Aging impairs human bone marrow function and cardiac repair following myocardial infarction in a humanized chimeric mouse

Abstract Ventricular remodeling following myocardial infarction (MI) is a major cause of heart failure, a condition prevalent in older individuals. Following MI, immune cells are mobilized to the myocardium from peripheral lymphoid organs and play an active role in orchestrating repair. While the effect of aging on mouse bone marrow (BM) has been studied, less is known about how aging affects human BM cells and their ability to regulate repair processes. In this study, we investigate the effect aging has on human BM cell responses post‐MI using a humanized chimeric mouse model. BM samples were collected from middle aged (mean age 56.4 ± 0.97) and old (mean age 72.7 ± 0.59) patients undergoing cardiac surgery, CD34+/− cells were isolated, and NOD‐scid‐IL2rγnull (NSG) mice were reconstituted. Three months following reconstitution, the animals were examined at baseline or subjected to coronary artery ligation (MI). Younger patient cells exhibited greater repopulation capacity in the BM, blood, and spleen as well as greater lymphoid cell production. Following MI, CD34+ cell age impacted donor and host cellular responses. Mice reconstituted with younger CD34+ cells exhibited greater human CD45+ recruitment to the heart compared to mice reconstituted with old cells. Increased cellular responses were primarily driven by T‐cell recruitment, and these changes corresponded with greater human IFNy levels and reduced mouse IL‐1β in the heart. Age‐dependent changes in BM function led to significantly lower survival, increased infarct expansion, impaired host cell responses, and reduced function by 4w post‐MI. In contrast, younger CD34+ cells helped to limit remodeling and preserve function post‐MI.


| INTRODUC TI ON
Successful infarct healing and scar formation is essential for preserving cardiac function following myocardial infarction (MI) (Prabhu & Frangogiannis, 2016). Although prompt reperfusion significantly reduces early mortality following acute MI, post-ischemia heart failure is a major contributor to reduced quality of life, increased healthcare burden, and mortality worldwide (Cooper et al., 2015). One proposed approach for limiting adverse remodeling post-MI has been stem cell transplantation. However, despite promising pre-clinical findings, clinical studies of direct stem cell transplantation to the ischemic myocardium have yielded mixed results (Marvasti et al., 2019). Recent studies have demonstrated that effective cardiac cell therapy relies on modulation of endogenous repair processes, as the transplanted cells are only transiently present in the infarcted myocardium (Vagnozzi et al., 2020). Therefore, if endogenous repair mechanisms are impaired, such as in elderly patients or in patients with co-morbidities (e.g., diabetes), it is anticipated that cell therapy will be less effective and less likely to yield beneficial results in these patient populations.
To overcome this limitation, our group has been investigating the potential of rejuvenation strategies to improve endogenous repair responses and stimulate cardiac repair post-MI. We previously demonstrated that reconstituting aged mice with younger bone marrow (BM) hematopoietic stem cells (HSC) leads to stable integration of these cells and improves cardiac repair post-MI (Li et al., 2017;Li et al., 2019;Tobin et al., 2020). This model is distinct from previous cell therapy approaches, as this approach requires stable engraftment of donor cells in the bone marrow which subsequently participate in the infarct healing process. Immune cells mobilize to the heart following infarction and orchestrate cardiac repair through several mechanisms. Although much is known about BM aging (Broxmeyer et al., 2020;Ho et al., 2019;Pang et al., 2011;Pritz et al., 2014), less is known about whether age-dependent changes in human BM cells contribute to impaired cardiac repair post-MI. Here, we establish a xenograft model to study the impact aging has on the function of human BM cells. We investigate the capacity of young and old patient BM cells to repopulate the BM of immune deficient NOD-scid-IL2rγ null (NSG) mice and examine the ability of these cells to participate in infarct healing post-MI.

