C-X-C Motif Chemokine Receptor 4–Targeted Radioligand Therapy in Hematological Malignancies—Myeloablative Effects, Antilymphoma Activity, and Safety Profile

Abstract

Background

After C-X-C motif chemokine receptor 4 (CXCR4)–directed radioligand therapy (RLT), lymphoma patients are scheduled for conditioning therapy (CON) followed by hematopoietic stem cell transplantation (HSCT). We aimed to determine whether CXCR4-RLT can achieve bone marrow ablation and direct antilymphoma activity independent from CON/HSCT and also evaluated the safety profile of this theranostic approach in an acute setting.

Patients and Methods

After CXCR4-directed ⁶⁸Ga-pentixafor PET/CT, 21 heavily pretreated patients with hematological malignancies underwent CXCR4-directed RLT using ⁹⁰Y-pentixather. The extent of myeloablative efficacy was determined by investigating hematologic laboratory parameters before RLT (day −1), at the day of RLT (day 0), 2 days after RLT (day 2), and before CON (median day 10). Serving as surrogate marker of antilymphoma activity, lactate dehydrogenase (LDH) levels were also assessed until CON. We also screened for laboratory-defined tumor lysis syndrome after the Cairo-Bishop definition and recorded acute laboratory adverse events using the Common Terminology Criteria for Adverse Events version 5.0.

Results

After RLT, we observed a significant decline of leukocyte levels by 79.4% ± 18.7% till CON (granulocytes, drop by 70.3% ± 21%; platelets, reduction by 43.1% ± 36%; P ≤ 0.0005 vs day 0, respectively). After RLT, LDH levels already reached a peak at day 2, which was followed by a rapid decline thereafter (peak vs day of CON, P = 0.0006), indicating that ⁹⁰Y-pentixather exhibits direct antilymphoma activity. At day of CON, LDH levels were also significantly lower when compared with day −1 (P = 0.04), suggestive for durable response mediated by RLT. No patient fulfilled the criteria of tumor lysis syndrome, whereas 25 laboratory adverse events attributable to CXCR4-directed treatment were identified (≥grade 3 in 2/25 [8%]). During further treatment course, all patients (100%) received HSCT.

Conclusions

CXCR4-directed RLT causes effective myeloablation, which allows for HSCT. In addition, it also exerts direct antilymphoma activity independent of subsequent therapeutic steps, whereas safety profile was acceptable.

The G-protein–coupled receptor C-X-C motif chemokine receptor 4 (CXCR4) and its ligand stromal derived factor 1 (SDF-1, CXCL12) are known to mediate retention of hematopoietic stem progenitor cells in stem cell niches of bone marrow (BM).¹ Moreover, hematological malignancies interact with their microenvironment via multiple chemokine receptors,² thereby rendering CXCR4 as an attractive target to achieve antilymphoma and myeloablative activity.

For this purpose, a novel CXCR4-targeted theranostic approach has been introduced.³ First, the retention capacities of the target are evaluated by injecting the PET agent ⁶⁸Ga-pentixafor, which allows to determine CXCR4 expression in sites of disease and BM.^4–6 In vivo assessments using CXCR4-directed molecular imaging revealed that the target is mainly upregulated in hematological neoplasms,⁴ including acute myeloid leukemia (AML), B-cell non-Hodgkin lymphoma (B-NHL), T-cell (TCL), and mantle cell lymphoma (MCL). These favorable imaging results fostered the concept to administer the therapeutic equivalent ⁹⁰Y-pentixather, as this compound delivers destructive β-emission to lymphoma manifestations, along with desired BM eradication.⁷ Radioligand therapy (RLT) is then followed by subsequent conditioning therapy (CON) and hematopoietic stem cell transplantation (HSCT) allowing for engraftment (Fig. 1). Long-term outcome benefits of this concept have already been reported, including complete remission in AML or TCL.^7,8 However, knowledge on myeloablative effects and antilymphoma activity directly mediated by RLT is rather limited, in particular as CON is also regularly applied after RLT.³ In addition, previous reports also raised concern on adverse events (AEs), including tumor lysis syndrome (TLS).³

FIGURE 1.

