Preclinical Characterization of DPI-4452: A 68Ga/177Lu Theranostic Ligand for Carbonic Anhydrase IX

Visual Abstract

Car bonic anhydrases (CAs) catalyze the reversible hydration of carbon dioxide to bicarbonate ions and protons.Of the 12 catalytically active CAs in humans, 4 are located on the extracellular surface.Two of these, CAIX and CAXII, are highly expressed in tumors (1,2) and contribute to maintenance of the acidic tumor microenvironment (3).Many healthy tissues express CAXII, whereas normal CAIX expression is restricted to the gastrointestinal epithelia (1,2,(4)(5)(6).

MATERIALS AND METHODS
Detailed information on materials and methods are available in the Supplemental Materials and Methods.

Immunohistochemistry
Staining of human healthy and tumor tissue microarrays (US Biomax Inc.) using an anti-CAIX antibody (clone M75; Creative Biolabs) was performed.Scoring by pathologists followed the methodology of Ilie et al. (14).

Surface Plasmon Resonance
Recombinant CA proteins (Sino Biologic) were immobilized on Biacore CM5 sensor chips or biotin CAPture chips (Biacore).DPI-4452 binding was measured using the Biacore Liquid System (Biacore), and binding kinetic parameters were calculated.

Animal Models
All animal experiments were performed by Minerva Imaging under a license approved by the National Animal Inspectorate under the Ministry of Environment and Food of Denmark and in accordance with ARRIVE guidelines (15).
For xenograft models, female Swiss nude mice were subcutaneously injected with between 2 3 10 6 and 5 3 10 6 human HT-29 CRC cells (HTB-28; American Type Culture Collection) or SK-RC-52 ccRCC cells (Memorial Sloan Kettering Cancer Center) (16).Mice were weighed and tumors were measured 2 or 3 times per week.Healthy beagle dogs and xenograft-bearing mice were injected intravenously with radiolabeled DPI-4452.Blood radioactivity counts or DPI-4452 plasma concentration data determined by liquid chromatography-tandem mass spectrometry were used for pharmacokinetic analyses.Biodistribution was assessed using PET/CT and SPECT/CT.

Statistical Analyses
Statistical analyses were performed using GraphPad Prism software.

CAIX Is Overexpressed in Multiple Human Tumor Types
The suitability of CAIX as a tumor biomarker and therapeutic target was investigated using immunostaining of human tissue microarrays.Healthy stomach and pericardial mesothelial tissues displayed the highest CAIX expression, with more modest staining in the skin, tumor-adjacent normal ovary tissue, and small intestine.High CAIX expression was observed in samples of ccRCC, CRC, squamous non-small cell lung cancer, and PDAC, with no or low expression in the respective normal tissues.Samples of squamous cell carcinomas of the head and neck and triple-negative breast cancer also had high CAIX expression (Table 1; Supplemental Fig. 2).

DPI-4452 Is Rapidly Eliminated and Well Tolerated in Mice
Single intravenous DPI-4452 injections of 0.7 and 5.5 mg/kg in male CD-1 mice were well tolerated.Pharmacokinetic analysis   revealed maximum plasma DPI-4452 concentrations (5 min after dose) of 1,380 and 9,270 ng/mL, respectively, and short elimination half-lives (0.28 and 0.47 h, respectively) (Table 4).

DPI-4452 Shows Selective and High Tumor Uptake in Xenograft Mouse Models
Radioactivity uptake was monitored up to 48 h after intravenous injection of 30 MBq of [ 111 In]In-DPI-4452 in mice with subcutaneous HT-29 CRC and SK-RC-52 ccRCC xenografts using SPECT/CT imaging (Fig. 1; Supplemental Fig. 7).Maximal tumor uptake was seen at the first time point (1 and 2 h for SK-RC-52 and HT-29 xenografts, respectively).Except for the earliest time point for HT-29 xenografts, tumor-to-kidney and tumor-to-liver ratios of greater than 1 were observed for both models throughout (Supplemental Tables 2 and 3).
In both xenograft models, strong and uniform epithelial CAIX expression was observed by immunohistochemistry (Supplemental Fig. 8).
Theranostic pairs must localize to the same tissues at comparable levels to maximize their diagnostic utility and ability to predict potential treatment benefit.
To address this issue, mice with HT-29 and SK-RC-52 xenografts were injected with [ 68 Ga]Ga-DPI-4452 and imaged using PET/CT 1 h later.One wk later, the same mice received [ 177 Lu]Lu-DPI-4452 and were imaged using SPECT/CT 4 h later.Comparably high tumor uptake and low background organ uptake were seen (Fig. 2).4).These findings were consistent with the improved survival seen in treated mice compared with control mice (Supplemental Fig. 14; Supplemental Table 5).SPECT/CT analysis of 3 mice with xenografts per treatment arm revealed high tumor uptake in both models that was not modulated by repeated [ 177 Lu]Lu-DPI-4452 treatments (Fig. 3).Similar radioactivity was found in SK-RC-52 tumors for all doses, suggesting that uptake was already maximal at the lowest dose of 33 MBq (Supplemental Fig. 15).These findings demonstrated a strong antitumor effect of [ 177 Lu]Lu-DPI-4452 on CAIX-expressing tumors and suggested dose fractionation may be beneficial.

