Enhanced Therapeutic Effects of 177Lu-DOTA-M5A in Combination with Heat Shock Protein 90 Inhibitor Onalespib in Colorectal Cancer Xenografts

Simple Summary Cancer treatment is hampered by the limitations of individual therapy modalities and the intricate nature of the disease. The administration of maximal monotherapy doses often leads to undesirable side effects and/or therapy resistance. As a result, there is a growing recognition of the importance of investigating combination therapy to effectively address these obstacles. In the present in vivo study, the therapeutic effects of combination therapy with the heat shock protein 90 inhibitor onalespib, a potential radiosensitizer, and 177Lu-DOTA-M5A for colorectal cancer (CRC) treatment were explored for the first time. The results demonstrated that the combination treatment was so effective that retained or even superior therapeutic effects could be achieved with only half the dose of administered 177Lu-DOTA-M5A, showing enhanced tumor growth suppression and increased apoptosis. Consequently, the combination therapy involving 177Lu-DOTA-M5A and onalespib constitutes a promising approach for treating metastatic CRCs. By enhancing therapeutic effects, minimizing therapy resistance, and reducing side effects, this approach has the potential to expand the patient population that can benefit from targeted treatment. Abstract Carcinoembryonic antigen (CEA) has emerged as an attractive target for theranostic applications in colorectal cancers (CRCs). In the present study, the humanized anti-CEA antibody hT84.66-M5A (M5A) was labeled with 177Lu for potential CRC therapy. Moreover, the novel combination of 177Lu-DOTA-M5A with the heat shock protein 90 inhibitor onalespib, suggested to mediate radiosensitizing properties, was assessed in vivo for the first time. M5A antibody uptake and therapeutic effects, alone or in combination with onalespib, were assessed in human CRC xenografts and visualized using SPECT/CT imaging. Although both 177Lu-DOTA-M5A and onalespib monotherapies effectively reduced tumor growth rates, the combination therapy demonstrated the most substantial impact, achieving a fourfold reduction in tumor growth compared to the control group. Median survival increased by 33% compared to 177Lu-DOTA-M5A alone, and tripled compared to control and onalespib groups. Importantly, combination therapy yielded comparable or superior effects to the double dose of 177Lu-DOTA-M5A monotherapy. 177Lu-DOTA-M5A increased apoptotic cell levels, indicating its potential to induce tumor cell death. These findings show promise for 177Lu-DOTA-M5A as a CRC therapeutic agent, and its combination with onalespib could significantly enhance treatment efficacy. Further in vivo studies are warranted to validate these findings fully and explore the treatment’s potential for clinical use.


Introduction
Colorectal cancer (CRC) is the third most prevalent cancer worldwide, with over 1.9 million cases diagnosed and 0.9 million deaths reported annually [1,2]. Carcinoembryonic antigen (CEA) is a cancer-associated biomarker and is highly expressed in colorectal tumors as well as other solid tumors such as gastric, pancreatic, breast, and lung cancer [3][4][5]. CEA undergoes post-translational modification known as glycosylphosphatidylinositol (GPI) anchoring, leading to the attachment of a phosphoglyceride moiety to its C-terminus. This specific molecular event renders CEA a membrane-bound protein with GPI-anchored localization. The biological, physiological, and pathological functions of CEA is not yet fully understood. There are studies suggesting a role in cell adhesion, where cells are more loosely associated with each other in normal tissues [6], and a potential role in innate immune defense of the colon and other tissues from microbial attack [7,8]. In the domain of pathological functions of CEA, there are studies suggesting a role in tumorogenesis function in vitro [9] and in transgenic mice [10,11]. CEA is a GPI membrane-bound protein that is minimally expressed on the apical surface of normal colorectal epithelial cells, but its expression is elevated on the basolateral surface of CRC cells that are accessible to the bloodstream [12][13][14][15][16]. Higher expression of CEA on cancer cells leads to higher levels of circulating CEA in blood, and the association between CEA serum level and prognosis of the disease has been widely studied [17]. Consequently, CEA has become a preferred marker for diagnosis and prognosis of CRC, and it is also a promising target for CRC research and therapy [12,[18][19][20].
Radioimmunotherapy (RIT) using radiolabeled conjugates to target CEA is a promising approach to the treatment of cancers including CRCs. Radiolabeled antibodies such as T84. 66 have shown positive outcomes in vitro, in vivo, and in clinical trials [21]. The chimeric cT84.66 has been studied extensively as a single agent and with chemotherapies [4,[21][22][23][24], and a humanized version (M5A) was developed to minimize immunogenicity [25]. A phase I clinical trial with 90 Y-labeled M5A in patients with advanced CEA-expressing tumors showed therapeutic effects, feasibility of therapy, and antibody tolerance [14]. 177 Lu is a promising therapeutic radionuclide for targeted therapy, and presents lower energy and a shorter emission range compared to 90 Y [26]. Recently, treatment of 3D in vitro models with 177 Lu-DOTA-M5A demonstrated promising therapeutic effects on colorectal cancer models [27]. The therapeutic effects of 177 Lu-DOTA-M5A have, however, not been studied in vivo, and this study constitutes the first in vivo assessment of the therapeutic effects of 177 Lu-DOTA-M5A.
A second target of interest in CRCs is heat shock proteins (HSPs), a family of molecular chaperones that play pivotal roles in the correct folding and function of client proteins [28]. In particular, HSP90 and its client proteins are involved in numerous cancer hallmarks, including proliferation, metastasis, and invasion [29,30]. The HSP90 client proteins comprise unstable signaling molecules such as kinases, transcription factors, and chromatin remodeling. This can eventually result in cancer cells gaining capabilities of self-sufficiency in growth signaling, insensitivity to anti-growth signaling, limitless replicative potential, and eventually invasion and metastasis [31,32]. High expression of HSP90 is associated with a poor prognosis in colorectal cancer patients [33] as well as other types of malignancies such as melanoma, leukemia, lung, and breast cancers [34][35][36][37]. By inhibiting HSP90, cancer progression can be attenuated, as many HSP90 client proteins and related pathways are required for cancer cell progression. Onalespib is a second-generation HSP90 inhibitor with better solubility and less hepatotoxicity than previous inhibitors such as 17-AAG and 17-DMAG. HSP90 inhibition has also been shown to delay DNA repair mechanisms and the repair of radiation-induced double strand DNA breaks (DSBs) [38,39].
The therapeutic effects of onalespib have been demonstrated in several preclinical and clinical studies. Combination therapies of onalespib and chemotherapy drugs/radiation in solid tumors have demonstrated greater therapeutic effects compared to monotherapies [40,41]. Onalespib, when combined with external beam radiation or 177 Lu-DOTATATE, suggested radiosensitizing properties in colorectal cancer and neuroendocrine cancer mod-els [38,42]. Moreover, an in vitro study using colorectal 3D spheroid models demonstrated synergistic therapeutic effects of onalespib combined with 177 Lu-DOTA-M5A [27]. Thus, the radiosensitizing properties of onalespib in preclinical studies and particularly in colorectal cancers, as well as the pronounced effects of HSP90 inhibitors on various cellular pathways motivate the investigation of the combination of 177 Lu-DOTA-M5A and onalespib in colorectal cancer.
The aim of the current study was to assess the in vivo effects of 177 Lu-DOTA-M5A therapy and HSP90 inhibition in colorectal cancer xenografts and to further investigate the radiosensitizing potential of onalespib when combined with 177 Lu-DOTA-M5A. The therapeutic effects and possible toxicities were assessed in mice bearing human colorectal cancer xenografts.

