CXCR4-targeted PET imaging using ⁶⁴Cu-AMD3100 for detection of Waldenström Macroglobulinemia

ABSTRACT

Objective:  Waldenström Macroglobulinemia (WM) is a rare B-cell malignancy characterized by secretion of immunoglobulin M and cancer infiltration in the bone marrow. Chemokine receptor such as CXCR4 and hypoxic condition in the bone marrow play crucial roles in cancer cell trafficking, homing, adhesion, proliferation, survival, and drug resistance. Herein, we aimed to use CXCR4 as a potential biomarker to detect hypoxic-metastatic WM cells in the bone marrow and in the circulation by using CXCR4-detecting radiopharmaceutical.

Methods: We radiolabeled a CXCR4-inhibitor (AMD3100) with ⁶⁴Cu and tested its binding to WM cells with different levels of CXCR4 expression using gamma counter in vitro. The accumulation of this radiopharmaceutical tracer was tested in vivo in subcutaneous and intratibial models using PET/CT scan. In addition, PBMCs spiked with different amounts of WM cells ex vivo were detected using gamma counting.

Results: In vitro, ⁶⁴Cu-AMD3100 binding to WM cell lines demonstrated a direct correlation with the level of CXCR4 expression, which was increased in cells cultured in hypoxia with elevated levels of CXCR4, and decreased in cells with CXCR4 and HIF-1α knockout. Moreover, ⁶⁴Cu-AMD3100 detected localized and circulating CXCR4^high WM cells with high metastatic potential.

Conclusions: In conclusion, we developed a molecularly targeted system, ⁶⁴Cu-AMD3100, which binds to CXCR4 and specifically detects WM cells with hypoxic phenotype and metastatic potential in the subcutaneous and intratibial models. These preliminary findings using CXCR4-detecting PET radiopharmaceutical tracer indicate a potential technology to predict high-risk patients for the progression to WM due to metastatic potential.

Introduction

Waldenström Macroglobulinemia (WM) is a B-cell malignancy, a type of a rare non-Hodgkin lymphoma, with an incidence of about three cases per million people per year in the United States, accounting for 1-2% of non-Hodgkin lymphomas.^1,2 WM is a low-grade lymphoma, characterized by secretion of monoclonal immunoglobulin M (IgM) in the peripheral blood (PB) and infiltrate of lymphoplasmacytic cells in the bone marrow (BM). The main risk factor for the development of WM is pre-existing IgM monoclonal gammopathy of undetermined significance (IgM MGUS) with 46-fold higher risk than for the general population.³ Patients with IgM MGUS can remain asymptomatic for years while the probability of IgM MGUS progressing to WM is 1.5% per year.⁴ However, the point at which asymptomatic disease evolves into symptomatic WM remains vague.⁵ Since treatment is reserved for symptomatic patients,^6,7 earlier detection of the disseminated form of WM implies earlier treatment and hopefully better patient response and survival. Therefore, there is an urgent need to develop a new system which can accurately evaluate and predict the progression from asymptomatic (local) IgM MGUS to symptomatic (disseminated) WM.

It was reported previously that hypoxia causes invasive behavior of cancer cells through HIF-1α-dependent C-X-C chemokine receptor type 4 (CXCR4) upregulation, thus CXCR4 is considered a cancer biomarker.⁸ CXCR4 is being exploited by various tumors contributing to increased survival, invasion, and homing to target organs demonstrated both in vitro and in vivo.^{9 –14} We reported that HIF-1α and CXCR4 are expressed in WM cells and increased in hypoxia,¹¹ and that CXCR4 was essential for the WM cell migration through direct interaction with its ligand, stromal cell-derived growth factor-1 (SDF-1).¹² Metastatic tumor cells expressing high levels of CXCR4 homed to tissues rich in SDF-1, suggesting that CXCR4 is a marker of WM aggression, invasion, and metastasis.^10–12 In addition, a small-molecule antagonist of CXCR4, AMD3100, or CXCR4 knockdown caused significant inhibition of cell migration and metastasis.⁹ We have also reported that hypoxia was the driving force for WM metastasis, and that hypoxic-CXCR4^high cells adhered less to the BM stroma and exhibited more chemotaxis and homing to the BM niche in vivo.¹¹ Therefore, hypoxia and CXCR4 play major roles in the dissemination of WM from a cancer-rich niche to a distant BM niche. Additionally, somatic mutation in CXCR4 correlated with poor overall survival in WM patients. CXCR4 mutation was shown in 20% of IgM MGUS patients and almost 30% of WM patients, with little or no mutations present in other B-cell diseases and healthy volunteers.^13,15 Moreover, CXCR4^high WM cells demonstrated excellent response to proteasome inhibitors, which was further enhanced by anti-CXCR4 monoclonal antibody, suggesting that CXCR4 is an important therapeutic target.¹³

