Development and preclinical evaluation of [68Ga]Ga-Alb-FAPtp-01, a novel tumour-associated broblast activation protein tracer for PET imaging of xenograft glioma

Dynamic changes in tumour-associated broblast activation protein (FAP) expression in tumours of different stages may be helpful for prognostic evaluation and treatment response monitoring, making this protein a promising surveillance biomarker for timely diagnosis of malignant tumours and effective planning of patient care. Thus, novel FAP-PET imaging tracers were developed and evaluated for the diagnosis of xenograft glioma tumours.


Introduction
Malignant tumours can usually be distinguished from nonmalignant cells, such as immune and endothelial cells, and extracellular matrix components by their ability to modulate the peritumoural environment [1][2][3][4]. Thus, the dynamic, constantly changing stromal tumour microenvironment (TME) provides cues for assessing tumour proliferation and invasion and the ability to metastasize to distant sites. In contrast, most localized tumours exhibit limited diffuse growth. Thus, the interactions between the tumour and the TME and even microvascular proliferation status, which indicates tumour growth and progression, are essential for tumour characterization.
Distinct from normal broblasts, cancer-associated broblasts (CAFs) in the TME express high levels of a unique protein called broblast activation protein (FAP), which has been recognized as a potential prognostic marker [5][6]. FAP, also known as FAPα or seprase, is a membrane-bound type II serine protease and a member of the S9 family of proline-speci c proteases, which includes dipeptidyl peptidase IV (DPPIV), DPP8, DPP9, and prolyl endopeptidase (PREP) [7]. FAP possesses both dipeptidy peptidase and endopeptidase activities. Importantly, the expression of FAP in adult mammals is usually low in most normal tissues but elevated in activated broblasts during tissue remodelling at sites of in ammation, such as sites of rheumatoid arthritis, osteoarthritis, brosis of the heart and liver, and in a variety of tumour-promoting environments [8][9][10][11].
Although the molecular mechanism underlying the contribution of FAP to tumour severity remains elusive, intense studies of potential FAP-targeted inhibitors, diagnostic agents and therapeutic interventions have been expanding in recent decades [12][13][14]. The rapidly growing eld of FAP-targeting agent development, including but not limited to antibodies, chimaeric antigen receptor T cells, immunoconjugates, peptide-drug complexes and small molecules for the inhibition of FAP enzyme activity, has received increasing attention. Among these agents, small molecules such as FAP inhibitors (FAPIs) have had the most promising results from the perspective of FAP-expressing tumour cell binding, internalization, e ux, tumour xenograft accumulation and theranostic application in patients [15][16][17][18][19].
Noninvasive positron emission tomography (PET) imaging with suitable radioligands may quantify the distribution of biomarkers of interest through the measurement of radioactivity pro les in living subjects.
Such radiopharmaceutic approaches may reveal the network of molecular mechanisms underlying diseases and assist in the development and evaluation of effective therapies. The development of optimal suitable PET tracers is especially challenging because they must satisfy a wide array of chemical, biochemical, and pharmacological requirments. To reduce unnecessary radioexposure, an important aspect of PET radiotracer development is to consider variations in speci c binding a nities and concentrations across species in relevant native tissues.
