99mTc-NTP 15-5 is a companion radiotracer for assessing joint functional response to sprifermin (rhFGF-18) in a murine osteoarthritis model

With the emergence of disease modifying osteoarthritis drugs (DMOAD), imaging methods to quantitatively demonstrate their efficacy and to monitor osteoarthritis progression at the functional level are urgently needed. Our group showed that articular cartilage can be quantitatively assessed in nuclear medicine imaging by our radiotracer 99mTc-NTP 15-5 targeting cartilage proteoglycans. In this work, surgically induced DMM mice were treated with sprifermin or saline. We investigated cartilage remodelling in the mice knees by 99mTc-NTP 15-5 SPECT-CT imaging over 24 weeks after surgery, as wells as proteoglycan biochemical assays. OA alterations were scored by histology according to OARSI guidelines. A specific accumulation of 99mTc-NTP 15-5 in cartilage joints was evidenced in vivo by SPECT-CT imaging as early as 30 min post-iv injection. In DMM, 99mTc-NTP 15-5 accumulation in cartilage within the operated joints, relative to contralateral ones, was observed to initially increase then decrease as pathology progressed. Under sprifermin, 99mTc-NTP 15-5 uptake in pathological knees was significantly increased compared to controls, at 7-, 12- and 24-weeks, and consistent with proteoglycan increase measured 5 weeks post-surgery, as a sign of cartilage matrix remodelling. Our work highlights the potential of 99mTc-NTP 15-5 as an imaging-based companion to monitor cartilage remodelling in OA and DMOAD response.

www.nature.com/scientificreports/ research on DMOAD, there is currently no pharmacological intervention approved for use in Europe or the USA demonstrating efficacy for modifying OA progression 5,6 . Even if promising DMOADs have emerged, the major challenge in demonstrating the proof of concept is to overcome (i) the absence of a precise assessment of the disease, particularly in the early stages, and (ii) the lack of consensus on how to detect joint changes and link them to clinically meaningful endpoints. Therefore, new treatments are needed, as well as new methods to evaluate their efficacy and to monitor OA progression. Authors indeed consider that research and development towards DMOADs is hampered by the lack of specific and sensitive imaging tools to reliably quantify OA progression, and monitor response to therapy, including radiographs and MRI 7,8 . Higher sensitive imaging approaches appear of real need to evaluate OA progression and to monitor response to innovative therapies 8 . Few DMOAD currently under development are aiming at preventing cartilage deterioration and/or restoring cartilage thickness 5,6 . Among them, sprifermin, a recombinant human Fibroblast Growth Factor 18 (rhFGF- 18), appeared as the most promising molecule. In rats 9 and in human explants 10,11 , it has been demonstrated that FGF-18 increased chondrocytes proliferation, as well as ECM deposition and PG production, leading to an increase in cartilage thickness. The strong anabolic effect of sprifermin seems to follow a sequential process, with an early increase of aggrecanase activity leading to aggrecan degradation, described as a prerequisite for cell proliferation, followed by ECM molecules production by newly produced chondrocytes, including sustained PG production, and ultimately cartilage regeneration 11 . To date, sprifermin is the only candidate DMOAD with a proven structural effect on cartilage. The clinical relevance of the results obtained is currently being discussed [12][13][14] . This makes it the only possible gold standard to date for a study to investigate a new surrogate marker of cartilage remodelling.
In such a context, our group aims at validating a nuclear medicine imaging strategy targeting cartilage PG in vivo, which could provide a functional access to joint and a new method to evaluate innovative therapies. Because cartilage contains up to 10% PG consisting of mainly chondroitin sulphate aggrecan which chains are negatively charged 15 , our strategy is based on the use of a bi-functional agent, the radiotracer 99m Tc-N-(triethylammonium)-3-propyl- [15]ane-N5 ( 99m Tc-NTP , that contains in its structure a positively charged quaternary ammonium function for binding to PG and a polyazamacrocycle to complex 99m Tc. Based on many preclinical studies, we believe that 99m Tc-NTP 15-5 and in vivo functional imaging of PG could provide a suitable set of criteria for quantifying cartilage functionality, and the efficacy of new emerging therapeutic strategies [16][17][18][19][20][21] .
