Introduction: Therapy resistance presents a significant challenge in cancer treatment. High Aldehyde dehydrogenase-1A1 (ALDH1A1) activity is closely linked to cancer stemness and drives cancer progression and chemotherapy resistance (Fig.1A) [1,2]. ALDH1A1 protects cancer cells by acting as a detoxifying enzyme to neutralise reactive aldehydes, thereby preventing oxidative stress and DNA damage [3,4]. The non-invasive identification of ALDH1A1 could improve cancer management through the prediction of relapse and associated therapy resistance and can provide a vulnerability for targeted radionuclide therapy (TRT).

Methods: To address this need, a theragnostic radiotracer was created for the non-invasive imaging of ALDH1A1 by PET, [¹⁸F]MILK-10, and TRT, [¹³¹I]MILK-11 (Suppl.Fig.1). MILK-10 and MILK-11 binding affinity and selectivity for ALDH1A1 was assessed using isolated enzymes.  In range of non-small cell lung cancer (NSCLC) cells (Suppl.Fig2A), A549 cells were chosen to provide a readout of ALDH1A1 status both in vitro and in vivo. The RNA levels of all ALDH isoforms were analysed in A549 WT and A549 ALDH1A1 knockout (KO) cells. [¹⁸F]MILK-10 and [¹²⁵I]MILK-11 cell retention was measured (2hrs uptake; 0.2 MBq; 37^oC). A549 WT and A549 ALDH1A1 KO tumours were grown s/c in mice to 100mm³ and imaged by [¹⁸F]MILK-10 PET (3.7 MBq; 2h dynamic) or by [¹³¹I]MILK-11 SPECT (5.6 MBq; 168h). Ex vivo biodistribution of [¹³¹I]MILK-11 was performed and data was used for dosimetry analysis (OLINDA /EXM).

Results: MILK-10 (K_i=13.6 nM) and MILK-11 (K_i=1.5 nM) had a higher affinity for ALDH1A1 than the gold standard inhibitor DEAB (K_i=214.2 nM) (Fig.1B).

ALDH1A1 expression was increased in A549 WT cells where ALDH3A1 was predominantly expressed in A549 ALDH1A1 KO cells (Fig.1C). [¹⁸F]MILK-10 and [¹²⁵I]MILK-11 uptake was 15× and 38× higher in A549 WT than A549 ALDH1A1 KO cells ([¹⁸F]MILK-10: 110.1±12.7% vs 7.6±3.1% radioactivity/mg, respectively; p=0.0002; Fig.1D; [¹²⁵I]MILK-11: 294.8±36.2% vs 7.7±2.1% radioactivity/mg, respectively; p<0.0001; Fig.1E). The uptake of both radiotracers was also higher in HEK293 overexpressing ALDH1A1 and in cisplatin-resistant SKOV3-TRip2 human ovarian cancer cells [5] in comparison to their matched counterpart HEK293 WT and the parental cisplatin-sensitive SKOV3-ip1 cells [5] (Suppl.Fig2B-E).

In mice bearing A549 WT tumours, [¹⁸F]MILK-10 retention was 6.4-fold higher compared to A549 ALDH1A1 KO tumours (9.1±2.3 vs 1.4±0.4 %ID/g, respectively; p<0.0001; n=8), corresponding with increased expression of ALDH1A1. [¹⁸F]MILK-10 uptake in A549 WT tumours was reduced 65% (p<0.0001) with administration of the competitive ALDH1A1 inhibitor NCT-501, confirming specificity of [¹⁸F]MILK-10 (Fig.1F-H). Unbound radioactivity was cleared through the liver and intestines. In mice bearing SKOV3-TRip2 resistant tumours, [¹⁸F]MILK-10 retention at 2h p.i. was 4.7-fold higher compared to SKOV3-ip1 tumours (6.6±1.4 vs 1.4±0.8 %ID/g, respectively; p=0.007; n=4) (Suppl.Fig.3).

[¹³¹I]MILK-11 was retained in A549 WT tumours in vivo for 1 week p.i (24h: 7.4±0.9 and 196h: 1.7±0.4 %ID/g, n=4), while no uptake was found in A549 ALDH1A1 KO tumours (24 hrs: 0.07±0.002 %ID/g, n=4) (Fig.1I). Ex vivo biodistribution of [¹²⁵I]MILK-11 showed a quick clearance, but persistent tumour retention in A549 WT tumours (Fig.1J) with a tumour residence time of 4.7 days (Suppl.Fig.4A). Blocking of [¹²⁵I] uptake by the thyroid was achieved with KI (Suppl.Fig.4B). The human absorbed doses for [¹³¹I]MILK-11 were highest in the intestine (0.73 mGy/MBq) and kidneys (0.0219 mGy/MBq) and the effective whole-body dose was 0.046 mSv/MBq for the human adult female phantom (Fig.1K).

Conclusion: Our study introduces the first successful example of an in vivo ALDH1A1 theragnostic agent enabling precise PET imaging and targeted therapy for ALDH1A1-positive, chemoresistant cancers. [¹⁸F]MILK-10 imaging may enable the identification of patients that are refractory to standard-of-care, allowing a shift to more effective treatments. Meanwhile, [¹³¹I]MILK-11 holds promise as a therapeutic option for ALDH1A1-positive patients.

Image/Figure:
Click to View

Image/Figure Caption:

Figure 1. [¹⁸F]MILK-10 and [¹³¹I]MILK-11 are specific markers of ALDH1A1 expression.

Schematic representation of ALDH1A1 in cancer stemness and resistance to therapy. (B) Western blot and real-time PCR analysis of ALDH enzymes expression in A549 WT and A549 ALDH1A1 KO cells. Actin was used as a loading control. (C) Concentration-dependent inhibition of ALDH1A1 activity by [F]MILK-10, [I]MILK-11 and DEAB. The results were fitted to the equation: Ki=IC501+[I]Km (solid line). (D) [¹⁸F]MILK-10 and (E) [¹²⁵I]MILK-11 retention in A549 WT and A549 ALDH1A1 KO cells. (F) Representative MIP of PET/CT images showing 105-120 min summed activity of [¹⁸F]MILK-10 in mice bearing subcutaneous A549 WT and A549 ALDH1A1 KO tumours. Tumour borders represented with dotted white lines. NCT-501 (competitive ALDH1A1 inhibitor) was administered i.p. 5 min before [¹⁸F]MILK-10. (G) Time versus radioactivity curves of A549 WT and A549 ALDH1A1 KO tumours normalised to the percentage injected activity. (H) The area under the time-activity curve (AUC) for the blood, bladder, liver and muscle of mice 2 hours after administration of [¹⁸F]MILK-10 in the presence or absence of NCT-501. (I) Representative axial SPECT/CT images in mice bearing subcutaneous A549 WT and A549 ALDH1A1 KO tumours (borders in dotted white lines) 24 and 168 hours post [¹³¹I]MILK-11 administration. (J) Ex vivo biodistribution for [¹²⁵I]MILK-11 in mice bearing subcut, A549 WT tumours 1, 6, 24, 48 and 168 hours post-injection. (K) Mice-to-human extrapolated dosimetry for [¹³¹I]MILK-11. Data are representative or the mean ± SD of n=4.

Author