Heart Failure Impairs Bone Marrow Hematopoietic Stem Cell Function and Responses to Injury

Abstract

Background

Heart failure (HF) is a clinical syndrome associated with a progressive decline in myocardial function and low‐grade systemic inflammation. Chronic inflammation can have lasting effects on the bone marrow (BM) stem cell pool by impacting cell renewal and lineage differentiation. However, how HF affects BM stem/progenitor cells remains largely unexplored.

Methods and Results

EGFP⁺ (Enchanced green fluorescent protein) mice were subjected to coronary artery ligation, and BM was collected 8 weeks after myocardial infarction. Transplantation of EGFP⁺ BM into wild‐type mice revealed reduced reconstitution potential of BM from mice subjected to myocardial infarction versus BM from sham mice. To study the effects HF has on human BM function, 71 patients, HF (n=20) and controls (n=51), who were scheduled for elective cardiac surgery were consented and enrolled in this study. Patients with HF exhibited more circulating blood myeloid cells, and analysis of patient BM revealed significant differences in cell composition and colony formation potential. Human CD34⁺ cell reconstitution potential was also assessed using the NOD‐SCID‐IL2rγ^null mouse xenotransplant model. NOD‐SCID‐IL2rγ^null mice reconstituted with BM from patients with HF had significantly fewer engrafted human CD34⁺ cells as well as reduced lymphoid cell production. Analysis of tissue repair responses using permanent left anteriordescending coronary artery ligation demonstrated reduced survival of HF‐BM reconstituted mice as well as significant differences in human (donor) and mouse (host) cellular responses after MI.

Conclusions

HF alters the BM composition, adversely affects cell reconstitution potential, and alters cellular responses to injury. Further studies are needed to determine whether restoring BM function can impact disease progression or improve cellular responses to injury.

Nonstandard Abbreviations and Acronyms

BM

bone marrow

HSC

hematopoietic stem cell

LSK

lineage negative, Sca‐1⁺, c‐Kit⁺

MPP

multipotent progenitors

NSG

NOD‐SCID‐IL2rγ^null

Clinical Perspective.

What Is New?

• Heart failure is associated with lower bone marrow cellularity and impaired CD34⁺ hematopoietic stem cell reconstitution potential.

• Bone marrow cells from patients with heart failure exhibit an elevated immune response following cardiac injury and adversely impact host immune responses when transplanted into healthy mouse recipients.

What Are the Clinical Implications?

• Perturbations in the bone marrow stem cell reservoir have a substantial impact on the function of multiple organ systems, including the heart.

• Remodeling of the bone marrow niche and changes in hematopoietic stem cell proliferation/differentiation dynamics likely have a major impact on immune function, therapies, such as cell therapy, that rely on immune responses to elicit beneficial effects, may not be effective in patients with heart failure with immune system/bone marrow dysfunction.

• Alternative approaches that improve the function of bone marrow stem cells are needed to restore the cellular responses required for effective tissue repair in patients with heart failure.

Heart failure (HF) is a clinical syndrome associated with the impairment of either systolic or diastolic function of the heart. Based on the trends from 1990 to 2017, >60 million individuals worldwide suffer from HF. When expanded to those living with cardiovascular disease, such as ischemic heart disease, the number is substantially higher. ¹ , ² The 5‐year survival rate for those living with HF is <50%, and the associated medical care needed to treat patients with HF places a large economic burden on the health care system. ³ Therefore, there is an urgent need to develop new therapeutic approaches that can improve the quality of life and reduce the mortality rate of patients with HF.

We previously demonstrated that factors such as aging can negatively impact bone marrow (BM) stem cell function and cellular responses to injury. ⁴ , ⁵ , ⁶ , ⁷ However, less is known about the impact other chronic conditions, such as HF, have on the function of BM stem cells and their derived cellular products. One common feature of both HF with reduced ejection fraction (either from ischemic or nonischemic causes), and HF with preserved ejection fraction is a correlation between serum proinflammatory cytokine levels and disease severity. ⁸ , ⁹ Although the cytokine levels observed in chronic HF are significantly less than what would be found in cases of acute infection, low‐grade chronic inflammation is an important factor that can contribute to the clinical deterioration of patients with established HF. ¹⁰ Chronic inflammation can also adversely affect BM hematopoietic stem cell (HSC) and progenitor cell populations, as a result of the excessive demand for cellular output. Interestingly, noncardiovascular‐related causes of mortality are common in patients with HF, with a recent report suggesting that sepsis is a major cause of mortality. ¹¹ Therefore, the deterioration of cellular function outside of the heart may be a critical factor in HF pathophysiology and mortality.

