FRα is highly expressed in non-small cell lung, ovarian, triple-negative breast, and endometrial cancers. In contrast, in normal tissues, FRα is only expressed at basal levels and is located at the apical portion of epithelial cells, making it less accessible [18]. FRα is a promising target for anti-cancer therapy; however, although several therapeutic strategies have been evaluated, only ADC mirvetuximab soravtansine received FDA approval. Although ADCs are promising, many challenges remain regarding suboptimal pharmacokinetics, hampered payload release, tumour targeting, and drug resistance [13]. That highlights the need for novel therapeutic strategies; therefore, we present the development of an anti-hFRα sdAb radioconjugate for RLT.
In this study, we generated a library of anti-hFRα sdAbs, of which 3A92 was selected as the lead compound in a first biopanning generation using a conventional surface-bound antigen approach. However, 3A92 was, despite good binding affinity of KD = 4.3 nM, deemed suboptimal for delivering a radiotherapeutic payload to the tumour, mainly because of the high dissociation constant of kd = 5.7 × 10− 3 s− 1 and, consequently, suboptimal tumour retention. Therefore, an off-rate selection step optimised the KD and kd of potential new binders in a second biopanning generation. This second generation resulted in 12 unique sdAbs, nearly all of which possessed lower kd values than the first generation. sdAb 2BD42 was deemed most promising with the picomolar binding affinity of KD = 39.8 pM and a kd of 1.4 × 10− 4 s− 1, increasing the binding affinity by 100-fold and the kd by 40-fold over the first-generation.
The lead sdAbs from first- and second-generation were successfully radiolabelled with 99mTc and evaluated in vivo and ex vivo for pharmacokinetics and tumour targeting in SKOV3 tumour-bearing mice. Remarkably, first-generation lead sdAbs showed tumour uptake values between 1–2%IA/g, while second-generation frontrunner sdAbs could nearly double the radiotracer tumour uptake (4–5%IA/g). Further evaluation showed high kidney uptake of 99mTc-sdAbs at early timepoints, but fast blood clearance, which is a typical feature of small compounds (< 50 kDa) with high hydrophilicity and labelled with radiometals [19]. However, while SKOV3 tumour-bearing athymic mice provided a quick and cost-effective first assessment of our radiotracer library, the biodistribution remained somewhat artificial because of the non-mouse cross-reactive characteristics of our sdAbs. To this end, we generated a hFRα knock-in transgenic mouse, in which we evaluated sdAb 2BD42, labelled with 99mTc, to enable micro-SPECT/CT imaging. For the brain (choroid plexus) and ovaries, a good correlation was observed between in vivo, IHC data and the existing literature [18]. On the other hand, kidney uptake presented conflicting results: while IHC showed FRα expression in tubular cells in homozygotes, heterozygotes, and wild-type animals using a cross-reactive antibody, radiotracer (non-cross-reactive) uptake remained similarly high for all groups (± 400%IA/g). Therefore, we hypothesised that kidney uptake is driven by the sdAb HIS6-tag and the radiolabel, which mask specific kidney uptake. High liver uptake was identified as an artefact of the 99mTc radiolabelling procedure owing to the innate low Tm (51.9 °C) of the sdAb (Supplementary Figure S6).
131I was selected as the therapeutic radionuclide of interest for the reasons described below. Iodine possesses a true theragnostic character, making it an attractive radionuclide for therapy (131I), SPECT imaging (131I, 123I), and PET imaging (124I). Imaging can contribute to a successful therapeutic effect with limited toxicity by performing patient selection, assessment of treatment delivery to tumours, calculation of dosimetry, and therapy response evaluation [20]. True, chemically identical, theragnostic radionuclides can be easily swapped in the construct, with limited changes to the radiolabelling procedures and no influence on biodistribution. Surrogate theragnostic radiometal pairs, such as the commonly used 177Lu/68Ga, may face additional challenges. For example, Umbricht et al. highlighted mismatches in pharmacokinetics between [177Lu]Lu-DOTA-PSMA-617 and [68Ga]Ga-DOTA-PSMA-617, related to the different coordination chemistry of the radiometals [21]. Furthermore, [131I]-SGMIB radiolabelling showed vast reduced kidney accumulation of sdAbs in mice, unmatched by any radiometal-sdAb [14, 22].
