^99mTc-labelled S-HYNIC certolizumab pegol in rheumatoid arthritis and spondyloarthritis patients: a biodistribution and dosimetry study

Synthesis of ^99mTc-S-HYNIC CZP

All preparations were carried out under aseptic conditions working in a LAF IIa cabinet.

Derivatisation of CZP for injection with S-HYNIC

A 200-mg lyophilized CZP vial was reconstituted with water for injection (B. Braun, Melsungen AG, Germany); 100, 50 and 25 mg CZP was transferred to a Slide-A-Lyzer with cut-off of 10 kDa (Pierce Protein Research Products, Thermo scientific, Rockford, IL, USA) and dialyzed against 500 ml of a mixture of a Dulbecco’s phosphate-buffered saline (Lonza, Verviers, Belgium) and a self-prepared 0.9% m/v sodium chloride solution (Riedel-deHaën, Seelze, Germany) in 1:2 v/v ratio. Dialysis was maintained for 4 h at 2–8 °C, with the buffer refreshed after 1.5 h. Subsequently, 0.5 ml of a 8.4% sodium hydrogen carbonate (Merck, Darmstadt, Germany) solution was added to the solution followed by 10 5.0 μl portions of 1.7, 0.86 and 0.43% m/v solution of S-HYNIC (ABX GmbH, Radeberg, Germany) in dry DMSO (Merck, Darmstadt, Germany) at a pace of 1 portion/min [5]. This yielded an average of 2.8 S-HYNIC groups per CZP. After 30 min incubation at room temperature in the dark, the reaction was quenched by adding 3.0 ml cooled 0.15 M acetate buffer pH 5.0 (Merck, Darmstadt, Germany). The unreacted S-HYNIC was removed by dialyzing the reaction mixture in a Slide-A-Lyzer (cutoff of 10 kDa) overnight at 2–8 °C against 500 ml acetate buffer, which was refreshed after 1, 2 and 3 h. The solution was diluted to 40.0 ml with 0.15 M acetate buffer pH 5.0 and membrane filtered (0.22 μm). Following dispensing into 1.0 ml portions, the glass vials were stored at −80 or 2–8 °C for 3 months. Three concentrations of CZP were obtained: 2.5, 1.25 and 0.625 mg of S-HYNIC-coupled CZP. Quality control was done by determination of the protein concentration (BCA protein reagent) and the p-NBA HYNIC assay to measure the number of S-HYNIC bifunctional chelator coupled to the protein.

Preparation of the co-ligand kit

A solution containing 4.66 mM tin(II) sulphate (Sigma Aldrich, Steinheim, Germany) and 55.81 mM tricine (Sigma Aldrich, Steinheim, Germany) dissolved in ultrapure sterile and pyrogen-free water was prepared.

Radiolabelling with ^99mTc

Fifty-microliter co-ligand kit and 925 MBq (±10%) ^99mTc pertechnetate were consecutively added to the S-HYNIC CZP vial (2.5, 1.25 and 0.625 mg). After 15-min incubation, physiological saline was added in order to obtain a volume of 3 ml. Quality control was carried out by instant thin layer chromatography (iTLC) with SilG as stationary phase and 0.9% NaCl solution as mobile phase. For the clinical study, the 1.25 mg S-HYNIC CZP vials stored at −80 °C were used and the radiochemical yield needed to exceed 90%.

Stability study

The impact of aggregation on the chemical stability and radiochemical yield during storage of the formulation at three different concentrations (2.5, 1.25 and 0.625 mg) was studied over a 3-month period. Aggregate formation was assessed by size-exclusion HPLC (Agilent Zorbax Diol guard column), 4 × 12.5 mm, in series with a GF450, 9.4 × 250 mm and a GF250 size exclusion analytical column, 9.4 × 250 mm (Agilent Technologies, Diegem, Belgium). The mobile phase was composed of a mixture of a 200 mM phosphate buffer pH 7.0 and ethanol 90:10 v/v (1 ml/min, 30 min). Influence on the radiochemical incorporation of ^99mTc was studied by iTLC as described earlier. Analyses were performed after preparation, at 2 weeks, 1 month and 3 months post production.