CXCR4-targeted theranostics protocol for hematological malignancies. CXCR4-directed PET/CT allows for identifying patients eligible for CXCR4-targeted RLT using ⁹⁰Y-pentixather, which exerts antilymphoma effect through β-radiation–induced double-strand breaks of the DNA. After desired aplasia due to CXCR4 RLT-mediated destruction of the stem cell niche, conditioning regimen (CON) allows for further aplasia, followed by HSCT and engraftment. Created with biorender.com.

As such, we aimed to evaluate the extent of BM eradication caused by RLT. We also determined direct antilymphoma activity attributable to ⁹⁰Y-pentixather by investigating the time course of lactate dehydrogenase (LDH), which served as surrogate marker of lymphoma cell kill. To determine the safety profile of CXCR4-RLT, we also investigated laboratory-based AEs, including occurrence of posttreatment TLS. Last, we also aimed to evaluate an association between baseline ⁶⁸Ga-pentixafor PET signal and cytotoxic effects.

PATIENTS AND METHODS

We retrospectively analyzed 21 patients with refractory and/or relapsed hematological malignancies treated with ⁹⁰Y-pentixather. All patients were heavily pretreated and had already exhausted previous standard therapies. CXCR4-directed RLT was conducted in accordance to the German Pharmaceutical Act (§13.2b). All subjects gave written informed consent for diagnostic and therapeutic procedures. As this was a retrospective investigation, the local ethics committee waived the need for approval (# 20220103 01). Parts of this cohort have been analyzed in previous studies,^7–12 without investigating myeloablative effects over time, direct antilymphoma efficacy right after radiotracer administration, occurrence of TLS in an acute setting, or associations of the ⁶⁸Ga-pentixafor PET signal for outcome. Table 1 provides an overview of included patients.

TABLE 1.

Patient Characteristics

Sex (n)
Male	12 (57.1%)
Female	9 (42.9%)
Age (y)
Median	52
Range	22–67
Diagnosis (n)
B-NHL	12/21 (57.1%)
Diffuse large B-cell lymphoma	8/12 (66.7%)
Follicular lymphoma	2/12 (16.7%)
MCL	1/12 (8.3%)
Burkitt lymphoma	1/12 (8.3%)
Acute myeloid leukemia	5/21 (23.8%)
T-cell lymphoma	4/21 (19%)
Previous treatment lines (n)
Median	5
Range	2–9
HSCT (n)
Allogeneic	17/21 (81%)
Autologous	4/21 (19%)

Pretherapeutic CXCR4-Directed ⁶⁸Ga-Pentixafor PET/CT

For all subjects, pretherapeutic ⁶⁸Ga-pentixafor PET/CT was performed. The PET radiotracer was synthesized in-house as described previously.^4,13 Sixty minutes after IV administration of median 130 MBq ⁶⁸Ga-pentixafor, PET/CT was carried out using a Siemens Biograph mCT 64 or 128, as described previously.⁴

Image Analysis

To determine associations between ⁶⁸Ga-pentixafor PET and RLT-induced antilymphoma activity, pretherapeutic PET/CTs were analyzed by a single reader (A.-L.D. or N.D.) and verified by experienced readers in inconclusive cases (S.E.S. or R.A.W.). A target lesion analysis was carried out by examining the visually most intense lesion on pretherapeutic imaging.⁴ SUV_peak was then obtained by applying 3-dimensional volumes of interest with an isocontour threshold of 40%,⁴ which allowed to correlate the PET signal with early antilymphoma effects (ie, LDH fluctuations after RLT).

CXCR4-Directed RLT

Before RLT, all patients underwent dosimetry using ¹⁷⁷Lu-pentixather, which was used to determine administered activity.¹⁰ Based on these calculations, we administered median 5.1 GBq of ⁹⁰Y-pentixather (range, 2.7–8.5 GBq). In addition, patients received a nephroprotective solution combining each 25 g/L arginine and lysine (2 L in total) according to a practical guidance for receptor-targeted radionuclide therapies.¹⁴ After respective guidelines for the management of TLS,¹⁵ we also applied a TLS protection protocol before treatment onset. Six patients (28.6%) were also treated with ¹⁸⁸Re-CD66, 4 patients (19%) with ⁹⁰Y–ibritumomab tiuxetan, and 1 patient (4.8%) with ¹⁵³Sm–ethylenediamine tetramethylene phosphonic acid.

