Targeted radionuclide therapy in patient-derived xenografts using ¹⁷⁷Lu-EB-RGD

Abstract

Currently, most patients with non-small cell lung cancer (NSCLC) are diagnosed in advanced stages with a poor five-year survival rate. Therefore, intensive research aimed at finding novel therapeutic strategies has been ongoing; experimental models that reliably emulate NSCLC disease are greatly needed to predict responses to novel therapeutics. Therefore, we developed patient-derived xenograft models of NSCLC, which we then used to evaluate the therapeutic efficacy of ¹⁷⁷Lu-EB-RGD, a peptide-based radiopharmaceutical with improved pharmacokinetics that targets integrin α_vβ₃. In this study, three different groups of NSCLC-PDXs were successfully established, all of which maintained the same immunohistochemical and genetic characteristics of the human primary tumour. The two NSCLC-PDX groups with intense and low expression of integrin α_vβ₃ (denoted as PDX_αvβ3+ and PDX_αvβ3-) were chosen as the experimental models to evaluate the in vivo biological behaviour of ¹⁷⁷Lu-EB-RGD. In SPECT imaging and biodistribution studies, ¹⁷⁷Lu-EB-RGD showed significantly higher accumulation in PDX_αvβ3+ and PDX_αvβ3- models than its corresponding monomer ¹⁷⁷Lu-RGD. A single dose of 18.5 MBq ¹⁷⁷Lu-EB-RGD was enough to completely eradicate the tumours in PDX_αvβ3+, with no sign of tumour recurrence during the observation period. Such treatment was also efficacious in PDX_αvβ3-: a single dose of 29.6 MBq ¹⁷⁷Lu-EB-RGD led to a significant delay in tumour growth as compared to that in the control or ¹⁷⁷Lu-RGD group. The preclinical data from the use of this model suggest that ¹⁷⁷Lu- EB-RGD may be an effective treatment option for NSCLC and should be further evaluated in human trials.

Introduction

Lung cancer is the leading cause of cancer-related death worldwide; approximately 2 million new cases and 1.8 million deaths were reported in 2018 by Global Cancer Statistics 2018¹. The two major forms of lung cancer are non-small cell lung cancer (NSCLC) and small cell lung cancer (SCLC). NSCLC accounts for approximately 80–85% of lung cancers, and adenocarcinoma is the most common histological subtype. Currently, most NSCLC patients are diagnosed in advanced stages with five-year survival rates of less than 10%². Chemotherapy based on platinum is the first-line treatment for NSCLC but only provides a modest survival benefit³. Therefore, a better understanding of the molecular basis of the NSCLC and novel therapies are needed to significantly improve patient outcome.

The last decade has witnessed a paradigm shift from cytotoxic drugs to targeted therapy in medical oncology and pharmaceutical innovation. Inspired by breakthroughs in molecular and cellular biology, several novel synthesised chemical compounds and recombinant antibodies have been developed to selectively target oncogenic signalling pathways in NSCLC⁴. Such targeted therapeutic agents show impressive clinical efficacy and often significantly prolong the overall survival of individuals with NSCLC. Inevitably, however, the treated tumours recur as resistance to these targeted therapies develops⁵. More recently, immunotherapies such as PD-1/PD-L1 immune checkpoint blockades (ICB) have shown promise, with their durable response in advanced NSCLC without driver oncogene mutations⁶. However, the low objective response rate is the main obstruction to using ICBs more widely.

Targeted radionuclide therapy (TRT), a branch of radiotherapy, uses isotope emitting charged particles after labelling with a targeting vector. A carrier molecule can seek special molecular or functional targets. Besides being utilised in the treatment of local tumours like external radiotherapy, TRT can treat metastatic tumours through systemic administration⁷. A unique feature of radionuclides is that they can exert the “cross-fire” effect, potentially destroying adjacent tumour cells⁸. On January of 2018, ¹⁷⁷Lu-labelled dodecanetetraacetic acid tyrosine-3-octreotate (DOTATATE) (trade name Lutathera) was approved by the FDA to treat patients with somatostatin receptor-positive gastroenteropancreatic neuroendocrine tumours (NETs)⁹.

