Effect of the CXCR4 antagonist plerixafor on endogenous neutrophil dynamics in the bone marrow, lung and spleen

Treatment with the CXCR4 antagonist, plerixafor (AMD3100), has been proposed for clinical use in patients with WHIM (warts, hypogammaglobulinemia, infections, and myelokathexis) syndrome and in pulmonary fibrosis. However, there is controversy with respect to the impact of plerixafor on neutrophil dynamics in the lung, which may affect its safety profile. In this study, we investigated the kinetics of endogenous neutrophils by direct imaging, using confocal intravital microscopy in mouse bone marrow, spleen, and lungs. Neutrophils are observed increasing their velocity and exiting the bone marrow following plerixafor administration, with a concomitant increase in neutrophil numbers in the blood and spleen, while the marginated pool of neutrophils in the lung microvasculature remained unchanged in terms of numbers and cell velocity. Use of autologous radiolabeled neutrophils and SPECT/CT imaging in healthy volunteers showed that plerixafor did not affect GM‐CSF‐primed neutrophil entrapment or release in the lungs. Taken together, these data suggest that plerixafor causes neutrophil mobilization from the bone marrow but does not impact on lung marginated neutrophil dynamics and thus is unlikely to compromise respiratory host defense both in humans and mice.


INTRODUCTION
The CXCR4 antagonist plerixafor (AMD3100) is used clinically for the acute mobilization of HSPCs for bone marrow transplants. 1 Plerixafor also causes a dose-and time-dependent blood leukocytosis in humans. 2 Studies in both mice and humans have shown that plerixafor also affects neutrophil dynamics, most notably increasing circulating neutrophil numbers as early as an hour after administration. [3][4][5] Initial work from our group showed that when plerixafor was infused directly mutations of CXCR4. 7,8 Indeed, patients and mouse models of this disease exhibit an increase in the numbers of mature neutrophils in the bone marrow reserve 7,8 and in patients with WHIM syndrome the blood neutropenia can be corrected by the administration of plerixafor. 9 Likewise, myeloid-specific deletion of CXCR4 in mice has been shown to result in reduced numbers of mature neutrophils in the BM and a blood neutrophilia. 10 Contrary to these findings, an alternative view is that the blood neutrophilia induced by plerixafor may be due to neutrophil demargination from the lung microvasculature and not due to mobilization from the BM. 11 Thus when Devi et al. imaged GFP + neutrophils that had been isolated from the BM of LysM-GFP + mice and adoptively transferred into WT recipients they observed neutrophil mobilization from the lung and not the BM. 11 Furthermore, they observed a transient (2-4 h) increase in neutrophil numbers in blood sampled from the carotid artery versus the vena cava following plerixafor treatment both in primates and mice, again consistent with mobilization of neutrophils from the lung. 11 These findings were subsequently challenged in a study by Liu et al., who quantified the numbers of endogenous neutrophils in tissues by flow cytometry. 12 They reported an increase in neutrophil numbers in the lung, spleen, and blood accompanied by a decrease in the BM following plerixafor treatment. 12 In addition, they reported that imaging of frozen sections of lung, post-plerixafor treatment, showed no evidence of neutrophil de-margination. Recent technological advances in intravital microscopy (IVM) now allow us to directly image endogenous neutrophil dynamics in the lung, spleen, and BM, thus providing the opportunity to characterize neutrophil migratory behavior within their anatomical location prior and subsequent to plerixafor administration.
Neutrophils in the pulmonary circulation play an important role in intravascular host defense. 13 Entrapment of systemically primed neutrophils within the lung microvasculature also facilitates de-priming, with these cells subsequently able to re-circulate. 14,15 In human studies, we have shown that radiolabeled neutrophils primed ex vivo with GM-CSF accumulate in the lung microvasculature, with 97% being retained in the lungs at first pass, versus <5% of un-primed neutrophils.
It is not known, however, whether the retention of primed neutrophils or their subsequent time-dependent de-priming in the lungs is affected by plerixafor. Similarly, a proportion of murine neutrophils present at a site of experimental liver injury have been reported to re-enter the systemic circulation and subsequently lodge in the pulmonary vasculature, where they up-regulate CXCR4 before trafficking back to the BM for final clearance. 16 In this context plerixafor impaired trafficking of neutrophils from the lung to the BM, suggesting that this process is also mediated by CXCR4. 16 Recently it was shown that chronic administration of plerixafor is a feasible strategy for long-term treatment of WHIM patients, reducing infection frequency and wart burden associated with the disease. 17 Chronic treatment with plerixafor is also being considered for additional clinical applications including pulmonary fibrosis, pancreatic cancer (Camplex-1 trial), and leukemia, where the mechanism of action appears independent of the effect of plerixafor on neutrophil For this reason, in this study we used IVM of BM, lung, and spleen and precision cut lung slices (PCLS) to interrogate the effect of plerixafor on the dynamics of endogenous neutrophils in the mouse, and gamma scintigraphy (SPECT/CT) of radiolabeled neutrophils to investigate whether plerixafor impacts the trafficking of primed neutrophils in the lungs of humans. Our studies reveal that plerixafor has no effect on neutrophil dynamics in this organ, including neutrophil number and cell velocity in the lung, suggesting that it will not impact directly on lung host defense in patients.