Assessment of Myeloablative Effects

The extent of myeloablative efficacy of CXCR4-directed RLT was determined by examining available hematologic laboratory parameters as follows: 1 day before treatment or at the day of RLT but before tracer injection (then referred to as day 0), 2 days after RLT (day 2), and before day of CON. Nineteen of 21 (90.5%) patients received CON, with a median of 10 days (range, 4–23 days) after RLT. In 2/21 (9.5%), no additional CON was applied, and thus, following the observed median of patients with CON, we recorded hematologic parameters accordingly at day 10 or 11. We considered both the percentage changes relative to values before RLT and absolute values at each time point. Blood samples for the determination of hemoglobin (in g/dL), leukocyte (in ×1000/μL), platelet (in ×1000/μL), and granulocyte (in ×1000/μL) levels were collected as part of clinical routine care using di-potassium-ethylenediaminetetraacetic acid tubes (Sarstedt, Nuembrecht, Germany). The samples were then analyzed using a fully automated modular analyzer (Sysmex XN-9000, Kobe, Japan), as described Hartrampf et al.¹⁶

Assessing Direct Antilymphoma Activity Until CON

To assess antilymphoma activity directly attributable to CXCR4-directed RLT, we analyzed repeated measurements of LDH levels (in U/I) starting 1 day before therapy (day −1) until the day of CON, if available. For those 2 patients not receiving CON, we again referred to available laboratory parameters at day 10 or 11. Blood samples were collected using serum-gel tubes (Sarstedt, Nuembrecht, Germany) and analyzed using a fully automated analyzer (Roche Cobas, Basel, Switzerland).⁷

Evaluating Laboratory-Based TLS and AEs Until CON

To detect relevant off-target effects, we applied the Common Terminology Criteria for Adverse Events version 5.0 (CTCAE v5.0)¹⁷ and determined potassium (in mmol/L), phosphate (in mmol/L), calcium (in mmol/L), uric acid (in mg/dL), renal (creatinine [in mg/dL], estimated glomerular filtration rates [GFR; in mL/min/1.73m²]), and liver functional parameters (glutamic oxaloacetic transaminase [GOT], glutamic-pyruvic transaminase [GPT], or γ-glutamyl transferase [GGT]; each in U/I), if available. Estimated GFR was estimated based on serum creatinine levels using the “Modification of Diet in Renal Disease” study equation.¹⁸

The Cairo-Bishop definition was used to screen for laboratory-based TLS, which is characterized by at least 2 of the following laboratory changes: 25% increase from baseline for either uric acid, potassium, or phosphate, or 25% decrease from baseline for calcium levels.¹⁹ Blood collection for the determination of TLS-associated and functional parameters was performed using serum-gel tubes (Sarstedt, Nuembrecht, Germany), again applying a Roche Cobas analyzer (Basel, Switzerland).⁷

Statistical Analysis

Statistical analyses were performed using GraphPad Prism version 9.5.0 (GraphPad Software, San Diego, CA). Descriptive results are displayed as mean ± standard deviation or median (including range or 95% confidence interval). Myeloablative effects and TLS-associated parameters were corrected for outliers using the ROUT-Method. As not all parameters were normally distributed, we performed the Wilcoxon matched-pairs signed rank rest to compare between baseline and follow-up values. To investigate a possible correlation between PET parameters and change of LDH, Spearman rank correlation coefficient (ρ) was calculated.¹² P values less than 0.05 were considered statistically significant.

RESULTS

Patient Characteristics

Four of 21 (19%) presented with TCL, 5/21 (23.8%) with AML. The remaining 12/21 (57.1%) were diagnosed with B-NHL, including diffuse large B-cell lymphoma in 8/12 (66.7%), follicular lymphoma in 2/12 (16.7%), and MCL and Burkitt lymphoma in 1/12 (8.3%), respectively. The median age was 52 years (range, 22–67 years; female, 9/21 [42.9%]). All patients were extensively pretreated with a median of 5 anteceding therapeutic drug regimens (range, 2–9), without further therapeutic options at time of RLT. After CXCR4-directed treatment and CON, HSCT was then performed in all patients with an autologous transplant in 4/21 (19%) or allogeneic HSCT in 17/21 (81%), respectively.

