CD45 limits early Natural Killer cell development

Abstract The clinical development of Natural Killer (NK) cell‐mediated immunotherapy marks a milestone in the development of new cancer therapies and has gained traction due to the intrinsic ability of the NK cell to target and kill tumor cells. To fully harness the tumor killing ability of NK cells, we need to improve NK cell persistence and to overcome suppression of NK cell activation in the tumor microenvironment. The trans‐membrane, protein tyrosine phosphatase CD45, regulates NK cell homeostasis, with the genetic loss of CD45 in mice resulting in increased numbers of mature NK cells. This suggests that CD45‐deficient NK cells might display enhanced persistence following adoptive transfer. However, we demonstrate here that adoptive transfer of CD45‐deficiency did not enhance NK cell persistence in mice, and instead, the homeostatic disturbance of NK cells in CD45‐deficient mice stemmed from a developmental defect in the progenitor population. The enhanced maturation within the CD45‐deficient NK cell compartment was intrinsic to the NK cell lineage, and independent of the developmental defect. CD45 is not a conventional immune checkpoint candidate, as systemic loss is detrimental to T and B cell development, compromising the adaptive immune system. Nonetheless, this study suggests that inhibition of CD45 in progenitor or stem cell populations may improve the yield of in vitro generated NK cells for adoptive therapy.


INTRODUCTION
5][6] To date, galectin-1 is the only physiological ligand that results in inhibition of PTPase activity upon CD45 engagement. 5,7onetheless, efforts to discover a ligand for CD45 that positively modulates PTPase activity are still ongoing.
][18] In addition to its role in TCR and BCR activation, [8][9][10][11] CD45 has been reported to activate Syk, JNK and p38, downstream of Ly49D activation in NK cells. 19,20Lastly, CD45 has been shown to suppress Janus kinase (JAK) activity, resulting in negative regulation of cytokine receptor signaling. 21n mice, the Ptprc gene consists of 34 exons that encode an extracellular domain (exons 1-12), interdomain (exons 13-14), transmembrane domain (exon 15) and an intracellular cytoplasmic domain (exons 16-34).T, B and NK cells express multiple CD45 isoforms, that are differentially expressed during developmental stages. 3CD45 isoform switching occurs as lymphoid cells differentiate into T, B and NK cells, with less mature populations expressing larger isoforms, suggesting a lower activation threshold compared with more mature populations expressing the smaller isoforms. 22hree CD45-deficient models have been generated in mice by independently targeting exons 6, 8 9 9 and 12. 10 All three models display defects in thymic development mediated by increased apoptosis and dysfunctional pre-TCR and TCR signaling. 11Deletion of exon 6 resulted in complete abrogation of CD45 expression on all B cells, while a small population of thymic T cells (3-5%) retained CD45 expression. 8Deletion of exon 9 resulted in the complete loss of CD45 on B and T cells. 9Lastly, targeting exon 12, which encodes part of the extracellular domain in all isoforms, resulted in the complete loss of CD45 surface expression on all cell types. 10Although this CD45 null mouse strain (Cd45 À/À ) lacks membrane bound CD45, immunoblotting revealed the presence of a truncated protein ($150 kDa). 10Despite this, the exon 12-targeted Cd45 À/À phenotype is identical to the exon 9-targeted model.
[10][11] T cell loss is due to dysfunctional TCR signaling leading to a reduction in double positive and single positive thymocytes, as well as a defect in negative selection following antigen stimulation. 23Thus, the CD45-deficient T cells that do develop and reach the periphery are highly autoreactive.B cell development halts at T2 transitional stage in the spleen as B cells fail to express IgD. 24lthough CD45 is required for T and B cell development, this does not appear to be the case for NK cells. 19,25Splenic NK cells in Cd45 À/À mice were increased 4-5-fold and actively dividing, compared with NK cells in control mice, suggesting that CD45 is a regulator of NK cell homeostasis, functioning to limit NK cell numbers. 19,26Additionally, splenic NK cells from Cd45 À/À mice retained cytotoxic ability but exhibited defective IFNγ production. 19,27he in vivo expansion and/or persistence of NK cells in CD45-deficient mice suggests that CD45 is a potential target for immunotherapies designed to enhance NK cell antitumor activity.Adoptive NK cell therapies are considered safe, with no cytokine release syndrome or graft-versus-host disease observed in clinical trials so far and have been successfully used against relapsed or refractory CD19positive cancers (non-Hodgkin's lymphoma or chronic lymphocytic leukemia. 28,29Targeting CD45 in NK cell-based immune cell therapy has the potential to bypass the severe T and B cell immunosuppression expected with whole body inhibition of CD45 activity.Here we investigated the role that CD45 plays in regulating NK cell homeostasis and development.