Patient-derived xenografts (PDXs), obtained by direct implantation of tissue fragments into immunodeficient mice, are of particular interest because they retain the morphology, architecture, and molecular signatures of the corresponding parental tumour more closely than is possible for in vitro established cell lines. PDXs have been successfully investigated to identify specific determinants of therapeutic response and to predict the therapy response of individual tumours to chemotherapy and targeted therapy. However, the potential of PDXs to predict the response of NSCLC to TRT has not been investigated. This has important clinical implications for patients with advanced NSCLC, especially for those who are ineffective or resistant to platinum-based therapy or epidermal growth factor receptor-tyrosine kinase inhibitors. In the present study, we investigated the therapeutic efficacy of peptide-based TRT in PDXs derived from NSCLC. The Arg-Gly-Asp (RGD) peptide targets the integrin α_vß₃ receptor on the tumour cells and neovasculature, which is highly expressed in many cancers¹⁰. Our previous investigation has demonstrated the positive expression of integrin α_vß₃ in patients with NSCLC and ⁶⁸Ga labelled RGD peptide (⁶⁸Ga-NOTA-PRGD2) could be of use for lesion detection and help stage patients with NSCLC¹¹. In this research, RGD was chemically conjugated with an albumin-binding moiety—truncated Evans Blue (denoted as EB-RGD). The major feature was the inclusion of the EB-derived albumin-binding moiety that affected the pharmacokinetics in a positive manner to elevate tumour uptake and tumour residence time^12,13. For TRT application, EB-RGD was labelled with ¹⁷⁷Lu, which emits β rays. Through this investigation, we aimed to investigate whether PDXs could reproduce the same characteristics of the corresponding parental NSCLC and explore whether the ¹⁷⁷Lu-EB-RGD could be used as a novel therapeutic approach in patients with advanced NSCLC.

Materials and methods

PDX establishment

Patient written informed consent was obtained, and the research protocol was approved by the Clinical Research Ethics Committee of the First Affiliated Hospital of Xiamen University (ID KYZ2017–001). All animal care and experimental procedures were reviewed and approved by the Animal Care and Use Committee of Xiamen University Laboratory Animal Center (ID XMULAC20170063). The tumour specimens were obtained from patients who underwent presurgical ¹⁸F-FDG PET/CT. After surgical resection, NSCLC specimens were immediately placed in DMEM (Cat. #C11995599BT, Gibco, USA) supplemented with 2% antibiotics (penicillin, streptomycin) and transferred into the ice box.

Immunodeficient BALB/c nude mice were bred under sterile conditions in Xiamen University Laboratory Animal Center from founders originally obtained from Charles River Co., Ltd (Beijing, China). To establish PDXs, a fresh tumour specimen was processed into approximately 30 mm³ for implantation subcutaneously (s.c.) at the level of trans-scapular brown fat of 4–6 weeks BALB/c Nude female mice within an average of 2 hours following the patient’s surgery. Tumour growth was monitored at least twice weekly using callipers until it reached a maximal volume of 1500 mm³, after which the mouse was euthanised by cervical dislocation, and the tumour was removed. The tumour was minced with scissors and tumour fragments were implanted in BALB/c Nude mice; the remaining fragments were used for histopathological analyses. For investigational purposes, PDXs were expanded in a higher number of mice to obtain statistically relevant results.

Radiochemistry

Cyclic-RGD peptide was purchased from C.S. Bio for research purpose. EB-RGD was synthesised as previously described¹³. Details of synthetic procedures and characterisation of chemical components are described in the supplemental materials (Supplemental Fig. 1). High-purity radionuclide ¹⁷⁷Lu was purchased from ITG (Germany) in a solution of ¹⁷⁷LuCl₃. For radiolabelling, 1.85 GBq of ¹⁷⁷LuCl₃ was diluted using 4 mL of 0.05 M HCl. EB-RGD (100 μg), or RGD (100 μg) was dissolved in 1 mL of 0.25 M NH₄OAc (pH = 5.6), and subsequently added to the ¹⁷⁷LuCl₃. The mixture was heated at 90 °C for 30 mins; a C₁₈ column was used for purifying and a 0.22-μm filtration membrane was used to ensure sterility. Next, 1 mL ethanol of 60% volume fraction was injected into the product container through a C₁₈ column and aseptic filtration membrane; further, 5 mL normal saline was added. The ethanol concentration in the solution was about 10% when the tail vein was injected. Analytical thin-layer chromatography (Bioscan, USA) was used for radiochemical purity control. Moreover, we evaluated the ¹⁷⁷Lu-EB-RGD stability both in vitro and in vivo; the methodology for the stability assays can be found in the supplemental material.

SPECT imaging and biodistribution study in NSCLC-PDX models

Single-photon emission computer tomography (SPECT) was performed with the Mediso (Hungary) multi-pinhole camera; 18.5 MBq of ¹⁷⁷Lu-EB-RGD or ¹⁷⁷Lu-RGD were intravenously injected into NSCLC-PDXs (n = 4). Whole-body SPECT imaging was performed as described previously¹². The acquisition times were at 4, 24, 48, 72, and 96 hours post injection (p.i.). Regions of interest (ROIs) of the tumour, liver, heart, and muscle were counted on the SPECT images to quantify the radioactive signals, which were recorded as blood-to-muscle (B/M), tumour-to-muscle (T/M), and liver-to-muscle (L/M) ratios.

In the biodistribution study, groups of NSCLC-PDXs were injected with the same batch (same specific activities) of 0.5 MBq ¹⁷⁷Lu-EB-RGD and were sacrificed at different times (4, 24, 48, and 72 hours p.i.; n = 3 for each time point, 12 mice for each group). The main organs and tumour were weighed and analysed. The ¹⁷⁷Lu-RGD group biodistribution was performed for contrasting as well. Groups of NSCLC-PDXs were sacrificed at 24 hours p.i. (n = 3 for each group). The γ-counter (WIZARD 2480; USA) was used for assaying radioactivity.

