulp (histology, Figure A) with fibronectin ED-A expression in the spleen of healthy (Figure B) and arthritic (Figure C) rats is shown. Fibronectin ED-A expression in spleen was ∼8× (p < 0.001) increased in arthritic rats compared with healthy rats (Figure D). Further, fibronectin ED-A expression in liver (hepatocytes) (histology, Figure E) of healthy rats (Figure F) and arthritic rats (Figure G) was observed. Fibronectin ED-A expression in liver was ∼7× (p < 0.001) increased in arthritic rats compared with healthy rats (Figure H).

Figure 5

Histopathology and immunofluorescence images of fibronectin ED-A stainings on arthritic and healthy rats spleen and liver sections. (A) HE staining of spleen (10× magnification). (B, C) Fibronectin ED-A staining (40× magnification) of healthy spleen (B) and arthritic spleen (C). (D) The total fluorescence intensity of healthy spleen vs the arthritic spleen (total fluorescence intensity E+06 ± SD). (E) HE staining of liver (10× magnification). (F, G) Fibronectin ED-A staining (blue channel, nucleus, and green channel, fibronectin ED-A) of healthy (F) and arthritic liver (G). (H) The total fluorescence intensity of healthy liver vs the arthritic liver (total fluorescence intensity E+06 ± SD).

Discussion

In this study, both targeting performance and in vivo biodistribution of the potential F8-IL10 antirheumatic drug were investigated using a translational approach. First of all, a feasibility PET-CT study in RA patients demonstrated that F8-IL10 targets clinically inflamed joints and also subclinically affected joints. Similarly, targeting was also found in arthritic joints of rats. The RA patient PET study revealed unexpected in vivo biodistribution. Very rapid blood clearance (less than 1% detectable in less than 5 min) and high liver and spleen uptake on PET-CT (within 1-h p.i. of [124I]I–F8–IL10) were observed. Subsequent animal experiments showed that increased tracer uptake in liver and spleen is associated with increased, arthritis related, expression of fibronectin ED-A in liver and spleen. The latter may be the main cause of local accumulation of the tracer in RA patients. In addition, a lower F8-IL10 administration dose (microdose of F8-IL10 administered in RA patients) may also result in relatively high uptake in liver and spleen, although our experiments did not allow definite conclusions for this mechanism.

Although our targeting data of F8-IL10 in arthritic joints were found in a small study of only 3 RA patients, these observations were supported by our arthritic rat data and are in line with targeting results previously found in a collagen-induced mouse model.[16] This provides further evidence of the feasibility of targeted delivery of IL10 by binding to fibronectin ED-A in arthritic joints, also reflected by the promising initial phase Ib therapeutic efficacy data of F8-IL10 in RA patients.[17,35]

The observed high uptake in liver (and to some extent in spleen), as well as the fast blood clearance of [124I]I–F8–IL10 in RA patients, was unexpected. In previous experiments in tumor-bearing mice, no increased F8-IL10 uptake in liver and spleen was observed, and blood clearance was clearly slower shortly after i.v. injection.[15] There are several potential explanations for the observed differences, (1) the use of different radioiodination methods (i.e., chloramine-T method in tumor-bearing mice and iodogen in RA patients), (2) effect of the RA disease on fibronectin ED-A expression in liver and spleen, (3) different injected doses of F8-L10, and (4) species differences.

First, the potential effect of the radioiodination method on the ex vivo biodistribution was addressed in healthy mice. F8-IL10 can potentially be modified by oxidation during the radioiodination procedure, which could result in rapid blood clearance and uptake in the mononuclear phagocyte system (liver/spleen).[36] However, no differences in liver and spleen uptake of F8-IL10 were found between both radioiodination methods. In fact, in healthy mice, liver and spleen uptake at 24 h p.i. were low, regardless of the radioiodination method. In addition, initial blood clearance was much slower (in line with previous observations in tumor-bearing mice[16]), with 15–25%ID/g present at 10 min p.i. in healthy mice as compared to <1% ID/L in the blood compartment of RA patients at 5 min p.i.

Second, hepatic and splenic F8-IL10 uptake of arthritic rats was more than twice as high as that in healthy rats. A high dose of F8-IL10 significantly decreased liver and spleen uptake of [124I]I–F8–IL10 in rats with experimental arthritis, indicating specific binding in these tissues. To explore whether this observation could be explained by increased fibronectin ED-A expression and hence increased F8-IL10 binding in arthritic rats, further immunohistochemical analyses were performed. Indeed, in arthritic rats, significantly higher fibronectin ED-A expression was found in liver and spleen as compared with healthy rats.

Third, a microdose of i.v.-administered agent may directly accumulate in high antigen expressing and well accessible organs upon the first passage through these organs (liver/spleen). This so-called antigen sink theory has also been observed by others in B-cell lymphoma patients injected with radiolabeled rituximab.[37] The RA patients were injected with a 500-fold lower dose of F8-IL10 than healthy mice (0.006 versus 3 mg/kg F8-IL10), where no accumulation of radioactivity in liver and spleen was observed. Thus, in the patient study, liver and spleen may have acted as antigen sinks for the low dosage of i.v.-administered F8-IL10, which could explain the fast blood clearance and relatively low uptake in arthritic joints. Upon increase of the F8-IL10 dose, liver and spleen may become saturated, resulting in higher plasma concentration, thereby leaving relatively more F8-IL10 available to bind to inflamed joints. To investigate this antigen sink theory, a high dose of F8-IL10 (∼2000-fold higher than that in humans) in arthritic rats was compared with the lowest feasible F8-IL10 dose for administration in rats (∼50-fold fold higher than in humans). Indeed, an uptake of radiolabeled F8-IL10 in the liver/spleen of the F8-IL10high group was significantly lower than that seen in the F8-IL10low group. However, uptake of radioactive F8-IL10 in several other tissues was decreased as well (data not shown), which presumably was, at least in part, related to blockade of local F8-IL10 binding sites. Therefore, these results could not definitely prove or rule out a possible role of liver and spleen as antigen sinks in the case of the F8-IL10low dose.

Finally, species differences between animals and humans cannot be ruled out. Findings in animals can be highly suggestive, but extrapolation to humans should be made with care. In the present study, direct dose comparisons were not possible since lowering the F8-IL10 dose (<0.3 mg/kg) in rats to allow for a direct comparison with the dose in RA patients (0.006 mg/kg) was technically not possible. Apart from tissue observations, differences in pharmacokinetics, stability of the tracer, and targeting between animals and patients may also play a role.

Conclusions

In conclusion, the imaging studies demonstrate that PET-guided drug development may provide significant information on the targeting potential of a new antirheumatic drug, although observations were made in a small group of 3 RA patients. [124I]I–F8–IL10 PET revealed promising targeting in arthritic joints in RA patients. In addition, PET whole body biodistribution data may point at tracer accumulation at other nontarget sites for F8-IL10, likely related to disease-induced antigen expression at these sites and possibly related to antibody dosing. These insights can add to optimal dosing schedules for targeting sites of interest as well as inform on potential sites of concern for side effects. So far, the human phase 1b data have not revealed major side effects. F8-IL10 was generally well tolerated up to a dose of 600 μg/kg.[35]