RESULTS
Mice lacking CD45 have reduced numbers of mature T and B cells, but increased numbers of mature M2 stage NK cells with enhanced in vitro cytotoxic capability. 19,26However, previous studies did not address whether the increased NK cell maturation in Cd45 À/À mice was related to IL-15 signaling, or at what stage in NK development CD45 limited expansion of the NK cell pool.
In vitro proliferation of Cd45 À/À NK cells is compromised in response to IL-15 To further assess the role of CD45 downstream of IL-15 signaling, splenic NK cells were purified from wild-type and Cd45 À/À mice, labeled with CTV, and NK cell division tracked over 8 days in the presence of various hIL-15 concentrations (Figure 1b-d and Supplementary figure 2b).Wild-type NK cells divided on average 6.3 times (50 ng mL À1 ; 8 days), consistent with previous data. 35In contrast, Cd45 À/À NK cells were not able to reach division 7, and instead divided $5.8 times, with most cells achieving fewer divisions (Figure 1b).The normalized Cd45 À/À NK cell cohort number (relative to initial number of cells that undergo division) was also decreased (0.226 AE 0.043 versus 0.917 AE 0 0.0367; Cd45 À/À versus wild-type).At days 5 and 8, the total numbers of Cd45 À/À NK cells were significantly reduced compared with wild-type NK cells ($4.6-fold in 50 ng mL À1 IL-15; Figure 1c, d).This result was not consistent with the described in vivo expansion of the NK cell pool in Cd45 À/À mice. 19
In contrast to the reduced in vitro proliferation of Cd45 À/À NK cells (Figure 1b-d), the enhanced in vivo expansion of the Cd45 À/À NK cell compartment was recapitulated in the mixed bone marrow chimeras.Changes in IL-15 levels or IL-15 receptor beta subunit (IL-2Rβ; CD122) expression had no impact on either the expansion of the Cd45 À/À NK cell compartment or the skewing towards mature M2 effector NK cells.