| Examination of human CD34 + cells frequency and function
A total of 88 patients were included in this study and were separated into middle aged (mean age 56.4 ± 0.97) and old (mean age 72.7 ± 0.59) donors. Patients' characteristics were reviewed to evaluate any underlying differences between the two cohorts ( Table 1).
Analysis of complete blood counts prior to surgery demonstrated that old patients exhibited greater levels of circulating immune cells, mainly driven by increased neutrophils and monocytes ( Figure S1a). However, total BM cellularity was not different between the middle aged and old cohorts (Figure 1a and Figure S1b). The frequency of CD34 + cells (Figure 1b and Figure S1c-d), a common progenitor cell marker, and the frequency of BM lineage progenitor populations  were also similar in the middle aged and old patient cohorts. However, BM from older patients did exhibit an increased frequency of CD33 + /CD14 + cells consistent with a greater frequency of myeloid cells in older individuals ( Figure S2b). Although progenitor cell frequencies were similar, CD34 + progenitor cells exhibited an age-dependent decline in colony formation capacity (Figure 1c) with significantly lower colonies formed with old versus middle aged CD34 + patient BM cells (Figure 1d). These data demonstrate that although the frequency of BM progenitor cells was similar between our middle aged and old patients, old cells exhibit impaired function.

| Aging reduces long-term reconstitution potential and lymphoid cell production of human CD34 + cells
To assess the reconstitution potential of CD34 + cells in vivo, NSG mice were reconstituted with 7 x 10 5 human CD34 + cells from middle aged or old patient donors, creating Y-CD34 + and O-CD34 + NSG chimera respectively (Figure 1e). CD34 − chimeras were created by reconstituting NSG mice with an equal concentration of cells from the CD34 depleted cell fractions ( Figure S3a). Three months after reconstitution, examination of BM repopulation by flow cytometry revealed that younger CD34 + cells have the greatest repopulation potential, as indicated by the highest frequency in BM human CD45 + cells and human CD34 + cells (Figure 1f,g) in comparison with CD34 − cell group. A similar trend was also observed in the blood and spleen ( Figure S3b). However, lineage progenitor frequencies were similar between younger and older CD34 + transplanted mice ( Figure 1h).
Hematopoietic stem cells trended to be more abundant in younger CD34+ transplanted mice ( Figure 1i); however, this did not reach statistical significance as HSC engraftment was variable.
Further examination of the lineages produced by engrafted cells demonstrated that the lymphoid production capacity of younger CD34 + cells was significantly greater compared with old CD34 + cells, as mice transplanted with younger CD34 + cells exhibited greater levels of B cells (CD19 + ) in the BM 3 months posttransplant ( Figure 1j). B-cell levels also trended higher in both the blood and spleen of NSG mice reconstituted with younger versus old CD34 + cells at baseline. Myeloid cell repopulation (CD33 + ) was similar between groups ( Figure 1k); however, due to a reduction in lymphoid cell production, the myeloid (CD33 + ):lymphoid (CD19 + ) cell ratio was higher in NSG mice transplanted with old CD34 + versus younger CD34 + cells (Figure 1l). Progenitor cell engraftment and lineage production were not assessed in CD34 − due to the low number of cells engrafted at 3 months post-transplant. Together, these data demonstrate that CD34 + cells from the middle aged patient cohort have increased colony formation in vitro, increased reconstitution potential in vivo, and greater lymphoid cell production in vivo compared with CD34 + cells from the old cohort. Next, we examined cardiac function using echocardiography over a 4-week period. (Table S1) (Figure 2h,i). Although Y-CD34 + mice exhibit the greatest improvement in outcome, old CD34 + cells appear to offer some benefits as cardiac function was significantly greater compared to CD34 − mice at 4 weeks post-MI. To further understand the effect reconstituted CD34 + cells have on the repair process, we also compared the functional outcome of reconstituted mice to WT NSG mice which were irradiated at 285cGy 3 months prior to LAD ligation but did not receive BM transplant ( Figure S4).

| Reconstitution of NSG mice with human CD34 + cells improves cardiac function post-MI
LV dilation and cardiac function of WT NSG mice was not different from O-CD34 + and CD34 − mice at 4 weeks post-MI and was significantly worse than Y-CD34+ mice. This is consistent with a beneficial effect of Y-CD34 + cells. Collectively, these data demonstrate that NSG mice engrafted with younger CD34 + cells exhibit significantly better functional outcome and survival post-MI.