HF has already been reported to adversely affect the function of endogenous stem/progenitor cell pools. For example, the number of circulating CD34⁺ cells in peripheral venous blood, which are typically classified as endothelial progenitors, has been shown to be significantly lower in patients with advanced HF compared with healthy controls. ¹² BM cells from patients with HF have also been reported to have diminished colony‐forming ability and potential for differentiation. For instance, Anand and colleagues established a relationship between HF and anemia, which resulted from blunted BM erythroid progenitor cell formation. ⁹ Patients with HF who present with anemia have a worse prognosis compared with nonanemic patients with HF, ¹⁰ supporting the notion that proper BM function is critical for maintaining patient health. Despite these findings, the effect HF has on BM function remains largely unexplored. Here, we investigate the impact HF has on BM stem cell function, using both mouse and human cells. First, we examined how HF affects BM stem cell populations and reconstitution potential using an experimental mouse model of HF. Next, we evaluated BM stem/progenitor cell frequency in sternal BM samples collected from HF and control patients. Lastly, we assessed the reconstitution potential of human CD34⁺ cells in vivo using immunodeficient NSG mice and examined cellular responses to injury following MI.

METHODS

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Patient Selection Criteria

The research ethics board of the University Health Network approved the investigation. This study also complies with the Declaration of Helsinki. Male and female patients >30 years of age who were scheduled for nonemergency cardiac surgery including coronary artery bypass graft or valve surgery or both consented for the study. Patients were recruited at the Toronto General Hospital between 2016 and 2020. Informed written consent was given before the inclusion of subjects in the study. Patients fulfilling the European Society of Cardiology diagnostic criteria for HF were included in this study and enrolled consecutively. Specifically, those with pulmonary edema, elevated jugular venous pressure, and peripheral edema, or any previous hospitalization for acute decompensated HF were identified by reviewing clinical charts. Among the identified patients, those with preoperative transthoracic echocardiography reports of reduced ejection fraction (<50%) or diastolic filling dysfunction (Doppler E wave/A wave ratio <0.75 or >2), unrelated to any underlying valvular pathology, were included. Exclusion criteria were as follows: patients with a history of congenital heart disease, chronic kidney disease, previous sternotomy, known malignancies, or on ventricular assist devices or extracorporeal membrane oxygenation. Overall inclusion and exclusion criteria are presented in Table S1, whereas analyses of patient medical history and circulatory markers are provided in Tables S2 through S4. There was no patient follow‐up aside from the clinical postoperative follow‐up to ensure proper recovery.

Sternal BM Harvest

Under general anesthesia, 5 mL of 10% heparin solution in a 20‐mL syringe was advanced slowly through the periosteum of the sternum. The solution was injected into the sternum, and the plunger was pulled back to aspirate between 15 and 20 mL of BM from the sternum using an 18‐gauge needle. The syringe containing the BM was transported into the laboratory, where mononuclear cells were separated by density gradient centrifugation using the Ficoll solution, according to the manufacturer's instructions (Pharmacia, Piscataway, NJ).

Human CD34 Cell Isolation and Hematopoietic Colony‐Forming Unit Assay

Mononuclear cells were separated into positively and negatively labeled CD34 fractions, using immunomagnetic activated cell sorting, according to the manufacturer's instructions (STEMCELL Technologies). The purity of positive cells was confirmed by flow cytometry. Functional capacity of CD34⁺ BM stem cells, derived from the patients, was evaluated using the colony‐forming unit assay for human hematopoietic cells (MethoCult H4034 Optimum; StemCell Technologies). One thousand CD34⁺ cells were plated per 35‐mm dish and cultured in the assay media. The number of colony‐forming unit‐granulocyte/macrophage (GM), burst‐forming unit erythroid (BFU‐E), and total colonies were quantified 14 days after plating using a Nikon light microscope.

Animals and BM Reconstitution

The Animal Care Committee of the University Health Network approved all experimental procedures, which were performed according to the Guide for the Care and Use of Laboratory Animals (National Institutes of Health, revised 2011). Animals were housed in a 12‐hour light:12‐hour dark light cycle and provided food and water ad libitum. All animals were euthanized by isoflurane overdose (minimum 5% flow rate), followed by cervical dislocation for all terminal experiments.

For BM transplant and MI studies involving mouse recipients and donors, 2‐month‐old female C57BL/6‐Tg(CAG‐EGFP)1Osb/J EGFP (enhanced green fluorescent protein) mice were used as donors, and 2‐month‐old female C57Bl/6J mice were used as recipients. EGFP mice were subject to sham or coronary artery ligation surgery, as detailed below. Eight weeks after surgery, animals were euthanized, and cells were flushed from BM in PBS using a 23‐gauge needle. Red blood cells were lysed, the cells washed, and passed through a 40‐μm filter. For BM transplant studies, BM cells were counted, and 5×10⁶ total EGFP⁺ BM cells were injected through the tail vein into lethally irradiated wild‐type (WT) recipients (9.5 Gy). EGFP MI donors (n=4) were used to reconstitute 7 WT mice, and EGFP sham donors (n=3) were used to reconstituted 3 WT mice. BM reconstitution potential was assessed 12 weeks after reconstitution.

For humanized mouse studies, 8‐ to 12‐week‐old female NSG mice were used (strain number 005557). NSG mice were irradiated at 285 cGy 24 hours before human BM stem cell injection using a Gammacell 40 Extractor Cesium‐137 Irradiator (Best Theratronics). The following day, animals received an intravenous (tail vein) injection of CD34 cells (0.7×10⁶) from each patient. BM reconstitution rate and experimental procedures were performed 12 weeks after reconstitution. Each n value from the humanized animal experiments refers to 1 patient's BM sample.