Adoptively transferred Cd45 À/À NK cells display reduced expansion in lymphocyte depleted mice
Although the competitive bone marrow chimeras confirmed that expansion of the NK cell compartment in Cd45 À/À mice was intrinsic to hematopoietic cells, the chimeras did not address whether expansion was intrinsic to the NK cell lineage or resulted from a defect at an earlier stage of hematopoiesis.Competitive, adoptive NK cell transfers were performed to determine whether CD45-deficient NK cells were able to expand and accumulate in vivo.Cd45 À/À and Cd45 +/+ NK cells were purified, labeled with CTV, mixed (1:1) and adoptively transferred into NK cell-deficient mice (Mcl1 fl/fl Ncr1 iCre/+ ), 36 T and B cell-deficient mice (Rag-1 À/À ) or completely alymphoid mice (Rag-2 À/À γc À/À ).
These observations were in sharp contrast to the original reported expansion of NK cells, 19 and the bone marrow chimera data in Figure 1, which suggested that deletion of CD45 conferred an intrinsic proliferative advantage to NK cells.Instead, these data indicated that CD45 acted prior to NK cell development.To test this hypothesis, mice were generated with conditional deletion of Cd45 in NK cells.
Conditional deletion of CD45 in NK cells does not disrupt NK cell homeostasis, but does skew maturation towards mature M2 cells CRISPR/Cas9 gene editing was used to generate mice carrying Cd45/ptprc conditional alleles (loxP).Mice carrying the floxed Cd45 allele (Cd45 fl/fl ) were crossed with mice containing the cre recombinase gene under control of the Ncr1 gene promoter (encoding NKp46; Ncr1 iCre/+37 ), to generate Cd45 fl/fl Ncr1 iCre/+ mice.This resulted in deletion of Cd45 exon 14 from the immature stage of NK cell development (Supplementary figure 1).Lack of CD45 surface expression on Cd45 fl/fl Ncr1 iCre/+ NK cells was confirmed by flow cytometry (Figure 3a).Surface expression of CD45 remained unaltered in the lymphocyte (Supplementary figure 3) and myeloid (Supplementary figure 4) compartments.Additionally, T cell (Supplementary figure 5), B cell (Supplementary figure 6) and NK cell (Supplementary figure 7) development was normal in Cd45 fl/fl Ncr1 iCre/+ mice.
In contrast to global deletion of Cd45, conditional deletion of Cd45 in NK cells did not result in expansion of the NK cell compartment.NK cell numbers in the spleen of Cd45 fl/fl Ncr1 iCre/+ mice were comparable to Cd45 fl/fl Ncr1 +/+ and Cd45 +/+ Ncr1 iCre/+ control mice, and within a normal range (Figure 3c, d).This was not due to an inability to detect NK cells, as a CD45 negative NK population was present in Cd45 fl/fl Ncr1 iCre/+ mice (CD3 À NK1.1 + NKp46 + ) throughout NK cell maturation (Figure 3a, e).This further suggested that the dramatic increase in NK cell numbers observed in Cd45 À/À mice resulted from changes in early NK development, prior to expression of NKp46.
Collectively, the data suggested that loss of CD45 likely impacted two stages in early hematopoietic development, resulting in a modest increase in LSK/LT-HSC and CLP populations, potentially contributing to the expansion of the NK cell compartment in Cd45 À/À mice.
Interestingly, although the NK cell expansion observed in Cd45 À/À mice was not present in mice with an NK cell specific-deletion of Cd45, the skewing of NK cell maturation towards a mature M2 stage was maintained.Again, this was independent of IL-15 levels and suggested that CD45 limited NK cell maturation towards an M2 effector stage, via a yet unknown mechanism.This observation may also account for the apparent reduction in NK cell proliferation observed ex vivo.Mature M2 (CD11b + KLRG1 + ) stage NK cells have limited proliferative capacity compared with immature or M1 (CD11b + KLRG1 À ) stage NK cells. 26,39iven M2 cells predominate in NK cell pools purified from Cd45 À/À mice, it is perhaps not surprising that we observed significantly reduced Cd45 À/À NK cell proliferation in vitro.Again, this is disconnected from the in vivo phenotype where Cd45 À/À M2 NK cells cycle faster than control M2 NK cells, as determined by BrdU uptake. 26he analysis of early progenitor populations revealed a potential role for CD45 in limiting the progenitor/stem pool.Previously, we did not observe a defect in the Cd45 À/À NK precursor population, 19 identified as CD122 + NK1.1 À DX5 À CD3 À CD19 À defined by Rosmaraki et al. 40 In the current study, we refined our characterization of the pre-NKP and rNKP populations using markers described by Carotta et al. 41 and Fathman et al. 42 to reveal a defect within the pre and restricted NK progenitor populations. 41,42In addition, we observed an increase within both LSK and CLP Cd45 À/À populations, which was not captured by hematopoietic analysis of exon 6-targeted CD45 deficient mice. 19However, exon 6-targeted CD45 deficient mice retain a small thymic population that expresses CD45, suggesting that the lack of perturbation in the lymphoid compartment may be due to residual CD45 expression. 8his study attributed the perturbed NK cell homeostasis in CD45-deficient mice to developmental defects impacting the LT-HSC and CLP populations.In CD45-deficient mice, most likely the increase in a common progenitor population, coupled with a faster transit through NK cell development and maturation, manifests as increased numbers of circulating mature NK cells.
Although the systemic use of CD45 inhibitors in the clinic would abrogate TCR and BCR signaling, 23,24 both of which are required for an effective anti-cancer immune response, this study identifies a potential clinical application in adoptive immunotherapy.Here we investigated the role that CD45 plays in regulating expression of NKp46, resulting in increased NK cell numbers, predicting that inhibition of CD45 or deletion of the ptrprc/Cd45 gene could enhance the yield of NK cells derived from CD34 + umbilical cord blood cells or inducible pluripotent stem cells.Improving the yield of NK cells would help to address one of the technical limitations currently associated with this approach.
While it is clear that deleting Cd45 will enhance the NK cell differentiation, this study reveals that CD45 is important for proper mature NK cell function.Indeed, other studies (e.g.Hesslein et al. 20 ) have shown impaired cytotoxic responses by CD45-deficient murine NK cells.Therefore, an optimal translational intervention would seek to impair CD45 function during cell development/differentiation, and then reactivate it in mature NK cells.Critically, further work is needed to confirm that these observations translate to human NK cell development and do not negatively impact NK cell cytotoxicity.