| Y-CD34 + cells enhance cardiac repair and minimize remodeling post-MI
To assess how CD34 + cells benefit infarct healing, we evaluated tis- to the vascularization process ( Figure S3d). Overall, these data indicate that engrafted human CD34 + cells modulate the infarct healing process post-MI to improve cardiac repair post-MI. Moreover, they demonstrate that Y-CD34 + cells more effectively reduce remodeling and stimulate infarct healing compared to old CD34 + cells.

| Human immune cell recruitment post-MI is reduced with age
Next, we examined early cellular responses post-MI. First, we assessed human CD45 + cell recruitment at 3 days (3d) post-MI by immunofluorescence to determine whether human immune cells also infiltrate the infarct post-MI. Interestingly, we found human cells in all groups, and Y-CD34 + hearts had the highest levels of human cells.
Human CD45 + cells were significantly higher in the infarcted myocar-

| Human CD34 + cells alter cytokine and MMP9 responses post-MI
Subtype analysis indicated that the T cells which infiltrate the heart post-MI are primarily T-helper cells, CD4 + (Figure 5a and Figure S5e).
Thus, we next assessed human cytokine levels in the heart at 3d  Human IL-10 and IL-4 were below the limits of detection, and human IL-1β and IL-2 were detected in low quantities and not different between Y-CD34 + and O-CD34 + hearts. Human IL-6 was higher in Y-CD34 + hearts; however, the response was variable as not statistically different between groups. We also assessed whole heart mouse cytokine levels at 3d post-MI to understand how human  (Figure 5f). This is consistent with increased remodeling in this group. MMP2 levels were not different between groups. Collectively, these data demonstrate that Y-CD34 + mice exhibit a greater human cell response primarily driven by the mobilization of T cells.

| Impact of transplanted cells on host cell responses post-MI
Although human cells are differentially recruited, our flow analyses indicate that the transplanted cells are a relatively smaller population compared to the host (mouse) cell population within the bone marrow 3 months after reconstitution ( Figure S2c). Therefore, we next examined whether the age of CD34 + cells can affect mouse cell re-   Although the molecular mechanisms underlying the beneficial effects of human CD4 + T cells require further investigation, one hypothesis is that CD4 + cells infiltrate the myocardium post-MI and modulate the cytokine/chemokine milieu to influence neighboring cell function. In F I G U R E 3 Younger CD34 + cells reduce remodeling and improve scar angiogenesis by 4 weeks post-MI. (a) WGA staining of younger CD34 + , old CD34 +, and CD34 − myocytes for cross-sectional area analysis at 4w post-MI, and quantification (52.  Using an IFNγ reporter, the authors also identified lymphocytes as a key source of IFN-y in the heart during early inflammatory phase post-MI. Interestingly, the loss of IFN-y was associated with lower neutrophil and monocyte cell numbers in the infarcted myocardium 3d post-MI, suggesting that this cytokine plays a role in the recruitment of these cells. As with any investigation involving a patient population, there are inherent limitations to this study. Here, the patient cohort was divided into "Middle Aged" or "Old" groups. In this investigation, the average age of the "Middle Aged" cohort is ~56 years old and the average age of the "Old" cohort is ~73 (Table 1). It is possible that the differences observed may be even more drastic if the "Middle Aged" cohort were 20-30 years of age, as in other investigations studying the effect aging has on BM function (Kuranda et al., 2011;Pang et al., 2011). However, our results support the notion that aging impairs Heart -mCD45 + Heart -Neutrophils (mLy6G + ) Heart -Monocytes (mLy6C hi ) Heart -Monocytes (mLy6C neg )

| Sternal bone marrow harvest
On the day of the surgery, an 18-gauge needle containing 5 ml of 10% heparin solution was advanced slowly through the peri-

| Bone marrow reconstitution
NOD-scid-IL2rγ null mice were irradiated at 285cGy twenty-four hours prior to injection using a Gammacell 40 Extractor Cesium-137 Irradiator (Best Theratronics). The following day, animals received an intravenous (tail vein) injection of freshly isolated 0.7 x 10 6 CD34 + or CD34 − cells. Cells used for all reconstitution studies were freshly isolated; cells were not frozen or expanded ex vivo. BM reconstitution analyses and myocardial infarction studies were performed 12 weeks after transplant.