DPI-4452 Binds cCAIX with Affinity Similar to hCAIX
Of cCAIX, rat CAIX, and mCAIX homologs, hCAIX shares the greatest amino acid homology with cCAIX.In cell-binding assays, hCAIX-and cCAIX-expressing Chinese hamster ovary cells displayed similar affinities for [ 111 In]In-DPI-4452 (K D , 0.3 nM); these affinities were significantly higher than those for mCAIX-expressing Chinese hamster ovary cells (K D , 63 nM) (Table 5).When surface plasmon resonance was used, cCAIX bound [ nat Lu]Lu-DPI-4452 with a subnanomolar K D and a 69-min dissociation half-life (Supplemental Table 6), revealing an affinity comparable to that of hCAIX.Overall, the cross-reactivity of DPI-4452 and its complexes with cCAIX and hCAIX showed that the dog is a suitable species for the preclinical assessment of DPI-4452.

DPI-4452 Is Rapidly Eliminated in Dogs
In a pharmacokinetic study after a single intravenous dose of 0.1 mg/kg in beagle dogs, the plasma half-life of DPI-4452 was 0.38 h, denoting rapid elimination of the peptide from the systemic circulation (Table 4).Plasma exposure (area under the concentration-time curve) increased more than dose-proportionally from 0.025 to 0.8 mg/kg in a dose-finding experiment (Supplemental Table 7), as well as in a good laboratory practice-compliant singledose study (Table 6).
Biodistribution of DPI-4452 in Beagle Dogs DPI-4452 biodistribution was evaluated in male and female beagle dogs.Each animal received a single dose of 200 MBq of [ 68 Ga]Ga-DPI-4452 followed, 14 d later, by a single dose of 1,000 MBq of [ 177 Lu]Lu-DPI-4452.Organ uptake of radioactivity over time was determined using PET/CT imaging and SPECT/CT imaging, respectively.At early time points, significant uptake occurred in the bladder, likely due to renal elimination.The highest sustained uptake was seen in the stomach and small intestine (Fig. 4), consistent with CAIX expression in these organs (1,2,5,18).No or low uptake was seen in the rest of the body, namely, the kidneys, bone marrow, blood, heart wall, and skin (Fig. 5; Supplemental Fig. 16).
Beagle dogs were injected with a single [ 111 In]In-DPI-4452 dose of 25 MBq/kg with a specific activity of 1.11, 3.33, or 9.99 MBq/mg, correlating to a ligand mass dose of 22.5, 7.5, or 2.5 mg/kg, respectively.No differences in the plasma profile of total radioactivity (Fig. 6) or organ uptake were seen (Supplemental Fig. 17), suggesting that the peptide mass dose did not affect uptake in healthy tissues in the tested dose range.

Dosimetry of [ 68 Ga]Ga-DPI-4452 and [ 177 Lu]Lu-DPI-4452 in Dogs and Extrapolation to Humans
Given the assumption that CAIX organ expression is similar in dogs and humans (18), the biodistribution data for [ 68 Ga]Ga-DPI-4452 were extrapolated, and the radiation doses absorbed in human organs per administered [ 68 Ga]Ga dose were estimated by dosimetry (Table 7) (19).This approach yielded a predicted effective dose of 3.60E202 mSv/MBq for [ 68 Ga]Ga-DPI-4452comparable to those of other 68 Ga-labeled PET tracers (20).
Radiation doses absorbed in human organs after [ 177 Lu]Lu-DPI-4452 administration were similarly extrapolated (Table 7).High values were observed, irrespective of sex, in the stomach wall and small intestine (0.824 and 0.695 Gy/GBq, respectively, in females); however, relatively low values were obtained in the other organs, such as the kidneys and red marrow (0.438 and 0.134 Gy/GBq, respectively, in females).To our knowledge, radiation dose limits for human organs have not been established for targeted radiotherapy.According to the estimated radiation doses absorbed in human organs, none of the limits set for external-beam radiation therapy (21) should be exceeded after a dose of up to 25.9 GBq of [ 177 Lu]Lu-DPI-4452.

