Modulating the Adhesion of Haematopoietic Stem Cells with Chemokines to Enhance Their Recruitment to the Ischaemically Injured Murine Kidney

Abstract

Introduction: Renal disease affects over 500 million people worldwide and is set to increase as treatment options are predominately supportive. Evidence suggests that exogenous haematopoietic stem cells (HSCs) can be of benefit but due to the rarity and poor homing of these cells, benefits are either minor or transitory. Mechanisms governing HSC recruitment to injured renal microcirculation are poorly understood; therefore this study determined (i) the adhesion molecules responsible for HSC recruitment to the injured kidney, (ii) if cytokine HSC pre-treatment can enhance their homing and (iii) the molecular mechanisms accountable for any enhancement.

Methods: Adherent and free-flowing HSCs were determined in an intravital murine model of renal ischaemia-reperfusion injury. Some HSCs and animals were pre-treated prior to HSC infusion with function blocking antibodies, hyaluronidase or cytokines. Changes in surface expression and clustering of HSC adhesion molecules were determined using flow cytometry and confocal microscopy. HSC adhesion to endothelial counter-ligands (VCAM-1, hyaluronan) was determined using static adhesion assays in vitro.

Results: CD49d, CD44, VCAM-1 and hyaluronan governed HSC adhesion to the IR-injured kidney. Both KC and SDF-1α pre-treatment strategies significantly increased HSC adhesion within injured kidney, whilst SDF-1α also increased numbers continuing to circulate. SDF-1α and KC did not increase CD49d or CD44 expression but increased HSC adhesion to VCAM-1 and hyaluronan respectively. SDF-1α increased CD49d surface clustering, as well as HSC deformability.

Conclusion: Increasing HSC adhesive capacity for its endothelial counter-ligands, potentially through surface clustering, may explain their enhanced renal retention in vivo. Furthermore, increasing HSC deformability through SDF-1α treatment could explain the prolonged systemic circulation; the HSC can therefore continue to survey the damaged tissue instead of becoming entrapped within non-injured sites. Therefore manipulating these mechanisms of HSC recruitment outlined may improve the clinical outcome of cellular therapies for kidney disease.

Figure 1. Adherent and free flowing HPC-7 s are increased in IR injured kidney.

A significant increase in HPC-7 to frozen sections of IR injured and non-injured CL tissue was observed compared to controls (A). Similarly, adhesion in vivo within renal peritubular capillaries was also significantly increased in IR injured mice compared to controls (B). Representative images of CFSE-labeled HPC-7 s in sham (C) and IR-injured (D) renal microcirculation are shown; scale bars shown are 200 μm. Both in focus and out of focus cells are counted. These events were paralleled by those occurring in other randomly selected regions of the kidney (E). Free flowing HPC-7 numbers were increased in IR injured mice at the point of infusion (F). Blood flow was significantly reduced in IR injured mice at the time of HPC-7 infusion (G). HPC-7 velocity in vivo was significantly reduced in IR injured renal microcirculation compared to sham (H). For all line graphs: sham control  =  solid line; IR-injured  =  dashed line. Results are presented as mean ± SEM (n≥4). *p<0.05, **p<0.01, ***p<0.001.

Figure 2. HPC-7 recruitment to IR injured kidney is dependant on CD44 and CD49d.

2×10⁶ HPC-7 were pre-treated with function-blocking monoclonal antibodies (80 µg/ml) against integrins CD18 and CD49d and the non-integrin CD44. Function blocking antibodies to endothelial VCAM-1 and CD44 and the enzyme hyaluronidase (to block HA) were administered in vivo at 1 minute post-reperfusion and 2×10⁶ naïve HPC-7 were infused at 60 minutes. No decrease in HPC-7 adhesion was observed with an anti-CD18 antibody compared to IgG control (A). Adhesion was significantly reduced by blocking CD49d (B) and also when its endothelial counter-ligand, VCAM-1, was blocked in vivo (C). Similarly, adhesion was significantly reduced by blocking CD44 on HPC-7 (D) and also when its endothelial counter-ligand, HA, was blocked in vivo (E). Blocking endothelial CD44 did not decrease HPC-7 adhesion (F). For all graphs: IgG controls  =  solid line; blocking treatments  =  dashed line. Results are presented as mean ± SEM (n≥4). *p<0.05, **p<0.01, ***p<0.001.