In this study, we determined the relevance and sensitivity of 99m Tc-NTP 15-5 imaging in the destabilization of the medial meniscus osteoarthritis murine model, for assessing the structural effect of sprifermin, with the opportunity to bridge the gap between preclinical and clinical testing.

Results
SPECT functional cartilage imaging with 99m Tc-NTP 15-5 in healthy mice (Fig. 1). 99m Tc-NTP 15-5 accumulation was evidenced from 30 min post-injection in many articular structures of the animals, such as knees, humeral head, and intervertebral disks (Fig. 1A). As evidenced on the coronal and sagittal slices, nonspecific accumulation was observed in liver, digestive tract, and bladder. At the knee level, scintigraphic images allow the discrimination of femoral condyle uptake from that of tibial plateau (Fig. 1B). Characterization of OA model: histology (Fig. 2). OA defects in DMM mice were confirmed by histology 12 weeks following DMM surgery. As illustrated in Fig. 2, cartilage erosion was clearly evidenced on the tibial plateau as a decrease in thickness and proteoglycan loss as a staining intensity decrease (Fig. 2).

99m
Tc-NTP 15-5 was previously demonstrated to be a candidate for the in vivo functional nuclear imaging of proteoglycan remodeling in degenerative pathologies of cartilage [16][17][18][19][20][21] . In the present study, we evaluated in the surgically induced DMM model the relevance of 99m Tc-NTP 15-5 nuclear imaging for monitoring in vivo cartilage remodelling in response to sprifermin (rhFGF-18), an anabolic agent which previously demonstrated cartilage repair properties [9][10][11][12][13][14] . The surgical model of destabilization of the medial meniscus (DMM) has become a gold standard for studying the onset and progression of osteoarthritis 22 . Some studies reported mild cartilage lesions in mice as early as 2 weeks after surgery, with regional proteoglycan loss, chondrocyte clustering and osteophyte formations, with lesions being deeper into the calcified zone 10 weeks after surgery 23 . In our work, joint alterations were scored at 12-and 24-weeks following surgery according to OARSI recommendations 24 and confirmed OA features in DMM, with OARSI scores higher than controls (Fig. 2, 4 and 5). In the DMM model, 99m Tc-NTP 15-5 accumulation in cartilage within the operated knee, relative to the contralateral one, was observed to change in the same animals as pathology progressed: PK/CK ratios of radiotracer uptake increased in the DMM group compared to the sham group at 7-and 12-weeks. Since 99m Tc-NTP 15-5 has been previously demonstrated, by Surface Plasmon Resonance to specifically bind to aggrecan (unpublished data) and was evidenced by autoradiographic studies to accumulate in knee cartilage of guinea pigs 16 , we consider that 99m Tc-NTP 15-5 imaging provides the assessment of joint integrity at the PG level. Such PK/CK scintigraphic ratio increase after surgery agrees with previously published results from our group in the meniscectomy and papain models of OA 16,25 . Hypertrophic changes in PG have been widely described in OA cartilage as a first line reparative reaction via www.nature.com/scientificreports/ an early increased synthesis of extracellular matrix (ECM) in damaged areas being followed at a later stage by degenerative destruction [26][27][28][29][30][31] .