MI Model and Echocardiography

Coronary artery ligation was performed in 2‐month‐old EGFP⁺ and WT C57Bl/6J mice or in NSG mice 12 weeks after BM reconstitution with human CD34⁺ cells. The mice were intubated and ventilated with 2% isoflurane. A left thoracotomy was performed, and the left coronary artery was permanently ligated 2 mm below the left auricle using 7‐0 proline suture. Infarction was confirmed using echocardiography. Cardiac function was measured with echocardiography, using a GE Vivid 7i ultrasound machine equipped with the I13L probe. Animals were anesthetized, shaved, and maintained under light anesthesia (1.5% isoflurane) for all functional measurements. Hearts were visualized on B‐mode at the level of the papillary muscle in left ventricular short‐axis view, and functional measurements were determined using M‐mode recording at the same level. At study end, the scar size (akinetic wall length) was measured on the B‐mode short‐axis recording at end‐diastole using echocardiography. The infarct wall was identified as the area of akinetic region, and the scar size was measured as the percentage of the endocardial length of the akinetic region over the left ventricular endocardial circumference.

Flow Cytometry Analysis

BM, blood, spleen, and heart were collected from mice 12 weeks after reconstitution for baseline analyses and at 3 and 7 days after MI for humanized mouse studies. BM, blood, and spleen cells were isolated, and the red blood cells were removed through red blood cell lysis. The hearts were minced and digested using 2‐mg/mL collagenase type II at 37 °C for 30 minutes. After isolation, the cells were washed in FACS buffer (PBS Ca²⁺/Mg²⁺ free +2% FBS) and filtered using a 70‐μm filter. Fc receptors were blocked using anti‐mouse CD16/32 (1:100; BioLegend) and human Fc block (1:100; BD Biosciences), after which cells were stained with primary antibodies for 1 hour at 4 °C in the dark. All antibodies used are listed in Table S5. Cells were washed and run on an LSRII equipped with a violet laser (25 mW), blue laser (50 mW), yellow laser (100 mW), and red laser (40 mW). Human BM cells were identified as previously described. ¹⁰ Human cell populations were identified as follows: human hematopoietic stem/progenitors (HSPCs): Lin⁻ CD34⁺ CD38⁻, human HSCs: Lin⁻ CD34⁺ CD38⁻ CD45RA⁻ CD90⁺, human multipotent progenitors (MPPs): Lin⁻ CD34⁺ CD38⁻ CD45RA⁻ CD90⁻, human multilymphoid progenitors: Lin⁻CD34⁺ CD38⁻ CD45RA⁺ CD90⁻, human early myeloid progenitors: Lin⁻ CD34⁺ CD38⁺ CD45RA⁻, human B cells: CD45⁺ CD19⁺, human T cells: CD45⁺ CD3⁺, human myeloid cells: CD45⁺ CD33⁺, human neutrophils: CD45⁺ CD33⁺ CD11b⁺ CD66b⁺ CD14⁻, and human monocytes: CD45⁺ CD33⁺ CD11b⁺ CD66b⁻ CD14⁺. Mouse lineage negative, Sca‐1+, c‐Kit+ (LSK) cells were identified as lineage (Ly6G, Ly6C, CD11b, CD3, Ter119, B220)‐negative, c‐Kit⁺ and Sca‐1⁺; granulocyte‐macrophage progenitors (GMPs) were identified as Lin‐negative, c‐Kit⁺, Sca‐1⁻, CD16/32⁺, CD34⁺, and HSPC populations were first gated on LSK cells and further divided by CD150 and CD48 staining (Figure S1D). Mouse immune cells were identified as follows: T cell (CD45⁺, CD11b⁻, B220⁻, CD3⁺), total myeloid (CD45⁺, CD11b⁺, B220⁻, CD3⁻), B‐Cell (CD45⁺, CD11b⁻, B220⁺, CD3⁻), neutrophil (CD45⁺, CD11b⁺, Ly6G⁺), monocyte (CD45⁺, CD11b⁺, Ly6G⁻, Ly6C^hi), and macrophages (CD45⁺, CD11b⁺, Ly6G⁻, F4/80⁺). For all experiments using EGFP⁺ cells, EGFP⁺ cells were first gated, after which the subsequent cell populations were identified as listed above. Gating was determined by fluorescent minus one controls, and all data were analyzed in FlowJo (TreeStar).

Statistical Analysis

All values are expressed as mean±SEM. Analyses were performed using GraphPad Prism 8.0 software. Statistical comparisons were performed with the Mann‐Whitney test when normality could not be tested. Otherwise, an unpaired 2‐sided Student's t test, unpaired t test with Welch correction when group variation was significantly different by an F test, or 2‐way ANOVA followed by Sidak multiple comparison test was used, as applicable. Survival was assessed using the log‐rank Mantel‐Cox test. Normality was tested using the Shapiro‐Wilk test. Analyses of categorical variables were performed using the χ² test. A value of P≤0.05 was considered statistically significant. All n values are provided in the figure legends.