Therefore, in the current investigation, a prototype peptide-based FAP-speci c PET tracer, [ 68 Ga]Ga-FAPtp, based on the structure of a potent FAP inhibitor (known as UAMC-1110) [20], were designed, characterized and optimized for use as novel FAP-targeted diagnostic agents and potential theranostic agents. In contrast to the original quinoline-based moiety, peptide amide bond core structures with an additional short PEGylating linker were contemporaneously adopted. Furthermore, to enhance the plasma retention time, the optimized [ 68 Ga]Ga-Alb-FAPtp-01 tracer with an endogenous albumin-binding moiety, 4p-chlorophenyl butyric acid [21], was designed to improve pharmacokinetics and was tested in glioma xenograft models. Characteristics including FAP binding speci city and a nity, were validated by in vitro cell uptake assays and biolayer interferometry (BLI), and preclinical pharmacokinetics were determined in a U87MG xenograft model by small-animal PET/CT imaging, ex vivo histochemical staining and biodistribution studies. These rst proof-of-concept tracers are unique in that they can be effectively synthesized through the standard solid-phase peptide synthesis (SPPS) procedure without any sophisticated chemical methods. Thus, the peptide probes can be easily accessed via custom order from any certi cated peptide vendor in any amount requested. Importantly, the cyclic chelator DOTA can be utilized as the functional component for 68 Ga isotope radiolabelling for PET imaging and, possibly, for endoradiotherapy interventions involving labelling with therapeutic isotopes, such as 90 Y and 177 Lu.

Radiochemistry
The details of the 68 Ga labelling procedures have been reported previously, with a radiochemical purity over 95% [22][23]. Brie y, [ 68 Ga]Ga-acetate was prepared by adding 555-370 MBq/mL 68 GaCl 3+ to 250 μL of 0.1 M ammonium acetate buffer (pH 5.5), which was then added to the DOTA-conjugated peptide solution (approximately 25-35 μg in 20 μL of 0.1 N ammonium acetate buffer). The vial was subsequently sealed and heated for 5-10 minutes at 95 °C to form the radioactive 68 Ga-DOTA-peptide complex. The overall synthesis time was approximately 20-25 minutes. The radiochemical yield of the tracers, which was determined by radio-TLC, was > 95%, and the molar activity of the solution was 185.09 ± 13.5 GBq/μmol. The radiochemical purity (RCP) was >95%, as determined by HPLC. Solution stability experiments were performed to demonstrate that 68 Ga-DOTA-peptide was intact before the tracer was injected into tumour-bearing mice. DOTA-conjugated peptides remained stable for at least 2 hours (only 5-6% of the 68 Ga was dissociated from the DOTA complex, supplementary Fig. S3).

Biolayer interferometry (BLI)
An eight-channel Octet RED96 BLI system (ForteBio, Menlo Park, California) was used to conduct the titration binding experiments at room temperature with shaking at 1000 rpm. Due to the small size of the peptide tracers, Super Streptavidin tips ("SSA," ForteBio, Pall Life Sciences) were used for biotinylated FAP protein loading in these binding assays; the tips were hydrated for at least 30 minutes in PBS buffer, loaded with biotinylated FAP in PBS buffer for 180 s, and moved to baseline buffer to wash unbound biotinylated FAPs as a baseline step (60 s). Then, the biotinylated FAP was associated with sequentially diluted tracers (from 5 μM to 150 μM) in a 200 µL volume in black 96-well plates (association step, 720 s). Finally, the tips were submerged in baseline buffer for dissociation analysis (300 s); the association and dissociation pro les were obtained for data tting analysis with Octet Data Analysis software (9.0.0.6.). The control study with tips with no peptides was also conducted in parallel, and the results showed low nonspeci c binding. The average values were subtracted for later kinetic binding analysis. The mathematical model used assumes a simple 1:1 stoichiometry, with only one analyte in solution binding to one binding site on the surface. A Savitzky-Golay lter was applied to smooth the data.
Cell line U87MG cells, a human glioblastoma cell line, were purchased from Bioresource Collection and Research Center (BCRC, Hsinchu, Taiwan) and maintained in minimum essential Eagle's medium (MEM) with 2 mM glutamine, 0.1 mM nonessential amino acids, 1 mM sodium pyruvate, 100 units/ml penicillinstreptomycin and 10% FBS. The cells were cultured as an adherent monolayer in an incubator with a 5% CO 2 /95% air atmosphere at 37 °C. All media and reagents were purchased from Thermo Fisher Scienti c (Thermo Fisher Scienti c, Waltham, MA, USA).
