Pre-clinical evaluation of biomarkers for the early detection of nephrotoxicity following alpha-particle radioligand therapy

Radioligand therapies (RLTs) have emerged as a form of therapy in oncology that has the potential to be transformative for cancer patients. These radiopharmaceuticals consist of a radionuclide coupled to a small peptide ligand that binds to cell membrane targets with high affinity and specificity, making these agents attractive as cancer therapeutics (Fig. 1a). Unlike external beam radiation therapy (EBRT), these targeted forms of systemic radiotherapy enable radionuclides to be delivered directly to cancer cells. Examples of FDA-approved radiopharmaceuticals that are significantly improving outcomes for cancer patients include beta (β)-particle-emitting RLT LUTATHERA® (i.e., [¹⁷⁷Lu]Lu-DOTATATE) for advanced neuroendocrine tumors (NETs) and PLUVICTO® (i.e., [¹⁷⁷Lu]Lu-PSMA-617) for metastatic castrate-resistant prostate cancer (mCRPC) targeting somatostatin receptor type 2 (SSTR2) and prostate-specific membrane antigen (PSMA), respectively. RLTs are also effective in other SSTR-expressing tumors, such as paraganglioma and pheochromocytoma, thyroid carcinoma, and meningioma [1]. The use of other ligands for overexpressed receptor targets has been demonstrated in breast, colon, myeloma, and ovarian cancer [1]; melanoma [2]; and hematological malignancies [3]. There is a significant difference in the observed efficacy of RLTs that is dependent on the radionuclide utilized. Compared to FDA-approved β-emitter drugs, recent studies using alpha (α)-emitter RLTs (e.g., ²¹²Pb, ²²⁵Ac, ²¹¹At) in patients with NET and mCRPC demonstrate greater tumor cell cytotoxicity, including patients with refractory disease following β-RLT [4]. Fundamentally, the advantages of α-emitters lie in the higher linear-energy transfer (LET) (~ 100 keV/µm), which induces double-strand DNA breaks (DSB) and a concomitant increase in ionizations (primary and secondary) along their short path length in the cellular tissue microenvironment [5]. The deposition of energy over extremely short path length results in a significantly increased incidence of lethal DSB, high relative biological effectiveness (RBE), improved tumor-cell-cytotoxicity, and activation of anti-tumor immunity [6]. Among commercially available α-emitters, [²¹²Pb] (half-life T_1/2 10.6 h) is produced using a [²²⁴Ra]/[²¹²Pb] generator system that enables nimble on-demand production of [²¹²Pb]Pb-based radiopharmaceuticals [7]. In the case of [²¹²Pb], the isotope [²⁰³Pb] (T_1/2 51.9 h) can be used as a single-photon emission computerized tomography (SPECT) imaging diagnostic companion to provide detailed dosimetry that guides [²¹²Pb] therapy dosing [8]. The [²⁰³Pb]/[²¹²Pb] radionuclide pair has the advantage that elementally identical imaging surrogate [²⁰³Pb] can guide [²¹²Pb] therapy (Fig. 1b). Here, we establish biomarkers that can be used in the pre-clinical setting to connect surrogate imaging information to the potential for acute kidney injury (AKI) to more effectively use the imaging data to establish appropriate therapeutic dosing strategies. It is envisioned that these data can lead not only to more effective dose planning for clinical translation of new RLTs but for a broader spectrum of pharmaceutical drug candidates for clinical introduction for cancer therapy.