The in vivo distribution pattern of EB-RGD and RGD was further evaluated via dynamic PET scan (using ⁶⁸Ga-EB-RGD and ⁶⁸Ga-RGD as the PET tracer) in healthy BALB/C mice. The half-life of EB-RGD and RGD were calculated by drug and statistics software (DAS version 2.0). The methodology for the dynamic PET scan can be found in the supplemental material.

In vivo therapy regimen

NSCLC-PDXs were used for in vivo therapy study when the tumour volume reached approximately 100 mm³. NSCLC-PDXs were divided into four groups (9 – 10 mice/group): Group A, 100ul saline; Group B, 29.6 MBq of ¹⁷⁷Lu-RGD; Group C, 18.5 MBq of ¹⁷⁷Lu-EB-RGD; Group D, 29.6 MBq of ¹⁷⁷Lu-EB-RGD (all drugs was injected through intravenous (i.v.).

Tumour volume and body weight were recorded each 2 days. The formula: (length × width × width)/2 was used to calculated tumour volume. Three days after the beginning of treatment, groups A–D (n = 3) from NSCLC-PDXs underwent ¹⁸F-FDG PET (Inveon, Siemens) imaging. The ¹⁸F-FDG PET imaging protocol was performed as previously described¹⁴.

The endpoint criteria for mice were as follows: tumour volume >1,500 mm³, weight loss >20%, or active tumour ulceration.

Histopathological staining

Animal tumour specimens from each group were sectioned at day 14 after treatment. CD31 immunohistochemistry (IHC) was used to picture the endothelial cells. The primary antibodies of this experiment were anti-mouse CD31 mAb (ab9498; Abcam, Cambridge, United Kingdom), Ki-67-specific mAb (ab15580; Abcam, Cambridge, United Kingdom), human integrin α_vβ₃ (ab7166; Abcam, Cambridge, United Kingdom), and murine integrin α_vβ₃ (CD61, Novus Biologicals, SJ19–09). Using these markers, routine IHC staining was performed according to our previous protocol¹³. The expression of integrin α_vβ₃ was further detected by Western blot analysis; the protocol can be found in the supplemental material. A commercially available kit (Roche Applied Science) was for immunofluorescent TdT-mediated dUTP Nick-End Labelling (TUNEL) analysis. Haematoxylin and eosin (H&E) staining was procured by Univ-Bio Inc.

Statistics

All quantitative data are expressed as mean ± SD. All statistical analyses were conducted by using SPSS 22.0 statistical analysis software (IBM, Armonk, NY, USA). One-way analysis of variance and Student’s t test were used to compare means. A P value of < 0.05 was considered statistically significant.

Results

Radiolabelled compounds preparation

Two radioligands with high affinity for integrin α_vß₃ were prepared in this study: ¹⁷⁷Lu-EB-RGD and ¹⁷⁷Lu-RGD, both of which had >95% radiochemical purity after purification. ¹⁷⁷Lu-RGD and ¹⁷⁷Lu-EB-RGD were radiolabelled at an average specific activity of 18.6 ± 4.7 and 55.85 ± 14.0 GBq/μmol, respectively.

Stability of ¹⁷⁷Lu-EB-RGD

¹⁷⁷Lu-EB-RGD was stable in mouse serum for up to 24 h with no significant de-metalation observed. In vivo, over 90% of the radioactivity was bound to the blood proteins and not extractable. The extracted radioactivity showed ¹⁷⁷Lu-EB-RGD with only a small amount of a more polar component appearing at 4 h after i.v. injection.

NSCLC-PDXs establishment and validation

Fragments of primary tumour tissues were collected from eight patients who underwent surgical resection, and the diagnosis of NSCLC was confirmed by histopathology. The small fragments of primary tumour tissues were subsequently implanted into eight groups of immunodeficient BALB/c Nude mice for propagation. Finally, three different groups of NSCLC-PDXs were successfully established, all of which were derived from adenocarcinomas. The three groups had different clinical stages (1 stage IB, 1 stage IIA, and 1 stage IIIA), different degrees of lymph node involvement (1 N0, 1 N1, and 1 N3) and different degrees of integrin α_vβ₃ expression levels (1 intense, 1 moderate, and 1 low expression). No driver mutations were observed in any of the three groups. The latency times (time interval from implantation to presence of palpable tumour) (range, 23.3–65.6 days post implantation) and tumour growth rate (Supplemental Fig. 2) varied in the three groups. The phenotype of each NSCLC patient is summarised in Supplemental Table 1. All these parameters were consistent between the NSCLC-PDXs and their corresponding primary NSCLC (Supplemental Table 2), confirming the unique nature of individual PDXs. The two NSCLC-PDXs groups with intense and low integrin α_vβ₃ expression (denoted as PDX_αvβ3+ and PDX_αvβ3-) were chosen as the experimental models to evaluate the in vivo biological behaviour of ¹⁷⁷Lu-EB-RGD. It is worth noting that integrin α_vβ₃ is expressed on both tumour cells and neovasculature. Although PDX_αvβ3- tumours have low expression of integrin α_vβ₃ on the tumour cells, they show high levels of murine vascular integrin expression (CD61, murine integrin α_vβ₃). The PET imaging and histopathology information from the two PDX groups and their parental tumours are presented in Figure 1.