⁹⁰Y-Pentixather Can Cause Direct Myeloablative Effects Early After Onset of RLT

⁹⁰Y-Pentixather led to a significant decline of leukocyte levels by 24.5% ± 46.8% from day 0 to day 2 (P = 0.003) and by 79.4% ± 18.7% from day 0 until CON (P = 0.0001). Granulocyte levels were also reduced by 16.9% ± 38.4% until day 2 (P = 0.08 vs day 0), which then further dropped by 70.3% ± 21% until CON (P = 0.0005 vs day 0). Platelets also declined by 19.2% ± 22.3% till day 2 (P = 0.002), which was again more pronounced at time point of CON (43.1% ± 36%, P = 0.0005 vs day 0). Hemoglobin levels were also lowered by 8.7% ± 11.4% till CON (P = 0.005 vs day 0; Fig. 2, Table 2).

FIGURE 2.

Myeloablative effects after CXCR4-directed RLT. Mean % changes from day 0 (time point of RLT) to day 2 and from day 0 to time point of CON. Conditioning therapy was performed 10 days (median) post-RLT, whereas in patients not receiving CON (n = 2), respective values at day 10 or 11 were recorded. All investigated blood values demonstrated a significant decline at time of CON, which was most pronounced for leukocytes. *P < 0.01, **P < 0.001, ***P < 0.0001.

TABLE 2.

Changes of Laboratory Parameters Relevant for Myeloablative Effects, TLS, and Off-Target Effects

Parameter	Day 0 to Day 2		Day 0 to CON
	% Change	P	% Change	P
Myeloablative effects
Leukocytes	−24.5	0.003	−79.4	0.0001
Hemoglobin	−9	<0.0001	−8.7	0.005
Platelets	−19.2	0.002	−43.1	0.0005
Granulocytes	−16.9	0.08	−70.3	0.0005
TLS
Uric acid	−3.4	0.70	−25.1	0.049
Potassium	−2.3	0.33	−0.8	0.66
Phosphate	12.5	0.2	−1.2	0.75
Calcium	−2.6	0.16	−2.3	0.02
Off-target effects
Creatinine	−2.9	0.31	−15.1	0.001
GFR	2.7	0.45	18.9	0.001
GGT	−2.8	0.32	−6.2	0.18
GOT	1.8	0.87	−25.3	0.003
GPT	−6.1	0.18	−29.5	0.0007

Mean percentage changes of laboratory parameters between the day of RLT (day 0), day 2 after RLT, and day of CON. Except for granulocytes, myeloablative effects with significant drop occurred already 2 days after RLT. None of the patients experienced laboratory-defined TLS according to Cairo and Bishop.¹⁹ Changes in renal and functional parameters were also not indicative for kidney or liver failure. Significant P values are marked in bold.

⁹⁰Y-Pentixather Can Exhibit Direct Antilymphoma Activity Before CON

The median LDH levels the day before therapy (day −1) were 497 (177–3566), followed by a peak at day 2 (633; range, 171–9460; P > 0.99 vs day −1). Subsequently, we observed a significant decrease to 316 (117–2441) until day of CON (P = 0.0006 vs day 2), suggestive for direct antilymphoma activity caused by ⁹⁰Y-pentixather before CON as the next therapeutic step. At day of CON, LDH levels were also significantly reduced when compared with day −1, suggestive for durable response (P = 0.04; Fig. 3).

FIGURE 3.

Time-course of LDH (in U/L) as an indicator of early response to CXCR4-directed RLT. Dotted lines indicate RLT and start of CON, occurring on 10 days (median) after RLT. LDH demonstrated a peak early after radiotracer injection, followed by a rapid decline thereafter. LDH levels at CON were significantly lower when compared with baseline and peak levels. The median and 95% confidence interval are displayed. *P < 0.05, **P < 0.001.