In vitro activity of ^99mTc-S-HYNIC CZP against TNF-induced cytotoxicity

Murine fibrosarcoma TNF-sensitive L929s cells were cultured for 24 h by seeding 20,000 cells/well in 96-well plates, incubated at 37 °C with 5% CO2. The culture medium consisted of DMEM (GIBCO-BRL, 41965-062) supplemented with 10% fetal calf serum, 400 μM sodium-pyruvate (Sigma) and non-essential amino acids (Lonza). The following day, 55 μl human TNF (6.8 × 10⁷ U/ml) produced by the Protein Service facility of VIB (Ghent, Belgium) was added at various concentrations (10 dilutions starting at 300 U/ml) to the test solutions at room temperature. The following solutions of antibodies directed against TNF were tested: CZP, S-HYNIC CZP, ^99mTc-S-HYNIC CZP and infliximab (Remicade, Merck, Johnson & Johnson). Dilutions of 250, 50 and 10 ng/ml anti-TNF antibodies (55 μl) were preincubated with the human TNF for 45 min, as well as control samples without TNF blocking agents. Actinomycin D (Sigma, 1 mg/ml dissolved in absolute ethanol), at a final concentration of 1 μg/ml was also added. Twenty-four hours following the exposure to the test solutions, TNF cytotoxicity was measured using the MTT-test by adding 20 μl filter sterilized 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT, Sigma, 5 mg/ml). After 4 h, 80 μl stopping solution (10% SDS, 0.01 M HCL) was added and the read out was performed by a multichannel plate reader at a wavelength of 595 nm with a reference at 655 nm. The percentage of cell survival was estimated according to the formula below:

% cell survival = (absorbance treated cells − background) / (average absorbance non treated cells − background)*100.

Patients

The study was conducted in accordance with the Helsinki declaration and was approved by the ethical committee of the Ghent University Hospital. All patients signed an informed consent. Patient selection was limited to subjects aged 18–70 years and meeting the ‘American College of Rheumatology’ (ACR) criteria for RA [6] or the ‘Assessment of SpondyloArthritis international Society’ (ASAS) criteria for axial or peripheral SpA [7, 8]. Patients were not allowed if they had previously been treated with CZP or any other biological treatment. For more detailed information on the clinical eligibility criteria, we refer to the online detailed study protocol (EudraCT number: 2009-017998-37, http://clinicaltrials.gov/show/NCT01590966).

Scan procedure, blood and urine sampling

All patients were scanned on the same double-headed gamma camera (BrightView, Philips Healthcare, Best, the Netherlands). First, an attenuation map was obtained from a whole-body (WB) scan using a Cobalt-57 flood source. Subsequently, the patient was injected IV with 10.6 MBq/kg ^99mTc-S-HYNIC CZP. The net IA was calculated following correction for rest activity in the syringe and exact time of administration (Veenstra ionization chamber VIK-202, Veenstra Instruments, the Netherlands). WB images (15 cm/min, 1024 × 512 matrix, pixel size 2.80 mm) were performed immediately following administration, at 1, 4–6 and 24 h post-injection (pi). A standard activity of approximately 5 MBq ^99mTc in an unshielded syringe was always in the field of view for quantification purposes. Static images (5 min, 256 × 256 matrix, pixel size 2.33 mm) of hands and feet were acquired immediately following the first WB scan, at 4–6 and 24 h.

All visualized organs (heart, lungs, liver, spleen, kidneys), as well as regions of interest (ROIs) for the WB, background and standard activity were delineated manually on the four geometric mean images, generated from the WB scans (Nuclear Diagnostics, Stockholm, Sweden). For each organ, attenuation correction factors were calculated based on measured conversion factors on our camera from ⁵⁷Co to ^99mTc. These conversion factors were derived from sequential attenuation scans with a ^99mTc flood source and a ⁵⁷Co flood source in the first patient. The activity measured in the urine collections between the scans was used to correct the estimated activity in the WB compartment. Cumulated activities in the WB and various organs were estimated using mono-exponential fitting and converted to time integrated activity coefficients to be processed by OLINDA/EXM 1.1 software (Stabin M. Vanderbilt University, Nashville) to estimate the absorbed organ doses. The adult male and female mathematical reference phantoms were applied. A bladder voiding interval of 4 h was included in the model. In OLINDA/EXM 1.1, effective doses (ED60) were directly calculated using ICRP 60 tissue weighting factors [9]. In order to account for the recent ICRP 103 recommendations, the effective dose was recalculated manually by combining the ICRP 103 tissue weighting factors and the organ doses from the OLINDA/EXM 1.1 output (ED103) [10]. As some organs contributing to ED103 are not listed in the OLINDA/EXM 1.1, the sum of tissue-weighting factors for the available target organs were calculated (0.913 for men and 0.922 for women). In the ED103 calculation, these factors were accounted for by scaling [11].