M5A Monoclonal Antibody, Radiolabeling, and Onalespib
The anti-CEA hT84.66-M5A mAb is a humanized IgG 1 [25]. To select and optimize the suitable xenograft model, the M5A mAb was labeled with 125 I using 12 × 75 mm pre-coated glass iodogen tubes (50 µg of iodogen/tube) according to the manufacturer's instruction (Thermofisher Scientific, Waltham, MA, USA). Briefly, the iodogen tube was washed with 1 mL of Phosphate Buffered Saline (PBS, St. Louis, MO, USA) followed by adding 100 µL of PBS. The 125 I was activated by adding 5 MBq of 125 I to PBS, mixing by pipetting and incubating for 6 min at room temperature (swirling every 30 s). The activated 125 I was then transferred to a new tube along with 50 µg of M5A antibody. The mixture of M5A and 125 I was stirred on a thermoshaker (TS-100C Smart, Biosan, Riga, Lativa) at 37 • C for 15 min (350 rpm), yielding the final specific activity of 100 kBq/µg. The extent of radiolabeling yield was assessed using instant thin layer chromatography (ITLC, Biodex Medical Systems, New York, NY, USA) and NaCl (0.9%) as a mobile phase, and the radiolabeling yield was quantified using a phosphorimager (BAS-1800II, Fujifilm, Tokyo, Japan). For the biodistribution experiment with 125 I-M5A, injected activity was 250 kBq and unlabeled M5A was added to obtain 150 µg/mL.
The M5A mAb was conjugated with NHS-DOTA, purified as previously described [47], and labeled with 177 Lu for biodistribution and therapy studies [27]. Briefly, DOTA-M5A mAb was buffer exchanged to sodium acetate (0.2 M, pH 5.5) using a spin column (Amicon Ultra-0.5 mL, 3 kDa NMWCO, Merck, Darmstadt, Germany). The mixture of DOTA-M5A (50 µg) dissolved in sodium acetate (0.2 M, pH 5.5) and 177 Lu (15 MBq, in 0.04 M HCl, ITG GmbH) was stirred on the thermoshaker at 42 • C for 1 h (350 rpm), yielding the final specific activity of 300 kBq/µg. The extent of radiolabeling yield was evaluated using ITLC and citric acid (0.2 M, pH 5.5) as a mobile phase and the radiolabeling yield was quantified using a phosphorimager. The 177 Lu-DOTA-M5A was purified using a spin column. A labeling yield above 96% or higher was used for animal experiments. 177 Lu-DOTA-M5A was further diluted in Dulbecco's Phosphate Buffered Saline (DPBS, Biowest, MO, USA) to 150 µg/mL and 100 µL was injected intravenously (IV) in the tail vein. For assessment of therapeutic effects, tumor growth and survival study, injected activity was either 4.5 or 10 MBq administrated and unlabeled M5A was added to obtain 150 µg/mL (15 µg of the total mass of antibody per animal). For the control group, unlabeled M5A (150 µg/mL, 100 µL) was administered. For the biodistribution experiments, a concentration of 150 µg/mL was obtained (15 µg of the total mass of antibody per animal).

