Transfusable neutrophil progenitors as cellular therapy for the prevention of invasive fungal infections

Abstract

The use of mature neutrophil (granulocyte) transfusions for the treatment of neutropenic patients with invasive fungal infections has been the focus of multiple clinical trials. Despite these efforts, the transfusion of mature neutrophils has resulted in limited clinical benefit, likely owing to problems of insufficient numbers and the very short lifespan of these donor cells. In this report, we employed a system of conditionally immortalized murine neutrophil progenitors that are capable of continuous expansion, allowing for the generation of unlimited numbers of homogenous granulocyte-macrophage progenitors (GMP). These GMP were assayed in vivo to demonstrate their effect on survival in two models of invasive fungal infection: candidemia and pulmonary aspergillosis. Mature neutrophils derived from GMP executed all cardinal functions of neutrophils. Transfused GMP homed to the bone marrow and spleen, where they completed normal differentiation to mature neutrophils. These neutrophils were capable of homing and extravasation in response to inflammatory stimuli using a sterile peritoneal challenge model. Furthermore, conditionally immortalized GMP transfusions significantly improved survival in models of candidemia and pulmonary aspergillosis. These data confirm the therapeutic benefit of prophylactic GMP transfusions in the setting of neutropenia and encourage development of progenitor cellular therapies for the management of fungal disease in high-risk patients.

Graphical Abstract

Summary Sentence:

Engineered neutrophil progenitors protect against the invasive fungal pathogens, Candida, and Aspergillus, in a neutropenic mouse model.

Introduction

The innate immune system functions to promptly recognize and eradicate microbial and fungal pathogens. Neutrophils comprise ~70% of circulating leukocytes under homeostatic conditions and function as crucial sentinels that act to eliminate pathogens and communicate with cells of the adaptive immune system [1]. Patients with abnormally low absolute neutrophil counts (ANC), such as those with congenital neutropenia, those receiving cytotoxic chemotherapy, and those on extended therapy with immunosuppressive drugs, face an increased risk of fatal infection by opportunistic pathogens [2],[3],[4]. Fungal infections are of growing concern in these patient populations due to their increasing prevalence, presence in healthcare settings, high lethality, and drug resistance, such as in the case of the emergent epidemic species Candida auris [5][6]. Here we focus on the most common human fungal pathogens, Aspergillus fumigatus and Candida albicans, which are ubiquitous [7],[8] and can cause severe opportunistic infections in neutropenic patients, especially in the settings of chemotherapy and allogeneic hematopoietic transplantation [9],[10].

Candida species are the fourth most common nosocomial bloodstream infection (of which C. albicans is the most common), with a mortality rate as high as 47% [11],[12]. A. fumigatus infections occur via the inhalation of spores that colonize the lung tissue in susceptible hosts [8]. Invasive pulmonary aspergillosis carries a mortality rate of up to 58%, even higher in bone marrow transplant recipients where mortality can be as high 86% [13]. Neutropenia is the single most significant risk factor in the development of bloodstream candidiasis and invasive aspergillosis [14],[10],[15].

Boosting a patient’s circulating neutrophil population with donor cells is an attractive prophylactic strategy to prevent fungal infections in at-risk patients. Pre-clinical animal models have demonstrated some efficacy in protecting against invasive infections [16]. Granulocyte transfusions have been attempted in clinical trials of neutropenic patients suffering from infections resistant to standard antimicrobial therapies [17]. In these trials, donors are typically treated with corticosteroids and/or granulocyte colony stimulating factor (G-CSF) to boost their circulating neutrophil count prior to donation. The collected neutrophils are transfused immediately into recipients as no capacity currently exists to preserve harvested cells [18],[19]. In one recent clinical phase 2, randomized, controlled trial, the use of allogenic myeloid progenitor cell product showed a decreased incidence of infections for neutropenic AML patients undergoing induction chemotherapy [20]. Small case reports and series have supported the efficacy of granulocyte transfusions [17],[21][22][23],[24],[25],[26]. However, larger trials [27],[28] have unfortunately failed to demonstrate a consistent and meaningful clinical benefit, and some have even suggested a potential detrimental effects of mature granulocyte transfusions [29].

Efficacy may be limited by two key hurdles: 1) the number of neutrophils that can be obtained from donors and 2) the effective function and longevity of the transfused neutrophil. The use of corticosteroids and G-CSF can boost the neutrophil harvest to as high as to 6.7x10¹⁰ cells [30], which approximates the estimated normal daily neutrophil production of a healthy adult [29]. However, the short lifespan of the mature neutrophil remains an obstacle for effective granulocyte transfusion, as transfused cells are quickly cleared in the recipient and multiple would be required on a daily basis to manifest a clinical benefit and demonstrate infection control [17]. In healthy individuals, peripheral blood neutrophils have a pre-programmed circulating lifespan estimated at 8-24 hours [31]. This duration may unfortunately be further reduced by the steroids and G-CSF required for mobilizing and the physical manipulation of the harvesting process. One recipient generally requires more than one granulocyte donor to provide the required number of daily transfusions, raising another major obstacle of donor availability. In addition to the challenges in obtaining sufficient neutrophil quantities, donated neutrophils may have reduced functional capacity, specifically with respect to phagocytic and anti-fungicidal activity [32], suggesting that transfused neutrophils are already compromised before reaching the vulnerable recipient.