DISCUSSION
As a direct transcriptional target of hypoxia-inducible factor 1a, CAIX is overexpressed in several hard-to-treat tumors, including ccRCC, CRC, breast cancer, and PDAC (2,(9)(10)(11), making it a promising therapeutic target.However, to date, the transition of CAIX-targeting therapies to clinical settings has been poor (7,13).Several small-molecule inhibitors of CAs, including CAIX, have been developed (7,13).The most promising of these, the sulfonamide derivative SLC-0111, was well tolerated in a phase 1 study  of advanced solid tumors, but no objective responses were observed (22).Modest responses were also reported in clinical trials with the anti-CAIX monoclonal antibody G250 ( 13), but no improvement in survival compared with placebo was observed in a phase 3 adjuvant trial in patients with high-risk ccRCC (23).Conjugation of G250 to the b-emitter [ 177 Lu]Lu resulted in improved preclinical activity (13,24,25); in 2 clinical trials of metastatic ccRCC, disease stabilization was reported for 9 of 14 and 17 of 23 patients (26,27), despite dose-limiting myelotoxicity, a common side effect of radiolabeled antibodies (28).These data suggest that although antibodies and inhibitors targeting CAIX have not, to date, shown clinically relevant antitumor efficacy, they can be used to deliver radioactivity to tumors.With CAIX expression in healthy tissues being largely restricted to the small intestine and stomach (1,2,5,18), peptide-based therapeutic agents targeting CAIX, such as [ 177 Lu]Lu-DPI-4452, provide the potential to target hypoxic or von Hippel-Lindau tumor suppressormutated tumors while featuring a favorable biodistribution (7,13).Moreover, peptide receptor radionuclide therapy agents are usually nonimmunogenic and display favorable pharmacokinetics, characterized by high target tissue concentration and rapid excretion from the body, which is expected to limit adverse effects from prolonged exposure to circulating radioactivity (29).Previously, CAIX-targeting peptides showed poor tumor uptake in preclinical studies (7).In the present study, DPI-4452 showed excellent tumor accumulation and lower uptake in healthy tissues.Also, [ 177 Lu]Lu-DPI-4452 showed antitumor efficacy in CRC and ccRCC xenografts.Moreover, the similar responses seen for 3 33-MBq doses versus a single 100-MBq dose suggest that dose fractionation could be a viable option (Fig. 3) for the reduction of potential dose-limiting organ toxicities (30).
Because of the lack of cross-reactivity of DPI-4452 with mCAIX, mouse models provide limited relevance for investigating radiotoxicity and ligand toxicity in organs naturally expressing CAIX, whereas the similar binding of DPI-4452 to hCAIX and cCAIX suggests that the dog is a suitable species for the preclinical safety assessment of this agent.
A single administration of DPI-4452 at up to 0.4 mg/kg was well tolerated in a good laboratory practice toxicity study in beagle    dogs (31).Accumulation was predominantly seen in CAIXexpressing tissues, such as the stomach and small intestine (18); in addition, transient distribution was seen in the urinary bladder, due to renal elimination.Importantly, low or no uptake was observed in other organs, including the kidneys and bone marrow.Coupled with high tumor-specific uptake in murine xenografts, this finding means that [ 68 Ga]Ga-DPI-4452 could serve as a PET-based tool for selecting patients for treatment with [ 177 Lu]Lu-DPI-4452.
Extrapolating [ 177 Lu]Lu-DPI-4452 dog dosimetry to humans suggested the small intestine and stomach to be potential doselimiting organs, considering the conservative limits set for externalbeam radiotherapy (21).Although the dog organs with the highest uptake are those associated with high CAIX expression in humans, differences in CAIX expression between these species remain uncertain.Therefore, the first biodistribution data from humans (e.g., with [ 68 Ga]Ga-DPI-4452) will provide useful insights.Accordingly, the present work provides a strong foundation for the initiation of human studies with this theranostic pair.