Fig. 1.

Histopathology and positron-emission tomography (PET) imaging of patient-derived xenografts (PDX) and parental tumours. (A) Haematoxylin and eosin (H&E) staining in human NSCLCs and corresponding PDXs; (B) Immunohistochemistry staining of integrin α_vβ₃ expression and CD31 in human NSCLCs and corresponding PDXs (tumour vasculatures were indicated by arrows); (C) The expression of integrin α_vβ₃ was detected by Western blot analysis; (D) ¹⁸F-FDG PET imaging in two patients with NSCLC and corresponding PDXs.

SPECT imaging and biodistributions of ¹⁷⁷Lu-EB-RGD in NSCLC-PDXs

Whole-body SPECT imaging studies were performed for exploring the in vivo behaviour of ¹⁷⁷Lu-EB-RGD. Representative whole-body Maximum Intensity Projections (MIP) of PDX_αvβ3+ and PDX_αvβ3- are presented in Figure 2.

Fig. 2.

Maximum intensity projections from single-photon emission computer tomography (SPECT) imaging studies in PDXs. (A) Projection SPECT images of mice in PDX_αvβ3+ models at 24 h after 18.5 MBq ¹⁷⁷Lu-RGD injection (n = 4) and at 4, 24, 48, 72, 96 and 120 h after 18.5 MBq ¹⁷⁷Lu-EB-RGD injection (n = 4). (B) Projection SPECT images of mice in PDX_αvβ3- models at 24 h after 18.5 MBq ¹⁷⁷Lu-RGD injection (n = 4) and at 4, 24, 48, 72, and 96 h after 18.5 MBq ¹⁷⁷Lu-EB-RGD injection (n = 4).

SPECT imaging of ¹⁷⁷Lu-EB-RGD in PDX_αvβ3+ is shown in Figure 2A. High tumour-to-background ratio was observed at 4 h p.i. (T/M: 7.34 ± 0.67), and elevated at 24 and 48h p.i., with tumour uptake peaking at 72 h p.i. (T/M: 15.99 ± 5.42) and reducing slightly from 96 to 120 h p.i. Blood ¹⁷⁷Lu-EB-RGD uptake was comparatively high at 1 h p.i. (B/M: 4.31 ± 1.22), and decreased from 24 to 120 h p.i. gradually. The kidneys also showed a ¹⁷⁷Lu signal, which gradually decreased from 24 to 120 h p.i. As shown in Figure 2B, ¹⁷⁷Lu-EB-RGD uptake in PDX_αvβ3- was significantly lower than that in PDX_αvβ3+ at all time points. The T/M ratio in PDX_αvβ3- from 4 to 96 h p.i were 4.18 ± 0.74, 5.17 ± 4.11, 6.97 ± 1.12, 6.62 ± 1.64 and 7.34 ± 0.67, respectively. ¹⁷⁷Lu-RGD SPECT imaging in both PDX_αvβ3+ and PDX_αvβ3- was also performed for comparison (Figure 2A, B). ¹⁷⁷Lu-RGD was shown to be cleared from the blood rapidly and showed a significantly low accumulation in both PDXs at 24 h p.i.

The biodistributions of ¹⁷⁷Lu-EB-RGD in the two PDX groups were obtained at different time points p.i. (Figure 3). The biodistributions results were generally in line with previous SPECT imaging studies. In the case of PDX_αvβ3+ (Figure 3A), ¹⁷⁷Lu-EB-RGD accumulated mainly in the blood (12.99 ± 1.54 %ID/g) at 4 h p.i. Moreover, ¹⁷⁷Lu-EB-RGD uptake was also observed in the kidney (9.81 ± 1.83 %ID/g), liver (7.56 ± 4.89 %ID/g), and spleen (7.87 ± 1.95 %ID/g). Tumour uptake at this time was 10.75 ± 1.17 %ID/g. ¹⁷⁷Lu-EB-RGD in the blood, kidneys, liver, and spleen had decreased to 4.41 ± 0.45, 8.99 ± 0.99, 6.06 ± 0.65, and 5.61 ± 2.67 %ID/g at 24 h p.i., respectively, while the tumour uptake had increased to 13.38 ± 1.04 %ID/g. ¹⁷⁷Lu-EB-RGD uptake in tumour had reached its peak (14.87 ± 6.71 %ID/g) at 48 h p.i., whereas the uptake in blood (1.71 ± 0.35 %ID/g), kidneys (7.78 ± 1.58 %ID/g), liver (5.60 ± 0.75 %ID/g), and spleen (5.23 ± 1.04 %ID/g) had continued to decrease. At 72 h p.i., ¹⁷⁷Lu-EB-RGD further clearance from all organs was observed. A similar in vivo distribution pattern was observed in PDX_αvβ3- (Figure 3B), except for the significantly reduced tumour uptake at all time points tested (¹⁷⁷Lu-EB-RGD uptakes in PDX_αvβ3- from 4–72 h p.i. were 4.89 ± 1.33, 6.25 ± 0.56, 6.47 ± 1.50, and 5.55 ± 2.05 %ID/g, respectively).