To circumvent these challenges and improve the potential success of a granulocyte cellular therapy, we hypothesized that the ideal transfusable product (a) would be a neutrophil progenitor, rather than a mature cell, (b) could be expanded to the requisite numbers in the progenitor state, (c) could be cryopreserved, and (d) could expand and differentiate to a fully mature and functional granulocyte within the recipient’s normal bone marrow environment.

The concept of ex vivo expansion or immortalization of hematopoietic stem cells (HSC) and progenitor cells has been explored [33], [34],[35]. Isolation and transfusion of precursor stem cells committed to the myeloid lineage (CMP, common myeloid progenitors) and GMP have successfully demonstrated protection in mouse models of neutropenia with bacterial and mold infection [36],[37],[38]. While these data provide the essential proof of concept showing precursor myeloid cells can provide protective benefit, there remain significant limitations mainly centered around the isolation of a rare cell population, with limited ex vivo expansion even under optimal conditions, resulting in a limited number of cells available for therapeutic purposes.

For these reasons, we investigated use of homeobox (Hox) family members, which are transcription factors with predictable expression patterns in HSCs and progenitor cells. Specific members of the HoxA and HoxB clusters are upregulated in HSCs and become silenced during the differentiation process to mature progeny [39]. Interestingly, the overexpression of certain homeobox genes has demonstrated the unique ability to stall hematopoietic stem cells in an undifferentiated progenitor state and allow for HSC expansion ex vivo [40]. A method to conditionally immortalize macrophage and neutrophil progenitors by transducing granulocyte-myeloid progenitors (GMP) with an estrogen receptor-Hoxb8 (ERHoxb8) fusion protein has been well-established [35]. Myeloblasts expressing the ERHoxb8 fusion protein remain in a perpetually and conditionally immortalized state of self-renewal in the presence of estrogen and functionally active Hoxb8, while the removal of estrogen inactivates the Hoxb8 activity and the GMPs resume their normal and synchronized differentiation to mature neutrophils.

Unlike primary CMP or GMP, these ERHoxb8 GMPs can be continually expanded ex vivo, can be cryopreserved to address the challenges of long-term product storage, can be transfused with the ability to reconstitute the recipient with a fully functional in vivo differentiated neutrophil. Here we demonstrate the utility of these conditionally immortalized GMPs as a transfusable cellular therapy, successful in the prophylaxis of two neutropenic mouse models of invasive fungal infection.

Material and Methods

Reagents

Reagents were purchased from Thermo Fisher Scientific except when noted.

Cell line production and culture

The ERHoxb8 cell line was generated as previously described [35] from bone marrow isolated from the UBC-GFP mouse [41] (JAX 004353). In brief, femurs were crushed, and the bone marrow filtered prior to Ficoll-Paque Plus density centrifugation to collect mononuclear cells [41]. Cells were placed into culture for ~36 hours (IMDM + 10% fetal bovine serum + 10 ng/ml stem cell factor + 10 ng/ml interleukin-6 + 5 ng/ml interleukin-3) to stimulate the cells into cycle. 250,000 cells were transduced with MSCVneo-HA-ERHoxb8 retrovirus (ecotropic) by spinoculation (1000g, 2 h, 22°C) using Lipofectamine (1:1000; Gibco) as a polycationic compound (can also be substituted by polybrene used at a final concentration of 8 ug/ml). Transduced progenitors were maintained in RPMI + 10% fetal bovine serum + stem cell factor (SCF) + 0.5 μM β-estradiol (E2, estradiol; Sigma E2758). SCF can be supplied as recombinant protein at a concentration of 50-100 ng/ml, though we typically use 1% conditioned media from a CHO cell line producing murine SCF to provide the same final concentration. Serial passaging of nonadherent cells over ~3-4 weeks enriched for conditionally immortalized progenitors while the proliferative capacity of normal and non-transduced cells were exhausted. These populations underwent single cell cloning in 96-well round bottom plates, and GMP clones that produced a homogenous population of mature neutrophils upon maturation were selected. To generate mature ERHoxb8 neutrophils ex vivo, cells were washed from of β-estradiol by pelleting (500g, 5 minutes) and washing once with PBS prior to seeding into media without β-estradiol for four days.