In Vivo Colorectal Xenograft Models
The in vivo studies were performed in accordance with current Swedish laws and regulations with the approval of the Uppsala Committee of Animal Research Ethics. Studies were designed to comply with replace, reduce, and refine (3R) animal study principles while maintaining the quality of data, with the approval of the Uppsala Committee of Animal Research Ethics. Female Balb/c nude mice (age = 6-8 weeks, weight = [16][17][18][19][20] were housed under standard conditions and fed ad libitum. A total of 70 mice were involved in the present study. Of these, 3 were used for the pilot study to select and optimize the tumor xenograft, 35 were used in the biodistribution study and molecular assessment, and 32 were used in the therapy study. Either 5 × 10 6 HT55 cells in PBS or 5 × 10 6 SNU1544 cells in PBS inoculated subcutaneously into the flank. The tumor takes in both xenograft models were 100%. After tumors were established, body weight and tumor growth were monitored every other day. Tumor diameter was measured using a digimatic caliper (Mitutoyo, Sweden) and volume was calculated as (Length × Width 2 ) × 0.5236. Mice with tumor sizes 60-300 mm 3 were selected for the studies. To verify antigen selectivity after labeling in physiological conditions as well as selecting the most suitable xenograft model, a small biodistribution study with iodinated M5A ( 125 I-M5A) was performed in mice bearing either HT55 or SNU1544 xenografts (N = 3). Eleven days after tumor cells inoculation, each animal was administered approximately 250 kBq 125 I-M5A (unlabeled M5A was added to obtain 150 µg/mL) by an IV injection of a volume of 100 µL into the tail vein (15 µg of total antibody/mouse). Twenty-four hours post-injection, animals were euthanized by a mixture of ketamine (10 mg/mL) and xylazine (1 mg/mL) at a dose of 0.2 mL per 10 g of body weight (IP), followed by heart puncture. Blood was aspirated, and organs of interest were excised and weighed. Average tumor weight was approximately 160 mg at the study endpoint. The radioactivity (uptake) of 125 I-M5A of blood and organs together with an injection standard of 100 µL (approximately 250 kBq) was measured in a gamma counter (1480 Wizard 3 , Wallace, Finland). The injected activity of 125 I-M5A for each animal was calculated by subtracting the residual activity in the syringe from the standard injected dose. Radioactivity uptake in the organ was calculated as percent of injected activity per gram of tissue (%ID/g). (150 µg/mL) by an IV injection of 100 µL into the tail vein (15 µg of total antibody/mouse). Animals were euthanized 3 h (n = 4), 24 h (n = 4), 48 h (n = 4), 96 h (n = 4), and 168 h (n = 3) post-injection, blood was aspirated and organs of interest were excised and weighed. The radioactivity (uptake) of 177 Lu-DOTA-M5A of blood and organs together with an injection standard of 100 µL (approximately 400 kBq) were measured in a gamma counter. The injected activity of 177 Lu-DOTA-M5A for each animal was calculated as described earlier.
The serum was separated on half of each blood sample using microvette tubes (Microvette ® 500 Serum, Merck, Darmstadt, Germany) for further assessment of soluble CEA level.
To further assess radioconjugate uptake in tumor and vital organs in animals bearing smaller tumors (seven days after tumor cells inoculation) and receiving either 177 Lu-DOTA-M5A (4.5 MBq), or the combination of 4.5 MBq of 177 Lu-DOTA-M5A and onalespib, a second biodistribution study was then performed, where the treatment scheme of the therapy study was imitated (N = 8) but all animals were euthanized (as described earlier) two days after the regime finished (96 h post-administration of 177 Lu-DOTA-M5A). Tumors were 137 ± 32 mm 3 in size at the time of radioconjugate administration. Blood was aspirated, tumor, liver, kidney, and spleen samples were collected. The radioactivity (uptake) of 177 Lu-DOTA-M5A of blood and vital organs together with an injection standard of 100 µL (approximately 4.5 MBq) were measured in a gamma counter. The injected activity of 177 Lu-DOTA-M5A for each animal was calculated as described earlier. The serum was separated from half of each blood sample using microvette tubes for further assessment of soluble CEA level. CT raw files were reconstructed by Filter Back Projection (FBP). SPECT raw data were reconstructed by Nucline software (3.00.020.000) using the TeraTomo 3D algorithm with 3 subsets, 48 iterations, and corrected for scatter and attenuation artifacts. SPECT and CT data were fused and analyzed using PMOD v3.508 (PMOD Technologies Ltd., Zurich, Switzerland). Coronal SPECT/CT images of the scans were presented as maximum intensity projections (MIP) in RGB color scale to obtain a visual confirmation of the biodistribution results.