Fig. 3.

Biodistribution of ¹⁷⁷Lu-EB-RGD in PDX groups. (A) PDX_αvβ3+ models at 4, 24, 48 and 72 h post-injection (n = 3). (B) Comparison of ¹⁷⁷Lu-EB-RGD and ¹⁷⁷Lu-RGD in PDX_αvβ3+ models at 24 h post-injection (n = 3/group). (C) ¹⁷⁷Lu-EB-RGD in PDX_αvβ3- models at 4, 24, 48 and 72 h post-injection (n = 3/group). (D) Comparison of ¹⁷⁷Lu-EB-RGD and ¹⁷⁷Lu-RGD in PDX_αvβ3- models at 24 h post-injection (n = 3/group).

The ¹⁷⁷Lu-RGD biodistribution was also investigated for comparison (Figure 3C–D). In accordance with the SPECT findings, ¹⁷⁷Lu-RGD demonstrated a much shorter blood half-life than ¹⁷⁷Lu-EB-RGD in both PDX groups. The blood uptake of ¹⁷⁷Lu-RGD was almost 15 times lower than ¹⁷⁷Lu-EB-RGD at 24 h p.i. Therefore, ¹⁷⁷Lu-EB-RGD demonstrated a significantly higher tumour uptake in both PDX_αvβ3+ (13.38 ± 1.04 vs. 2.59 ± 1.60 %ID/g) and PDX_αvβ3-; (6.26 ± 0.56 vs. 1.56 ± 0.35 %ID/g). The promising high tumour uptake of ¹⁷⁷Lu-EB-RGD and prolonged tumour retention time resulted in high exposure to the radioactive isotope. The area under the curve calculated from the radioactivity accumulation in tumours over time is shown in Supplemental Fig. 3.

Furthermore, the pharmacokinetic parameters of EB-RGD and RGD were calculated by dynamic PET scan (Supplemental Fig. 4), which demonstrated that intravenously administered RGD was cleared from the circulation rapidly with a t_1/2 of 33.4 ± 0.9 min. The t_1/2 of EB-RGD was 167.8 ± 19.7 min, a 5-fold longer half-life than that of RGD.

¹⁷⁷Lu-EB-RGD demonstrates significant antitumoral radiotherapy efficacy in NSCLC-PDXs

To investigate whether ¹⁷⁷Lu-EB-RGD TRT would be efficacious in NSCLC-PDXs, the antitumoural radiotherapy efficacy of ¹⁷⁷Lu-EB-RGD was first evaluated in the PDX_αvβ3+ subjects, which were divided into four groups (A–D) and monitored for 50 days post treatment. The Figure 4A showed therapy protocols along with the mice relative tumour volumes and body weigh.

Fig. 4.

Therapy protocol and outcomes in PDX_αvβ3+ mice. (A) Design of therapy protocol in PDX_αvβ3+ models (n = 9 or 10 per group). (B) Change in body weight after treatment. (C) Change in tumour volume after treatment. (D) Survival of mice after therapy (Groups C and D were overlapped; Group D was indicated by the blue arrow). (E) Projection PET images of PDX_αvβ3+ models injected with ¹⁸F-FDG at 3 days after treatment. (F) PET quantification of the images in (E). ** P < 0.01

Transient weight loss was observed in this PDX, and the mice regained their starting weight by the end of the study. (Figure 4B). Group A (saline) and group B (29.6 MBq of ¹⁷⁷Lu-RGD) showed rapid tumour growth. All PDXs in groups A and B had to be euthanised by day 36 because of excessive tumour volume (Figure 4C, pink and purple lines). To evaluate the appropriate therapy dose of ¹⁷⁷Lu TRT, ¹⁷⁷Lu-EB-RGD was delivered in a radioactive dose of either 18.5 MBq (Group C) or 29.6 MBq (Group D). In groups C and D, significant inhibition of tumour growth could be seen (Figure 4C). ⁷⁷Lu-EB-RGD and saline-treated groups showed significant difference in tumour volume starting from day 4 (P = 0.007), while ¹⁷⁷Lu-EB-RGD and ¹⁷⁷Lu-RGD group demonstrated significant difference starting from day 6 (P = 0.001). These differences increased over time; most tumours in the ¹⁷⁷Lu-EB-RGD therapy groups began to shrink from day 6 and completely disappeared at day 24. Although Group D showed a faster tumour inhibition rate as compared to Group C, all of the PDXs (9/9, 100%) in both groups finally showed a complete response (CR), with all of these CR maintaining for the duration of the 50 days study. Kaplan-Meier curves based on the protocol endpoints showed significant differences (P < 0.001) between groups A–B and groups C–D, but no significant difference was observed between groups C and D (Figure 4D). The survival rate in the two ¹⁷⁷Lu-EB-RGD-treated groups were both maintained at 100% up to day 50 after the initial treatment. Therefore, a therapy dose of 18.5 MBq ¹⁷⁷Lu-EB-RGD may be strong enough for tumour elimination in the PDXs with intense integrin α_vβ₃ expression.