We then evaluated sprifermin treatment on OA evolution in DMM treated at doses of 1-or 10-µg respectively to DMM receiving PBS, DMM without any treatment, and shams. Although not significant at 12 weeks, mean OARSI score in the DMM SPRI10 group was lower than DMM, DMM PBS and DMM SPRI1 (only 2 mice were   The progressive changes in PG content of the extracellular matrix were monitored in vivo by 99m Tc-NTP 15-5 SPECT-CT imaging at 7-, 12-and 24-weeks post-surgery. An increase in PK/CK ratio was observed in sprifermin treated DMM respectively to the other DMM groups. From W7, sprifermin (3 × 10 µg) induced a significant increase of 99m Tc-NTP 15-5 ratios respectively to DMM PBS, as a sign of early cartilage remodelling. Considering DMM SPRI1, scintigraphic ratio increase became statistically different from controls only from 12 W. This tends to suggest that sprifermin effect on tracer accumulation is related to the dose that was administered. From these results, it seems interesting to refer to those of Moore and al. obtained in a rat model of OA and observing a significant effect of sprifermin on cartilage thickness at W6 after intra-articular injections of sprifermin, 5 µg  www.nature.com/scientificreports/ bi-weekly over 3 weeks, while no significant effect was observed with the 1-µg dose 9 . DMM SPRI1 showed a higher mean effect on scintigraphic ratio at 24 W. However, differences were not statistically significant. It would be of interest to increase the DMM SPRI1 sample size to further study this. Finally, using a dedicated cohort of mice receiving sprifermin 10 µg or PBS, imaging parameters and joints PG content were determined on the same animals, at 5 weeks after DMM surgery. For these animals, increase in scintigraphic ratios in response to sprifermin was concomitant with proteoglycan ratio increase. Such relation between 99m Tc-NTP 15-5 scintigraphic ratio and proteoglycan content was previously demonstrated in an experimental model of rheumatoid arthritis, under meloxicam treatment thus reinforcing the specificity of 99m Tc-NTP 15-5 to monitor the effect of treatments as well as monitoring disease progression 21 .
From our results, 99m Tc-NTP 15-5 radiotracer demonstrated its ability to assess in vivo cartilage repair by the anabolic agent Sprifermin, which has emerged as one of the most promising DMOAD [12][13][14][32][33][34] . For clinical development programs for DMOADs, the current regulatory guidance determined that approval requires inhibition of loss in joint space width (JSW) on plain radiograph, associated with symptomatic benefit 35,36 . 99m Tc-NTP 15-5 imaging has the opportunity to be a complementary method to conventional anatomical imaging such as radiography and MRI, used in routine to evaluate loss in JSW and cartilage. However, due to the slow progression of the disease, these anatomical techniques have a limited value for short duration DMOAD trials 37 .
For many years, nuclear medicine applications for OA progression evaluation have been under investigations 38 . To our knowledge, the radiotracers currently available for clinical OA imaging provide indirect evaluations of the pathology, such as bone and inflammation, which are useful for guiding choice of therapy but not for providing an early index of OA. Using PET in patients, 18 F-NaF has recently been used to assess bone remodelling in knees of healthy and OA subjects 37 and has been suggested for the evaluation of bone involvement in OA progression, in complement of other imaging techniques, such as MRI, which could measure other parameters of interest (cartilage thickness). 18 F-Fluoro-Deoxy-D-Glucose ( 18 F-FDG) has also been proposed to investigate metabolic changes in knee OA 39 and to identify inflammatory regions of the joints in symptomatic patients 40 , thus limiting its potential applications to symptomatic OA. Regarding SPECT-CT studies, many addresses quantitative bone metabolism imaging with 99m Tc-DPD uptake being highly correlated with other imaging parameters for OA activity of the medial compartment of the knee 41 . Other radiotracers have been proposed to monitor the progression of OA disease. Activated macrophages were observed in rat models of OA using a folate receptor targeting radiotracer and SPECT 42,43 . Sobal et al. demonstrated increased joint to background ratio of 99m Tc labelled chondroitin sulphate ( 99m Tc-CS, which accumulates in cartilage inversely to negative charges of GAGs present in the ECM of the tissue) in dogs with OA with a significant positive correlation between grade of disease and scintigraphic ratio 44 . In this context, SPECT-CT imaging with 99m Tc-NTP 15-5 offers the unique advantage to be directly related to the negatively charged GAGs of PG [16][17][18][19][20][21] , responsible for the fixed charged density involved in biomechanical properties of articular tissue 45 . Conclusions. Our experimental results in the DMM model of OA treated by sprifermin demonstrate the proof of concept that 99m Tc-NTP 15-5 imaging could provide an early and sensitive tool to evaluate cartilage remodelling and DMOAD response.