Mice
C57Bl/6J female mice between 6 and 8 weeks old were used in all the experiments. All mice were housed in specific pathogen free conditions

Lung IVM
This method was first described in 21 with modifications. 22 Imaging was performed on an upright Leica SP5 confocal microscope using a 25 × 0.95na water immersion objective.

Spleen IVM
This method is described in ref. 23  Imaging was performed on an upright Leica SP5 confocal microscope using a 25 × 0.95na water immersion objective. Images were acquired in 3 z-slices 5 µm apart. In both lung and spleen IVMs, neutrophils were labeled with 3 µg/mouse of Ly6G (clone 1A8)-PE, the vasculature was labeled with CD31 (clone 390)-488, fluorescent Abs were injected intravenous (i.v.) in a maximal volume of 50 µl ≈10-20 min before imaging commences.
While imaging both lung and spleen, i.p. injections of plerixafor at 5 mg/kg (AMD3100, Sigma-Aldrich) in vehicle PBS were performed and imaging continued non-stop until 90 min after treatment. At the end of the imaging session, mice were humanely killed by anesthetic overdose (Sodium-Pentobarbital) and blood was collected by cardiac puncture and lung and spleen were harvested.

Cell tracking
BM IVM 3D time-series in .czi format were saved as 16bit TIFF and imported into NIS-Elements (Version 4.50, Nikon Instruments, UK).
Files were processed with Advanced Denoising and saved as Maximum Intensity Projections. Neutrophils were tracked using the Spot Tracking plugin 24 in Icy, an open-source platform for bioimage analysis. 24 Tracks in the 200 frames sequence were checked manually to ensure they were correct. 25 Lung IVM and spleen IVM 3D time-series in .lif format were analyzed using Imaris software (Bitplane, Oxford Instruments). The video was cropped in time to analyze 60 frames before and 30, 60, and 90 min after AMD3100/PBS application. Neutrophil tracking was performed automatically on Ly6G-positive cells transformed in spots. XYZ data were exported and track mean speed was plotted.

Study participants
The human study was approved by the Cambridgeshire and Hert- ≤80% of predicted or a FEV1 to forced vital capacity (FVC) ratio ≤70%.

Administration of plerixafor or placebo in human
Volunteers received either plerixafor or placebo, and neutrophils were reinfused 60 min following plerixafor administration at the circulating pharmacological T max for the drug. Plerixafor was administered in the clinically effective dose of 0.24 mg/kg (s.c.), which is used for mobilization of hematopoietic stem cells, which resulted in the expected marked leukocytosis at 3 h post-injection. 1 Re-infusion of neutrophils took place 60 min after plerixafor/placebo administration and was undertaken while the patients were on the SPECT-CT to allow for immediate imaging.

Statistical analysis
Statistical analysis was performed using GraphPad Prism 5 (GraphPad Software, Inc). A P-value of less than 0.05 was considered significant: P < 0.05 *, P < 0.01 **, P < 0.001 ***, NS, not significant. Statistical tests used are as detailed in the figure legends.

IVM reveals direct mobilization of neutrophils from the BM and increased velocity of neutrophils after plerixafor treatment
In mice, circulating blood neutrophil numbers increased significantly 60 min after i.p. injection of plerixafor (Fig. 1A); this is accompanied by a concomitant reduction in neutrophil numbers in the BM (Fig. 1B).
To directly test whether the blood neutrophilia was due to neutrophil release from the BM, we undertook IVM of the mouse calvarium, identifying BM vasculature by i.v. injection of Cy5-Dextran and endogenous neutrophils by i.v. injection of low dose anti-Ly6G-PE mAb. The use of low dose anti-Ly6G mAb (2-5 µg/mouse), as an imaging tool for neutrophil dynamics, has been widely reported and several stud-ies have shown that it does not compromise neutrophil dynamics such as rolling, adhesion, and intravascular crawling in a number of tissues including the lung. 28,29 Although the same anti-Ly6G mAb is used to cause neutrophil-depletion, the dose to achieve this effect is much higher (100-500 µg/mouse). 13,30,31 Moreover, a previous study has compared the dynamics of LysM-GFP + and fluorochrome-conjugated Ly6G mAb neutrophils and results showed that Ly6G mAb did not cause change in their migratory behavior or in their recruitment during inflammation. 32 The mean speed of Ly6G neutrophils within the calvarium bone marrow (BM) parenchyma was calculated (Fig. 1C) and a significant increase in neutrophil speed was observed 20 min after a single i.p. injection of plerixafor (Fig. 1D). Tracking individual neutrophils in the BM showed that ≈60% of neutrophils under homeostasis have a speed ranging from 2 to 4 µm/min (Fig. 1E). Thirty minutes after injection of plerixafor, the percentage of neutrophils migrating at such speed significantly decreases, ≈40% with a concomitant significant increase in the percentage of neutrophils with a higher speed ranging from 4 to 6 µm/min, ≈30% (Fig. 1E). Moreover, the percentage of neutrophils with the lower speed ranging from 0 to 2 µm/min did not change after plerixafor, ≈30% (Fig. 1E).