Molecular Assessment and Short-Term Toxicity via Ex Vivo Immunohistochemistry
Ex vivo immunohistochemistry (IHC) was performed to assess molecular response. In this experiment, the treatment scheme of the therapy study was imitated (N = 16, 4 per treatment group), all animals were euthanized (as described earlier) two days after the regime finished, and tumor, liver, kidney, and spleen samples were collected. Tissues were fixed in 4% buffered formaldehyde (VWR, Gdańsk, Poland), paraffin-embedded, sectioned, and deparaffinized. Embedding, sectioning, and Hematoxylin and eosin (HE)/IHC-staining was performed by the Research and Development department, Clinical Pathology at the Uppsala University hospital, Sweden. Briefly, staining was performed with a Dako Autostainer 48 (Agilent, Santa Clara, CA, USA). Antigen retrieval was performed with retrieval solution low pH (Dako K8005, Agilent, Santa Clara, CA, USA) or high pH (Dako K8004, Agilent, Santa Clara, CA, USA). Sections were immunostained with antibodies targeting cleaved PARP1 (rabbit monoclonal antibody, ab32064, Abcam, Cambridge, UK) and CEA (rabbit monoclonal antibody, ab133633, Abcam, Cambridge, UK). Signals were detected using the Envision Flex kit (Dako K8010, Agilent, Santa Clara, CA, USA). Counterstaining with hematoxylin (Histolab, Gothenburg, Sweden) was done in a Tissue-Tek Prisma (Sakura, Alphen aan den Rijn, The Netherlands). Immunohistochemical sections were manually scored according to staining intensity (negative −, weak (Grade 1, +), moderate (Grade 2, ++), or strong (Grade 3, +++) for CEA. Figure S1 illustrates the reference images. For cleaved PARP1, positive cells of four random areas of 0.2 mm 2 per section were counted using ImageJ (1.53t, Java 1.8.0_172, NIH, Bethesda, MD, USA). In short, the areas where color deconvolution, signals in threshold window of 0-100 and area of >5 µm were counted. All histological analyses were performed blinded to the treatment group of the animal. For IHC staining, stroma and necrotic areas were excluded and only tumor tissue was scored and analyzed. For toxicity assessment in vital organs via HE staining, presence of fibrosis, necrosis, and inflammation were assessed.
For toxicity analyses, three Hematoxylin and eosin-stained sections of 0.2 mm 2 each on the liver, kidney, and spleen were randomly chosen and presence/absence of damage was determined in a blinded manner, with consideration for the presence of necrosis and fibrosis in liver, glomeruli contraction in kidney, and necrosis and congestion in the spleen.

Therapy Regime, Tumor Growth, and Survival
Seven days after tumor cell inoculation, mice were randomized and individually earmarked to start the therapy regime. Animals were weighed and received one of the following treatments: (1) Table 1. There were no significant differences in initial tumor volume between the groups, with an average size of 166 ± 49 (range 69-274) mm 3 for all groups. The therapy regime was four consecutive days of daily IP injection of 30 mg/kg onalespib or DMSO (control). On the second day of the therapy regime, a single dose of either 4.5 or 10 MBq 177 Lu-DOTA-M5A (15 µg of total antibody/mouse) or DOTA-M5A (15 µg of total antibody/mouse) was injected once into the corresponding groups. Tumors were measured three times/week as the therapy started until either the study endpoint of tumor size 1.2-1.3 cm 3 or consistent weight loss of more than 10% compared to the day of treatment initiation was reached. At the study endpoint, animals were euthanized by a mixture of ketamine (10 mg/mL) and xylazine (1 mg/mL) at a dose of 0.2 mL per 10 g of body weight (IP), followed by heart puncture. Blood was aspirated and serum was separated using microvette tubes. Tumors and livers were also harvested and preserved in formalin (VWR, Darmstadt, Germany). To evaluate possible long-term drug-induced hepatotoxicity, serum was separated from the collected blood samples to evaluate the long-term effects of therapy on the aforementioned hepatotoxicity-related enzymes at the end-point. Collected serum samples were tested for mouse aspartate aminotransferase (AST) using an enzyme-linked immunosorbent assay (ELISA) kit (Catalog number EKU11950, Biomatik, Wilmington, DE, USA) and mouse alanine transaminase (ALT) ELISA kit (Catalog number EKN43321, Biomatik, DE, USA). ELISA assays were performed according to the manufacturer's instructions.

Tumor Growth Rate Analysis
Tumor growth rates were assessed based on tumor volume measurements as described earlier. Daily tumor volume data were log-transformed and used as the dependent variable in a linear mixed model allowing for varying initial tumor size and tumor growth rate per mouse. A bootstrap procedure was used to estimate 95% confidence intervals and test the statistical significance of between-group differences [48]. Log-scale model coefficients were back-transformed to represent average daily growth rates in each treatment group for presentation. See Supplementary Document S1 for details of the analysis underlying the analysis of tumor growth rate.

Statistical Analysis
Statistical data analysis for experiments was performed using GraphPad Prism Version 9.4.0 (GraphPad Software Inc., La Jolla, CA, USA) and R (v 4.2.0) using the lme4 [49] and lmeresampler [49] packages to estimate mixed effects models and perform bootstrapping. Statistical comparisons between two groups were performed using Student's t-test and statistical analysis comparing three or more groups was performed using analysis of variance (one-way ANOVA) with Tukey's multiple comparison test. Data are presented as means ± standard deviation (SD), if not otherwise stated. p < 0.05 was considered significant (* p < 0.05, ** p < 0.01).