METHODS Mice
All animal experiments followed the National Health and Medical Research Council (NHMRC) Code of Practice for the Care and Use of Animals for Scientific Purposes guidelines and were conducted in accordance with the regulatory standards approved by the Walter & Eliza Hall Animal Ethics Committee (AEC2019.034,AEC2018.040,AEC2021.011 and SABC) and Monash University Animal Ethics Committee (25 004, 22 111).All strains (Table 1) were maintained on a C57BL/6 background and bred at either the Walter & Eliza Hall Institute or Monash University.

Generation of mice with conditional deletion of Cd45 (Ptprc) in NK cells
Mice with a floxed Cd45 allele (Cd45 ). 37These strains were maintained on a C57BL/6 background and bred at Monash University animal facilities (Clayton).

Organ processing
Blood, BM, lungs, thymus and spleen were collected from 6-8-week-old mice, each organ was processed to a single cell suspension for further analysis as follows.

Blood
Retro-orbital, mandibular or cardiac bleeds were collected in Microvette ® 500 K3 EDTA tubes (Sarstedt, Nümbrecht, Germany), transferred to a 5-mL polypropylene Falcon tube (Corning, New York, NY, USA), and adjusted to a total volume of 1 mL with phosphate-buffered saline (PBS; Gibco, New York, NY, USA).The cell suspension was under-layered with 2 mL of Histopaque-1077 (Sigma Aldrich, St Louis, MO, USA) and centrifuged at 300 g for 15 min at room temperature (RT).
Leukocytes can be found at the interface layer, these were then transferred to a 10-mL Falcon tube, washed twice with ice-cold PBS and resuspended in FACS buffer [PBS, 2% FBS (Bovogen Biologicals, Lot#2009A; Clayton, VIC, Australia), 1 mM ethylenediaminetetraacetic acid (EDTA; Invitrogen™, MA, USA)].Enriched leukocytes were used in subsequent experiments.

Bone marrow
Unless otherwise specified, femurs and tibias were collected, and BM flushed into a 10-mL Falcon tube using a PBS filled syringe.Cell suspensions were passed through a 70-μm cell strainer and centrifuged at 300 g for 5 min at 4°C.The cell pellet was resuspended in 1 mL of red cell removal buffer [RCRB; 156 mM NH 4 Cl (Sigma Aldrich), 11.9 mM, NaHCO 3 (Sigma Aldrich) and 0.097 mM, EDTA] and incubated for 5 min at RT.The cells were washed twice with ice-cold PBS and resuspended in FACS buffer.BM single cell suspensions were used in subsequent experiments.

Lungs
Lungs were minced at RT into small fragments.Minced lungs were thoroughly mixed with 5 mL digestion buffer (1 mg mL À1 collagenase IV and 30 μg mL À1 DNase I in PBS) and incubated for 30 min at 37°C.To further dissociate cells, the digested tissue was forcefully passed through a 70-μm cell strainer with PBS, and the cell suspension was centrifuged at 300 g for 5 min at 4°C.The cell pellet was resuspended in 5 mL of RCRB, incubated for 5 min at RT and then washed twice with ice-cold PBS.

Thymi
Thymi were forcefully passed through a 70-μm cell strainer with PBS, collected in a 10-mL Falcon tube and centrifuged at 300 g for 5 min at 4°C.The cell pellet was resuspended in 5 mL of RCRB, incubated for 5 min at RT and then washed twice with ice-cold PBS.