RESULTS

HF Alters BM Stem Cell Function in Mice

The objective of this study was to determine whether HF affects BM stem cell function, relating to engraftment and lineage differentiation; therefore, we first aimed to establish whether HF in an experimental animal model is associated with changes in BM stem cell function. We characterized BM stem cell populations in mice subjected to MI. Flow cytometry analysis of BM populations revealed that HF mice had significantly more LSK cells compared with sham mice at 8 weeks after MI (Figure S1A and Figure S1B). Further examination revealed a generalized increase in HSPC (Figure S1C and Figure S1D) and GMP (Figure S1E and Figure S1F) cell numbers in HF versus sham mice, indicating expansion of BM stem/progenitor cell populations.

Figure 1. Heart failure (HF) reduces engraftment potential in mice.

A, Experimental design for the bone marrow transplant (BMT) study. EGFP⁺ (Enhanced green fluorescent protein) mice were subject to sham or left anterior descending coronary artery (LAD) ligation (HF), and 8 weeks (8w) later, bone marrow (BM) was collected. 5×10⁶ BM mononuclear cells were transplanted into wild‐type C57BL/6J mice, and BM populations assessed 3 months later. B, BM cellularity 3 months after reconstitution with sham or HF BM. C, Engrafted EGFP⁺ lineage negative, Sca‐1⁺, c‐Kit⁺ (LSK) cells in sham‐BMT or HF‐BMT mice 3 months after reconstitution. D, Representative flow plots of EGFP⁺ LSK cells. E, EGFP⁺ hematopoietic stem/progenitor cell (HSPC) populations in BMT‐sham and BMT‐HF mice. F, Relative abundance of EGFP⁺ blood B cells (B220⁺), T cells (CD3⁺), and myeloid cells (CD11b⁺), as well as (G) myeloid cell subpopulations in mice transplanted with sham or HF BM. H, Representative flow cytometry plots of blood T‐cell and myeloid lineages (n=3–7 per group). *P≤0.05 by Mann‐Whitney test. Values are mean±SEM. LAD indicates left anterior descending coronary artery; LT‐HSC indicates long‐term repopulating hematopoietic stem cells; and MPP, multipotent progenitor.

To understand how HF affects HSC function, we used EGFP⁺ mice to assess BM reconstitution potential. EGFP⁺ mice were infarcted, and BM collected 8 weeks after MI for transplantation into irradiated WT mice (Figure 1A). Reduced cardiac function of donor HF EGFP⁺ mice was confirmed by echocardiography before BM isolation (Figure S2). Three months after BM reconstitution, EGFP⁺ cell engraftment and cell lineages in recipient WT mice were examined by flow cytometry. Despite having an increased frequency of LSK and HSPC populations in the BM at 8 weeks after MI (Figure S1), HF BM showed reduced reconstitution potential, as HF BM‐transplanted mice had significantly lower total BM cellularity (Figure 1B) and EGFP⁺ LSK cell numbers (Figure 1C and 1D) at 3 months after transplantation compared with BM‐transplanted sham mice. Although all HSPC populations assessed trended lower, only MPPs were significantly lower in BM‐transplanted HF mice (Figure 1E). BM cells from HF mice also exhibited different lineage repopulation potential, as there was a higher frequency of circulating blood myeloid (CD11b⁺) cells in BMT‐HF mice compared with BMT‐sham mice (Figure 1F). This was driven by an increase in circulating monocyte and neutrophil populations (Figure 1G and 1H). Collectively, these data indicate that in mice, experimentally induced HF is associated with a change in BM stem cell function characterized by reduced reconstitution potential and greater myeloid lineage cell production. Next, we aimed to determine whether BM stem cells isolated from patients with HF also exhibit the same functional changes, as observed in our mouse model.

Analysis of BM Stem Cells Isolated From Patients With HF

Ninety patients undergoing elective cardiac surgery consented to participate in the study. Patients for whom preoperative blood work or echocardiography reports were not available were excluded (Figure 2). Out of the remaining 71 patients undergoing elective cardiac surgery, 20 patients demonstrated clinical symptoms of HF, such as elevated B‐type natriuretic peptide, elevated jugular venous pressure, and respiratory congestion. Analysis of their preoperative echocardiography report showed reduced left ventricular ejection fraction (<50%) or diastolic filling dysfunction (E/A ratio <0.75 or >2) not related to any underlying valvular pathology. Clinical characteristics of the study population are presented in the Table as well as Tables S2 through S4. As expected, the patients with HF were more frequently prescribed diuretics and presented with clinical manifestations of HF, such as dyspnea and/or peripheral edema (Table). They were also prescribed more angiotensin receptor blockers (40% versus 9.8%) and β‐blockers (90% versus 39.2%; Table S3) compared with the control patient group. Moreover, mean NT‐proBNP (N‐terminal pro‐B‐type natriuretic peptide) level was moderately‐to‐severely elevated in the HF group (92.17±20.30 versus 40.65±10.84 pg/mL; Table S4), indicating circulatory volume overload. We also examined circulating blood populations by chart review of complete blood counts collected during patient admission to determine how HF affected circulating cell numbers in our patient cohort. As expected, patients with HF demonstrated lower levels of circulating erythrocytes (Figure 3A) and reduced hemoglobin levels (Figure 3B) compared with the control group. Preoperative leukocyte count in the absence of active infection was significantly higher in the HF cohort (Figure 3C). This increase in white blood cell (WBC) count was driven by a significantly greater number of circulating neutrophils (Figure 3D) and monocytes (Figure 3E). There were no significant differences in circulating lymphocyte cell numbers between the 2 groups (Figure 3F).