Urinary tracer excretion was measured in sequential urine collections at 1, 4–6 and up to 24 h pi. Total volumes were recorded. Blood sampling was performed immediately following IV tracer injection from a vein in the contralateral arm and subsequently at 1, 4–6 and at 24 h pi. ^99mTc activity was measured in 2 ml blood and urine samples in triplicate on a calibrated NaI(Tl) 3″ × 3″ gamma well counter (Cobra II, PerkinElmer, USA). The raw count data generated from the NaI(Tl) 3″ × 3″ detector were converted to kBq by a calibration curve. The standard samples—in the same geometry as the urine and blood samples—were prepared from a stock solution which was measured in the same calibrated dose calibrator (ionization chamber) used to measure the ^99mTc-HYNIC CZP syringes. Results were decay corrected and expressed as kBq/ml and as %IA.

To assess the tracer accumulation in the peripheral joints, we scored each joint visually and semiquantitatively. The latter was done by manually drawing ROIs around each joint (Hermes, Nuclear Diagnostics, Sweden). For each individual scan, a background ROI was defined within the field of view, e.g. left supraclavicular region on WB, right forearm on static images of the wrists and hands, right distal tibia for the static images depicting the ankles and feet. These results were compared with the findings on clinical examination for each assessable joint. The nuclear medicine physician reading the immunoscans and the clinician performing the clinical examination were blinded to each other’s observations. Groups were compared by Mann-Whitney U test using IBM SPSSv21 software (NY, USA).

Preparation of S-HYNIC CZP

After preparation, the 2.5-, 1.25- and 0.625-mg vials contained respectively 105, 98.5 and 99.8% of the label claim for the total amount of protein indicating that no protein losses were observed during the different handling steps. The results of the p-NBA S-HYNIC incorporation test showed a respective ratio of S-HYNIC/CZP of 1.2, 1.5 and 2.0. No substantial losses were observed for the vials stored at −80 °C. In contrast, all formulations of the vials stored at 2–8 °C suffered from a decrease of the S-HYNIC/CZP ratio <0.1 after 3 months of storage.

The influence of storage temperature on aggregation and radiochemical yield

At −80 °C, no change in the amount of aggregation was observed over time; nevertheless, the amount of aggregation (mean for the 3 formulations: 1.7%) was somewhat higher compared to the CZP blank formulation (<0.5%). At 2–8 °C, however, aggregation increased over time with the highest amount (20%) of aggregation observed for the 2.5 mg CZP formulation after 1 month of storage. All formulations stored at-80 °C had radiochemical yields >95%. For the storage at 2–8 °C, all radiochemical yields were >95% except for the 2.5 mg formulation of which the yield decreased to 55%.

In vitro testing of the specific TNF neutralizing activity of the ^99mTc-S-HYNIC CZP formulation

In the samples exposed to ^99mTc-S-HYNIC CZP, the cytotoxic activity of TNF was completely blocked at doses as low as 10 ng/ml (Fig. 1). The TNF neutralizing effect of the ^99mTc-S-HYNIC CZP formulation was comparable to the effect obtained with the unlabelled CZP with or without the HYNIC linker. Infliximab, another TNF blocker that was used as a positive control, also showed good neutralizing activity, although it was less active at the doses used (Table 1).

	Mean % cell survival (SD)
Tested TNF blockers at 3 concentrations	99mTc-S-HYNIC CZP	S-HYNIC CZP	CZP	Infliximab
Control (no TNF blocker)	0.0 (0.0)	0.0 (0.0)	0.0 (0.0)	0.0 (0.0)
250 ng/ml	109.1 (2.6)	111.9 (0.5)	81.8 (1.5)	81.1 (3.9)
50 ng/ml	113.8 (4.2)	111.0 (2.2)	96.1 (23.2)	57.4 (1.4)
10 ng/ml	115.2 (0.6)	88.5 (0.3)	56.6 (2.8)	5.2 (0.0)

Patient study

We analysed eight patients (four males and four females): one RA patient, two peripheral SpA patients and five axial SpA patients. The median age was 44 years (range 26–68). A mean (standard deviation, SD) activity of 690 (SD 35, range 665–728) MBq ^99mTc-S-HYNIC CZP (1.07 mg, SD 0.04, range 1.01–1.10) was injected. There were no adverse or clinically detectable pharmacologic effects in any of the subjects.