Surface fluorescence labeling and flow cytometry

Since the ERHoxb8 cells were generated in the background of the UBC-GFP mouse [41] the cells constitutively express GFP at all stages of differentiation. GMP clones and mature neutrophils were incubated with anti-GR-1 and CD11b (Biolegend) in FACS buffer (2% FBS, 1 mM EDTA in PBS) for 45 min (4°C), washed and analyzed on an Accuri or BD Calibur flow cytometer. For CFSE labeling, 0.125 μM of CFSE was added to GMP cell suspension for 20 min and then washed. CFSE was assessed by flow cytometry and MFI derived to determine number of cellular divisions and expressed on a log2 scale.

Candida albicans and Aspergillus fumigatus preparation

Candida albicans SC-5314 (ATCC, Manassas, VA) and Caf2Dtom C. albicans [42] stocks were maintained at −80°C. Cultures were initiated from frozen stock and grown in YPD media with ampicillin (100 ug/ml) and shaken overnight at 30°C. Yeast were washed and resuspended in PBS. Serial dilutions were prepared and counted using an automated cell counter (LUNA, Logosbio, Annandale, VA). Heat-killed C. albicans were prepared using incubation at 95°C for 10 min. For A. fumigatus (gift from Robert Cramer, Dartmouth College), frozen stock of Aspergillus was made onto YPD agar and grown for 1 week at 30°C. Conidia were harvest by introducing PBS with 0.1% Tween-20 to the surface of the YPD growth and gently brushed with a cell scraper. Conidia were harvested, washed in PBS and counted using the LUNA automated counter.

Fungal cocultures and confocal imaging

GMPs were deprived of E2 for 4 days to induce maturation into neutrophils. Fungi were pre-grown on poly-D-Lysine coverslips prior to coculture with neutrophils. C. albicans SC-5314-dTomato was grown for 2 hours at 37˚C 5% CO₂ in RPMI 10% FBS. A. fumigatus Af293::RFP was grown for 9 hours at 37˚C 5% CO₂ in RPMI. Fungi were cocultured with mature neutrophils for 1 hour, stained with DAPI, and imaged on a Zeiss 780 confocal microscope. Z-stack images were collected, and the stacks merged using the ImageJ Maximum Intensity Projection algorithm.

In vivo experiments

Following approval from the MGH Institutional Animal Care and Use Committee, mice were housed in the Thier Research Building Animal Facility. Mice were irradiated using either a fixed-cesium source irradiator or an x-ray irradiator. An irradiation dose of either 450 cGy or 900 cGy (450 cGy x 2 doses, with 4 hours between doses) was used. Mice were rested for one day following radiation and were then transfused with GMPs via retro-orbital injection under isoflurane anesthesia. Transfusions were provided as a maximum of 20 million cells in a volume of 200 μL of PBS, given 1-2 times daily as indicated with additional infusions provided as single daily doses for the two days prior to pathogen challenge and the two days following pathogen challenge.

To perform the GMP distribution at homeostasis, mice were transfused with a loading dose of ERHoxb8 UBC-GFP GMPs which were labelled with a far-red fluorescent succinimidyl ester dye. Mice were then euthanized at the time points indicated and bone marrow and spleen harvested for flow cytometry analysis.

To identify ERHoxb8 cell expansion and maturation in vivo, irradiated mice were transfused with either 20 million ERHoxb8 UBC-GFP GMPs or with 20 million primary neutrophils harvested from the bone marrow using a negative isolation kit (Mouse Neutrophil Enrichment Kit, Stem Cell) from CD45.1^STEM mice [43]. Mice were then euthanized at day 1,2,3, or 4 post transfusion and peripheral blood, spleen, and bone marrow, cells were harvested from each. Samples were processed as described below. Primary CD45.1^STEM mouse neutrophils were labelled with CD45.1-AF488, CD11b-PE, and Ly6G-BV421 while UBC-GFP neutrophils were identified by GFP fluorescence and confirmed to be neutrophils by CD11b-PE, and Ly6F-BV421 (BioLegend).

To determine in vivo homing of GMP cells to sterile inflammation, mice were then transfused with a single dose of 20 million ERHoxb8 UBC-GFP GMPs one day after irradiation. Mice were then challenged with 1 ml thioglycolate in the intraperitoneal cavity one day prior to peritoneal lavage and cell harvest at the points indicated [44],[45]. Peritoneal fluid was collected by lavage by instilling 10 mLs FACS buffer and aspirating contents. Resulting cells suspenstions were immunostained and analyzed by flow cytometry.

For pathogen challenge, on day 4 after UBC-GFP GMP infusion, C. albicans or A. fumigatus, at the indicated inoculum, were introduced via tail vein injection or intratracheal instillation, respectively.