The selected sdAbs were further radiolabelled with 131I, and saturation binding assays confirmed previously acquired data using 99mTc-sdAb counterparts. Time-dependent biodistribution revealed that [131I]-SGMIB-2BD42 outperformed [131I]-SGMIB-2CD8 at all timepoints with respect to higher tumour uptake, but also for kidney retention. Indeed, [131I]-SGMIB-2BD42 had the best characteristics with values < 1%IA/g in kidneys at 3 h p.i. This biodistribution profile is in line with what was previously reported by our group and underlines the attractive characteristics of sdAbs labelled with [131I]-SGMIB, which give rise to rapidly clearing catabolites after renal filtration [14, 16]. This unique characteristic is not present in metal-based sdAbs. For example, [225Ac]Ac-DOTA-4AH29 (anti-FAP sdAb), still shows ~ 3%IA/g kidney retention at 96 h p.i. and [225Ac]Ac-DOTA-2Rs15d (anti-HER2 sdAb) has kidney retention of ~ 7%IA/g at 48 h p.i [22, 23].
Kidney clearance is important for anti-FRα RLT because FRα is also expressed in the proximal tubules of the kidneys; therefore, RLT may be prone to nephrotoxicity. Kidney retention and nephrotoxicity are the main reasons why the field has made limited progress in developing anti-FRα RLT. However, some attempts have been made to develop albumin-DOTA-folic acid for RLT and to reduce kidney retention substantially. In our study, we observed kidney retention for [131I]I-GMIB-2BD42 at 24 h p.i. of 0.34 ± 0.03%IA/g and 1.03 ± 0.06%IA/g in SKOV3 tumour-bearing athymic mice and LLC-OVA-fLuc-hFRα tumour-bearing hFRα knock-in C57BL/6 transgenic mice, respectively. For comparison, others report kidney retention at 24 h p.i. of 30.09 ± 4.04%IA/g for [177Lu]Lu-DOTA-cm09, and of 35 ± 10%IA/g for [177Lu]Lu-DOTA-cm10 [12, 24]. These unique characteristics of [131I]-GMIB-2BD42 help fill the gap in anti-FRα RLT. Furthermore, co-infusion with positively charged amino acids can reduce the kidney retention even more of [131I]-GMIB-2BD42 [25].
Toxicity assessment of [131I]-GMIB-2BD42 as a single-bolus injection (55.5 MBq) showed no signs of long-term toxicity with all animals completing the 180-day post-injection study endpoint and no morphological changes in organs as assessed by IHC. Furthermore, [99mTc]-DMSA SPECT scans showed no difference in dosing groups’ uptake values, indicating similar kidney function. Previous studies also showed 100% overall survival in animals receiving a repeated dosing regimen of anti-FAP sdAb [131I]-GMIB-4AH29 (6 × 37 MBq) for up to 180 days post-injection [22]. Also, our HER2 targeting sdAb [131I]-GMIB-2Rs15d, which progressed into clinical trials, was considered safe at a therapeutic dose level of 5.55 GBq [26]. Therefore, in our therapeutic design, we suggested a dosing regimen of 10 × 37 MBq (high dose) and 10 × 18.5 MBq (low dose), leading to an absorbed dose of 69.93 Gy and 34.97 Gy to the tumour. The kidneys act as a dose-limiting organ and receive absorbed doses of 15.76 Gy (high dose) and 7.88 Gy (low dose). These calculated values are only valid when administering the dose in one bolus injection; however, we split the drug administration into 10 injections in our design to reduce kidney toxicity further. This adsorbed dose is far below the generally accepted threshold of 23 Gy for kidney toxicity [27]. Remarkably, even our low dosing regimen showed therapeutic efficacy, allowing us to maximise the therapeutic window.
Repeated injections, as proposed in our therapeutic regimen, could lead to the generation of anti-drug antibodies. However, sdAbs have the innate characteristics to be low immunogenetic making them ideal for numerous applications. Clinical trials using single injections of various [68Ga]Ga-NOTA-sdAb showed minimal to no immunogenic responses [28,29,30]. Recently, a phase 2 clinical trial using [68Ga]Ga-NOTA anti-HER2 sdAb, included 2 repeated tracer injections within 8 days without reporting any immunogenicity [31]. Most clinical-trials involving sdAbs have reported minimal immunogenic responses comparable to fully humanized IgGs with neutralising anti-drug antibodies in < 3% of the cases [32].