Spleen is not a source of blood neutrophilia following plerixafor treatment
The spleen is an alternative site of significant neutrophil margination. 33,34 Mobilization of neutrophils from the spleen might therefore also contribute to the increase in circulating neutrophil numbers seen after plerixafor treatment. 33,34 However, when quantifying cell numbers by flow cytometry, we noted a significant increase in the absolute number of splenic neutrophils 60 min after plerixafor administration ( Fig. 2A); this is in agreement with previous reports. 12 Moreover, analysis of neutrophil dynamics using spleen IVM showed that as early as 30 min after plerixafor injection there was an increase in the number of splenic neutrophils and that this increase was maintained up to 90 min after treatment ( Fig. 2B and C; Supplementary Video 2). Tracking individual neutrophils in the spleen showed that <40% of neutrophils under homeostasis have a speed ranging from 0 to 1 µm/min while >40% of neutrophils have a speed ranging from 1 to 2 µm/min and 20% from 2 to 3 µm/min (Fig. 2D). Sixty minutes after injection of plerixafor, the percentage of less motile neutrophils decrease significantly with a concomitant increase in the percentage of cells with a greater migratory speed (Fig. 2D). and decrease expression on CD62L suggesting that circulating neutrophils did not get activated by this drug (Supplementary Fig. 1E).

F I G U R E 1 Plerixafor treatment causes neutrophil mobilization and blood
These results suggest that plerixafor treatment causes an increase in the velocity of splenic neutrophils but not neutrophil activation.
Of note, CXCR4 levels were significantly higher on neutrophils remaining in the BM 60 min after plerixafor administration (Supplementary Fig. 1F) suggesting that plerixafor does not mobilize these senescent neutrophils. These data suggest that plerixafor did not cause activation of mobilized neutrophils nor change in their phenotype. Together, these data suggest that the spleen functions to 90 min following i.p. administration of plerixafor ( Fig. 3B and C).
Furthermore, plerixafor did not influence the velocity of lung intravascular marginated neutrophils, suggesting that CXCR4 signaling is not involved in homeostatic neutrophil migration within the murine lung (Fig. 3D). By IVM, an increase in circulating neutrophils flowing within and not getting in contact with the lung microvasculature was apparent 20 min following plerixafor treatment consistent with the blood neutrophilia (Supplementary Video 3). PCLS allowed us to image neutrophils located deeper in the lung, but this again showed that marginated neutrophil numbers were not significantly affected by plerixafor treatment (Fig. 3E). Thus, direct imaging of endogenous marginated neutrophils in the pulmonary vasculature of the mouse indicates that contrary to Devi et al., plerixafor does not cause the de-margination of neutrophils from the lung microvessels nor a change in migratory behavior. However, it is beyond the technical capacity of our current IVM system to phenotypically correlate migratory behaviors. This is an area that clearly warrants further investigation given the current interest in neutrophil subsets. 35,36

Low dose anti-Ly6G mAb does not interfere with neutrophil redistribution after plerixafor treatment
In a model of arthritis, Cunin et al. have shown that Ly6G ligation has no effect on the integrin-independent migration of neutrophils but attenuates integrin-dependent migration. 37 To test whether Ly6G ligation has any impact on neutrophil redistribution following plerixafor treatment in our system, we i.v. injected low dose (3 µg/mouse) anti-Ly6G mAb or IgG2A mAb as a control prior to plerixafor administration.
Our data show that the low dose of anti-Ly6G mAb has no effect on neutrophil mobilization from the BM and the increase in circulating and splenic neutrophils after plerixafor treatment (Supplementary Fig.   1H-J). These data suggest that low dose anti-Ly6G mAb, used in this study, does not interfere with neutrophil redistribution after plerixafor treatment. More studies are needed to directly prove whether integrins are involved in this response.