Figure 2. Patient selection flowchart.

Patient enrollment for the current study. A total of 90 patients were recruited, and 71 met the study inclusion criteria listed in Table S1. HF indicates heart failure.

Table 1.

Baseline Patient Characteristics

Characteristic	HF	Controls	P value
Patient demographics
Enrolled	20	51
Age, y	67.7 (2.0)	63.6 (1.5)	0.11
Sex, men	15 (75%)	42 (82%)	0.52
BMI, kg/m2	30 (1.3)	28.5 (0.6)	0.18
Medical conditions
Diabetes	14 (70%)	22 (43%)	0.07
Hypertension	13 (65%)	38 (74%)	0.56
Chronic kidney disease	3 (15%)	3 (6%)	0.34
Diuretic use
Furosemide	6 (30%)	NR
Hydrochlorothiazide	5 (25%)	NR
Spironolactone	2 (10%)	NR
Indapamide	1 (5%)	NR
LVEF
>50%	12 (60%)	51 (100%)	<0.0001*
30%–49%	7 (35%)	0	<0.0001*
<30%	1 (5%)	0	<0.0001*
Heart failure types
HFrEF	8 (40%)	0	<0.0001*
HFpEF	12 (60%)	0	<0.0001*
eGFR
>60	14 (70%)	45 (88%)	0.084
30–59	5 (25%)	5 (10%)	0.131
<30	1 (5%)	1 (2%)	0.487
Physical examination
Peripheral edema	9	NR
Dyspnea	11	NR

Qualitative data are presented as n (% of group). P values are from χ ² test for categorical variables. P<0.05 indicates that there is a significant difference between HF and control groups. BMI indicates body mass index; eGFR, estimated glomerular filtration rate; HF, heart failure; HFpEF, heart failure with preserved ejection fraction; HFrEF, heart failure with reduced ejection fraction; LVEF, left ventricular ejection fraction; and NR, not reported in electronic patient records.

^* P<0.0001.

Figure 3. Complete blood count analysis of control patients and patients with heart failure (HF).

When compared with the control group, patients with HF have significantly lower levels of erythrocyte (A) and hemoglobin (B). Patients with HF also show significantly increased numbers of circulating leukocytes (C), which is driven specifically by increased neutrophils (D) and monocytes (E). F, No differences in total lymphocytes were observed (n=20–51 per group). *P<0.05 control vs HF by unpaired 2‐tailed Student's t test. Values are mean±SEM. RBC indicates red blood cell; and WBC, white blood cell.

Despite the increased circulating levels of WBCs, the total BM cellularity of patients with HF, as measured by mononuclear cell concentration derived from sternal BM aspirate, was significantly reduced (Figure 4A). Further examination of the CD34⁺ stem cell/progenitor population revealed that the frequency of CD34⁺ cells was not different between patients with HF and control patients (Figure 4B). However, examination of CD34⁺ cell colony formation in vitro revealed that CD34⁺ cells isolated from patients with HF exhibited reduced colony‐forming unit‐GM and BFU‐E progenitor colony formation, suggesting a reduction in their functional capacity (Figure 4C). A more detailed examination of CD34⁺ cell subpopulations revealed that patients with HF have an increased frequency of the CD34⁺/CD38⁻ population (Figure 4D). This was primarily driven by a greater number of MPPs, as the frequency of this population was significantly higher in HF compared with control BM (Figure 4E). We did not detect any differences in myeloid progenitor or mature populations (Figure 4F through 4H). Collectively, these results demonstrate that patients with HF have reduced BM cellularity, increased frequency of specific HSPC populations, and reduced CD34⁺ cell colony formation.

Figure 4. Analysis of patient bone marrow (BM).

Total mononuclear cellularity (MNC) (A) and CD34⁺ cell frequency (B) in BM from control and patients with heart failure (HF). *P<0.05 by unpaired two‐tailed Student's t test. C, Colony formation ability of BM CD34⁺ cells from control patients and patients with HF. **P<0.01 by unpaired 2‐tailed Student's t test. D, Frequency of CD34⁺/CD38⁻ cells was increased in BM of patients with HF. P=0.055 by unpaired t test with Welch correction. E, This increase appears to be driven by a greater number of multipotent progenitors (MPPs), whereas hematopoietic stem cells (HSCs) and multipotent lymphoid progenitors (MLPs) were not different (n=5–6 per group). *P≤0.05 by unpaired t test with Welch correction. F, Myeloid progenitor relative abundance was not different between BM of control patients and patients with HF (n=5–10 per group). Representative gating for identification of BM neutrophils and monocytes (G) and quantification of the populations in the BM (H) (n=8–12 per group). Values are mean±SEM.