These promising results are in favour of reinforcing first into humans transfer for evaluating the sensitivity of 99m Tc-NTP 15-5 imaging in patients with OA, that we have recently initiated (NCT04481230).

Animals.
Animal experiments were conducted in compliance with the 2010/63/UE European Directive and with the relevant guidelines and regulations. In addition, all experiments were authorized by the French Ministry of Research after approval by the local ethical committee of Clermont-Ferrand (C2EA-02, approval n°04.623.02 and n°12160-2017111317154154v3). Experiments were conducted on a total of 104 C57BL/6j male mice (Charles River Laboratories). Mice (12 weeks old at time of surgery) were housed in standard conditions (n = 5 per cage, with enrichment), in ventilated racks, with a 21-24 °C temperature comprising 60% humidity and a 12 h light / 12 h dark cycle with access to food and water ad libitum. Animal health and wellbeing was monitored daily by competent persons. Each cage was allocated to a single treatment group. In addition, mice were ear marked under general gas anaesthesia using isoflurane at 2.5 per cent and 2:3 oxygen. All intravenous injections were made in lateral tail vein of vigil mice. For in vivo imaging, mice were placed under general gas anaesthesia. Euthanasia was performed by cervical dislocation after isoflurane gas overdose.
Destabilization of medial meniscus experimental OA model. The DMM surgery has been performed according to already described procedures 46 . Briefly, a skin incision was performed, before opening the right knee joint capsule of the hind legs. The medial meniscus of the right knee was destabilized by cutting the medial menisci-tibial ligament. The "sham" intervention corresponds to the same procedure without meniscal destabilization. Afterwards, skin was closed with stitches, the joint capsule healing spontaneously. During the procedure, mice were placed under general gas anaesthesia (induction with isoflurane 4%, then 1.5% during the procedure). Anaesthesia induction was followed by an intraperitoneal injection of Buprenorphine (Buprecare, 0.05 mg/kg) for analgesia induction. Buprenorphine injections were performed following surgery every 8-12 h for 24-48 h.

Study design. Model characterization.
A first cohort of mice was used to verify the phenotype of DMM mice on knee joint lesions accordingly to the paper of Sonia Glasson 24 . We characterized osteoarthritic structural lesions in the DMM model at 12 weeks post-surgery using histology: www.nature.com/scientificreports/ • Shams: n = 1 experiment corresponding to a total of 6 animals, • DMM: n = 1 experiment corresponding to a total of 8 animals.
Sprifermin therapeutic protocol. Sprifermin was administered 2 weeks after OA surgical induction by injections into the joint cavity of knees through a 33-G needle of either 1 or 10 µg in 4 µl Phosphate Buffer Saline (PBS) over 3 weeks (1 injection per week, i.e., up to 5 weeks post DMM surgery). As a control for treatment, PBS alone was administered according to the same schedule as sprifermin.
Monitoring of sprifermin effect. The monitoring of sprifermin effect using SPECT-CT imaging and histology started 7 weeks post-surgery, i.e., 2 weeks after the end of treatment, and ended 24 weeks post-surgery. Two independent experiments were conducted, except for the sprifermin 1 µg condition. At initiation of each experiment, animals were randomly divided in 5 groups to have large enough data set to determine statistical significance up to 24 weeks. Groups were set as follows: • Mice were imaged at 7-, 12-and 24-weeks post DMM surgery. Among these mice, a total of 12 Shams, 10 DMM, 19 DMM PBS, 10 DMM sprifermin 1 µg, and 14 DMM sprifermin 10 µg were used for histology and OARSI scoring at 12-and 24-weeks post-surgery.