Positron emission tomography (PET) imaging is used successfully in a BM malignancies such as multiple myeloma (MM) and lymphomas.^16,17 PET has a number of advantages including the possibility of detecting BM involvement with sensitivity ranging from 80% to 100% in MM patients.^18,19 PET results correlated well with the achievement of high-quality responses to therapy in MM;^20,21 however, there is a narrow utility of PET in MGUS and smoldering myeloma, which usually appear negative.¹⁷ It is noteworthy that combined PET and computed tomography (CT) imaging showed affirmative results in 83% of WM patients with abnormal BM uptake in 43% of patients;²² however, the studies are limited. Therefore, we hypothesized that imaging hypoxic-metastatic cells using a CXCR4-targeted PET tracer will provide a powerful tool for evaluation of disseminated and more aggressive WM. A radiopharmaceutical agent, ⁶⁴Cu-AMD3100, was previously shown to accumulate in CXCR4-expressing tumors in vivo.^14,23 Visualization of CXCR4-positive tumors using ⁶⁴Cu-AMD3100 and a noninvasive PET imaging was able to detect both local and metastatic tumors,^21,24 which provided the reason for using ⁶⁴Cu-AMD3100 in current study.

Herein, we developed a molecularly targeted system which accurately detected metastatic potential of WM cells based on the CXCR4 expression status in the BM and circulation. ⁶⁴Cu-AMD3100 reflected active disease utilizing PET/CT, suggesting a potential technology to predict high-risk patients for WM dissemination and thus progression.

Results

We have previously shown that CXCR4 plays a key role in the hypoxia-induced dissemination of WM and blocking it consequently inhibited migration and homing of WM cells to the new BM niches.^11,12 Herein, we showed that WM cell lines have differential basal expression of CXCR4 with RPCI-WM1 cells having the highest CXCR4 expression, and BCWM1 cells expressing half the amount of CXCR4 protein compared to RPCI-WM1 (Figure 1a; Supplementary Figure 1A). In addition, hypoxia (1% O₂) increased the expression of CXCR4 in WM cells in vitro, with the two-fold increase in RPCI-WM1 and ~1.5 fold increase in BCWM1 cell line (Figure 1b; Supplementary Figure 1B). Since RPCI-WM1 cells had higher expression of CXCR4 and grow as a single cell suspension, we chose this cell line to perform knock out (KO) studies. We developed a RPCI-WM1 cell line with CXCR4 KO and HIF-1α KO using CRISPR-Cas9 technology providing permanent downregulation of the genes. The gRNAs activity was validated in the K562 cells for CXCR4 (Figure 2ai) and HIF-1α (Figure 2bi). Using NGS, the KO of the genes was confirmed at the DNA level in the four alleles for both CXCR4 (Supplementary Figure 2A) and HIF-1α (Supplementary Figure 2B) with 0% WT reads and almost 100% −1 and −2 reads, which suggests out of frame editing and thus complete deletion. However, the lack of complete loss of mRNA might be due to mRNA from the out of frame CXCR4, which will not necessarily be degraded. In addition, CXCR4 deletion creates a truncated protein and does not lead to non-sense mediated decay. Subsequently, we validated the KO of CXCR4 at the protein level using flow cytometry, demonstrated as a histogram with nearly complete KO of CXCR4 (Figure 2aii; Supplementary Figure 1C). Also, we exposed the CXCR4-KO cells to hypoxia for 24 h, and showed that CXCR4 cell surface protein expression was downregulated in both conditions by 86% and 90% in normoxia and hypoxia, respectively, compared to RPCI-WT (Figure 2aiii; Supplementary Figure 1C). HIF-1α was more difficult to KO; intracellular protein expression was partially downregulated in the HIF1-KO cell line as shown on the histogram (Figure 2bii), and by 65% and 35% in normoxia and hypoxia, respectively, compared to RPCI-WT (Figure 2biii).