To measure kidney organ involvement with C. albicans, mice were euthanized at the indicated times following challenge. Kidneys were dissected and placed into buffered formalin for histopathology analysis. The contralateral kidney was homogenized in 1 mL of PBS using a homogenizer (TH 115, Omni International). Serial dilutions of the kidney homogenate were performed and 100 μL of each dilution was plated onto agar plates and spread using sterile glass beads. Plates were incubated at 30°C for 2 days prior to colony counting.

Histopathologic analysis for tissue sections

Following formalin fixation, kidney were processed for pathology evaluation and stained with H&E for inflammation and Grocottt’s methenamine silver (GMS) for identification of C. albicans hyphae. Slides were serially coded without revealing the treatment groups and a pathologist performed a blinded evaluation. The extent of inflammation was estimated as average of percent neutrophilic inflammation at 20x magnification of all fields of the kidney section. Fungal burden was estimated using the average of fungal involvement at 20x magnification in all fields of a kidney section. Average number of fields examined per section is 10.

Bone Marrow, Spleen, Peripheral Blood, and Peritoneal Fluid Processing

Bone marrow cells were prepared by crushing both femurs in ~8 ml of FACS buffer and filtering through a 40 μm cell strainer. 2 mL of the remaining solution was transferred to a 50 ml conical and 8 mL of ACK lysis buffer (ThermoFisher) was added for red blood cell lysis for 5 minutes at room temperature. FACS buffer was added to 50 ml and the cells pelleted (500g, 5 minutes) and then resuspended in 1 mL of FACS buffer.

Spleens were prepared by macerating through a 40 μm cell strainer into a 6 cm dish using the plunger from a 5 ml syringe into a volume of ~4 mL of FACS buffer. One mL of cells was transferred to a 50 ml conical and 4 mL of ACK lysis buffer (ThermoFisher) was added for red blood cell lysis (5 minutes at room temperature). FACS buffer was added to 50 ml and the cells pelleted (500g, 5 minutes) and then resuspended in 1 mL of FACS buffer.

Peripheral blood was collected either by retro-orbital bleeding immediately before euthanasia (under isoflurane anesthesia) or via cardiac puncture immediately after euthanasia; whole blood was collected into EDTA tubes. 200 μL of blood were transferred to a 1.5 mL tube containing 1 mL of ACK lysis buffer and slowly mixed on a rocker for 5 minutes at room temperature. The cells were pelleted (2000g, 5 minutes) and resuspended in ~200 ul of FACS buffer using a P200 pipet.

Peritoneal fluid samples were collected by injecting 8 mL of PBS into the peritoneal cavity of the mouse immediately after euthanasia using a ~20g needle. The mouse belly was then massaged for 20 seconds to mobilize cells into the instilled fluid and the peritoneal fluid was recovered using a 10cc syringe and a 16G needle (the larger caliber needle seems to help with recovery and avoiding clogs). The cells were pelleted and resuspended in FACS buffer for flow cytometry.

Statistics

Statistical calculations were performed using GraphPad Prism 7 software (La Jolla, CA). Data was analyzed using two-tailed unpaired t-test or by one-way analysis of variance (ANOVA), and survival studies by Log-Rank test. Conclusions of significant differences were determined as a p value ≤0.05.

Study Approval

Study approval from the MGH Institutional Animal Care and Use Committee (2017N000058) was obtained.

Results

Fully functional neutrophils are derived from ERHoxb8 granulocyte-macrophage progenitors (GMP)

To develop a conditionally immortalized GMP line, primary bone marrow from a UBC-GFP (Ubiquitin promoter driving the expression of GFP) mouse was transduced with retrovirus to constitutively express the ERHOXb8 fusion protein. Here, the hormone binding domain (HBD) of the estrogen receptor alpha is fused with full length Hoxb8. In the presence of β-estradiol, ERHoxb8 is functional and results in differentiation arrest and continued expansion of GMPs (Figure 1A). In the absence of β-estradiol, the ERHoxb8 is sequestered in the cytoplasm, is non-functional, and the GMPs synchronously resume their normal maturation process to terminally differentiated neutrophils.

Figure 1.

Granulocyte monocyte progenitors (GMPs) transduced with the ERHoxb8 system mature and respond to C. albicans in vitro. (A) Schematic illustrating the ERHoxb8 cassette insertion into murine mononuclear cells which results in the conditional immortalization of GMPs in the presence of estrogen and stem cell factor. With the removal of estrogen from the media, GMPs mature into functional neutrophils in vitro. (B) Representative Wright Giemsa images of GMPs and polymorphonuclear cells (PMNs) demonstrating morphological differences between cell states. Scale bar = 10 μm. (C) PMN maturation as demonstrated by CD11b and GR-1 staining by flow cytometry. (D) UBC-GFP expressing neutrophils were co-incubated with C. albicans expressing dTomato for 1 hour. Neutrophils were then labelled with DAPI and images were taken with a Zeiss 780 confocal microscope. Scale bar 20 μm.