HT55 Is a Suitable Xenograft Model
Biodistribution of the anti-CEA 125 I-M5A mAb was first performed to evaluate the tumor uptake in HT55 and SNU1544 xenografts (Figure 1a). 125 I-M5A mAb tumor uptake 24 h post-injection in HT55 and SNU1544 tumor xenografts were 36 ± 6 injected dose per gram of tissue (% ID/g) and 41 ± 3%ID/g, respectively. The tumor uptake in the two xenograft models was comparable, and HT55 was chosen for continued in vivo assessments.

In Vivo Tumor Targeting was Confirmed
An extended biodistribution study was conducted using 177 Lu-DOTA-M5A mAb on HT55 xenografts (409 ± 0.24 mg, Figure 1b). Tumor uptake was 17 ± 1%ID/g at 3 h post-injection and peaked at 49 ± 4%ID/g at 48 h, remaining steady until 168 h. Blood activity decreased with time. Uptake in organs peaked at 3 h for the kidney and spleen, and at 48 h for the liver, decreasing at later time points. Tumor to blood ratio was 0.35 ± 0.035 at 3 h and peaked at 39.3 ± 6.9 at 96 h ( Figure 1c).

In Vivo Tumor Targeting was Confirmed
An extended biodistribution study was conducted using 177 Lu-DOTA-M5A mAb on HT55 xenografts (409 ± 0.24 mg, Figure 1b). Tumor uptake was 17 ± 1%ID/g at 3 h postinjection and peaked at 49 ± 4%ID/g at 48 h, remaining steady until 168 h. Blood activity decreased with time. Uptake in organs peaked at 3 h for the kidney and spleen, and at 48 h for the liver, decreasing at later time points. Tumor to blood ratio was 0.35 ± 0.035 at 3 h and peaked at 39.3 ± 6.9 at 96 h ( Figure 1c).

Low Uptake in Organs at Risk, SPECT/CT Imaging
Since previous studies have demonstrated an association between tumor burden and amount of soluble CEA in the blood [50], which can eventually alter the biodistribution profile, a second biodistribution study was performed on mice bearing smaller tumors (137 ± 32 mm 3 , 103 ± 0.02 mg, Figure 1d), representative of the therapeutic setting. As expected, uptake in liver, kidneys, and spleen was significantly lower in this experimental setting, with subsequently higher tumor uptake. Liver, kidney, and spleen uptakes 96 h post-injection were 8.7 ± 0.7%, 8 ± 0.5%ID/g, and 18.7 ± 4.8%ID/g, respectively. The average tumor uptake was 116 ± 19%ID/g, and did not differ significantly between two groups. ELISA assessment of blood CEA levels demonstrated a lower blood CEA level (12.6 ± 1.8 ng/mL) in animals with smaller tumor burden (0.1 ± 0.02 g) compared to animals with larger tumor burden [(0.64 ± 0.1 g, 1074 ± 700 ng/mL), Figure 2]. SPECT/CT imaging of animals receiving 4.5 MBq of 177 Lu-DOTA-M5A or combined administration of 177 Lu-

Low Uptake in Organs at Risk, SPECT/CT Imaging
Since previous studies have demonstrated an association between tumor burden and amount of soluble CEA in the blood [50], which can eventually alter the biodistribution profile, a second biodistribution study was performed on mice bearing smaller tumors (137 ± 32 mm 3 , 103 ± 0.02 mg, Figure 1d), representative of the therapeutic setting. As expected, uptake in liver, kidneys, and spleen was significantly lower in this experimental setting, with subsequently higher tumor uptake. Liver, kidney, and spleen uptakes 96 h post-injection were 8.7 ± 0.7%, 8 ± 0.5%ID/g, and 18.7 ± 4.8%ID/g, respectively. The average tumor uptake was 116 ± 19%ID/g, and did not differ significantly between two groups. ELISA assessment of blood CEA levels demonstrated a lower blood CEA level (12.6 ± 1.8 ng/mL) in animals with smaller tumor burden (0.1 ± 0.02 g) compared to animals with larger tumor burden [(0.64 ± 0.1 g, 1074 ± 700 ng/mL), Figure 2

Therapeutic Effects at Molecular Level and No Detected Short-Term Toxicity
Immunohistochemical-stained sections demonstrated a significant increase in cleaved PARP1+ cells in the combination-treated tumors compared to onalespib-treated tumors 48 h after end of treatment (96 h post-radioconjugate injection). No significant differences were observed between 177 Lu-DOTA-M5A and combination therapy (Figure 4). CEA was strongly expressed in the tumors of all groups according to CEA expression grades (Figures 4a and S1) and no CEA expression alteration attributable to the administered therapeutic regimen was detected. Hematoxylin and eosin staining of the vital organs, liver, kidney, and spleen, assessing the presence of necrosis and fibrosis in liver, glomeruli contraction in kidney, and necrosis and congestion in spleen ( Figure S2) at the same time point did not indicate any toxicity. Immunohistochemical-stained sections demonstrated a significant increase in cleaved PARP1+ cells in the combination-treated tumors compared to onalespib-treated tumors 48 h after end of treatment (96 h post-radioconjugate injection). No significant differences were observed between 177 Lu-DOTA-M5A and combination therapy (Figure 4). CEA was strongly expressed in the tumors of all groups according to CEA expression grades (Figure 4a and Figure S1) and no CEA expression alteration attributable to the administered therapeutic regimen was detected. Hematoxylin and eosin staining of the vital organs, liver, kidney, and spleen, assessing the presence of necrosis and fibrosis in liver, glomeruli contraction in kidney, and necrosis and congestion in spleen ( Figure S2) at the same time point did not indicate any toxicity.