Figure 1.

CXCR4 is highly expressed in WM cell lines and increases in hypoxia. Basal CXCR4 protein expression in RPCI-WM1 and BCWM1 cell lines demonstrated as relative mean fluorescent intensity (RMFI) (a). CXCR4 protein expression in normoxia (21% O₂) and hypoxia (1% O₂) at 24 h in RPCI-WM1 and BCWM1 cell lines demonstrated as fold change normalized to normoxia (b). The experiments were performed in triplicates and repeated at least 3 times. Results are shown as mean ± s.d.; the statistical significance was assessed by unpaired Student’s t-test (** p < .01; *** p < .001).

Figure 2.

Validation of CRISPR knock out of CXCR4 and HIF-1α in RPCI-WM1 cell line. Editing efficiencies of gRNA activity shown as % normalized to non-homologous end joining (NHEJ) (ai). Validation of CXCR4 knock out (KO) in RPCI-WM1 cell line tested at the protein level using flow cytometry and shown as a histogram (aii), and as a fold change expression normalized to normoxic cells (aiii). Editing efficiencies of gRNA activity shown as % normalized to NHEJ (bi). Validation of HIF-1α KO in RPCI-WM1 cell line tested at the protein level using flow cytometry and shown as a histogram (bii), and as a fold change expression normalized to normoxic cells (biii). The experiments were performed in triplicates and repeated at least 3 times. Results are shown as mean ± s.d.; the statistical significance was assessed by unpaired Student’s t-test (* p < .05; ** p < .01; *** p < .001).

Next, we performed radiolabeling of AMD3100 with ⁶⁴Cu (t_1/2 = 12.7 h, β⁺ = 17%, β⁻ = 39%, EC = 43%, E_max = 0.656 MeV) as described in the methods section, with the UV spectra and retention time for ⁶⁴Cu-AMD3100 of 5.90 min, and 6.00 min for cold AMD3100 shown in Figure 3a. A radiochemical yield of greater than 97% was achieved for the labeled compound (not shown) and therefore was used without further purification. The specific activity of the ⁶⁴Cu-AMD3100 was 1.0 Ci/µmol (37.02 GBq/µmol). We then demonstrated a direct correlation between binding of ⁶⁴Cu-AMD3100 (shown as CPM) to WM cells the level of CXCR4 expression (shown as RMFI) confirming that RPCI-WM1 cells expressed twice as much of CXCR4 as BCWM1 cells, thus bound twice as much of ⁶⁴Cu-AMD3100 (Figure 3b). To confirm the specificity of the binding of ⁶⁴Cu-AMD3100 to CXCR4 we performed the binding assay in the presence of large excess of cold AMD3100 (blocking) and found that pre-treatment with AMD3100 significantly blocked ⁶⁴Cu-AMD3100 binding to CXCR4 (Figure 3c). In addition, ⁶⁴Cu-AMD3100 bound to hypoxic RPCI-WT cells 3.5-fold more than to normoxic cells using gamma-counting, which was significantly decreased by 65% and 82% in CXCR4-KO cells in normoxia and hypoxia, respectively (Figure 3d). Also, HIF1-KO cells demonstrated 58% lower binding of ⁶⁴Cu-AMD3100 in hypoxic conditions only, which implies that CXCR4 is a HIF-1α target gene. These results suggest that ⁶⁴Cu-AMD3100 binds preferentially to RPCI-WT cells and even further to hypoxic cells with high expression of CXCR4 and a high metastatic potential compared to normoxic and low expressing-CXCR4 cells (i.e., KO cells) with low metastatic potential.

Figure 3.