Pretherapeutic ⁶⁸Ga-Pentixafor PET Is Associated With Peak LDH Levels After RLT

On pretherapeutic ⁶⁸Ga-pentixafor PET, we observed a median SUV_peak of 7.98 (range, 3.09–25.08) in target lesion. Correlating this baseline uptake with LDH levels, SUV_peak was significantly associated with LDH levels at day 2 post-RLT (ρ = 0.47, P = 0.03), indicative for predictive information obtained from CXCR4-directed PET.

Safety Profile of RLT Is Acceptable

None of the patients fulfilled the criteria of laboratory-based TLS. Of note, uric acid even significantly decreased by 25.1% ± 16.6% from baseline until CON (P < 0.05), indicating effectiveness of the applied TLS protection protocol (Table 2, Fig. 4A). Independent of an early laboratory-based TLS, however, one patient affected with TCL and high tumor burden developed a clinical TLS resulting in grade 3 kidney failure early after RLT.

FIGURE 4.

Mean % changes to baseline after RLT for (A) blood-based biomarkers relevant for TLS and (B) organ function parameters. None of the patients fulfilled the Cairo-Bishop criteria for laboratory-based TLS.¹⁹ Hepatic and renal functional parameters did not indicate renal or liver failure (both organs with physiological excretion of ⁹⁰Y-pentixather). *P < 0.05, **P < 0.01, ***P ≤ 0.001.

Based on CTCAE v5.0 assessment, we identified 25 laboratory AEs. Twenty-three of 25 (92%) were classified as grade 1 or 2 (most often, hyperphosphatemia in 5/23 [21.7%], followed by hypophosphatemia and hypokalemia in 4/23 [17.4%] each, and hypocalcemia in 3/23 [13%]). Moreover, in 2/25 (8%), grade ≥3 events were recorded (GFR decrease and GOT elevation in 1/2 [50%], respectively; Table 3). There was no increase in creatinine, decrease in GFR, or increase in GGT indicative for renal or liver failure (Fig. 4B).

TABLE 3.

Laboratory-Defined AEs Occurring Between Baseline (Before Radioligand Therapy) Until Start of Conditioning Therapy (Based on CTCAE v5.0)

	Grade 1/2	Grade 3/4	Grade 5
All events	23/25 (92%)	2/25 (8%)	0
Hyperphosphatemia	5/23 (21.7%)	0	0
Hypophosphatemia	4/23 (17.4%)	0	0
Hypokalemia	4/23 (17.4%)	0	0
Hypocalcemia	3/23 (13%)	0	0
Hypercalcemia	1/23 (4.3%)	0	0
Hyperkalemia	0	0	0
Creatinine	1/23 (4.3%)	0	0
GFR	0/23 (0%)	1/2 (50%)*	0
GOT	2/23 (8.7%)	1/2 (50%)†	0
GPT	1/23 (4.3%)	0	0
GGT	1/23 (4.3%)	0	0
Hyperuricemia	1/23 (4.3%)	0	0

There were only 2/25 (8%) ≥grade 3 and no grade 5 AEs.

*GFR decrease.

†GOT elevation.

DISCUSSION

Investigating patients affected with refractory/relapsed hematological malignancies and treated with CXCR4-targeted ⁹⁰Y-pentixather, we observed relevant decline of leukocyte and granulocyte levels, along with successful HSCT in all subjects. Serving as a surrogate marker of lymphoma cell kill, LDH levels demonstrated a peak 2 days post-RLT, with decrease thereafter, indicating that CXCR4-RLT may exhibit direct antilymphoma activity. The observed rapid LDH peak early after RLT, however, did not cause laboratory-defined TLS. The safety profile of CXCR4-directed treatment was further corroborated by assessment of other laboratory-based AE, with grade ≥3 in 2/25 (8%) of instances. Last, we observed an association between pretherapeutic PET signal in lymphoma manifestations and LDH levels post-RLT, which may provide a rationale for future studies to determine whether PET quantification may allow to identify candidates that benefit from treatment.