Of note, the ERHoxb8 was engineered with a single amino acid mutation within the hormone binding domain (G400V) which greatly reduces the affinity to estradiol. This change ensures that the physiologic (e.g., picomolar) concentrations of estrogen (in both male and female mice) are incapable of activating the ERHoxb8 protein. Supraphysiologic concentrations of E2 (e.g., 0.5 μM) are required to maintain activation of the ERHoxb8 in vitro [46],[35].

Following single cell cloning, GMP clones committed to differentiation along the neutrophil lineage were selected. As previously reported, some clones produce a mixed population of neutrophils and monocytes, and a rare number of clones are specifically monocyte/macrophage biased. Neutrophil maturation and function were confirmed using several approaches: morphology by Wright-Giemsa staining, detection of cell surface markers of maturity by flow cytometry, and recognition and binding of C. albicans. Immature progenitors and mature neutrophils exhibited their characteristic morphology. Immature cells demonstrated large nuclei and sparse cytoplasm, while the mature cells formed the condensed and circular nuclei typical of mature mouse neutrophils with a pale cytoplasm and granules (Figure 1B) [35]. As ERHoxb8 progenitors matured to neutrophils, they demonstrated the expected upregulation of cell surface markers CD11b (Mac1) and Ly6C/Ly6G (Gr-1) (Figure 1C). One functional hallmark of mature neutrophils is their ability to recognize and bind fungal hyphae [47]. Following co-culture, confocal microscopy confirmed that the UBC-GFP ERHoxb8 neutrophils readily engaged C. albicans hyphae (hyphae are red fluorescent-labeled, Figure 1D).

ERHoxb8 GMPs undergo a predictable pattern of expansion in vitro and in vivo

Normal GMPs undergo multiple cell divisions during their differentiation process, giving rise to many neutrophils. Here we assessed the kinetics of GMP cell division as well as the number of divisions en route to mature neutrophils. ERHoxb8 cells were labeled with carboxyfluorescein succinimidyl ester (CFSE), a dye which is then divided equally between daughter cells such that fluorescence intensity is inversely proportional to the number of divisions (Figure 2A) [48]. Using flow cytometry to quantify the slope of fluorescence decay, we found that this clone of GMPs divided approximately once every 10 hours (Figure 2B). GMPs divided continuously as self-renewing progenitors in the presence of E2. Once out of E2, GMPs completed an estimated 8 divisions during the maturation process prior to cell cycle arrest as terminally differentiated neutrophils.

Figure 2.

GMPs undergo cell division in vitro and in vivo to mature into functional PMNs. (A) Schematic demonstrating the loss of fluorescence in GMPs labelled with the cell tracing dye CFSE. (B) GMPs and PMNs were stained with CFSE for 20 minutes before washing. Mean fluorescence intensity was measured via flow cytometry. The relationship between cell division and time was calculated for GMPs in media containing estrogen, R² = 0.9998. (C) Wild type C57BL/6 mice were irradiated and injected with 20 million UBC-GPF GMPs labelled with a far-red fluorescent CFSE dye. ERHoxb8 progenitors undergo 5 cell divisions as demonstrated by CFSE labelling from cells harvested from the bone marrow and spleen. Mean fluorescence intensity measured by flow cytometry. (D) Wild type C57BL/6 mice were irradiated as described and injected with 20 million UBC-GFP or 20 million primary PMNs harvested from CD45.1^STEM mice. Each day following transfusion, cells from the bone marrow (BM), spleen, and peripheral blood (PB) were harvested from each group. Flow cytometry was used to identify CD45.1^STEM and UBC-GFP PMNS. N = 4 mice with UBC-GFP each day, 3 mice with CD45.1^STEM on day 1 and 2, 2 mice on day 3. Mean and SEM shown, analyzed by one-way ANOVA with Kruskal-Wallis post-test. (E) To examine UBC-GFP PMN recruitment, thioglycolate was injected in the intraperitoneal cavity as a method of inflammation one day prior to PMN harvest from the peritoneal fluid (PF). Fluorescence was measured by flow cytometry. N = 5 mice on day 2 and day 6, 4 mice on day 4. Mean and SD shown, analyzed by one way ANOVA with Turkey’s post test **p = 0.0025, ***p = 0.0006

To determine if the same kinetics were maintained in vivo, irradiated mice were transfused with CFSE-labeled ERHoxb8 GMPs. The ERHoxb8 cells in the bone marrow, spleen, and peripheral blood were analyzed 1, 2, 4, 8, 16, and 28-days following transfusion. Compared to in vitro division, the slope of CFSE fluorescence decay was steeper in vivo suggesting the possibility of an even more rapid cell divisions en route to mature neutrophils (“expected line” Figure 2C). The transfused cells differentiated and were cleared from the bone marrow and spleen by day 8 such that GFP+ cells could no longer be detected (at a sensitivity of approximately 1 in 10,000 cells) suggesting that these GMP do not permanently engraft in the bone marrow but transiently exist until neutrophil maturation is complete. Furthermore, the ERHoxb8 GMPs could be recovered from the bone marrow and peripheral blood by day 4 and to a smaller extent from the spleen (Figure 2D). Primary neutrophils transfused from CD45.1^STEM mouse [43] diminished rapidly over time; whereas the UBC-GFPs continued to expand and were recovered in greater numbers in the bone marrow and peripheral blood by day 4 post transfusion reflecting the potential of precursor GMP additional expansion in vivo and enhanced cell longevity.