Differential expression of CXCR4 in WM cell lines renders differential binding of ⁶⁴Cu-AMD3100 to WM cells in vitro and ex vivo. High performance liquid chromatography (HPLC) traces of ⁶⁴Cu-AMD3100 complex (a). Correlation between CXCR4 expression (relative mean fluorescent intensity, RMFI) and binding of radiolabeled ⁶⁴Cu-AMD3100 to RPCI-WM1 and BCWM1 cell lines detected using gamma counting and shown as count per minute (CPM) (b). Binding of ⁶⁴Cu-AMD3100 to RPCI-WM1 and BCWM1 with and without previous blocking of CXCR4 with cold AMD3100, demonstrated as fold change normalized to control (c). Binding of ⁶⁴Cu-AMD3100 to RPCI-WT, CXCR4-KO, and HIF1-KO in both normoxia and hypoxia demonstrated as fold change normalized to normoxia (d). Levels of ⁶⁴Cu-AMD3100 binding to peripheral blood mononuclear cells (PBMCs) spiked with 1%, 5% and 10% of hypoxic RPCI-WT and CXCR4-KO cells (cultured at 1% O₂ for 24 h) and measured using gamma counting, demonstrated as CPM. The results are shown as mean ± s.d. from 5 replicates (e). The experiments were performed in quadruplets and repeated at least 3 times. Results are shown as mean ± s.d.; the statistical significance was assessed by unpaired Student’s t-test (* p < .05; ** p < .01; *** p < .001).

Furthermore, we found that the number of circulating WM cells was in a direct linear correlation with the level of hypoxia in the BM, indicating that the metastatic potential of WM cells is related to their hypoxic status emphasizing the role of hypoxia in the dissemination of WM cells.^9,11 Therefore, we tested whether radiolabeled ⁶⁴Cu-AMD3100 could detect WM in the circulation using a surrogate experiment. For this purpose, we cultured RPCI-WT and CXCR4-KO cell lines in hypoxia for 24 h, in order to increase the CXCR4 expression. PBMCs were spiked with 1%, 5% and 10% of hypoxic RPCI-WT and CXCR4-KO cells, followed by ex vivo binding of ⁶⁴Cu-AMD3100 using gamma counting. As demonstrated in Figure 3e, we were able to detect increasing radioactive signal due to binding of ⁶⁴Cu-AMD3100 to CXCR4 in the hypoxic RPCI-WT, but not in the CXCR4-KO cells. The radioactive signal was in a linear correlation with the amount of WM cells in the blood.

Subsequently, we used subcutaneous WM mouse model to demonstrate the ⁶⁴Cu-AMD3100 binding. Nine (9) days after WM cell implantation of RPCI-WT (T1), CXCR4-KO (T2), HIF1-KO (T3) and BCWM1-WT (T4), tumor growth was validated by BLI (Figure 4ai). At day 15, IV administration of ⁶⁴Cu-AMD3100 was performed, and the radioactivity was tested at 1 and 4 h post-injection to monitor the tracer clearance. As demonstrated on Figure 4aii, ⁶⁴Cu-AMD3100 accumulated specifically in the WT tumors (CXCR4^high) such as RPCI-WT (T1) and BCWM1-WT (T4) with sparse accumulation in CXCR4-KO (CXCR4^low) and HIF1-KO (CXCR4^intermediate). The uptake of ⁶⁴Cu-AMD3100 was quantified and demonstrated as SUV_mean at these two-time points (Figure 4b). In addition, biodistribution of the radioactivity at 4 h post-injection was tested in harvested tumors (Figure 4c) and other organs (Supplementary Figure 3A), which confirmed the PET analysis results.

Figure 4.

PET imaging with ⁶⁴Cu-AMD3100 detects localized WM tumor cells with high CXCR4 expression in vivo. A representative image of localization of WM cells lines injected subcutaneously into NCG mice (T1: RPCI-WT, T2: CXCR4-KO, T3: HIF1-KO and T4: BCWM1) using bioluminescent imaging (BLI) (ai). A representative PET/CT images demonstrating ⁶⁴Cu-AMD3100 accumulation in subcutaneous tumors at 1 and 4 h post-tracer injection (aii). Quantification of ⁶⁴Cu-AMD3100 uptake demonstrated as standard uptake values (SUV) mean [calculated as nCi/cc×animal weight/injected dose] in tumors (T1, T2, T3, and T4) at 1 and 4 h post injection from n = 3 mice (b). Post-PET biodistribution of ⁶⁴Cu-AMD3100 at 4 h post-injection in harvested tumors (T1, T2, T3, and T4) demonstrated as % of injected dose per gram of tissue (%ID/g) from n = 3 mice (c). A representative image of localization of WM cells lines injected intratibially into NCG mice (T1: RPCI-WT and T2: CXCR4-KO) using BLI (di). A representative PET/CT images demonstrating ⁶⁴Cu-AMD3100 accumulation in tumors (T1 and T2) at 1 and 4 h post-tracer injection (dii). Quantification of ⁶⁴Cu-AMD3100 uptake demonstrated as intensity signal in tumors (T1 and T2) at 1 h and 4 h post injection from n = 3 mice (e). The statistical significance was assessed by unpaired Student’s t-test (* p < .05; ** p < .01).