Combination Therapy Decreased Tumor Growth Rate
In a xenograft model with human colon carcinoma HT55, the therapeutic effects of 177 Lu-DOTA-M5A and onalespib were investigated alone and in combination. The tumor growth rates demonstrated therapeutic benefits for 177 Lu-DOTA-M5A and the combination therapy groups. The daily tumor growth rates are listed in Table 2, the daily tumor growth curves and a comparison summary are illustrated in Figure 5. The daily tumor growth rate did not exhibit any statistically significant differences among the control groups. As a result, they were consolidated into a singular group for further analysis. While monotherapy with onalespib reduced the tumor growth rate, the impact did not attain statistical significance compared to controls. However, monotherapy with 4.

Combination Therapy Decreased Tumor Growth Rate
In a xenograft model with human colon carcinoma HT55, the therapeutic effects of 177 Lu-DOTA-M5A and onalespib were investigated alone and in combination. The tumor growth rates demonstrated therapeutic benefits for 177 Lu-DOTA-M5A and the combination therapy groups. The daily tumor growth rates are listed in Table 2, the daily tumor growth curves and a comparison summary are illustrated in Figure 5. The daily tumor growth rate did not exhibit any statistically significant differences among the control groups. As a result, they were consolidated into a singular group for further analysis. While monotherapy with onalespib reduced the tumor growth rate, the impact did not attain statistical significance compared to controls. However, monotherapy with 4.5 MBq and 10 MBq of 177 Lu-DOTA-M5A reduced tumor growth rates by a factor of 2.4 and 3.4 times, respectively. Combination therapy of 4.5 MBq of 177 Lu-DOTA-M5A and onalespib significantly reduced tumor growth rate compared to monotherapies and controls by a factor of 4, 1.6, and 3.5 compared to control group, 177 Lu-DOTA-M5A (4.5 MBq), and onalespib, respectively. The tumor growth rates of combination therapy (4.5 MBq of 177 Lu-DOTA-M5A and onalespib) and 10 MBq of 177 Lu-DOTA-M5A were not significantly different (Figure 5d,e). No complete tumor remission was observed. Serum levels of hepatotoxicity-related enzymes were below the detection limit for all samples (0.312 ng/mL for AST and 1.56 ng/mL for ALT).      The same color/shape codes were applied (mean of tumor size at each time, error bars represent SEM. n ≥ 3). The graph was cut when the first animal in the group was euthanized. Graphs display tumor volume (mm 3 ) on y axis and days after treatment on x axis.

Combination Therapy Increased the Median Survival
Next, the in vivo data was further analyzed for median and maximum survival (Figure 6a and Table 2). Onalespib alone increased animal survival by 25%, while 4.5 or 10 MBq of 177 Lu-DOTA-M5A increased median survival by 175% and 260% compared to control. Combination therapy of 4.5 MBq 177 Lu-DOTA-M5A and onalespib improved median survival by 265% compared to control. The median survival of combination treated animals (4.5 MBq of 177 Lu-DOTA-M5A and onalespib) and monotherapy with 10 MBq of 177 Lu-DOTA-M5A was comparable; moreover, the maximum survival of combinationtreated animals was 100 days, while it was 80 days for animals treated with 10 MBq of 177 Lu-DOTA-M5A. Thus, onalespib potentiated the therapeutic effects of 4.5 MBq 177 Lu-DOTA-M5A to that of more than the double dose (10 MBq) of 177 Lu-DOTA-M5A as a monotherapy. The individual mouse weights are presented in Figure 6b.

Combination Therapy Increased the Median Survival
Next, the in vivo data was further analyzed for median and maximum survival (Figure 6a and Table 2). Onalespib alone increased animal survival by 25%, while 4.5 or 10 MBq of 177 Lu-DOTA-M5A increased median survival by 175% and 260% compared to control. Combination therapy of 4.5 MBq 177 Lu-DOTA-M5A and onalespib improved median survival by 265% compared to control. The median survival of combination treated animals

Discussion
RIT has demonstrated significant potential for cancer therapy [51]. The anti-CEA T84.66 monoclonal antibody (mAb) is a high affinity mAb that targets CEA-positive disease, and radiolabeled variants have demonstrated specificity in clinical trials alone and when combined with chemotherapy drugs [13][14][15]52]. A humanized version, hT84.66-M5A, has been developed and a clinical trial with beta-emitter 90 Y has shown therapeutic potential [14]. While 177 Lu is a well-established radionuclide for cancer therapy, it has not previously been assessed with M5A for potential in RIT of solid tumors.
Despite its promise, RIT has some limitations for solid tumors due to abnormal blood flow and heterogeneous antigen expression. As a result, combination therapies have been explored to enhance therapeutic effects and minimize toxicity [53,54], where CEA targeting combined with chemotherapy or immunocytokines has shown some promise [15,18,19,22,23]. Moreover, HSP90 inhibitors have been studied in preclinical and clinical trials, either alone or with other therapies such as chemotherapy or radiation [55]. Onalespib, a second-generation HSP90 inhibitor, has demonstrated radiosensitizing properties in preclinical studies when combined with external beam radiation, 177 Lu-DOTA-M5A (in vitro) [27,38], or 177 Lu-DOTATATE in tumor xenografts [56]. The combination of CEA-targeted RIT with HSP90 inhibition using onalespib is, thus, an intriguing therapeutic strategy that warrants further investigation. In this paper, following the successful in vitro characterization of 177 Lu-DOTA-M5A [27], we have assessed the therapeutic strategy of