ERHoxb8 GMP neutrophils respond to inflammatory stimuli in vivo

Having demonstrated that GMPs mature into fully differentiated neutrophils, we assayed if these cells could respond to inflammatory cues in vivo. Four days following transfusion, (to allow for maturation of the UBC-GFP GMPs within the irradiated mice), thioglycolate broth was instilled intraperitoneally one day prior to cell harvest as a model of sterile inflammation at indicated time points. The presence of transfused neutrophils within the marrow, spleen, and peritoneal lavage fluid was assessed at day 2,4 or 6 post transfusion of GMPs. The transfused ERHoxb8 progenitors underwent complete maturation and were recruited to the intraperitoneal cavity by 6 days post GMP transfusion (Figure 2E). This data suggests that the conditionally immortalized progenitors underwent complete and terminal differentiation to mature cells and they are recruited in an in vivo sterile inflammation challenge in a similar fashion to the primary neutrophils.

ERHoxb8 GMP transfusions prolong survival in a model of disseminated Candida infection

After establishing that ERHoxb8 neutrophils engage with Candida in vitro (Figure 1D), and can respond to sterile inflammation in vivo, we sought to determine if UBC-GFP GMP cells were protective in a disseminated model of live Candida albicans. Following irradiation, mice were transfused with ERHoxb8 GMPs along two schedules: two GMP transfusions on day 1 or two transfusions on day 1 followed by subsequent daily GMP doses. On day four, the mice were challenged with live C. albicans delivered by tail vein injection and followed for survival (Figure 3A). Mice that received the single loading dose of GMPs on day 1 showed increased survival (p=0.0004), as did mice that received the loading dose and a daily cell transfusion (p=0.0001) (Figure 3B). The loading dose followed by daily doses improved survival when compared to the no transfusion group but was not significantly different from the single loading dose of GMPs (p=0.25).

Figure 3.

GMPs differentiate into functional neutrophils in vivo and extend host survival during C. albicans infection. (A) Schematic illustrating murine challenge. Briefly, wild type C56BL/6 mice were irradiated (XRT) as described and injected with 20 million UBC-GFP GMPs retro-orbitally with a single loading dose or with the initial loading dose and an additional daily dose thereafter. Mice were challenged with 800 C. albicans intravenously four days after the loading dose of GMPs. (B) Survival curve of mice infected with C. albicans. N=4-8 mice per group. Images created with BioRender.com

The kidney is a common and reproducible site of fungal involvement in disseminated C. albicans infection [49]. To further elucidate the survival advantage following GMP transfusion, we examined fungal infiltration of the kidneys. Histopathologic examination of tissue revealed substantial areas of necrosis and loss of normal kidney architecture in non-GMP transfused animals. These areas of necrosis were paired with extensive Candida abscess formation noted by large areas of hyphae infiltration as visualized by Grocott’s Methenamine Silver (GMS) stain (Figure 4A). In contrast, in animals who received GMP transfusions, the areas of kidney necrosis were significantly smaller and more restricted, with fewer Candida visualized with silver stain (Figure 4A). Of interest, in the transfused recipients, neutrophils can be seen accumulating at the border of Candida abscesses and are conspicuously absent in the neutropenic non-transfused control mice.

Figure 4.

Transfused neutrophils respond to C. albicans in the kidneys but do not eliminate fungi. C57BL/6 mice were irradiated and injected with GMPs as described previously. Mice were then infected with 500 C. albicans intravenously and kidneys were harvested on day 2 and day 3 post infection. (A) Histology sections of infected kidneys with and without GMP transfusion. (B) Percent scoring of inflammation and (C) cortical fungal burden from kidney histology sections. Pooled kidney sections from 3 experiments, 12 mice per group. Unpaired t-test, p = 0.0011 in inflammation, p = 0.0058 in cortical fungal burden. (D) Colony-forming units (CFUs) of C. albicans from kidneys of mice with and without GMP supplementation. Represents data from one experiment, 6 mice per group. Mann-Whitney test, n.s. between plus and minus GMPs at day 2 and day 3.