Figure 4.

(Continued).

Consequently, we performed similar study with a more physiological intratibial WM model, where we injected 0.25 × 10⁶ WM cells (in 25 µL) into the tibias of NCG mice. We confirmed the tumor localization and growth using BLI at day 9 post cell implantation as demonstrated on a representative image (Figure 4di). Next, after ⁶⁴Cu-AMD3100 injection, PET/CT was performed and the binding was shown on representative images at 1 and 4 h post-injection to monitor the tracer clearance (Figure 4dii). Quantitative analysis demonstrated significantly lower uptake of ⁶⁴Cu-AMD3100 in the CXCR4-KO tumor compared to the RPCI-WT (Figure 4e). Additionally, a post-PET biodistribution study of the radioactivity in harvested organs was shown in Supplementary Figure 3B at 4 h post-injection.

Discussion

We have previously demonstrated that WM progression in the BM induced hypoxic conditions which directly correlated with tumor burden and metastasis.¹¹ Hypoxia modulates gene expression mainly by HIF-1α, a potent transcription factor and a biomarker, which makes it a negative prognostic factor for cancer treatment efficacy and patient survival.^25,26 Hypoxia/HIF-1α pathway is a driving force of cancer cell metastasis; it was shown that tumor cells cultured in hypoxic conditions are more invasive and home faster than the cells cultured in normoxic conditions.^11,27,28 HIF-1α inhibition significantly reduced the metastasis of cancer cells in vivo, and HIF-1α deficient cells were less motile, invasive and adhesive in vitro.²⁹ CXCR4 plays a major role in the hypoxia-induced dissemination of WM and inhibition of HIF-1α prevented the increase of CXCR4 in hypoxic WM, and consequently inhibited migration and homing of WM cells to the new BM niches.^11,12 Therefore, we investigated whether we could identify WM cells with a metastatic potential by measuring biological changes happening in the cancer cells by detecting the level of hypoxia and CXCR4 using chemical tracer that can be used clinically for PET imaging as a novel strategy to identify WM cells with a high risk for metastasis.

We developed biological and chemical tools. The biological tools included development of WM cell lines with differential expression (knock out) of CXCR4 and HIF-1α, and the chemical tool included production of radiolabeled tracer which detects changes in CXCR4 levels using ⁶⁴Cu-AMD3100. It was reported previously that incorporation of copper ion (Cu²⁺) into AMD3100 did not change its ability to inhibit cellular migration in response to SDF-1 and could be used as a radiopharmaceutical agent.^14,24 We found that the ⁶⁴Cu-AMD3100 showed specific binding to WM cells with high expression of CXCR4, especially when the cells were hypoxic. Moreover, KO of CXCR4 in the WM cells significantly reduced the binding of ⁶⁴Cu-AMD3100, as well as HIF-1α KO due to the fact that CXCR4 is a HIF-1α downstream target. Thus, ⁶⁴Cu-AMD3100 accurately detected changes in the nature of the WM cells, based on CXCR4 and HIF-1α expression, which are very important factors for the understanding of metastasis.

Additionally, we were able to detect circulating tumor cells using ⁶⁴Cu-AMD3100 ex vivo, where we spiked PBMCs with increasing number (to recreate different amounts of circulating tumor cells) of RPCI-WT and CXCR4-KO cells exposed to hypoxia, and we found that the radioactive signal detected from CXCR4 expressing RPCI-WT was proportional to the percentage of spiked cells. We found that ⁶⁴Cu-AMD3100 has a technical ability to bind to circulating CXCR4^high cells with significant signal-to-noise ratio, and with the linear correlation between the amount (%) of cells in the blood and the radio-signal. These results indicate that detection of ⁶⁴Cu-AMD3100 binding to circulating cells with high metastatic potential (CXCR4^high) is a sensitive and promising biomarker for cancer dissemination.