Cell-based uptake and competition studies Brie y, U87MG glioma tumour cells were seeded in 24-well plates. When the cells reached 80 to 90% con uence, 18.5 kBq of radiolabelled tracers were added to each well in 0.5 ml of Hank's buffered salt solution (HBSS) for different times (15 min, 30 min, 60 min, 90 min and 120 min). Furthermore, competition assays were performed by incubating 2×10 5 U87MG cells in 24-well plates with 68 Ga-labelled tracers (∼18.5 kBq) and various concentrations of cold FAPI-04 tracers in the medium at 37 °C for 1 h. In all experiments, the cells were washed twice with 1 mL of ice-cold PBS and dissolved in lysis buffer containing 0.2 M NaOH and 1% SDS. The remaining cell-bound radioactivity after washing was determined by gamma counting (Cobra II, Packard Instruments, Downers Grove, IL) as counts per minute (CPM) and calculated as the percentage of the applied dose. All experiments were performed in triplicate and repeated three times. The binding a nity value was calculated by tting the data with nonlinear regression using Prism6 (GraphPad software).

Subcutaneous tumour xenografts
All experiments were conducted according to a protocol approved by the Institutional Animal Care and Use Committee (IACUC) of the Laboratory Animal Center, Chang Gung Memorial Hospital, Taiwan. Female BALB/c nude mice, 6 to 8 weeks old, were obtained from the National Laboratory Animal Center (Taipei, Taiwan) and housed under a regular 12 h light/dark schedule. Food and water were available ad libitum. hours and 3 hours post-injection. Animals were anaesthetized using 2.0% isoflurane (Benson Medical Industries, Markham, ON, Canada) prior to and during imaging. All images were generated with scatter correction, random correction and attenuation correction (AC). CT images were acquired on a smallanimal Nano CT (Mediso Medical Imaging Systems) scanner with an 8.5-cm axial-by-5.0-cm transaxial eld of view. Coregistered PET/CT images were recorded and mapped to a matrix with PMOD image analysis software, version 4.0 (PMOD Technologies Ltd., Zurich, Switzerland). We determined the regions of interest (ROIs) by determining 50% of the maximum minus minimum activity level in the coronal plane of fused CT images, and we determined the ROIs of muscles by analysing the shoulder muscle using PMOD image analysis software, version 4.0 (PMOD Technologies Ltd., Zurich, Switzerland). The average radioactivity concentration within the tumour or organ of interest was calculated semiquantitatively as a percentage of the administered activity per gram of tissue (% ID/g) or as the standard uptake value (SUV) (SUV mean ). Maximum intensity projection (MIP) images were created by collapsing across one dimension with the maximum value in each column of voxels. The imaging ROI-derived tumour-to-normal organ radioactivity ratios, which were calculated for organs such as the muscles, heart, lung, liver, kidney and brain, were also calculated for comparisons.

Ex vivo histological analysis and autoradiography
To investigate the expression level of FAP histologically, immunohistochemistry and autoradiography were conducted to con rm the colocalization of activated FAP and the accumulation of the imaging agents. The tumour tissues were embedded in OCT compound (Tissue-Tek), snap-frozen in a dry iceisopentane slurry and stored at -80 °C. The tissue samples were cut into 10 μm thick sections, which were then stained with primary antibodies against FAP (rabbit anti-human; AB28244). Staining for FAP was performed by immunohistochemistry on histologic sections that were xed in 4% paraformaldehyde and incubated with anti-FAP monoclonal antibodies (dilution, 1:200 in Tris buffer) at 4 °C overnight. After washing with PBS, the slides were incubated with a biotinylated goat anti-rabbit IgG secondary antibody for 1 h at room temperature. Streptavidin-conjugated peroxidase (ABC Elite, Vector Laboratories) was applied, and the reaction was developed by staining with 3,3'-diaminobenzidine (Sigma). The autoradiogram images were collected by a Cyclone Pulse Storage Phosphor system (Perkin-Elmer, USA) with an exposure time of 24 h. The white-light pictures of these tumour sections were taken by a Leica MZ75 (Leica Inc., Germany) high-performance stereomicroscope equipped with a 2.5 x Plano objective.