On the WB images, we observed high tracer uptake in the bloodpool and liver and less pronounced uptake in the spleen and kidneys. WB scintigraphy and static images of the hands of a woman suffering axial SpA are depicted in Fig. 2 and illustrate the ‘normal peripheral distribution’ of ^99mTc-S-HYNIC CZP. Patients with peripheral joint involvement showed typical patterns on immunoscintigraphy with a poly-articular pattern in RA and involvement of the distal interphangeal joints, dactylitis and tenosynovitis in psoriatic arthritis (Figs. 3 and 4). The tracer uptake in the affected joints persisted over time. In the patient with RA, 16 joints showed marked tracer accumulation and 4 of these joints were clinically swollen. Out of the 50 negative joints on scintigraphy, all but 1 were also negative on clinical examination for swelling. Semiquantitative analysis yielded a median uptake ratio in swollen joints of 3.3 (range 0.6–7.5) vs 1.3 (range 0.8–5.0) in joints negative for swelling on clinical examination (p < 0.001 Mann-Whitney U test). In the 2 patients with peripheral SpA, 29 joints were visually scored positive on the immunoscan of which 20 joints were swollen on clinical assessment. Amongst the 147 joints negative on scintigraphy, 144 were also clinically classified as not swollen. Semiquantification yielded a median uptake ratio of swollen joints of 2.4 (range 1.0–6.3) vs. 1.6 (range 0.3–3.3) in non-swollen joints (p < 0.001 Mann-Whitney U test).

According to our dosimetry study, the highest organ doses were attributed to the kidneys, spleen and liver with estimates of respectively 38.7 (SD 5.1) mGy, 23.4 (SD 4.3) mGy and 22.6 (SD 4.8) mGy or 56 (SD 7), 34 (SD 6) and 33 (SD 7) μGy/MBq. Table 2 summarizes our findings for the various organs included in the analysis. The estimated mean ED60 and ED103 were 8.93 (SD 1.14) μSv/MBq and 8.89 (SD 1.31) μSv/MBq, respectively. Taking into account the administered activity of 690 (SD 35) MBq, a mean ED60 and ED103 of 6.2 (SD 0.9) mSv and 6.1 (SD 1.0) mSv is obtained. Figure 5 shows the average time-activity curves as a percentage of the IA for the different organs. It can be observed from this figure that the mono-exponential fit adequately models the data. The urinary excretion was 15.1% (SD 8.1) of the IA at 22.5 h. The cumulative urinary excretion is tabulated in Table 3. The blood clearance shows a biphasic bi-exponential shape. Blood samples were fitted to a two-compartment model by the least square method. As illustrated in Fig. 6, a distribution half-life of 1.2 h (SD 1.5) and an elimination half-life of 26.9 h (SD 2.7) were estimated.

Organs	Organ dose μGy/MBq Mean and SD		Total organ dose mGy Mean and SD
Brain	1.3	0.5	0.9	0.3
Skin	1.8	0.3	1.2	0.2
Thyroid	2.1	0.4	1.4	0.3
Uterus	4.6	0.5	3.3	0.4
Ovaries	3.9	0.5	2.7	0.4
Testes	1.7	0.5	1.1	0.4
Breasts	3.7	0.5	2.5	0.4
Red marrow	3.8	0.4	2.6	0.3
Muscle	3.5	0.5	2.4	0.3
Small intestine	4.3	0.6	3.0	0.4
ULI wall	4.9	0.7	3.4	0.5
LLI wall	3.1	0.7	2.2	0.5
Stomach wall	6.3	1.0	4.3	0.7
Thymus	6.9	0.8	4.7	0.5
Osteogenic cells	7.4	1.4	5.1	1.0
Pancreas	10.8	1.6	7.4	1.1
Adrenals	11.2	1.7	7.7	1.2
Gallbladder wall	11.9	1.8	8.2	1.2
Lungs	18.6	2.7	12.8	1.9
Urinary bladder wall	19.9	8.3	13.7	5.8
Heart wall	30.6	5.1	21.1	3.5
Liver	32.7	6.9	22.5	4.7
Spleen	34.0	6.2	23.4	4.3
Kidneys	56.1	7.3	38.7	5.1

Interval p.i Mean (SD)	Cumulative urinary excretion Mean (SD) %IA
1.5 h (0.3)	4.3% (2.1)
4.5 h (0.4)	8.5% (3.2)
22.5 h (1.4)	15.1% (8.1)