Two days after Candida infection, animals were euthanized, and kidney sections processed for blinded pathology assessment of recruited inflammatory cells (H&E) and fungal involvement (GMS). In the mice who had received GMP transfusions, there was a dramatic increase in recruited immune cells as measured by the percent area of inflammatory infiltrate (p=0.0011) (Figure 4B). The cortical fungal burden was also significantly reduced in mice who had received GMP transfusions (p=0.0058) (Figure 4C). Overall, histopathologic analysis of the kidney demonstrated a robust neutrophil response to disseminated Candida following transfusion of ERHoxb8-derived GMPs. Given absence of neutrophils in this lethally irradiated model, the presence of recruited neutrophils are the result of matured cells following transfusion of Hoxb8-derived GMPs. No neutrophils are seen in the control, mock treated mice following fungal challenge.

The kidneys of infected mice were homogenized and plated for formal quantification of fungal burden by colony forming unit (CFU). Despite the histopathologic improvement in organ samples, the C. albicans CFU were not significantly different between transfused and non-transfused animals at both day 2 and day 3 post infection (number of yeasts per gram of tissue, Figure 4D). These data suggest that the transfused neutrophils achieve can control C. albicans hyphal germination, a major factor in tissue destruction, but do not eradicate the initial blastospore inoculation in the target organ.

ERHoxb8 GMP transfusions protect against pulmonary Aspergillus fumigatus infection

After establishing a role for ERHoxb8 GMP transfusions in the protection against disseminated C. albicans, we sought to determine if this protection extended to other fungal species such as the obligate mold, Aspergillus fumigatus. Using confocal imaging, we demonstrated that ERHoxb8 derived neutrophils were capable of recognizing and engaging A. fumigatus hyphae in vitro (red fluorescent), in a similar fashion as primary cells (Figure 5A).

Figure 5.

GMPs extend host survival in an inhalation model of Aspergillus fumigatus infection. (A) UBC-GFPs allowed to mature into functional neutrophils were challenged with A. fumigatus for 1 hour in vitro. Images demonstrate neutrophils interacting with A. fumigatus filaments. Scale bar = 20 μm. (B) Schematic demonstrating wild type C57BL/6 murine challenge with A. fumigatus conidia. Briefly, mice were irradiated (XRT) and transfused with GMPs with one loading dose or with an initial loading dose plus an additional daily dose for four days. Mice were then anesthetized and given 200 A. fumigatus conidia via inhalation and mice were followed for survival. (C) Survival curve of mice infected with A. fumigatus. n=5-6 mice per group. Images created with BioRender.com

In vivo, we used an established inhalation model of A. fumigatus [50],[51],[52]. Irradiated, neutropenic mice were transfused with ERHoxb8 UBC-GFP GMPs. At day 4, A. fumigatus spores were introduced into anesthetized mice via oral tracheal aspiration (Figure 5B). Transfusion with a loading dose of GMP followed by subsequent daily doses significantly increased overall survival as compared to mock transfusion (p=0.0012) (Figure 5C). In contrast to the in vivo Candida model, the daily dosing significantly improved survival when compared to the single loading dose (p=0.0049). In fact, survival from the single loading dose of GMPs was not significantly different from mock transfusion (p=0.73). These data suggest that there may be a fungal-specific, dose-dependent protective effect of the ERHoxb8 GMPs that reflects unique invasive fungal pathogenesis.

Discussion

The concept of granulocyte transfusions to address serious infections in neutropenic patients is highly appealing, though it has proven to be practically challenging resulting in inconsistent clinical benefits. The barriers to granulocyte transfusions include the inherently short lifespan of the mature neutrophil, the reduced functional capacity following mobilization and collection, and the difficulty in achieving adequate numbers of neutrophils (per transfusion as well as the need for daily transfusions) to demonstrate a significant impact on clinical outcome. Here we describe an approach that addresses and circumvent these obstacles via the conditional immortalization of neutrophil precursors that maintain the capacity to mature into functional neutrophils in vitro and in vivo.

The ability to produce unlimited quantities of murine neutrophils is possible [35] using conditional immortalization using a system of enforced expression of ERHoxb8. Upon inactivation of ERHoxb8, these GMPs differentiate into mature neutrophils, demonstrating the expected morphology of primary mouse neutrophils [53]. Along with those morphologic changes, trademark cell surface marker (e.g., CD11b and Gr-1) expression changes accompany differentiation to mature neutrophils. Additionally, these mature neutrophils can perform hallmark functions of a mature neutrophil, including the recognition and killing of target pathogens such as C. albicans and A. fumigatus. This intact functionality is notable, as even donor granulocyte neutrophils may possess inherent deficiencies in the ability to respond to pathogens [54].