Statistical analysis
Quantitative data are expressed as the mean with standard deviation (± SD). All data were analysed with GraphPad Prism 6 statistical software. Student's t test was used to analyse the differences between the control group and the experimental group. A p-value less than 0.05 was considered statistically significant.

Results
Chemistry and radiochemistry 68 Ga 3+ was used under previously reported radiolabelling conditions to radiolabel FAPI-04 and generate FAP-targeting peptides with the chelator DOTA, with high radiochemical purity (> 95%), as validated by analytic radio-HPLC. The molar radioactivity of all tracers was estimated to be ~185.09 ± 13.5 GBq/μmol.
The octanol/water partition coe cient (logP) was calculated as -2.92 (FAPI-04), -2.64 (FAPtp) and -2.88 (Alb-FAPtp-01), and the negative log p values indicate the high hydrophilicity of all tracers. Higher hydrophilicity is desirable for PET tracer design to reduce radioexposure, since the radioligands will quickly wash out through the urine. The proposed FAP tracers all demonstrated low liver uptake, which might be due to the PEGylating linker and highly hydrophilic peptide residues. The purity of all radiolabelled tracers was su cient; thus, the tracers were passed through a 0.22 µm syringe lter for the following in vitro cell binding a nity assay or in vivo animal imaging studies without further puri cation.
Binding speci city/a nity To elucidate and quantify the kinetic binding ability of FAP with the proposed peptide tracers, a fast, realtime, high-throughput and label-free analysis method based on BLI was performed (Fig. 2). In these experiments, the association pro le of FAP-04 was determined as an inhibitor, and the binding signals gradually increased over time and had a minimal dissociation drop pattern. The K d value was determined as 17.4±3 x 10 -6 M with BLI. However, the Alb-FAPtp-01 peptide tracers demonstrated the typical quick upand-down pro le pattern of small molecules with fast association and dissociation rates, and the K d value was calculated as 18.5±2.4 x10 -6 M. Due to the potential cleavage site (-Gly-Pro-) of FAP in the FAPtp peptides, the association rate of FAP is quite different than those the other two tracers. The association rate quickly increased, similar to Alb-FAPtp-01, at the early time points, but instead of reaching a plateau, the signals gradually increased over time but were not as high as those of FAP-04. This complex association pattern might be due to the peptide being cleaved by the FAP enzyme and cannot be simply determined by a 1:1 or 1:2 stoichiometry. These results were also con rmed with proteomic analysis experiments, in which the FAPtp peptide was found to be cleaved into two fragments when incubated with FAP enzymes, but the other two tracers were found to be intact under the same processing conditions (Supplementary Fig. S4). When DPP4 was used in this binding assay, only FAPI-04 demonstrated low binding and a quick dissociation pro le, with a Kd value of 2.55 x 10 -4 M (R 2 =0.96) (Fig. 2b). FAPtp and Alb-FAPtp-01 both demonstrated nonspeci c binding throughout the experimental period. All these data indicated the speci city of the developed peptide tracers.

In vitro cell uptake and competition assays
The upregulated FAP expression in U87MG cells was con rmed by western blotting (Fig. 3 (a)) Then, the cellular uptake of [ 68 Ga]Ga-FAPI-04, [ 68 Ga]Ga-FAPtp and [ 68 Ga]Ga-Alb-FAPtp-01 into U87MG cells was examined. In the cellular uptake experiment, U87MG cells were seeded into 24-well plates at a density of 0.5 million cells per well and incubated overnight. The cellular uptake was normalized to total added radioactivity. All tracers demonstrated comparable and speci c cell uptake within 30 min of incubation and plateaued afterward, and [ 68 Ga]Ga-Alb-FAPtp-01 showed slightly higher cellular uptake (> 2%) at 120 min but did not reach statistical signi cance (Fig. 3b). Note that the cellular uptake protocol used in this study does not distinguish between cell surface-bound and internalized radioactivity. In the competition  (Fig. 3c). The inhibition value of [ 68 Ga]Ga-FAPI-04 was similar to that reported in previous papers [28]; however, a 10-fold higher concentration of FAPI-04 was needed to block the cellular uptake of [ 68 Ga]Ga-Alb-FAPtp-01, indicating that Alb-FAPtp-01 may have a stronger ability to bind to U87MG cells than FAPI-04.