Discussion
RIT has demonstrated significant potential for cancer therapy [51]. The anti-CEA T84.66 monoclonal antibody (mAb) is a high affinity mAb that targets CEA-positive disease, and radiolabeled variants have demonstrated specificity in clinical trials alone and when combined with chemotherapy drugs [13][14][15]52]. A humanized version, hT84.66-M5A, has been developed and a clinical trial with beta-emitter 90 Y has shown therapeutic potential [14]. While 177 Lu is a well-established radionuclide for cancer therapy, it has not previously been assessed with M5A for potential in RIT of solid tumors.
Despite its promise, RIT has some limitations for solid tumors due to abnormal blood flow and heterogeneous antigen expression. As a result, combination therapies have been explored to enhance therapeutic effects and minimize toxicity [53,54], where CEA targeting combined with chemotherapy or immunocytokines has shown some promise [15,18,19,22,23]. Moreover, HSP90 inhibitors have been studied in preclinical and clinical trials, either alone or with other therapies such as chemotherapy or radiation [55]. Onalespib, a secondgeneration HSP90 inhibitor, has demonstrated radiosensitizing properties in preclinical studies when combined with external beam radiation, 177 Lu-DOTA-M5A (in vitro) [27,38], or 177 Lu-DOTATATE in tumor xenografts [56]. The combination of CEA-targeted RIT with HSP90 inhibition using onalespib is, thus, an intriguing therapeutic strategy that warrants further investigation. In this paper, following the successful in vitro characterization of Initial evaluation of tumor uptake using 125 I-M5A in human colon adenocarcinoma SNU1544 and human colon carcinoma HT55 xenografts demonstrated that both models had high and comparable tumor uptake values above 35%ID/g (Figure 1). An extended biodistribution study was then performed using 177 Lu-DOTA-M5A in HT55 tumor-bearing mice, illustrating high and stable tumor uptake over seven days, in line with previous studies [25,57]. The HT55 tumor xenograft model was selected for subsequent therapeutic investigations of 177 Lu-DOTA-M5A and/or onalespib treatments.
In the context of a biodistribution study using 177 Lu-DOTA-M5A, the liver exhibited the highest percentage injected dose per gram of tissue (%ID/g) compared to other organs, which is consistent with the fact that the liver's reticuloendothelial system is responsible for clearing circulating CEA and IgGs [58]. The liver's high internalization and catabolism is further demonstrated by the comparison of liver uptake between 125 I-and 177 Lu-labeled antibodies, where the latter had much higher uptake values than iodine ( Figure S3). It is likely that the higher levels of soluble circulating CEA in animals with larger tumor burdens ( Figure 2) may have led to more formation of 177 Lu-DOTA-M5A:CEA complexes, resulting in higher uptake in the liver (approximately 32%ID/g at 96 h post-injection) than in animals with smaller tumors. This theory is supported by the lower liver uptake observed in mice with smaller tumors, and consequently lower and more clinically relevant CEA blood levels [59,60], resulting in lower than 10%ID/g liver uptake (as seen in Figure 1b,d). As smaller tumors were used in the subsequent therapy study, these results suggest a low risk of 177 Lu-DOTA-M5A induced hepatotoxicity in the therapy study.
In the therapy study, the effects of 177 Lu-DOTA-M5A and HSP90 inhibitor onalespib, alone or in combination, were evaluated ( Figures 5 and 6 and Table 2). The results showed that monotherapies reduced tumor growth rate and increased median and maximum survival compared to controls, with 177 Lu-DOTA-M5A (in both doses) being more effective than onalespib in the present settings. Moreover, tumor growth rate was significantly decreased in the combination therapy compared to the corresponding monotherapies, further reducing tumor growth rates by 1.6 times compared to monotherapy of 4.5 MBq of 177 Lu-DOTA-M5A group and 3.5 times compared to HSP90 inhibitor onalespib group. Animals treated with 10 MBq of 177 Lu-DOTA-M5A exhibited comparable suppression of tumor growth to those receiving a combination therapy with a significantly lower dose (4.5MBq) of 177 Lu-DOTA-M5A. It should be noted that this analysis assumes constant growth rates, which appeared to be the case in control and onalespib mice. In mice treated with 177 Lu-DOTA-M5A, however, tumor growth tended to be suppressed in early measurements, while accelerating further along in the experiment (see Figure 5). Immunohistochemical analysis of harvested tumors showed significant expression of CEA (in all groups) and increased apoptosis and karyorrhexis in 177 Lu-DOTA-M5A or the combination treatment groups. These findings support previous studies showing that radiation-induced DNA damage causes apoptosis and karyorrhexis [61].
As for survival data for monotherapies, onalespib alone slightly increased median survival (25 days) compared to control (20 days), while 4.5 or 10 MBq of 177 Lu-DOTA-M5A substantially increased median survival by 175% (55 days) and 260% (72 days), respectively ( Figure 6). These beneficial radioimmunotherapy results are in line with the initial phase I clinical trials using 90 Y-labeled chimeric antibody T84.66 [24,62] and the recent phase I clinical trial using the same radionuclide for labeling a humanized version of the antibody ( 90 Y-DOTA-M5A) for advanced CEA producing cancers [14]. Although RIT with antibodies against CEA shows promise, the therapeutic effects can be diminished by varying levels of CEA expression in CRCs and heterogeneity in tumor bulk [63,64]. Hence, a more potent therapy approach is needed to overcome these limitations and a suitable combination therapy could be transformative.
Even though monotherapies were effective, the real promise was shown in the combination treatments. In the present study, combination therapy demonstrated a 265% increased median survival (73 days) compared to control, corresponding to a 192% improvement in survival compared to onalespib monotherapy and 33% compared to 4.5 MBq of 177 Lu-DOTA-M5A monotherapy. Molecular assessment demonstrated that the number of apoptotic cells was greater in tumors treated with 177 Lu-DOTA-M5A either as a standalone therapy or in combination, significantly higher for the combination group compared to the onalespib monotherapy and control groups. Moreover, the combination therapy was so effective that retained therapeutic effects could be achieved with only half the dose of administered 177 Lu-DOTA-M5A, potentially decreasing toxicity and improving therapeutic efficiency. These findings support previous studies combining 90 Y-cT84.66 with chemotherapy agents in CEA-expressing tumors [19,22,23] and suggest that 177 Lu-DOTA-M5A and onalespib may be a promising combination therapy for CRCs.
Using full-sized antibodies like M5A in RIT can increase tumor uptake [65], but it also increases the risk of high radiation dose to normal organs including bone marrow and liver, as liver processes antigen-antibody complexes [66]. Moreover, while onalespib is a second-generation HSP90 inhibitor with relatively low hepatotoxicity, hepatotoxicity as a dose-limiting factor must nonetheless be considered in combination therapies [67]. In the present study, AST and ALT enzyme levels were, therefore, closely investigated, demonstrating that the values of both hepatotoxic-related enzymes in blood were below the detection limit of the ELISA assays. The study found no hepatotoxicity or significant weight loss in animals receiving monotherapies and the combination therapy thereof, indicating the potential benefit of this combination therapy for future therapies.