Considering changes in CD34⁺ colony formation, we further assessed CD34⁺ cell function in vivo using our established xenotransplant model in NSG mice. ¹⁰ Mice were reconstituted with CD34⁺ cells isolated from the BM of control and patients with HF, and cell engraftment/lineage differentiation potential was assessed 3 months later. Interestingly, animals reconstituted with HF BM exhibited significantly reduced survival over the 3‐month period following cell transplantation (Figure S3A). Although the cause of death was not determined, mice presented with splenomegaly and pronounced weight loss. Analysis of cell engraftment at 3 months after reconstitution revealed similar levels of human CD45⁺ (hCD45) cells in the BM of mice reconstituted with control and HF CD34⁺ cells (Figure 5A). However, the frequency of CD34⁺ cells were significantly reduced in mice reconstituted with HF BM cells, indicating reduced reconstitution potential (Figure 5B). Further analysis of engrafted HSPC subpopulations revealed that a significantly greater number of CD34⁺ cells in HF‐NSG reconstituted mice were MPPs (Figure 5C). This is consistent with our observed increase of MPPs in BM from patients with HF (Figure 4E). Next, we assessed the cell lineages that arose from engrafted human cells to determine whether HF alters cell lineage production. The frequency of human CD45⁺ cells in the blood and spleen did not differ between mice reconstituted with the HF and control BM (Figure 5D). However, mice reconstituted with HF BM exhibited a significantly lower level of B cells in the spleen compared with mice reconstituted with control BM (Figure 5E). This may indicate lower levels of engrafted lymphoid progenitors or reduced lymphoid cell potential of engrafted BM HSCs. We did not detect differences in human‐derived B‐cell or myeloid populations in the blood between HF and control reconstituted mice (Figure 5F).

Figure 5. Reduced engraftment of bone marrow (BM) CD34⁺ cells from patients with heart failure (HF) in NOD‐SCID‐IL2rγ^null (NSG) mice.

A, Human CD45 (hCD45) and (B) human CD34⁺ (hCD34) cells in the BM of NSG mice 3 months after reconstitution (n=7–25 per group). *P<0.05 by Mann‐Whitney test. C, Frequency of BM progenitor populations 3 months after reconstitution (n=6–9 per group). *P≤0.05 by Mann‐Whitney test. D, hCD45 frequency in the blood and spleen in reconstituted animals (n=6–15 per group). B‐cell and myeloid‐cell frequency in the spleen (E) and blood (F) (n=6–15 per group). *P≤0.05 by unpaired t test with Welch correction. Values are mean±SEM. HSC indicates hematopoietic stem cell; MLP, multilymphoid progenitor; MPP, multipotent progenitor; and MNC, mononuclear cell.

Engrafted Cells From Patients With HF Exhibit Altered Responses After MI

We then investigated whether the altered function of HF BM cells included changes in the response to injury. Using the permanent left anterior descending coronary artery ligation MI model, we assessed cardiac repair responses in HF and control BM reconstituted NSG mice (Figure 6A). Throughout the first week following MI, we found that HF‐NSG mice exhibited significantly greater mortality compared with mice reconstituted with control BM (Figure S3B). The cause of death is not known but was not related to ventricular rupture, as animals that died did not present with this phenotype upon examination by necropsy. To assess how functional outcomes were impacted, we examined cardiac function 4 weeks after MI by echocardiography and the cellular responses of reconstituted human cells at days 3 and 7 after MI by flow cytometry. By 4 weeks after MI, mice reconstituted with HF BM cells exhibited greater infarct length, as measured by akinetic wall length (Figure 6B). Mice reconstituted with HF BM also trended to have lower contractile function, as indicated by a lower fractional shortening (Figure 6C through 6E). However, no significant differences were detected in any of the echocardiography parameters (Figure 6B through 6E). Given differences in early survival, it is important to note that a survivorship bias could have impacted our functional data set.

Figure 6. Altered cardiac immune responses after myocardial infarction (MI) in mice reconstituted with bone marrow (BM) from patients with heart failure (HF).

A, Schematic illustrating experimental design. B, Akinetic wall length/left ventricular (LV) endocardial length at 4 weeks after MI in reconstituted mice (n=5–11 per group). Unpaired 2‐tailed Student's t test. Echocardiographic analysis of LV dimensions at diastole (LVIDd) (C), at systole (LVIDs) (D), and fractional shortening (FS) (E) at baseline (0 weeks), 1 week, 2 weeks, and 4 weeks after MI (n=5–11 per group). Flow cytometry quantification of human (h) CD45⁺ (F), T cell (CD3⁺) (G), and myeloid (CD33⁺) (H) cell infiltration in the heart at baseline, 3 days, and 7 days after MI (n=3–14 per group). *P≤0.05 by 2‐way ANOVA followed by Sidak multiple comparison test. Flow cytometry quantification of mouse (m) CD45 (I), monocytes (J), neutrophils (K), and macrophages (L) in the heart at 3 days after MI (n=5–6 per group). *P≤0.05 by unpaired 2‐tailed Student's t test. Values are mean±SEM. LAD indicates left anterior descending coronary artery; and NSG, NOD‐SCID‐IL2rγ^null.