Dynamic PET imaging analysis of [ 68 Ga]Ga-FAPI-04 and [ 68 Ga]Ga-DOTA-FAPtp
Armed with promising data from in vitro assays, we were able to proceed with radiolabelling and evaluation of the FAP tracers in animal models with PET. [ 68 Ga]Ga-FAPI-04 is believed to be a highly selective and potent FAP ligand (IC 50 = 6.5 nM) and has emerged as a lead PET imaging agent [15][16] that might compensate for the functional role that has been dominated by 18 F-FDG in recent decades. . After intravenous injection of both tracers, very good overall tumour radioactivity uptake and rapid whole-body washout patterns were observed (Fig. 4a). The time-activity curve of U87MG tumours displayed rapid peak uptake of both tracers, with approximately 1 standardized uptake value (SUV) by 10 min, but [ 68 Ga]Ga-DOTA-FAPtp showed a delayed signal decline during the 20-40 min period, with a tumour uptake value of 0.5 SUV compared with 0.42 SUV for [ 68 Ga]Ga-FAPI-04 (Fig. 4b). In addition to the higher tumour uptake observed for [ 68 Ga]Ga-FAPtp, there was a much faster washout rate in the two major excretion organs, kidneys and liver, especially in the liver, and the clearance rate of [ 68 Ga]Ga-FAPtp was 3-fold faster than that of [ 68 Ga]Ga-FAPI-04, which may be due to the hydrophilic PEGylation linker of [ 68 Ga]Ga-FAPtp that accelerated the tracer excretion time (Fig. 5). In addition, the higher heart accumulation signals indicated the longer plasma half-life of [ 68 Ga]Ga-FAPtp. For brain and muscle tissues, the potential adverse side effects were minimal, as evidenced by indistinguishably low tracer uptake (SUV= 0.03 and 0.042, respectively) at late time points. These preclinical imaging results proved that the proposed FAP-targeting peptide tracer may be an effective positron emission tomography radioligand for quantifying FAP expression levels in U87MG tumours. The continuous washout of radioactivity occurred mostly through the renal pathway. Collectively, [ 68 Ga]Ga-FAPtp peptide tracers bound reversibly to FAP and demonstrated comparable or even higher tumour targeting capability and more favourable biodistribution pro les than [ 68 Ga]Ga-FAPI-04 and thus warrant further improvements in follow-up tracer development studies.

In vitro albumin binding a nity and in vivo pharmacokinetics of [ 68 Ga]Ga-Alb-FAPtp-01
The major focus of this work was to modify and optimize current FAPI derivative-based tracers, such as FAPI-04, FAPI-21, and FAPI-46, to prolong tumour retention and improve the tumour-to-normal ratio based on the structural framework of the peptides, eventually making them potential theranostic agents. Therefore, we designed and synthesized prototypic [ 68 Ga]Ga-FAPtp-targeted PET imaging agents with critical features of the peptide substrate and experimentally substantiated these agents. However, the prototype [ 68 Ga]Ga-FAPtp tracer exhibited a rapid washout rate; thus, an intriguing in situ albumin binding strategy was incorporated into the tracer design with 4-p-chlorophenyl butyric acid. The concept of optimizing the tracer pharmacokinetic pro le and prolonging tracer stability in the plasma has been proven to be effective for several published PSMA-relevant tracers. With adoption of the albumin-binding agent 4-p-chlorophenyl butyric acid, [ 68 Ga]Ga-Alb-FAPtp-01 was rst validated via in vitro FAP binding (Fig. 2), tumour cellular uptake (Fig. 3-b), albumin binding assays and in vivo half-life experiments (Fig.   6). The HSA binding a nity can reach a Kd value of 1.9 μM, and the half-life of the [ 68 Ga]Ga-Alb-FAPtp-01 tracer in living animals was calculated to be approximately 120 mins, which re ected a signi cantly prolonged circulation time for such small peptide constructs that usually have circulation times of less than 10 min. Overall, these results demonstrated the high selectivity and albumin binding ability of the tracer, which may substantially enhance the opportunity for higher tumour uptake due to the long retention time in the plasma.