The transfusion of neutrophil progenitors rather than mature cells potentially provides a distinct advantage as it allows for the differentiation of neutrophils within the host. Using CFSE tracing, we confirmed that the GMPs undergo a predictable number of divisions in vivo, yielding multiple mature CD11b-positive, c-KIT-negative neutrophils for every transfused GMP cell. These GMPs matured and expanded in the bone marrow and mature neutrophils were eventually found circulating in peripheral blood and in peripheral organs such as spleen. Our data also confirms the advantage of using a precursor cell by comparing GMP to primary neutrophils. Transfusion of harvested primary neutrophils from healthy naïve wildtype mice when compared to GMP, did not persist in the bone marrow, spleen, or peripheral blood demonstrating these primary mature cells were cleared or senesced rapidly.

The in vivo functional capacity of transfused GMPs was demonstrated in several ways. Transfused progenitors homed to the bone marrow, where they underwent differentiation and were then capable of mobilizing in response to sterile inflammation. The response to sterile inflammation (thioglycolate) was most pronounced between day 4 and day 6, consistent with the time required to complete the GMP to neutrophil maturation process and in turn respond with the expected stages of extravasation from circulation towards an area of inflammation. Beyond day 6, few transfused cells could be recovered, consistent with the expected pre-programmed short lifespan of neutrophils. Together, these data suggest that the complex multistep in vivo process of neutrophil development, homing, and blood vessel extravasation are fully intact in these ERHoxb8 engineered cells.

The transfused ERHoxb8 GMPs were protective in two in vivo models of lethal fungal challenge: disseminated candidiasis and inhalational aspergillosis. Our two transfusion strategies (one time vs. loading dose + daily) suggest that like the human experience, the number of transfused cells is very important. Here we were not able to serially monitor the peripheral absolute neutrophil count, though that will be of great interest in future studies to ensure that we are achieving specific ANC goals in the neutropenic recipients. In this regard, the pre-clinical animal model is somewhat more challenging and may not reflect the experience of a human patients who can easily be transfused once, or even multiple times a day. That said, based on these data, the protective effect of transfused GMP is evident in the control of hyphae formation, which are a main driver of tissue destruction and pathogenesis. In fact, Candida that are not able to generate hyphae or are engineered to lack the factors required to transition to hyphal morphotype have significantly less virulence [56],[57]. It is noteworthy to mention that invasive Candida infecteion in mouse models has been shown to result in significant kidney damage, which can result in the impairment of normal kidney function (elevated BUN and creatinine levels) [55]. Thus, future pre-clinical animal models may consider use of biomarkers of renal function to assessment of Candida control by GMP.

Building on these data, future experiments can focus on other pathogens, including gram negative and gram-positive microbes, as well as different neutropenia etiologies such as chemotherapy rather than radiation exposure. We recognize immune cells other than neutrophils are affected by irradiation, both macrophages and dendritic cells are known to be critical in the eradication of fungal infections and may be playing a role in conjunction with neutrophils through paracrine effect [58][59]. The dramatic sensitivity of irradiated animals, succumbing to disease within days, underscores the importance of innate immune cells, particularly neutrophils, to address disease control in disseminated fungal infections [60].

The essential role of neutrophils in the clearance and eradication of invasive fungal infections (IFIs) has been well documented [2]. However, traditional granulocyte transfusions have not proven to be realistic, durable, nor safe for the prevention or treatment of neutropenic patients. Our studies suggest that the use of neutrophil progenitors may be a viable alternative, and possibly more effective as compared to traditional mature, adult donor granulocyte as transfusion therapy to protect against IFIs, including C. albicans and invasive molds including A. fumigatus. Future studies will focus on better defining the optimal transfusion schedule and number of progenitors with the knowledge that the GMPs will engraft and expand in vivo on their way to terminal differentiation. While we did not formally compare the feasibility of transfusing progenitors and mature neutrophils, the human complications of transfusion mature granulocytes (including febrile transfusion reactions and respiratory complications) have been well documented [3]. Progenitor transfusions may obviate many of the practical limitations of transfusing a terminally differentiated effector cell with an already short lifespan that is already activated by the mechanical stimulation of isolation and purification. We hope that the further development of precursor neutrophil cellular therapy may provide the ultimate prophylactic agent against fungal and bacterial agents for those patients facing extended periods of profound neutropenia.

Abbreviations:

ANC

Absolute neutrophil count

BM

Bone marrow

CFSE

Carboxyfluorescein succinimidyl ester

CFU

Colony-forming unit

CMP

Common myeloid progenitors

ERHoxb8

Estrogen receptor-HoxB8

G-CSF

Granulocyte colony stimulating factor

GMP

Granulocyte-macrophage progenitors

GMS

Grocott’s methenamine silver

HBD

Hormone binding domain

Hox

Homeobox

HSC

Hematopoietic stem cells

IFIs

Invasive fungal infections

PF

Peritoneal fluids

PMN

Polymorphonuclear cells

SP

Spleen