During treatment, early therapeutic response was evaluated by ¹⁸F-FDG PET at day 3 after treatment. The representative ¹⁸F-FDG PET images of four therapy groups are presented in Figure 4E. Compared to the saline control and ¹⁷⁷Lu-RGD-treated groups, significantly decreased ¹⁸F-FDG uptake was observed in the ¹⁷⁷Lu-EB-RGD-treated groups (Figure 4F), indicating reduced tumour metabolism following ¹⁷⁷Lu TRT.

We also tested the therapeutic efficacy of ¹⁷⁷Lu-EB-RGD in the NSCLC-PDXs with low expression (PDX_αvβ3-), using the same therapy protocol among four groups (Figure 5A). Once again, transient weight loss due to ¹⁷⁷Lu TRT was observed in these PDX_αvβ3-, although the extent of wight loss was lower than that of PDX_αvβ3+. (Figure 5B). The tumour volumes in the two ¹⁷⁷Lu-EB-RGD-treated groups were similar to those in the control and ¹⁷⁷Lu-RGD-treated groups up to day 8 after treatment, and then a minor tumour volume reduction was observed in both of the ¹⁷⁷Lu-EB-RGD-treated groups, which lasted for several days. However, the tumour volumes in the ¹⁷⁷Lu-EB-RGD-treated groups increased once more from day 14 onward but grew at a slower rate than in the control and ¹⁷⁷Lu-RGD-treated groups. In other words, ¹⁷⁷Lu-EB-RGD-targeted radionuclide therapy led to a delay in tumour growth in the PDX_αvβ3-, rather than tumour shrinkage, as shown in PDX_αvβ3+ (Figure 5C).

Fig. 5.

Therapy protocol and outcomes in PDX_αvβ3- mice (A) Design of therapy protocol in the NSCLC-PDXs with low integrin expression (PDX_αvβ3-). (B) Change in body weight after treatment. (C) Change in tumour volume after treatment.

Tumour biology evaluated by immunohistochemistry

As shown in Figure 6, Groups C and D (18.5 and 29.6 MBq ¹⁷⁷Lu-EB-RGD) displayed a much lower extent of tumour vasculature than did Groups A and B (saline and 29.6 MBq ¹⁷⁷Lu-RGD). Groups A and B demonstrated a relatively higher percentage of cells that stained positively for Ki-67, while groups C and D demonstrated reduced cell proliferation. Regarding TUNEL staining, both ¹⁷⁷Lu-EB-RGD groups showed considerably more cell apoptosis compared to groups A and B. H&E staining showed extensive necrotic areas for the ¹⁷⁷Lu-EB-RGD-treated groups, whereas tumours from the saline control and ¹⁷⁷Lu-RGD-treated groups only showed necrotic areas in the middle of the tumour. To further test the possible toxicity effects of ¹⁷⁷Lu-EB-RGD, pathological examination of H&E staining of selected non-cancerous organs was performed, which revealed no obvious differences between the ¹⁷⁷Lu-treated and control groups (Supplemental Fig. 5).

Fig. 6.

Immunohistochemistry (IHC) staining of excised tumours for CD31, Ki-67, TUNEL, and haematoxylin and eosin (H&E) after treatment with saline (group A), 29.6 MBq of ¹⁷⁷Lu-RGD (group B), 18.5 MBq of ¹⁷⁷Lu-EB-RGD (group C), and 29.6 MBq of ¹⁷⁷Lu-EB-RGD (group D) in PDX_αvβ3+ models.

The IHC results in the PDX_αvβ3- groups were similar to those in the PDX_αvβ3+ groups, except that fewer necrotic areas of PDX_αvβ3- were observed in ¹⁷⁷Lu-EB-RGD-treated groups as compared to those in PDX_αvβ3+ (Supplemental Fig. 6).

DISCUSSION

Current chemotherapy regimens have unsatisfactory results in most advanced NSCLC. Targeted therapies and immunotherapies represent two of the most researched types of therapy under current investigation for the treatment of NSCLC⁴. However, over 80% of patients with advanced NSCLC do not have oncogenic driver biomarker status and resistance is inevitable⁴. Immunotherapy, especially ICB therapy, provides the hope that long-term survival is achievable in a subset of patients, but the ICB therapy could only benefit to a small population of patients¹⁵. It is therefore imperative to develop novel therapeutic strategies and to optimize selection of patients, identifying early those who could benefit from available treatments.