In vivo PET imaging evaluation of [ 68 Ga]Ga-Alb-FAPtp-01
Therefore, intravenous administration of the optimized albumin binding [ 68 Ga]Ga-Alb-FAPtp-01 tracer for PET/CT imaging was conducted to validate the FAP-targeting speci city. The imaging results showed profound rapid tumour uptake followed by moderately slow washout compared with the non-albuminbinding [ 68 Ga]Ga-FAPI-04 and [ 68 Ga]Ga-FAPtp peptide tracers (Fig. 7). The peak tumour uptake reached an SUV of 1.775 and subsequently slightly declined to 1.425 SUV at 180 min after injection. The percentage of radioactivity in the heart represented by the circulating tracers decreased to approximately 63.154±1.779% within the initial 30 min scan interval and thereafter slowly decreased tõ 12.334±3.052% at the end of the scan (180 min), indicating that 4-p-chlorophenyl butyric acid played a role in optimizing the pharmacokinetics of the tracers. The tumour margins were visualized at early time points but may be confused with nonspeci c perfusion imaging; thus, the persistent tracer signals detected at 2 or 3 hours post-injection in [ 68 Ga]Ga-Alb-FAPtp-01 PET imaging re ect the speci city of tumour lesion detection. Furthermore, due to the low normal uptake background, the tumour contrast ratio of tumour/muscle increased signi cantly from 5.904±0.824 (1 hour) to 6.704±0.903 (2 hour) and 9.459±1.330 (3 hour), which indicates precise whole-tumour localization for preoperative planning and delineates the tumour margins from coregistered PET/CT imaging data with high sensitivity and high spatial resolution.
In addition, both the [ 68 Ga]Ga-FAPI-04 and [ 68 Ga]Ga-FAPtp tracers were quickly washed out through the kidney within 60 min. The higher blood circulation half-life of [ 68 Ga]Ga-Alb-FAPtp-01 was attributable to the albumin-binding ability, which maintained a greater availability of radioligands in the plasma.
Although there was somewhat higher liver uptake, the tumour uptake of the [ 68 Ga]Ga-Alb-FAPtp-01 peptide tracer was signi cantly improved due to the prolonged and increased availability of stable radioligands in the plasma. These results validated the FAP-targeting speci city in animals injected with [ 68 Ga]Ga-Alb-FAPtp-01 tracers.
Longitudinal monitoring of tumour growth with FAPI-targeted PET imaging An elevated FAP expression level in tumour neoplasia was correlated with high tumourigenic and metastatic potential and a high rate of mortality in clinical patients, justifying its use as a potential prognostic marker. To further con rm the application of imaging with tracers for staging, longitudinal PET scans were conducted. The stable [ 68 Ga]Ga-Alb-FAPtp-01 PET tracer demonstrated prominent longitudinal tumour uptake in subcutaneous U87MG xenografts with sequential imaging acquisition time points at different tumour sizes/stages (tumour size from 200-870 mm 3 ). The tracer uptake in malignancies was 0.540+/-0.096, 1.198+/-0.155 and 1.829+/-0.100 (SUV mean ) at the small, medium and large stages (late stage), respectively. The increased tumour uptake observed in PET/CT imaging was consistent with the immunostaining pro le of FAP, and tumour growth was accompanied by upregulation of FAP expression levels. Overall, the developed albumin-binding FAP-targeting [ 68 Ga]Ga-Alb-FAPtp-01 tracer was able to differentiate the malignant neoplasia of U87MG tumours, and this feature may serve as a prognostic marker to identify the speci c tmour phenotype for anti-FAP therapies. These pilot studies demonstrated that [ 68 Ga]Ga-Alb-FAPtp-01 tracers may be superior to FAPI-04 PET for phenotypic tumour imaging for patient distraction diagnosis. Furthermore, it is worth noting that albumin-binding FAPtp-01 tracers are more suitable for treatment response monitoring, with potential applications for anti-FAP interventions or combination therapy with conventional treatments, such as radiotherapy and chemotherapy drugs. These follow-up studies are currently being performed.