While emerging research demonstrates the potentially transformative impact of α-particle RLT for cancer treatment, kidneys are often treatment-limiting due to their role in the bioelimination of unbound therapeutic agents. While agents can be pharmacologically designed to minimize the reabsorption, biomarkers that provide information on the potential for nephrotoxicity are needed because current evaluation is limited to late post-therapy histological findings and estimation of glomerular filtration rate (eGFR) using the endogenous marker of kidney function serum creatinine (sCr). Fortunately, due to their relatively small size (< 5.5 kDa) and hydrophilic nature, radiolabeled peptides are rapidly filtered through glomeruli. However, a fraction can be reabsorbed in tubular cells, which may lead to injury [9]. Studies in tubule-specific megalin knockout mice demonstrate that proximal tubules are primarily, but not exclusively, responsible for radiopeptide uptake within the kidneys [10]. Because of the short travel distance of emitted particles (low-energy β < 1.7 mm¹; α < 80 µm), the radiation exposure from RLT, especially α-RLT, may be highly localized in the proximal tubular area rather than glomeruli. This suggestion is further supported by prior pre-clinical studies evaluating nephrotoxicity of alpha- and beta-emitter RLT that demonstrated the presence of tubulointerstitial disease as early as 10 weeks, without significant glomerular changes, which progressively worsens and leads to severe chronic kidney disease (CKD) 7 months after therapy [11, 12]. This observation is consistent with the known pathophysiology of AKI-induced CKD, wherein primary tubular cells that fail to repair after severe or recurrent injuries adopt a pro-fibrotic and pro-inflammatory phenotype that drives progressive kidney damage resulting in structural abnormalities including fibrosis, vascular rarefaction, tubular loss, glomerulosclerosis, and a chronic inflammatory infiltrate, with subsequent renal function decline [13]. Thus, primary tubular damage alone could be the etiology of RLT-induced long-term kidney toxicity.

Current nephrotoxicity screening for assessing pharmaceutical-treated patients and mice for establishing potential kidney injury depends mainly on prospective sCr measurement and its routine use to calculate eGFR. Kidney toxicity in patients receiving RLT is expected to be associated with patchy loss of tubule cells, focal areas of proximal tubular dilation, and distal tubular casts, together with measurements of cellular regeneration [14]. Serum creatinine is a delayed and insensitive biomarker of kidney function decline that does not reflect the degree of tubulointerstitial damage. Additionally, long-term toxicity evaluation is limited by the patient population’s poor survival. In comparison, urinary biomarkers such as neutrophil gelatinase-associated lipocalin (NGAL), kidney injury molecule 1 (KIM-1), and epidermal growth factor (EGF) have been used to detect AKI early in patients with normal sCr (subclinical AKI) and to evaluate overall tubular health [15]. The change in their urinary secretion pattern has been used to identify patients at risk of progressive CKD [16]. Predictive biomarkers could accurately identify patients with subclinical kidney injury and those at risk of AKI and CKD following RLT treatment.

Regulatory bodies currently employ 23 Gy in kidneys as the absorbed dose limit, derived from EBRT, despite significant differences in the form and interaction of radioactivity emissions associated with RLT vs. EBRT. For example, the typical dose rate for EBRT is ~ 1 Gy/min, whereas the absorbed dose from RLT is delivered over multiple days, even weeks, depending on the half-life of the radionuclide, resulting in a slower dose rate (mGy/min). Although conventional EBRT dose limits may differ from those encountered with RLT, dosimetry-based therapy remains a potentially powerful tool for personalizing RLT. Specifically, in RLT, a biologically effective dose of 40 Gy to the kidney has been suggested to be safe in patients without risk factors for kidney disease and 28 Gy in those at risk of kidney injury [17]. A more precise understanding of the nature of kidney-delivered doses may help maximize cancer therapy and reduce long-term kidney toxicity according to individual patient tolerability. However, this approach is limited by the need for more information on tubular damage caused by a specific kidney-absorbed dose not provided by sCr measurement. Thus, the purpose of this study was to investigate the urinary secretion pattern of tubular injury biomarkers following injection of a [²¹²Pb]Pb-based RLT in a murine model. In this dose-escalation study, a bifunctional cyclic peptide, melanocortin 1 ligand (MC1L), was labeled with [²⁰³Pb]Pb-MC1L to determine the distribution and absorbed dose of [²¹²Pb]Pb-MC1L at levels intended to induce AKI, followed by a comprehensive nephrotoxicity analysis including novel urine secretion pattern of kidney injury biomarkers in CD-1 Elite mice. To explore potential absorbed dose tolerated in humans, we performed human dosimetry scaling up from mice. It is envisioned that combining biomarkers and dosimetry can lead to a more detailed understanding of the use of [²⁰³Pb]-based imaging as a quantitative guide for [²¹²Pb] therapy planning and a personalized approach to cancer therapy.