CONCLUSION
This is the first report of preclinical biodistribution and pharmacologic proof-of-concept studies involving DPI-4452.[ 68 Ga]Ga-DPI-4452 and [ 177 of [ 177 Lu]Lu-DPI-4452 were assessed in HT-29 and SK-RC-52 xenografts.Tumor-bearing mice were randomized into 4 groups and injected with single doses of vehicle, 100 or 33 MBq of [ 177 Lu]Lu-DPI-4452, or 3 once-weekly 33-MBq doses of [ 177 Lu]Lu-DPI-4452.Treatment was well tolerated, including no significant changes in kidney function biomarkers (Supplemental Figs.9-11).All treatment groups, except HT-29 xenograft-bearing mice receiving a single dose of 33 MBq, had significantly reduced tumor volumes compared with vehicle controls by day 16 (HT-29) or day 13 (SK-RC-52) (Fig. 3; Supplemental Figs. 12 and 13; Supplemental Table 4).Both 100-MBq and 3 once-weekly 33-MBq doses produced maximal tumor growth inhibition at day 23 for HT-29 xenografts and day 36 for SK-RC-52 xenografts, although 3 once-weekly 33-MBq doses produced a more sustained effect than a single 100-MBq dose in both models.No significant difference in tumor size for SK-RC-52 xenografts was observed at day 36 after treatment with single 100-or 33-MBq doses of [ 177 Lu]Lu-DPI-4452 (Supplemental Table FIGURE 1. [ 111 In]In-DPI-4452 tissue pharmacokinetics in mice bearing HT-29 and SK-RC-52 xenografts.Mean uptake (radioactivity concentration as percentage injected dose per gram of tissue [%ID/g]) over time in key organs after injection of 30 MBq of [ 111 In]In-DPI-4452 into mice with human CRC HT-29 (A) and human ccRCC SK-RC-52 (B) subcutaneous xenografts.Shown are measurements for individual mice.Line represents mean.n 5 3.

TABLE 1 CAIX Expression Measured by Immunohistochemistry in Multiple Tumor and Control Tissues
Data are reported as percentages, with number of positive samples/total number of samples assessed in parentheses.Sq.NSCLC 5 squamous non-small cell lung cancer; PDAC 5 pancreatic ductal adenocarcinoma; SCCHN 5 squamous cell carcinoma of head and neck; TNBC 5 triple-negative breast cancer; ND 5 not determined. †

TABLE 4
Plasma Pharmacokinetics of DPI-4452 in Mice and Dogs area under plasma concentration-time curve extrapolated to infinity; C 5 min 5 measured concentration at 5 min after injection; t last 5 time to last measurable concentration; V ss 5 volume of distribution at steady state.

TABLE 6
Mean Plasma Pharmacokinetics of DPI-4452 After Single Injection of DPI-4452 in Beagle Dogs Reported as median, with range (minimum to maximum) in parentheses.C 5 min 5 measured concentration at 5 min after injection; t last 5 time to last measurable concentration; AUC inf 5 area under plasma concentration-time curve extrapolated to infinity. † Lu]Lu-DPI-4452 showed tumor-specific uptake, high tolerability, and rapid systemic elimination.[ 177 Lu]Lu-DPI-4452 also demonstrated strong antitumor efficacy in 2 murine xenograft models.[ 68 Ga]Ga-DPI-4452 and [ 177 Lu]Lu-DPI-4452

TABLE 7
Calculated Absorbed Doses to Organs of [ 68 Ga]Ga-DPI-4452 and [ 177 Lu]Lu-DPI-4452 for Human ICRP-89 Phantoms a promising theranostic pair for targeting CAIX-expressing tumors.A clinical phase 1/2 study is currently underway to assess the safety, tolerability, and imaging characteristics of [ 68 Ga]Ga-DPI-4452 and the efficacy of [ 177 Lu]Lu-DPI-4452 in patients diagnosed with CRC, PDAC, or ccRCC and having unresectable locally advanced or metastatic solid tumors (ClinicalTrials.govidentifier: NCT05706129).received financial support from Debiopharm International SA.Franck Brichory, Fr ed eric Massi ere, Norbert Wiedemann, and Antoine Attinger are current employees and Inês Borrego was a former employee of Debiopharm International SA.Aileen Hoehne, Frank Osterkamp, Matthias Paschke, Dirk Zboralski, Anne Schumann, and Anne Bredenbeck are employees of 3B Pharmaceuticals GmbH.No other potential conflict of interest relevant to this article was reported.