Proteoglycan content and imaging. An additional protocol was especially dedicated to compare in the same animals, imaging parameters and proteoglycan dosage 5 weeks after surgery. A total of 10 animals was included in this cohort, consisting of: • DMM with PBS treatment: n = 1 experiment corresponding to a total of 5 animals, • DMM with sprifermin treatment at 10 µg per injection: n = 1 experiment corresponding to a total of 5 animals.
In vivo SPECT-CT functional imaging of joint with 99m Tc-NTP 15-5. SPECT-CT imaging was performed 30 min after the intravenous injection of 20 MBq of 99m Tc-NTP 15-5. Multimodal SPECT-CT imaging was performed using a NanoScan SPECT-CT camera (Mediso Ltd) equipped with four detectors and multi pinhole collimation (APT62). Mice were anesthetized with isoflurane (4% for induction, and 1.5% during images acquisition) and placed in a Multicell Mouse L bed (Mediso Ltd) with temperature control (37 °C). Nucline software (Nucline 3.00.018, Mediso Ltd) was used for image acquisitions and reconstructions. CT parameters: helical scan with 480 projections (300 ms per projection), 50 kV, 590 μA, pitch 1.0, binning 1:4 and field of view: max. SPECT acquisitions parameters: images were acquired within the CT scan range, with resolution set to "standard". The time per projection was determined in accordance with the detected radioactivity (most frequently used: 30 s). SPECT images reconstruction were conducted using TeraTomo3D (Nucline v3.00.018, Mediso Ltd) with a normal dynamic range. Regularization filters, reconstruction resolution and iterations were set to "medium". Additional corrections were performed during reconstruction: Monte Carlo correction quality was set to "high"; Attenuation: based on CT attenuation map and scatter corrections; Activity decay correction: during acquisition time lapse. 99m Tc-NTP 15-5 activity was quantified in a volume of interest delineated over the operated (pathologic) and non-operated (contralateral) knees. The activity ratios between pathologic over contralateral knees were calculated for each animal at each time point.

Histology.
Histology was performed at 12 and/or 24 weeks (W) after DMM surgery. After sacrifice, whole knee joints were dissected free of soft tissues. The murine knee joints were fixed in 4% paraformaldehyde (PFA; pH 7.4) for 48 h and decalcified with ethylenediaminetetraacetic acid (EDTA) 0.5 M for 15 days at room temperature (RT) with the solution changed twice a week. After dehydration in a graded series of alcohol, murine knees were embedded in paraffin at 60 °C. Serial Sects. (5 µm thick) were cut in the medial femorotibial compartment using a Polycut E microtome Leica RM2135 (Leica, Wetzlar, Germany) and then mounted on slides. Osteoarthritis scoring was based on Safranin O-Fast green (0.02% fast green for 30 min, 1% acetic acid for 10 s, and 1.5% Safranin O for 3 min) stained-sections according to OARSI (The Osteoarthritis Research Society International) recommendations with a total severity score ranging from 0 to 6 24 . For each animal, the OA score was the highest score obtained at one of the 3 levels. All sections of each model were blindly scored by the same 2 readers. All images were taken using a Olympus DP73 microscope.
Proteoglycan content of whole knee extract. Proteoglycan content of DMM mice knees was measured straight after 99m Tc-NTP 15-5 imaging, 5 weeks post-surgery. Hind legs were removed, and knees were isolated and incubated for 16 h at 65 °C with ethylenediaminetetraacetic (EDTA) buffer solution containing Papain 0.60 mg/ml. Digests were processed for proteoglycan content determination using the 1,9-dimethyl- www.nature.com/scientificreports/