In this study, we report for the first time the use of urinary biomarkers of tubulointerstitial disease to identify subclinical AKI and predict CKD associated with α-emitter RLT in a murine model. Using [²⁰³Pb]Pb-MC1L to characterize the distribution and absorbed dose of [²¹²Pb]Pb-MC1L, kidneys appear to have the highest uptake. The dose depositions in kidneys from injection of 0.9, 3.0, and 6.7 MBq [²¹²Pb]Pb-MC1L were, on average, 2.8, 8.9, and 20 Gy, respectively. The injected activities were selected to cover low, medium, and high kidney doses intended to induce AKI. A dose-escalation toxicity study in mice was well tolerated based on body weight measurement with lack of significant hematotoxicity and liver toxicity. However, we observed the development of late CKD in mice treated with higher [²¹²Pb]Pb-MC1L doses based on the profile change of kidney function biomarkers and the kidney histological alterations found 28 weeks after treatment. There was a significant dose-dependent urinary excretion of the tubular injury biomarker uNGAL the day after [²¹²Pb]Pb-MC1L treatment that highly correlated with the severity of tubulointerstitial injury and late kidney function biomarkers. Despite the lack of early changes in conventional endogenous kidney function biomarkers (BUN, sCr, and serum cystatin C), the early concurrent increase in tubular injury suggests that early subclinical tubular injury could explain the long-term renal dysfunction observed at 28 weeks.

Dosimetry is a strategy that facilitates the determination of doses administered to normal organs, including those more sensitive to RLT toxicity. In the present study, we administered [²¹²Pb]Pb-labeled peptide [²¹²Pb]Pb-MC1L via i.v. injection to deliver α-radiation. MC1L is a synthetic peptide of 7 amino acids cyclized via Cu-mediated click chemistry. As an analog of alpha-melanocortin stimulating hormone (α-MSH), MC1L binds with melanocortin receptor 1 (MC1R) overexpressed in metastatic melanoma. Within this paradigm, the 10.6-h half-life of [²¹²Pb] decay matches the relatively shorter biological half-life of small peptides compared with the long biological half-lives of antibodies [24]. Cyclotron-produced gamma-ray-emitting radionuclide [²⁰³Pb] was used as an elementally identical imaging surrogate for [²¹²Pb] [7]. [²⁰³Pb]Pb-MC1L surrogate was injected in the same animal model to determine the in vivo biodistribution for calculating the dosimetry of [²⁰³Pb]Pb-MC1L in normal organs. Our prior studies have demonstrated excellent chelation kinetics and stability with PSC for [²¹²Pb] and [²¹²Bi] [8]. However, previous literature has shown that during a fraction of [²¹²Pb] disintegrations, the daughter [²¹²Bi] cations are released from the chelator [25]. Based on the known kinetics of bismuth, it is likely that some released ions will accumulate and decay in the liver and kidneys. Therefore, doses to these structures may be underestimated when utilizing [²⁰³Pb] as a dosimetric surrogate without accounting for these effects.

Due to its rapid clearance, we observed low [²⁰³Pb]Pb-MC1L liver and blood uptake and retention. We did not observe significant liver toxicity after [²¹²Pb]Pb-MC1L treatment like in previous reports using [¹⁷⁷Lu]Lu-DOTATATE [26]. Some cases of hematological toxicity (grade 3/4) and dose-dependent acute hematological toxicity have been reported in patients receiving [¹⁷⁷Lu]Lu-DOTATATE [27] and [²²⁵Ac]Ac-DOTATOC [28] therapy. Hematological toxicity after RLTs increases in patients with baseline nephrotoxicity (transient or persistent elevation in sCr), probably due to the prolonged circulation time of the radiopeptide in patients with poor kidney function, which increases bone marrow exposure to radiation [29]. Our study showed a late reduction in hemoglobin and RBC 28 weeks after [²¹²Pb]Pb-MC1L treatment in mice with evidence of kidney dysfunction based on BUN and sCr measurements. Because anemia is common in patients with CKD due to reduced kidney erythropoietin production, our findings suggest the late reduction in hemoglobin and RBC after [²¹²Pb]Pb-MC1L treatment was secondary to anemia of CKD.