To assess the immune cell responses of engrafted patient BM cells after MI, we analyzed cell populations in the heart at baseline, 3 days, and 7 days post‐MI by flow cytometry. As expected, both control and HF reconstituted mice exhibited a temporal cellular response after MI with the greatest cell numbers in the heart at 3 days after MI. HF BM reconstituted NSG mice had significantly greater human CD45⁺ cell (Figure 6F) and human T‐cell (Figure 6G) infiltration into the heart at 3 days after MI. Although both B cells and myeloid cells exhibited a temporal response after MI, no significant differences in total cell abundances (cells per milligram of tissue) were detected within the infarcted myocardium (Figure 6H, Figure S3C and S3D). Consistent with the profile of immune cell recruitment to the infarcted myocardium, human cells in the blood and spleen also exhibited temporal responses after MI (Figure S3E and S3F). Most notably, the frequency of T cells in the spleen dramatically increased in both control and HF BM reconstituted mice, whereas the frequency of B cells decreased, suggesting there is preferential expansion of T‐cell populations in reconstituted mice after MI (Figure S3G and S3H). To further characterize immune responses after MI, we also assessed mouse (host) myeloid cells in the heart at 3 days after MI to understand how BM reconstitution impacted the host cell responses. Interestingly, HF BM reconstituted mice had significantly lower mouse CD45⁺ cells in the heart at 3 days after MI (Figure 6I). This was driven by lower levels of all myeloid cells detected, including monocytes, neutrophils, and macrophages (Figure 6J through 6L). This suggests that the engrafted HF BM also impacts host myeloid cell responses after MI. Overall, in this investigation, we demonstrate that HF is associated with changes in BM stem cells, which include changes in HSPCs population frequencies, reduced colony formation in vitro, reduced reconstitution potential in vivo, and altered cellular responses after MI.

DISCUSSION

Following an MI, innate and adaptive immune cells are recruited to the heart to help repair the damaged tissue; by participating in this process, BM‐derived cells play a fundamental role in myocardial repair after MI. Even after the repair process is complete, the body undergoes several compensatory changes in an attempt to maintain cardiac function. However, these compensatory changes can affect multiple organ systems, as changes in sympathetic activity, vascular function, and immune cell function have all been documented. ¹² Moreover, chronic inflammation, which is often observed in patients with HF, is linked to disease progression and lower patient survival. ¹³ Thus, there is a need to understand the progressive changes that occur throughout the body during HF, as this can help with the development of novel therapeutic approaches that limit disease progression.

In this study, we aimed to characterize the impact HF has on BM stem cell function. Patients with HF often present with comorbidities that may impact the BM cell phenotype; therefore, we first used an experimental mouse model of HF to determine whether BM HSCs show a change in reconstitution potential and cell lineage production. This animal model provided a simplified experimental system to investigate changes in BM function. Compared with mice subjected to the sham surgery, HF BM showed reduced reconstitution potential evidenced by lower BM cellularity and fewer EGFP⁺ LSK cells in the BM of recipient WT mice at 3 months after transplant. Interestingly, HF mice had a greater number of LSK cells in the BM at 8 weeks after MI before BM isolation; therefore, differences in cell numbers following reconstitution were not due to lower LSK levels in the transplanted cell population. We also investigated the effect HF has on human BM stem cell function using CD34⁺ cells isolated from sternal BM samples. Analysis of the total mononuclear and CD34⁺ cell populations in the BM collected from patients with HF indicated lower BM cellularity and reduced CD34⁺ cell colony forming ability compared with BM collected from control patients. Despite lower overall cellularity, HF‐BM had an increased frequency of CD34⁺/CD38⁻ cells, a key cell population responsible for long‐term multilineage engraftment. ¹⁴ We further assessed the in vivo reconstitution potential of the human BM CD34⁺ cells using our established xenotransplant model. ⁵ Interestingly, HF‐CD34⁺ reconstituted mice demonstrated significantly reduced BM CD34⁺ cell engraftment at 3 months after reconstitution, mirroring the findings in our EGFP mouse model.

The lower number of HF CD34⁺ cells in the BM of reconstituted mice could be related to impaired retention in the BM or an increase in their mobilization from the BM to the circulation. It should be noted that the reconstitution potential of BM stem cells can be impacted by several factors, including homing to the BM following intravenous injection, engraftment within the conditioned niche, cell survival, proliferation and mobilization, as well as the frequency of short‐ and long‐term repopulating cells within the transplanted population. Changes in any one of these can impact the final cell numbers in the BM when examined 3 months after reconstitution. Although the molecular mechanisms underlying reduced reconstitution potential were not investigated in this study, it has been documented that HF can impact key regulators of HSC function. For example, patients with HF and mice with experimentally induced HF exhibit changes in BM niche cell function, which in turn impacts HSCs within the BM. In the study by Hoffmann and colleagues, the effect was primarily attributed to HF‐induced changes in BM endothelial cells. ¹⁵ MI‐induced remodeling of the BM niche also impacts HSC responses to injury, as cellular responses to a recurrent MI were significantly reduced compared with responses in mice with no prior injury. ¹⁶ In addition to changes in HSC:niche cell interactions, HSCs may also be directly affected by secreted factors commonly found in patients with HF. For example, cytokines such as IL (interleukin)‐1β, IL‐6, or interferons can impact HSC apoptosis, proliferation, and subpopulation frequency. ¹⁷ , ¹⁸ In our study, it is likely that a combination of changes in secreted factors and BM niche cell function in patients with HF contributed to differences in BM stem cell function. Future studies should be conducted, via lineage tracing, to determine the fate and turnover of HF‐ and control‐BM cells in BM, blood, and spleen in the reconstituted mice. Additionally, genomic/proteomic profiling‐based studies will be pivotal for identifying the exact molecular changes that occur in BM stem cells of patients with HF.