Next, we characterized the binding of ⁶⁴Cu-AMD3100 in subcutaneous WM model, and we found minimal uptake (2%ID/g) in the normal BM in the immune-deficient SCID mice. This was in agreement with previously published studies by Nimmagadda et at. showing no (or minimal) uptake in the bone marrow in subcutaneous brain cancer model and around 3% in the bone in subcutaneous breast cancer model in immune-deficient NOD/SCID mice.²³ Also, PET images in ovarian cancer model in immune-deficient nude mice did not show accumulation of ⁶⁴Cu-AMD3100 in the bone (and the BM) as shown by Weiss et al.¹⁴ On the other hand, Jacobson et al. showed 14%ID/g marrow uptake in naïve immune-competent C57BL/6 mice (with no tumor).²⁴ Therefore, the possible explanation for the discrepancy in BM accumulation is due to using immune-deficient mice (without immune cells in the BM) versus immune-competent mice with immune cells in the BM, which were previously shown to highly express CXCR4 on T cells and B cells. Moreover, in the first-in-human study, part of the vertebral column which was irradiated (depleting the immune cells in this area), resulted in lack of radio-signal in this area compared with some uptake in the non-irradiated marrow (with immune cells).³⁰

We found that in WM-bearing mice ⁶⁴Cu-AMD3100 accumulated very specifically only in the tumors expressing high CXCR4, while the cells with CXCR4-KO were barely detectable using microPET imaging. We used the subcutaneously injected tumors with different levels of CXCR4 expression, and we found that ⁶⁴Cu-AMD3100 bound as soon as 1 h post injection specifically to CXCR4^high WT cells. Finally, we performed a similar study with physiologically relevant model, i.e., intratibially injected WM cells and observed increased accumulation of ⁶⁴Cu-AMD3100 in wild type RPCI-WM1 while the signal was significantly lower in the CXCR4-KO tumor. It is noteworthy that the biodistribution values of 4%ID/g uptake in the WM tumor and 40%ID/g in the liver that we observe in our Waldenström’s tumor model for ⁶⁴Cu-AMD3100 are typical and similar to the biodistribution in other tumor models. Nimmagadda et al. have shown uptake values of 3% and 7% in two breast cancer cell lines in mouse models, and around 5% in glioma tumor model; while in both models the liver uptake was 40%.²³ Similarly, Weiss et al. showed accumulation of 40% in the liver, with around 10% accumulation in the tumor (in CHO tumor cells artificially over-expressing CXCR4).¹⁴ Other studies performed in naïve mice with no tumor have also shown similar liver accumulation of 40%.²⁴ Binding specificity of 64Cu-AMD3100 shown herein by the blocking studies in vitro using unlabeled AMD3100 and by knockout of CXCR4, and aforementioned blocking studies in vivo demonstrated binding of ⁶⁴Cu-AMD3100 through CXCR4-mediated mechanism.

In conclusion, these results provide a strong preclinical basis for future clinical studies to identify metastatic tumor cells in WM by utilizing PET/CT imaging with ⁶⁴Cu-AMD3100 as a radiotracer to detect localized and circulating tumor cells with high metastatic potential. The future goal of these findings is to indicate asymptomatic WM patients with local stage of the disease and to predict a high(er) risk for development of symptomatic and disseminated disease.

Materials and methods

Cell culture

The WM cell lines BCWM1, RPCI-WM1, RPCI-eGFP-Luc-wild type (RPCI-WT), and the knock out (KO) cell lines RPCI-eGFP-Luc-CXCR4-KO (CXCR4-KO) and RPCI-Luc-HIF1α-KO (HIF1-KO) were cultured in RPMI-1640 media (Corning CellGro, Mediatech, Manassas, VA), supplemented with 10% fetal bovine serum (Gibco, Life Technologies, Grand Island, NY), 2 mmol/L of L-glutamine, 100 U/mL Penicillin and 100 μg/mL Streptomycin (Corning CellGro). Cells were cultured at 37°C and 5% CO₂ in a humidified tissue culture incubator (21% O₂, NuAire water jacket incubator, Plymouth, MN) or hypoxic chamber (1% O₂, Coy, Grass Lake, MI).