Mouse tumour models are the most valuable tool for preclinical evaluation of novel therapeutic strategies in cancer¹⁶. However, some drugs that show therapeutic efficacy against human tumour xenografts have often failed to show activity in human clinical trials^17,18, resulting in an imperfect correlation between preclinical and clinical results in terms of therapeutic efficacy. Recently, some studies have demonstrated that PDXs have great potential in drug development studies since they reliably reproduce the patient’s parental tumour for both IHC markers and genetic alterations as well as for evaluating the response to the corresponding therapeutic regimens. Moreover, PDXs accurately reproduce the heterogeneity of human cancers and maintain at least some aspects of the human microenvironment; therefore, they represent a promising preclinical model for studies of tumour heterogeneity and tumour-microenvironment studies^19,20. As such, we investigated the establishment of PDXs from NSCLC as a source of therapeutically relevant information. As expected, our results demonstrated that PDXs derived from NSCLC can maintain the same immunohistochemical and genetic characterization of the human primary tumour, confirming the robustness of this model to test the efficacy of new treatments for NSCLC. Moreover, in our experience, the growth features of PDX (such as latency time and tumour growth rate) reach stability after 3–4 passages in mice; it is therefore advisable to carry out drug efficacy evaluation with PDXs that have reached a certain stability in their growth characteristics.

We have further confirmed the robustness of the PDXs by using PET imaging, which showed concordant uptake of ¹⁸F-FDG between primary NSCLC and corresponding PDXs. Interestingly, lower ¹⁸F-FDG uptake was observed in PDXs with lower integrin α_vβ₃ expression. This may be partially explained by the finding that both integrin α_vβ₃ expression and ¹⁸F-FDG uptake are correlated with tumour aggressiveness and prognosis in NSCLC²¹. However, it cannot be concluded that ¹⁸F-FDG provides information similar to that of an RGD-based tracer, because both tracers have completely different pharmacodynamic properties. In a previous study²², Beer et al. compared the ¹⁸F-RGD uptake and ¹⁸F-FDG uptake on PET in primary and metastatic tumour lesions in cancer patients, and found no substantial correlation between ¹⁸F-RGD and ¹⁸F-FDG uptake, suggesting that α_vβ₃ expression and glucose metabolism are not closely linked in malignant lesions.

Targeted radionuclide therapy (TRT) is the use of radioisotopes to deliver ionization energy to tumour cells⁸. TRT is particularly useful in cases of metastatic disease where conventional treatments are no longer effective^7,23. According to previous clinicals trials in NETs and CRPC, ¹⁷⁷Lu TRT results in high responses, a low toxicity profile, and improves overall survival^23,24. As such, TRT may also provide a novel strategy for the treatment of advanced NSCLC. Although several preclinical studies have demonstrated the significant anti-tumour efficacy of TRT in lung cancer mouse models⁴, no clinical trial until now has provided a clue as to whether this approach holds promise or not. In our present study, we chose RGD peptide as the TRT vector for targeting the integrin α_vβ₃ receptor, which has been validated to be overexpressed in NSCLC and to have a crucial role in tumour angiogenesis¹¹. More importantly, the pharmacokinetics of the targeting vector RGD were further improved by conjugation of RGD with an albumin binding moiety, the truncated EB. The reversible binding of EB to albumin has been well documented, with a stoichiometry of 8–14 dye molecules per molecule of albumin²⁵. In our previous study¹³, we demonstrated that albumin can serve as a pool and allow the slow release of EB-RGD from albumin. The slow release of drug over time would allow continual uptake at the biological target (integrin α_vβ₃ on tumour cells and vasculature). Furthermore, the internalization assays in our previous study have found that most of the EB-RGD was internalised into tumour cells (due to α_vβ₃ receptor-mediated internalization), while the albumin was not internalised in the presence of EB-RGD.

The improved pharmacokinetics of EB-RGD, demonstrated by both a longer-half life and higher retention in the blood pool, resulted in significantly higher accumulation of ¹⁷⁷Lu-EB-RGD and enhanced tumour to background ratios as shown in SPECT imaging and biodistribution studies. The anti-tumour efficacy of ¹⁷⁷Lu-EB-RGD in NSCLC-PDXs is also quite impressive: a therapeutic dose of 18.5 MBq ¹⁷⁷Lu-EB-RGD was enough to completely eradicate the NSCLC-PDXs with high integrin α_vβ₃ expression, with no sign of tumour recurrence during the observation period. Considering integrin α_vβ₃ is expressed on both tumour cells and neovasculature, ¹⁷⁷Lu-EB-RGD was able to target both tumour cells and tumour vasculature, as confirmed by the reduced expression of CD31 (a vascularization marker) and Ki-67 (a proliferation marker) after treatment.