The observed BM ablation with a drop of leukocyte and granulocyte levels by at least 70% is substantial to prepare patients for subsequent transplant. As an alternative, total body irradiation (TBI) has been extensively investigated since the early 1980s due to its immunosuppressive and cytotoxic properties. Relative to the systemic administration of ⁹⁰Y-pentixather, this approach acts independent from vascular supply. Thus, even tissues that are normally not reached by the bloodstream, such as the central nervous system, still receive high tumor doses.²⁰ Of note, morbidity and mortality associated with TBI remain high,²⁰ and thus, its use may be restricted to younger and healthier individuals, whereas ⁹⁰Y-pentixather may also be used in the elderly and in heavily pretreated patients as demonstrated in the present investigation (median 5 prior treatment lines). Nonetheless, combination approaches of TBI and CXCR4-directed RLT are feasible and may be promising also in refractory/relapsed AML.²¹

Focusing on an early time window after injection of ⁹⁰Y-pentixather until CON, we observed a substantial increase of LDH with a peak within 48 hours post-RLT. Converting lactate to pyruvate, cancerous tissue is characterized by increased LDH activity, and thus, this biomarker has been commonly assessed in blood cancers to monitor therapeutic efficacy.²² For instance, in patients with TCL neoplasms, elevated LDH levels reflect extensive disease and are also linked to decreased overall survival.²³ Thus, for patients with TCL, regular assessment of LDH after therapy is also recommended by consensus guidelines.²⁴ In the present study, we observed a peak of LDH 2 days after RLT, and thus, CXCR4-targeted treatment may cause antilymphoma activity directly attributable to injection of the radiopharmaceutical. Of note, those LDH fluctuations were also associated with the pretherapeutic ⁶⁸Ga-pentixafor PET signal. As such, further studies should also include other biomarkers such as β2-microglobulin,²⁴ which will then define the predictive, independent role of the pretherapeutic PET signal relative to other clinical parameters for outcome, preferably in a multivariate setting. Nonetheless, the substantial rise of LDH may also be associated with tumor cell lysis, but none of our patients fulfilled the Cairo-Bishop criteria of an acute, laboratory-defined TLS.¹⁹ As such, the applied preventive protocol even led to a significant decrease of uric acid until initiation of CON. Of note, the present investigation focused on an acute damage of proliferating tumor cells shed into the patient’s bloodstream. Independent of such a laboratory-defined assessment, one TCL patient with massive tumor load still developed a clinically relevant TLS post-RLT (long-term safety not shown),⁷ indicating that subjects with bulky disease should also be closely monitored in a longitudinal setting. Nonetheless, no acute laboratory-based TLS occurred, indicative for a high efficacy of the implemented protocol. Thus, given the potential effectiveness of this preventive measure, laboratory-based TLS under CXCR4-targeted RLT may rather be seen as a manageable adverse effect.

Biodistribution of both the diagnostic PET agent ⁶⁸Ga-pentixafor and its therapeutic equivalent ⁹⁰Y-pentixather include renal and hepatic excretion, and thus, those organs may be at increased risk for radiation-induced damage.^11,25,26 Focusing on an acute toxicity until CON, monitoring of renal and hepatic functional parameters did not indicate acute nephrotoxicity or hepatotoxicity (Fig. 4B). In addition, we also observed a substantial rate of CTCAE-based AEs, including hypokalemia, hyperphosphatemia, or hypophosphatemia. Nonetheless, we observed only 2 patients with grade ≥3 events, and thus, the herein presented CXCR4-directed theranostics concept may also have a high safety profile.

The present study has limitations. Its retrospective nature should trigger future studies in a prospective setting and corroborate our initial findings, for example, by results of the planned phase I/II COLPRIT trial investigating the efficacy of CXCR4-RLT for lymphoproliferative disease (Eudra-CT 2015-001817-28).^13,26

CONCLUSIONS

In patients with refractory and/or relapsed hematological malignancies scheduled for CXCR4-RLT, a rapid increase of LDH as surrogate marker of lymphoma cell kill along with myeloablative effects was observed, which led to HSCT in all subjects. We recorded no laboratory-defined TLS and a low rate of grade ≥3 AE. The observed association of pretherapeutic PET with LDH levels after RLT may provide a rationale to further investigate the predictive value of the PET signal for outcome and appropriate patient selection.