Histopathologic analysis
The in vivo three-dimensional (3D) PET imaging is presented as a movie (see supplementary S3), and the PET imaging was strongly correlated with ex vivo autoradiography images, suggesting that FAP can be accurately validated by the proposed [ 68 Ga]Ga-Alb-FAPtp-01 PET tracer (Fig. 9). We next explored the relationship between radioactivity signal images and the ex vivo histopathologic analysis of FAP and vascular maturity (CD31-positive blood vessels) on a per-tumour basis in tumour xenograft models (Fig.  9b). U87MG xenograft tumours were collected and stained with antibodies against FAP and CD31 to evaluate FAP and vascular density by histochemistry. The PET imaging and ex vivo autoradiography images were compared spatially with the immunostaining data. The results demonstrated that the vascular density and resulting tumour perfusion were not directly correlated with [ 68 Ga]Ga-Alb-FAPtp-01 tracer uptake, as shown in Fig. 9. After investigating the histopathologic data in greater depth, it was obvious that the tumour regions with higher FAP expression levels but relatively low vascular density (weaker CD31 immunostaining) have signi cantly high tracer uptake in both in vivo 3D volume-rendered PET imaging and ex vivo autoradiography. The spatial colocalization and cooccurrence of the histopathologic staining with tracer accumulation con rmed the accurate registration of FAP expression status. However, vascular function and perfusion status were not signi cantly related to tracer accumulation but to FAP-binding speci city. Overall, in addition to FDG, the albumin tracer [ 68 Ga]Ga-Alb-FAPtp-01 might provide complementary information for the noninvasive assessment of tumour status and thus support better patient strati cation and subsequent monitoring of therapeutic response. Discussion FAP, a recognized hallmark of tumours, is a cell-surface serine protease that acts on various hormones and extracellular matrix components. FAP expression is highly upregulated in a wide variety of cancers, especially epithelial tumours, of which 90% overexpress FAP [24]. To date, massive efforts has been put into targeted diagnostic imaging and peptide receptor radionuclide therapy (PRRT) based on the identi ed upregulated FAP [13][14][15][16][17][18][25][26]. However, despite this growing eld of research and the nding that tumour-derived activation of FAP signalling augments aggressive behaviours in certain types of tumours, the molecular mechanisms underlying the functional roles of FAP in various tumours remain elusive. Cancer biology is far more complex than originally expected, and a deeper understanding of the causes of FAP expression and relevant complications during different stages of tumour progression must be obtained to meet the unmet needs of early tumour diagnosis, therapeutic prognosis prediction and metastasis prevention.
However, the development of FAP-targeting small molecules requires extensive organic synthesis knowledge and skills. The whole panel of small modi cations of the original parent structures has to be designed and ne-tuned, and then extensive follow-up experimental tests have to be carried out both in vitro and in vivo [27][28]. For most investigators, accordingly, the more effective and easy-to-use peptidebased FAP-targeted tracer proposed here would offer an amiable choice or a key noninvasive FAP imaging strategy for tumour phenotyping, staging and/or intervention monitoring studies.