Such TRT treatment was also efficacious even in the NSCLC- PDXs with low integrin α_vβ₃ expression; a therapeutic dose of 29.6 MBq ¹⁷⁷Lu-EB-RGD led to a significant delay in tumour growth as compared to control or ¹⁷⁷Lu-RGD monomer group. The modest therapeutic effect observed in this PDX may be due to the following reason: the tumour uptake of ¹⁷⁷Lu-EB-RGD can be divided into three portions, the expression of integrin α_vβ₃ on tumour cells and on neovasculature, as well as the reinforced vascular permeability and retention (EPR) effect. Although PDX_αvβ3- models have low expression of integrin α_vβ₃ within the tumour cells, they exhibit high levels of murine vascular integrin α_vβ₃ expression (as confirmed by CD61 IHC staining). From the biodistribution study, ¹⁷⁷Lu-EB-RGD uptake in PDX_αvβ3- was about half that of PDX_αvβ3+ tumour uptake (13.38 %ID/g vs. 6.26 %ID/g), which means 50% of tumour uptake may have been due to uptake by the neovasculature and EPR effect. The remaining 50% of ¹⁷⁷Lu-EB-RGD uptake in tumour could explain the observed therapeutic effect on PDX tumours with low integrin expression. Furthermore, we could also observe the reduced expression of CD31 in the PDX_αvβ3- tumours, suggesting that the modest therapeutic effect is partly due to the damage to the tumour vasculature.

In future clinical translation, ⁶⁸Ga/¹⁷⁷Lu-RGD theranostics could enable a highly personalised approach using ⁶⁸Ga-RGD PET/CT to non-invasively image and quantify integrin α_vβ₃ expression, and thereby select the patients most likely to benefit from treatment. ⁶⁸Ga-RGD PET/CT can also be used to evaluate the response to therapy after ¹⁷⁷Lu-EB-RGD treatment; further therapy cycles can be cancelled if there is no or low uptake on post-treatment PET imaging (indicative of a complete or partial response). The principles and expertise required in RGD-based NSCLC theranostics broadly parallel those using ⁶⁸Ga/¹⁷⁷Lu-labelled DOTATATE for metastatic NETs^26,27 or PSMA for metastatic CRPC^24,28. However, based on previous trials using peptide receptor radionuclide therapy (PRRT) or peptide radioligand therapy (PRLT), three to four cycles of ¹⁷⁷Lu-DOTATATE/¹⁷⁷Lu-PSMA are usually required for an entire treatment paradigm, and each cycle requires a dose of at least 1.85–7.4 GBq of ¹⁷⁷Lu, resulting in a total dosage of 7.4–29.6 GBq ¹⁷⁷Lu^26,29–31. Such a huge radiation dosage may result in treatment-related toxic effects on normal organs, such as salivary gland dysfunction, loss of kidney function, and haematological toxicity. However, according to our experience using ¹⁷⁷Lu-EB-TATE or ¹⁷⁷Lu-EB-PSMA^32,33, a single dose of 1.85–3.70 GBq ¹⁷⁷Lu provides comparable therapeutic efficacy due to the enhanced tumour accumulation and prolonged retention of such EB derivatives. Unexpectedly, a single imaging dose (0.67–0.8 GBq) of ¹⁷⁷Lu-EB-TATE /¹⁷⁷Lu-EB-PSMA provides some therapeutic efficacy in some patients with NETs or CRPCs. The approximately 6-fold lower dosage of ¹⁷⁷Lu and reduced dosing frequency would greatly decrease the potential side effects and increase patient compliance.

Besides the pharmacokinetics of radiotherapeutic agents, barriers to the success of TRT include the heterogeneous expression of molecular markers on cancer cells and poor vasculature of the tumour, which would prevent delivery of the agents³⁴. In addition, the biological response to TRT of individual cells within a tumour may vary depending on cell-specific radiosensitivity to the radiotherapeutic agent³⁵. Other variables that are important in TRT include the biological effectiveness of the emitted particles, dose rate and the induction of bystander responses³⁶. These properties of TRT may be exploitable in rationally designed combination protocols that aim to complement the strengths of TRT with other therapeutics and take into consideration optimal timing and sequencing. In our latest study, the combination of TRT (¹⁷⁷Lu-EB-RGD) and immune checkpoint inhibitors (anti-PD-L1 antibodies) in MC38 colon cancer xenografts enhanced mice survival significantly as compared with that of TRT or immunotherapy alone¹². This combination may represent a promising approach for metastatic tumours in which TRT can be used. Therefore, a concentrated effort is needed to shift research from preclinical studies to randomised trials of combination treatments if the potential of TRT is to be fully exploited. The integration of TRT into first-line therapy, rather than as a treatment for end-stage disease (which has been the norm to date) is predicted to be beneficial in the future.

In conclusion, we have designed a theranostic agent, based on an RGD peptide, that targets integrin α_vβ₃-expressing NSCLC and demonstrates improved pharmacokinetics. Using the β-emitting radionuclide ¹⁷⁷Lu, we obtain improved contrast for SPECT imaging and effective TRT for NSCLC-PDXs. Furthermore, our PDX models maintain the same immunohistochemical and genetic characteristics of the human primary tumour. These preclinical data suggest that ¹⁷⁷Lu- EB-RGD may be an effective treatment option for NSCLC that should be further evaluated in human trials.