Spleens
Spleens were forcefully passed through a 70-μm cell strainer with PBS, collected in a 10-mL Falcon tube and centrifuged at 300 g for 5 min at 4°C.The cell pellet was resuspended in 1 mL of PBS and passed through a 70-μm cell strainer with PBS, adjusting the volume to 10 mL in a 10-mL Falcon collection tube, prior to centrifugation at 300 g for 5 min at 4°C.

Flow cytometry analysis
Cell proliferation and intracellular IFNγ data were collected on a FACSVerse (BD Biosciences, Franklin Lakes, NJ, USA) using BD FACSuite software.NK cell progenitor data were collected on a FACSymphony (BD Biosciences) using BD FACS Diva software.Data for all other experiments were collected on a BD LSR Fortessa X-20 using BD FACS Diva software.All analysis and statistics were performed using FlowJo v10 software and Prism GraphPad, respectively.Flow cytometric analysis of viable cells was performed by excluding debris (FSC-A; Forward Scatter-Area low events), doublets (FSC-A versus FSC-H; Forward Scatter-Height) and dead cells (dead-cell indicator dye negative).Cell subsets were gated based on surface marker expression (Table 2).

In vitro proliferation assays
NK cell proliferation studies, including cohort number and mean division number determination, were performed as described, 35 and based on previously published methods using T cells. 50In short, purified NK cells are incubated with 0.1 nM of CellTrace Violet (CTV; Invitrogen™) in phosphate buffered saline (PBS) supplemented with 0.1% bovine serum albumin (Sigma Aldrich) for 20 min at 37°C.Labeled cells were subsequently washed twice with ice-cold NK complete media [IMDM (Gibco) containing 10% v/v heat-inactivated FBS, 1% v/v penicillin-streptomycin, 1× GlutaMAX™ (Gibco), 55 μM β-mercaptoethanol (Sigma Aldrich)] to quench the labeling reaction.Labeled NK cells (4000-10 000 per well) were seeded into 96-well round-bottom plates and cultured at 37°C in 5% CO 2 .A mixture of propidium iodide (PI; 200 nM; Sigma Aldrich) and 123count eBeadsTM (5005 beads/well; Invitrogen™) was added to cultures prior to flow cytometric analysis.Cells were analyzed daily by flow cytometry.

Generation of bone marrow chimeras
Bone marrow chimeras were generated by lethally irradiating (2 × 550 rads) host mice and reconstituting by I.V. injection into the tail vein with 6 × 10 6 donor BM cells for 100% chimeras, or 3 × 10 6 control donor BM and 3 × 10 6 experimental donor BM for mixed chimeras.Generally, host and donor mice express allelic variants of the panhematopoietic cell marker CD45, also known as CD45.1 (Ly5.1) and CD45.2 (Ly5.2).Donor BM was collected from femur, tibia, pelvis, radius ulna, humerus and cervical vertebrae by crushing the bones in a mortar with a pestle and PBS.BM suspension was passed through a 70-μm cell strainer, centrifuged at 300 g for 5 min at 4°C, resuspended in PBS and the cells counted, prior to injection.Immediately postirradiation, neomycin treated water (2 mg mL À1 neomycin sulfate in drinking water; Sigma Aldrich) was made available to mice for up to 3 weeks.Reconstitution of the hematopoietic compartment was checked at 6-week post-bone marrow transplantation.Peripheral blood was obtained by retro-orbital bleeding using a sterile hematocrit capillary tube (Sarstedt).Post-processing, blood samples were stained with anti-CD45.1 and anti-CD45.2for flow cytometric analysis of the leukocyte compartment.

Table 1 .
Genetically modified mouse strains a Strain name as it appeared in first publication.b Common strain name.c Strain generation and characterization.

table 1 .
Cre-lox deletion of exon 14 results in a premature stop codon in exon 15.Furthermore, translation of the compromised mRNA sequence predicts a truncated Cd45 protein, which lacks a transmembrane and cytoplasmic domain and is expected to be non-functional.Established Cd45 fl/fl mice were crossed with Ncr1 iCre/+ mice to conditionally delete Cd45 from mature NK cells (Cd45 fl/fl Ncr1 iCre/+