Plerixafor does not perturb pulmonary sequestration of primed neutrophils in humans
Undertaking similar experiments in a human setting is essential, however obviously more challenging. Neutrophil priming, occurring either systemically or ex vivo under experimental conditions results in neutrophil retention in the lungs. 38 This process is transient and proposed to be driven by priming/activation-related changes in neutrophil shape and deformability. 14,15 Hence, to investigate whether plerixafor administration interfered with either the initial entrapment of primed neutrophils within the pulmonary circulation, or the subsequent depriming and release events, we examined the effect of plerixafor on the pulmonary sequestration of autologous radiolabeled neutrophils that had been primed ex vivo with GM-CSF (1 or 100 ng/ml). 14 Neutrophils from healthy volunteers were isolated, dual radio-labeled, and re-infused (Fig. 4). Immediate dynamic planar gamma scintigraphy was undertaken to monitor early neutrophil bio-distribution, with lung, spleen, and liver time-activity curves generated. Volunteers received either plerixafor (0.24 mg/kg s.c.) or placebo (double-blinded) and autologous radiolabeled neutrophils were injected as a single bolus at the T max (60 min) following plerixafor administration. Plerixafor resulted in the expected marked leucocytosis in all subjects at 3 h postinjection. Specifically, neutrophils increased from 4.3 × 10 9 /L ± 0.7 (mean ± SEM) before injection of plerixafor to 9.9 × 10 9 /L ± 1.1 (mean ± SEM), 180 min after injection. Robust and immediate sequestration of radiolabeled neutrophils was seen in the lungs of both groups, with 63.5 ± 4.3% versus 65.6 ± 3.5% (plerixafor vs. placebo; mean ± SEM) of the peak pulmonary signal still present at 40 min ( Fig. 4A-C). No difference in the neutrophil signal could be distinguished when quantifying transaxial 45 min SPECT/CT of spleen and liver ( Fig. 4B and C) in either saline or plerixafor-dosed subjects. The proportion of cells remaining within the left and right lung compared with peak levels of 99m Tc-neutrophils, again showed no difference between plerixafor and saline treated subjects (Fig. 4D). In addition, the percentage recovery of the injected radiolabeled neutrophils from the peripheral blood at 40 min was identical 6.9 ± 3.7% versus 6.1 ± 1.9% (plerixafor vs. placebo; mean ± SEM; Fig. 4E). Taken together these data show that pulmonary sequestration of primed neutrophils in humans is not affected by plerixafor administration.

DISCUSSION
The CXCR4 antagonist, plerixafor, is used clinically as a single dose to mobilize stem cells for bone marrow transplants. Recently, plerixafor has been trialed in WHIM patients, 17 where it has been shown that chronic administration reverses the blood neutropenia and reduces the high rate of infections in these patients. Preclinical studies in mice also indicate that chronic administration of plerixafor reduces lung fibrosis. 39 It is important to note that unlike other CXCR4 antagonists, plerixafor binds to the transmembrane region of CXCR4 40  playing an essential role in bacterial surveillance in the lung. 13,42 Additionally, it has also been proposed that the lung constitutes a unique niche where neutrophil de-priming takes place under inflammatory conditions. 14    Liu reported that plerixafor increased neutrophil numbers in the spleen and they also showed that plerixafor could increase circulating neutrophil numbers in splenectomized mice, indicating that the spleen was not a source of mobilized neutrophils. Our data are consistent with these finding in which IVM of the spleen revealed that plerixafor caused a significant increase in splenic neutrophil velocity, while flow cytometry showed an increase neutrophil numbers in the spleen, suggesting that the spleen may function as a pool for excessive numbers of neutrophils in the blood.
Critically, although it is not possible to recapitulate this exact approach in humans, we have been able to address whether plerixafor affects neutrophil de-priming in human lungs, using nuclear imaging techniques to follow the trafficking of GM-CSF primed neutrophils through the lungs. Our data show that plerixafor does not impact the kinetics of retention or release of primed neutrophils in humans.
This is the first in vivo imaging study to comprehensively assess the effect of plerixafor on neutrophil kinetics in humans and mice. Our IVM data in mice show that plerixafor increases neutrophil motility and mobilization them from the bone marrow, and causes neutrophil accumulation in the spleen, while have no effect on numbers or migratory behavior of marginated intravascular neutrophils in the lung. Further in the human lung plerixafor did not affect the accumulation and release of GM-CSF-primed neutrophils. This study therefore adds to our knowledge of how plerixafor redistributes neutrophils from the BM into the blood with the resulting pooling of excess numbers of neutrophils in the spleen. Importantly while we observed an increase in the number of circulating neutrophils there was no evidence that these leukocytes were activated or primed and no evidence that plerixafor stimulates neutrophil de-margination in the lung.
In conclusion, our results suggest that it is unlikely that plerixafor will compromise respiratory host defense, however further experiments assessing the impact of chronic plerixafor treatment in models of respiratory disease are required to further determine the safety of this drug in the context of its clinical use in WHIM patients 17  wrote the manuscript, which was edited by all authors.