In vitro cell uptake and competition assays and in vivo translational imaging evaluation studies of FAPtargeting ligands were the major focus of the current tumour diagnosis studies. In this study, the structure and radioligand properties of FAP-targeted peptide tracers were fully elucidated. The rationale of the proposed FAP-targeted PET agents is based on the natural FAP dipeptide substrate ligand sequence (-Gly-Pro-), which can be synthesized by a facile SPPS procedure, as a surrogate biomarker for early detection of the functional status of FAP, with the aim of determining whether such a biomarker can be used for early detection and to phenotypic differentiation of tumours.
Unlike most published papers using transgenic FAP-expressing HT-1080 cells, however, the permanent glioma tumour cell line U87MG was used here for in vitro and in vivo tracer evaluation. Because U87MG cells evolved from a rather heterogeneous cell population, the expression pro le of FAP, tumour perfusion and interactions with surrounding tissues may be closer to those of real human tumours than those of the homogeneous HT-1080 cell population, which possesses consistent FAP protein expression. In an in vivo xenograft model, U87MG cells recruited murine broblast cells, which were FAP positive and resulted in speci c FAP-targeted PET imaging. FAP expression is also observed in tumour vessels, and this may facilitate and enhance FAP tracer binding to the tumour vasculature, contributing to higher tracer uptake in tumours. Overall, the heterogeneous nature of U87MG cells may assist the interpretation of the tumourbinding and pharmacokinetic pro les of different tumour entities in general. This valuable information may be extrapolated to support clinical translation of FAP-based imaging evaluation or even endoradiotherapy applications.
To con rm the advantage of the newly developed tracers over widely used FAP inhibitor tracers, the targeted to FAP but also steadily diffused and transported into the tumour by the intrinsic blood vessels or internal uid pressure due to its albumin-binding capability. Thus, the [ 68 Ga]Ga-Alb-FAPtp-01 targeted tracer not only provides a facile strategy for tracer synthesis but also improves the pharmacokinetic pro le of the [ 68 Ga]Ga-FAPtp tracer with a 4-p-chlorophenyl butyric acid component for endogenous albumin-binding capability in vivo. This feature enables prolonged stability in the systemic circulation and precise marker targeting in the blood, highlighting its potential for clinical application in theranostic agent development. However, it is a concern that [ 68 Ga]Ga-Alb-FAPtp-01 uptake does not necessarily re ect only the FAP expression level but also the neovasculature of U87MG tumours due to the intrinsic blood-targeting agent in [ 68 Ga]Ga-Alb-FAPtp-01. In fact, the kinetic pro le of tracer distribution clearly demonstrates two regions: "rapid washout" and retaining regions. The higher tracer uptake in the latter region was con rmed by the immunostaining results, showing that FAP expression is upregulated but not associated with the immature and leaky vasculature; thus, the increase in tracer accumulation over time may be due to speci c FAP binding (Fig. 9). Consequently, with high overall scores for pharmacokinetic performance, suitability for tumour diagnosis, and accessibility to the wider clinical and research communities, the [ 68 Ga]Ga-Alb-FAPtp-01 tracer has emerged as a tracer of choice for further studies of potential clinical application. The tracer may play a role in ascertaining the links between FAP expression, disease state, and disease progression in the near future. The use of a DOTA chelator as a dual functional component that may be labelled with therapeutic radioisotopes, such as Y-90 and Lu-177, shows promise for theranostic applications, and related therapeutic effect experiments are currently underway.

Conclusion
Dynamic changes in tumour FAP expression are a characteristic marker of cancer; however, precise the protein uctuation changes that occur and their correlations with tumour progression are not known.
Thus, we try to provide an effective and noninvasive imaging strategy to advance the understanding of the functions of FAP in cancer, particularly its prognostic value for various tumour types, and as a complement to diagnostic uses, its potential theranostic application is under investigation in preclinical investigations. We highlight this unique and accessible solution to the existing limitations in the eld, which could shape preclinical and clinical studies of FAP tracer development.