Protein expression

WM cell lines were stained with anti-human CXCR4-APC antibody with isotype control (purchased from BD Biosciences, San Jose, CA) for 1 h on ice, and the protein expression was analyzed using MACSQuant Flow Cytometer (Miltenyi, San Diego, CA). HIF-1α intracellular level was measured using HIF-1α antibody (Novus Biologicals, Littleton, CO) following fixation and permeabilization steps using BD Cytofix/Cytoperm Solution (BD Biosciences) according to the manufacturer protocol. The expression was demonstrated as histograms, as relative mean fluorescent intensity (RMFI), or as a fold change normalized to normoxic conditions. The experiments were performed in triplicates and replicated independently at least three times.

Cell line engineering

The RPCI-WM1 cell line stably expressing eGFP-Luciferase (Luc) fusion was first generated by the Genome Engineering & iPSC Center (GEiC) of Washington University in St. Louis via random integration. After three rounds of FACS sorting, the population contained over 97.7% of GFP positive cells (not shown). The CXCR4 and HIF-1α genes were disrupted independently with CRISPR gRNAs (CACTTCAGATAACTACACCGAGG and TGTGAGTTCGCATCTTGATAAGG, for CXCR4 and HIF-1α, respectively) designed to cleave early in the coding sequences. Expression constructs were assembled and validated in K562 cells before nucleofected to eGFP-Luc RPCI-WM1 cells. The transfected cells were then single cell sorted, and single cell clones were screened using Next Generation Sequencing (NGS) for harboring only out-of-frame alleles.

Radiolabeling

⁶⁴Cu was produced by a (p,n) reaction on enriched ⁶⁴Ni on a TR-19 biomedical cyclotron (Advanced Cyclotron Systems Inc., Richmond, B.C., Canada) at Mallinckrodt Institute of Radiology, Washington University School of Medicine and purified with an automated system using standard procedures.^31,32 A stock solution of ⁶⁴CuCl₂ was diluted with a 10-fold excess of 0.1 M ammonium acetate (NH₄OAc), pH 5.5 for radiolabeling. Labeling of AMD3100 with ⁶⁴Cu was achieved by adding 1 or 2 μg to 7.2 MBq (200 μCi) of ⁶⁴CuCl₂ in 100 µL of 0.1 M NH₄OAc (pH 5.5). The reactions were incubated on a thermomixer with 800 rpm agitation at 60°C for 20 min. Radiolabeled complexes were analyzed by high-performance liquid chromatography (HPLC) with a mobile phase of water (0.1% TFA) and acetonitrile (0.1% TFA), 0-100% acetonitrile over 10 min with a 1 mL/min flow rate. A radiochemical yield of greater than 99% was achieved for all labeled compounds and therefore used without further purification.

Binding of ⁶⁴Cu-AMD3100 in vitro

About 20 μL of radiolabeled mixture was diluted in 2 mL of 1 x PBS to prepare for binding assay. For blocking, cold doses of AMD3100 (1 μg in 10 μL) per tube were used and 10 μL of PBS/0.1% BSA were used as a control. 1 × 10⁶ of cells in a volume of 500 μL of PBS with 0.1% BSA were applied to 1.5 mL vials. Triplicates were used for each compound and the experiments were performed at least 3 times. Approximately 50,000 counts per minute (CPM) of ⁶⁴Cu labeled AMD3100 were added in a volume of 10 μL to each tube. These were incubated at room temperature for 1 h on a shaker, and then centrifuged and carefully removed the supernatant. Cell pellets were washed twice with cold PBS and the measurement of ionizing radiation was registered in dry pellets using automated Packard II gamma counter (Perkin Elmer, Boston, MA). The bound radioactivity was expressed as CPM or fold change relative to normoxic or wild-type cells.

Binding of ⁶⁴Cu-AMD3100 to circulating WM cells ex vivo

Peripheral blood mononuclear cells (PBMCs) freshly isolated from a healthy patient were spiked with 1%, 5% and 10% of RPCI-WT and CXCR4-KO cells cultured in hypoxia for 24 h, and ex vivo binding of ⁶⁴Cu-AMD3100 was measured in quadruplets as described above.