The understanding of RLT-associated nephrotoxicity in patients is limited and primarily evaluated by measuring endogenous kidney functional biomarkers, like sCr, that do not assess the severity or localization of tubulointerstitial damage. During steady-state conditions, sCr is commonly used to estimate GFR. More than half of the kidney mass must be impaired to detect an increase in sCr as a kidney dysfunction biomarker [30], explaining the late increase in functional biomarkers after [²¹²Pb]Pb-MC1L treatment as evidence of the histological CKD progression. Moreover, sCr levels are affected by tubular secretion and reabsorption, kidney generation, extrarenal elimination, certain medications, and patient gender, age, diet, and muscle mass, which results in an overestimation of GFR and does not reflect accurately intrinsic kidney injury [31]. The reports of kidney toxicity in α-emitters are limited, and sCr restrictions may explain the controversy in nephrotoxicity development. In a study of 39 patients receiving [²²⁵Ac]Ac-DOTATOC α-RLT, only 22 were suitable to evaluate long-term kidney function due to tumor progression or change in therapy. Over a median of 57 months (range 18–90 months), they observed an average eGFR-loss of 8.4 mL/min (9.9%) per year, and treatment-related kidney failure occurred in two patients after a delay of > 4 years [32]. The lack of accuracy in nephrotoxicity detection, mainly by sCr-based eGFR, led to the conclusion that kidney failure is independent of administered radioactivity and other clinical risk factors are significant or the main contributors [32‐34]. Following findings in humans, we did not observe any early changes in the conventional kidney function biomarkers BUN and creatinine. We only found late significant changes at 28 weeks in mice treated with [²¹²Pb]Pb-MC1L at 3.0 and 6.7 MBq, consistent with a kidney-absorbed dose of ~ 8.9 and ~ 20 Gy, respectively. Even though no RBE scaling was used in this work to avoid ambiguity, and all doses are presented as physical absorbed dose to the structure of interest, these results are in accordance with pre-clinical studies using [²¹³Bi]Bi-DOTATATE (RBE of 4–5 for acute renal toxicity), where the calculated renal LD₅₀, at which 50% of the treated animals would develop acute nephrotoxicity, was 20 ± 8 Gy [35]. Kidney functional biomarkers (BUN and sCr) do not directly detect injury in the tubular epithelium, the leading site of injury in most forms of AKI, that, without treatment, can progress to CKD [13]. In line with radiopeptide reabsorption by proximal tubular cells, pre-clinical data in mice and rat models have shown evidence of long-term tubulointerstitial disease after β-RLTs and α-RLT in the absence of severe glomerular pathology or sCr changes [11, 12]. In this context, the delayed change in kidney function biomarkers and the lack of information on tubular injury result in poor accuracy in diagnosing acute structural kidney damage associated with RLT treatment and could result in missing valuable treatment windows to prevent CKD progression.