In addition to baseline differences between the control and HF CD34⁺ reconstituted mice, we also identified several changes in the cellular responses after MI. Although the HF‐NSG mice had lower CD34⁺ engraftment at baseline, analysis of the infarcted myocardium of the reconstituted mice at 3 and 7 days after MI revealed increased inflammatory cell recruitment in the hearts of HF‐NSG mice compared with control‐NSG mice. This was primarily driven by greater T‐cell infiltration at 3 days after MI. T cells have emerged as critical regulators of cardiac repair, ¹⁹ and clinical studies have observed that the levels of circulating T cells in patients with HF correlate with disease severity. ²⁰ Future studies examining the effect HF has on lymphocyte progenitors, T‐cell thymic maturation, and peripheral proliferation will be crucial for understanding how manipulation of this cell population can affect disease progression. It is also important to note that the strain of NSG mice used in this study do not have the cytokine milieu needed for complete multilineage development. ²¹ For example, human myeloid cells require human colony‐stimulating factor for the differentiation, proliferation, and maintenance of mature cell populations; this effect cannot be fulfilled by the mouse homolog. Therefore, in our current model, the effect of HF on myeloid cells is likely underrepresented. However, we did observe a significantly reduced mouse (host) myeloid response after MI. Although the mechanisms underlying these observations were not explored, these data suggest that cells derived from HF BM can alter mouse myeloid responses to injury. Differences in the cellular responses after MI were associated with worse survival and a trend of greater cardiac remodeling at 4 weeks after MI in NSG mice reconstituted with HF versus control patient BM cells. Future experiments, involving a larger cohort of reconstituted mice, will be needed to detect significant changes in cardiac structure and function.

As with any investigation involving a patient population, there are inherent limitations to this study. Of note, the prevalence of type 2 diabetes is higher in the HF cohort (70% versus 43%; P=0.07). This increased prevalence of type 2 diabetes in the HF group may have contributed to the functional differences observed in BM stem cells obtained from patients with HF compared with those from control patients. Type 2 diabetes has been identified to be an independent risk factor for HF development, ²² and impaired endothelial progenitor cell function has been reported in patients with type 2 diabetes. ²³ Future investigations should identify the impact HF has on BM stem cell function in both diabetic and nondiabetic patient populations. A higher percentage of patients with HF were also on angiotensin receptor blockers (40% versus 9.8%) and β‐blockers (90% versus 40%) in this study. This significant difference is due to the importance of guideline‐directed medical therapy for patients with HF with reduced ejection fraction. β‐blockers and renin‐angiotensin blockade have been considered the main pillars of HF management and have aided in reducing the mortality rate among these patients. However, these medications may have independently impacted the quality of BM stem cells derived from patients with HF, and thus impacted our findings. Another limitation was that although we have documented circulating and cardiac inflammatory cellular responses, the underlying molecular basis was not fully investigated. Notably, the increased mortality observed among the HF BM‐reconstituted NSG mice during the first week after MI could be attributed to those heightened inflammatory responses. Future studies should examine the changes in cytokine and other cardiac inflammatory markers, to understand how differences in inflammatory responses could contribute to the increased mortality observed in this study. Lastly, it is important to note that our control cohort of patients consisted of individuals requiring nonemergency cardiac surgery. Although they were not diagnosed with HF, this is not a population without any underlying medical conditions. Therefore, the effect sizes observed in this study may be different when patients with HF are compared with a healthy control population. Moreover, given that the number of enrolled patients for this study was limited by practical considerations of consent for invasive BM sampling, it was not feasible to perform subgroup analyses based on demographic, pathologic, comorbidity, or other potential modifiers, or to perform propensity‐score matching, as these subanalyses would have had too small numbers to yield any conclusions. Therefore, further studies will be needed to determine whether the results presented here are generalizable for a larger diverse population of patients with HF.

The BM contains a critical reservoir of stem cells that are responsible for producing blood and immune cells over the course of our lifespan; perturbations in this cell population can have substantial impacts on the function of multiple organ systems. Remodeling of the BM niche and changes in HSC proliferation/differentiation dynamics in patients with HF are likely to have a major impact on immune system function. Therapies, such as cell therapy, ²⁴ , ²⁵ which in part rely on an immune response to elicit beneficial effects, may not be effective in  patients with HF with advanced immune system/bone marrow dysfunction. The development of therapies that improve the function of BM stem cells in patients with HF may restore the cellular responses required for effective tissue repair and limit disease progression.

Sources of Funding

This work was supported by the Canadian Institutes of Health Research (332652 awarded to R.‐K.L.), and Heart and Stroke Foundation (G‐22‐0032119 awarded to R.‐K.L.).

Disclosures

None.

Supporting information

Tables S1–S5

Figures S1–S3