Subcutaneous and intratibial WM mouse models

All animal experiments were performed in compliance with the Guidelines for Care and Use of Research Animals established by the Division of Comparative Medicine and the Animal Studies Committee of Washington University School of Medicine.

Four (4) different cell lines mixed with Matrigel (Corning) at the Matrigel:cell suspension ratio of 2:1 were injected subcutaneously into NCG mice (n = 3; strain 572; male; 9 weeks old; Charles River Lab, Wilmington, MA) in four different spots (2 x 10⁶ cells per spot): RPCI-WT (T1), CXCR4-KO (T2), HIF1-KO (T3) and BCWM1-WT (T4). Additionally, RPCI-WT cells (T1) and CXCR4-KO cells (T2) were injected into the tibias (0.25 x 10⁶ cells in 25 µL per tibia) of NCG mice (n = 3) using Hamilton syringe. Tumor localization and growth was validated using bioluminescent imaging (BLI) at day 9 post cell implantation. Briefly, mice were first injected intraperitoneally with D-Luciferin (150 mg/kg; Gold Biotechnology, St. Louis, MO), anesthetized with 2% isofluorane vaporized in oxygen and imaged with IVIS 50 (PerkinElmer, Waltham, MA). Total photon flux per second was measured from fixed regions of interest (ROIs) using Living Image 2.6.

MicroPET/CT

The distribution of ⁶⁴Cu-AMD3100 in mice was evaluated after intravenous (IV) injection with 100 µCi of ⁶⁴Cu-AMD3100. In both cases the mice were anesthetized with 1-2% isofluorane/oxygen and imaged on an Inveon small animal PET/CT scanner (Siemens Medical Solutions) at 1 and 4 h after injection. Static images were collected at 1 h post injection for 15–30 min and reconstructed with the maximum aposteriory probability (MAP) algorithm followed by CT co-registration with the Inveon Research Workstation (IRW) image display software. ROIs were selected from PET images with the CT anatomical guidelines, and the associated radioactivity was measured using Inveon Research Workstation software. Standard uptake values (SUV) were calculated as nCi/cc×animal weight/injected dose.

Biodistribution study

At 4 h post ⁶⁴Cu-AMD3100 injection, the mice were euthanized and subjected to biodistribution studies. Blood, lung, liver, spleen, kidney, bladder, muscle, heart, brain, stomach, bone, marrow, and tumors were harvested, and the amount of radioactivity in each organ was determined by gamma counting; percent injected dose per gram of tissue (%ID/g) was calculated.

Statistics

Quantitative data are expressed as mean ± s.d. Statistical analysis was performed using one-way analysis of variance and Student’s t-test. Differences at the 95% confidence level (* p < .05; ** p < .01; *** p < .001) were considered statistically significant.

Funding Statement

This work was supported by the International Waldenstrom’s Macroglobulinemia Foundation; National Center for Advancing Translational Sciences [U54CA199092]. The study was supported partially by a research grant from IWMF and by the award from the National Center for Advancing Translational Sciences (NCATS) of the National Institutes of Health (NIH) and the National Cancer Institute of the NIH under (U54CA199092). The content is solely the responsibility of the authors and does not necessarily represent the official view of the NIH.

Acknowledgments

We thank GEiC at Washington University in St. Louis for gene editing of RPCI-WM1 cells using CRISPR-Cas9 technology. We thank Lynne Marsala from Molecular Imaging Center, Department of Radiology for performing in vivo BLI imaging. We thank Amanda Klaas, Margaret Morris, Lori String, and Nicole Fettig from Preclinical Imaging Facility of Mallinckrodt Institute of Radiology for performing the PET imaging and biodistribution study on mice. The imaging studies were supported by NIH P50 CA094056 (Molecular Imaging Center) and NCI P30 CA091842 (Siteman Cancer Center Small Animal Cancer Imaging shared resource). Kinan Alhallak was supported by the National Center for Advancing Translational Sciences of the NIH under Award Number TL1TR002344.

Conflict of Interest Disclosure Statement

Dr. Azab receives research support from Arch Oncology and Cantex Pharmaceuticals; and is the founder and owner of Targeted Therapeutics LLC and Cellatrix LLC; however, these have no contribution to this study. Dr. Rogers states no conflicts of interest. Other authors state no conflicts of interest.

Supplemental Material

Supplemental data for this article can be accessed on the publisher’s website.