Biomarkers of tubular injury and health [36] are more frequently used to diagnose AKI before the loss of filtration function (kidney function biomarkers changes), supporting their potential incorporation into nephrotoxic evaluation in RLT patients. Urinary biomarkers of tubular health (EGF) and injury (KIM-1 and NGAL) measurements in mice treated with [²¹²Pb]Pb-MC1L showed an early increase in NGAL and EGF excretion. NGAL is a protein reabsorbed by proximal tubules and released by the damaged distal tubules in response to acute tubular injury, which can be detected in the urine within hours of tubular injury [37]. Thus, an increase in urine NGAL reflects both distal tubular injury that causes an increase in NGAL production to the urine and systemic circulation and proximal tubular damage that results in a lack of NGAL reabsorption by proximal tubular cells [38]. We demonstrate that uNGAL excretion was significantly increased after [²¹²Pb]Pb-MC1L treatment with a dose-dependent pattern on day 1 after treatment. Moreover, its increased urinary excretion was highly correlated with tissue damage at 28 weeks, similar to the long-term serum NGAL increase observed in mice receiving the α-emitter [²¹³Bi]Bi-DOTATATE [35]. The early elevation in uNGAL indicates a dose-dependent [²¹²Pb]Pb-MC1L-induced tubular injury that correlates with long-term kidney toxicity. The observed early increase in tubular injury biomarkers with a lack of changes in kidney function biomarkers is used to diagnose a subclinical AKI condition [39]. This condition is usually asymptomatic and has been associated with a two- to threefold increased risk of death or the need for kidney support therapy compared to patients with normal levels of sCr and tubular damage biomarkers [40]. EGF expression is specific for the ascending limb of Henle and distal tubules and has been associated with enhanced tubular cell regeneration and repair, accelerating kidney recovery after injury [23, 41]. uEGF is decreased in various kidney diseases in humans, including patients with CKD [42]. We found a significant initial increase in uEGF excretion in the 3 MBq and 6.7 MBq that was then significantly reduced at 28 weeks after treatment, suggesting an adequate initial response to injury with a lower late tubular reserve 28 weeks after treatment compared to the control group. Finally, KIM-1 is a protein expressed in proximal tubular epithelial cells rarely expressed in organs other than kidneys, whose increased urine level is a sensitive indicator for early proximal tubular injury, particularly for ischemic or nephrotoxic AKI patients [43]. We found a trend toward an increase in uKIM-1 after [²¹²Pb]Pb-MC1L treatment that persisted and was significant in the mice treated with 6.7 MBq after 28 weeks. The urinary biomarker patterns suggest that proximal tubule cells might not be the only tubular segment affected by [²¹²Pb]Pb-MC1L treatment.

We recognize some limitations in our study design. Studies evaluating long-term radiation-induced toxicity can be extended up to 36 weeks post-therapy [44], and our study concluded at 28 weeks (~ 7 months). However, the average mouse lifespan is about 24 months [45], and a follow-up of 7 months post-treatment represents approximately one-third of their lifespan. In murine CKD models, where it is expected to have a gradual renal function decline, mice are followed between 1 and 6 months post-injury to evaluate chronic changes in tissue, including fibrosis [46]. Because mice treated with 6.7 MBq had significant changes in kidney function biomarkers suggesting clinically meaningful renal function decline, we concluded our study at 28 weeks to ensure adequate survival in each group to compare histological findings. We found significant fibrosis in that group (score 3 of 4), supporting adequate time to identify fibrosis in this model.

The results from this dose-escalation pre-clinical toxicity study propose that increased uNGAL secretion could be an additional diagnostic tool to identify patients developing α-RLT subclinical AKI and at risk of CKD. The changes in uEGF suggest that proximal tubule cells might not be the only tubular segment affected by α-RLT and highlight how potential early tubular damage could be an essential driver of long-term α-RLT-related nephrotoxicity. Together with the conventional clinical predictors of kidney health (sCr, cystatin C, UPC), urine biomarkers of tubular injury (i.e., NGAL) and tubular health (i.e., EGF) in combination with dosimetry could be used to personalize α-RLT and provide higher doses to those patients without evidence of tubulointerstitial injury. On the contrary, in patients where therapy is still required and highly beneficial, it would empower practitioners to discuss the long-term risk with patients and ensure early and close CKD follow-up to improve their outcomes, as previously proposed. More studies using dosimetry and evaluating biomarkers of tubulointerstitial disease in RLTs are needed to identify the timing of tubular injury at specific renal dosages and guide future toxicity mechanistic studies. This approach could facilitate tumor treatment maximization, development of nephroprotective therapies, and identification of patients at risk of long-term toxicity.