Conclusions
The 177 Lu-DOTA-M5A is a promising novel radioconjugate with potential for RIT of CEA-expressing CRCs. In addition, onalespib can further potentiate the therapeutic effects of 177 Lu-DOTA-M5A, prolong the survival, further suppress the tumor growth and enhance tumor cells apoptosis. Henceforth, using 177 Lu-DOTA-M5A with HSP90 inhibitors as combination therapies may be a feasible therapy approach for metastatic colorectal cancers to overcome the antitumor therapy resistance, and may have the potential to widen the targetable patient population.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/cancers15174239/s1, Figure S1: IHC grading reference. This figure displays manually graded immunohistochemical staining sections targeting CEA. The scale bar is 50 µm; Figure S2: Hematoxylin and eosin staining of vital organs. This figure displays representative images of liver, kidney and spleen of untreated animals (control), animals treated with 177 Lu-DOTA-M5A (4.5 MBq), onalespib or combination of 177 Lu-DOTA-M5A (4.5 MBq) and onalespib, n = 3 per treatment group.; Figure S3: Biodistribution of 177 Lu-DOTA-M5A and 125 I-M5A as %ID/g in Balb C nude mice bearing HT55 xenografts 24 h post-injection as %ID/g. Error bars represent the means ± SD), n ≥ 2; Supplementary Document S1: Tumor growth analysis notebook. This document provides an accounting of the analysis steps relating to tumor growth rates in the paper "Enhanced therapeutic effects of 177 Lu-DOTA-M5A in combination with heat shock protein 90 inhibitor onalespib in colorectal cancer xenografts".
Author Contributions: T.M. contributed to experimental studies with focus on animal experiments, analyzing and interpreting the data, and drafting and revising the manuscript. D.S. contributed to experimental studies with focus on statistical analysis using R software, analyzing and interpreting the data, and drafting and revised the manuscript. S.C. contributed to experimental studies with focus on animal experiments, analyzing and interpreting the data, and drafting and revising the manuscript. F.Y.F. contributed to the data interpretation, and revising the manuscript. P.J.Y. contributed to the data interpretation, and revising the manuscript. M.N. initiated and designed the study, contributed to data analysis and interpretation, and drafting and revising the manuscript. All authors have read and agreed to the published version of the manuscript.

Institutional Review Board Statement:
This study was performed according to the FELASA recommendations. All experiments complied with Swedish law and were performed with permission from the Uppsala Committee of Animal Research Ethics (C33/16). The authors have nothing to disclose.

Informed Consent Statement: It is not applicable.
Data Availability Statement: The datasets generated for this study are available on request to the corresponding author.