Figure 3. KC and SDF-1α mediate HPC-7 recruitment to the healthy and IR injured kidney.

HPC-7 expressed the main KC and SDF-1α receptors, CXCR2 (A) and CXCR4 (B) respectively. Topically treating a healthy kidney with KC [200 ng/ml; C] or SDF-1α [200 ng/ml; D] for 4 hours led to significant HPC-7 recruitment compared to the PBS control. Blocking CXCR2 (E) and CXCR4 (F) on HPC-7 prior to administration also decreased HPC-7 adhesion in vivo within the IR injured kidney. For all figures: PBS controls  =  solid line; kidney pre-treatments/mAb treated HPC-7 =  dashed line. Results are presented as mean ± SEM (n≥4). *p<0.05, **p<0.01.

Figure 4. HPC-7 adhesion can be enhanced by pre-treating them with KC and SDF-1α.

Only HPC-7 pre-treatment with SDF-1α (25 ng/ml; 5 minutes) significantly increased HPC-7 adhesion to IR injured frozen renal tissue when compared to PBS control (A). Pre-treatment with KC (25 ng/ml; 5 minutes) and SDF-1α (25 ng/ml; 5 minutes) significantly increased HPC-7 adhesion to TNFα (100 ng/ml; 4 hours) activated murine renal endothelium compared to the PBS control (B). Pre-treatment of primary murine lineage negative cells also yielded similar results. Both KC and SDF-1α significantly increased Lin⁻ cell adhesion to TNFα activated renal endothelium (C). Pre-treating with KC (D), SDF-1α (E) or KC+SDF-1α (F) significantly increased HPC-7 adhesion within IR injured kidney in vivo, although dual pre-treatment did not confer a greater effect. For line graphs: PBS pre-treated HPC-7+ IR kidney  =  solid line; KC and/or SDF-1α pre-treated HPC-7 s + IR injured kidney  =  dashed line. Results are presented as mean ± SEM (n≥4). **p<0.01.

Figure 5. SDF-1α pre-treatment increases numbers of free flowing HPC-7 in IR injured kidney.

Following KC pre-treatment, no difference in free flowing HPC-7 numbers was observed at anytime point in IR injured kidney (A). However, SDF-1α pre-treatment significantly increased free-flowing HPC-7 s at the point of administration only (B). Dual pre-treatment with both KC+SDF-1α not only increased free-flowing cells at the point of infusion but sustained this increase in HPC-7 trafficking around the renal microcirculation throughout the entire period of observation (C). The first one minute recording was broken down into 6 second intervals to illustrate each full circulatory pass. With SDF-1α pre-treatment, more HPC-7 continue to flow between 6–24 seconds suggesting less are lost to the circulation (D). HPC-7 velocity in vivo was not affected by any pre-treatment (E). HPC-7 became significantly more deformable with SDF-1α and KC+SDF-1α pre-treatment as demonstrated by reduced time taken to aspirate into a glass capillary (F). A photograph illustrating the method is shown and 50 cells were tested/group (G). For all line graphs; PBS treated HPC-7+ IR kidney  =  solid line; KC and/or SDF-1α treated HPC-7 s + IR injured kidney  =  dashed line. Results are presented as mean ± SEM (n≥4). *p<0.05, **p<0.01.

Figure 6. KC and SDF-1α pre-treatment increases adhesion of HPC-7 surface adhesion molecules for their endothelial counter-ligands.

Pre-treatment with KC or SDF-1α did not increase surface expression of CD49d (A) or CD44 (B) on HPC-7 s. PBS-treated HPC-7 =  black line, grey fill; KC-treated HPC-7 =  blue line; SDF1α-treated HPC-7 =  orange line. Both KC and SDF-1α pre-treatment significantly increased HPC-7 adhesion to immobilised VCAM-1 (C) and HA (D) when compared to PBS pre-treated control cells. SDF-1α pre-treatment only increased the number of clusters of CD49d on the HPC-7 surface (Ei–ii). However, KC pre-treatment only increased the number of clusters of CD44 on the HPC-7 surface (Fi–ii). Results are presented as mean ± SEM (n≥3). *p<0.05, **p<0.01, ***p<0.001.