| Myocardial infarction (MI) model and functional measurements
Myocardial infarction was performed 12 weeks after bone marrow

| Flow cytometry
The BM, blood, spleen, and heart were collected from mice 12 weeks post-reconstitution for baseline analyses as well as 3 and 7 day post-MI for infarction analyses. The BM, blood, and spleen cells were isolated, and red blood cells removed through red blood cell lysis. Hearts were minced and digested using collagenase type II (Worthington, 2 mg/ml) at 37°C for 30 min and filtered through a 70 µm filter. Single cell suspensions were re-suspended in FACS Buffer (PBS Ca2 + /Mg2 + free +1% FBS). Cells were blocked with anti-mouse CD16/32 (1:100, Biolegend) and Human Fc block (1:100, BD Biosciences) for 15 min on ice, after which cells were stained with the corresponding primary antibodies (Table S2). All primary antibody incubations were carried out for 1 h at 4°C in the dark after which cells were washed and events collected on a LSRII flow cytometer equipped with a violet laser (25 mW), blue laser (50 mW), yellow laser (100 mW), and red laser (40 mW). For identification of mouse cells, monocytes were CD45 + /CD11b + /F4/80 − /Ly6G-/ Ly6C hi/neg , neutrophils CD45 + /CD11b + /F4/80 − /Ly6G + , and macrophages CD45 + /CD11b + /F4/80 + . Gating was determined by FMOs, and all data were analyzed in FlowJo (TreeStar). For experiments using patient mononuclear cells, the isolated cells were stained for cell surface markers as described above, after which cells were analyzed on a LSR II flow cytometer. Gating used is shown in Figures S1, S2, and S7.

| Gelatin zymography
Infarcted hearts were homogenized in liquid nitrogen using mortar and pestle and kept on ice for an hour in zymography lysis buffer (5% glycerol, 0.1% TritonX-100 in 120 mM Tris buffer (pH 8.7)). After centrifugation (10,000 g), the supernatant was assayed for protein concentration using the DC protein assay (Bio-rad) according to the manufacturer's instructions. Twenty microgram of protein was electrophoresed in an 8% gelatin gel under non-reducing conditions. After electrophoresis, the gel was washed in 2.5% Triton X-100 for three times and incubated in developing buffer (49.1 mM Tris, 4.81 mM CaCl2, and 0.002% NaN3 at 37 °C for 24 h). The gel was stained with 0.5% Coomassie Blue R-250 and de-stained using 50% methanol and 10% acetic acid solution in water. The bands representing MMP9 and MMP2 were quantified using the ImageJ software.

| Protein isolation and cytokine analysis
Heart samples homogenized in liquid nitrogen using mortar and pestle. The total protein was extracted from powdered tissue using tissue lysis buffer (5% glycerol, 0.1% TritonX-100 in 120 mM Tris buffer [pH 8.7]). For analysis of human cytokine, abundance protein was quantified using the DC protein assay (Bio-rad) and suspended at a concentration of 900 µg/ml in lysis buffer supplemented with 0.5% BSA. Cytokines were quantified using the MILLIPLEX MAP Human TH17 Magnetic Bead Panel for anti-human IFNγ, IL-10, IL-1β, IL-6, IL-4, and IL-2. All samples and standards preparation as well as data acquisition was performed by the Princess Margaret Genomics Centre core facility (University Healthy Network, Toronto, Canada).

| Statistical analysis
All values are expressed as mean ± SEM. Analyses were performed using GraphPad Prism 8.0 software. Statistical comparisons were done using, an unpaired two-sided Student's t test, one-way, or two-way analysis of variance followed by a Tukey's post hoc for multiple com- parisons. Values of p ≤ 0.05 were considered statistically significant.

ACK N OWLED G M ENTS
This work was supported by grants from CIHR to RKL [332652].
TBM is a recipient of a CIHR Vanier Doctoral Award. FJA is a recipient of a CIHR Post-Doctoral Fellowship. Schematics in Figures 1e and 2a were prepared in BioRender.

CO N FLI C T O F I NTE R E S T S
None.

AUTH O R CO NTR I B UTI O N S
TBM, FJA, and RKL conceptualized and designed the study. TBM, FJA, AF, SHL, LW, and JW performed experiments. TBM, FJA, and LW analyzed data. RJC, MO, and TY performed patient sample collection. All authors drafted and edited the manuscript. All authors approved the final manuscript.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon request.