Murine marrow cellularity and the concept of stem cell ... - Nature

Leukemia (2004) 18, 575–583 & 2004 Nature Publishing Group All rights reserved 0887-6924/04 $25.00 www.nature.com/leu

Murine marrow cellularity and the concept of stem cell competition: geographic and quantitative determinants in stem cell biology GA Colvin1,5, J-F Lambert1,2,5, M Abedi1, C-C Hsieh3, JE Carlson3, FM Stewart4 and PJ Quesenberry1 1 Department of Research, Roger Williams Medical Center, Providence, RI, USA; 2Service of Hematology, University Hospital of Geneva, Geneva, Switzerland; 3Cancer Center, University of Massachusetts Medical School, Worcester, MA, USA; and 4Seattle Cancer Care Alliance, Seattle, WA, USA

In unperturbed mice, the marrow cell numbers correlate with the stem cell numbers. High levels of long-term marrow engraftment are obtained with infusion of high levels of marrow cells in untreated mice. To address the issue of stem cell competition vs ‘opening space’, knowledge of total murine marrow cellularity and distribution of stem and progenitor cells are necessary. We determined these parameters in different mouse strains. Total cellularity in BALB/c mice was 530720 million cells; stable from 8 weeks to 1 year of age. C57BL/6J mice had 466748 million marrow cells. Using these data, theoretical models of infused marrow (40 million cells) replacing or adding to host marrow give chimerism values of 7.5 and 7.0%, respectively; the observed 8-week engraftment of 40 million male BALB/c marrow cells into female hosts (72 mice) gave a value of 6.9170.4%. This indicates that syngeneic engraftment is determined by stem cell competition. Our studies demonstrate that most marrow cells, progenitors and engraftable stem cells are in the spine. There was increased concentration of progenitors in the spine. Total marrow harvest for stem cell purification and other experimental purposes was both mouse and cost efficient with over a four-fold decrease in animal use and a financial saving. Leukemia (2004) 18, 575–583. doi:10.1038/sj.leu.2403268 Published online 29 January 2004 Keywords: hematopoietic stem cells; cell differentiation; bone marrow transplantation; total marrow cellularity

Introduction A general assumption in stem cell transplantation is that the percentage of donor cells after transplantation correlate directly with the percentage of donor stem cells in the host. This has been the basis for most short- or long-term engraftment studies where the percentage of differentiated progeny has been ascribed to be a measure of the stem cell potency of the infused cells. Generally, in quantitative stem cell transplant models, a mixture of distinguishable marrow or marrow subset cells (purified populations) are infused into irradiated animals. Donor marrow cells with a discernible specific sex or congenic marker, for example, male DNA or CD45.1, are mixed with distinctly marked cells, for example, female DNA or CD45.2, and infused into lethally irradiated mice. The percent test donor cells at different points after cell infusions (transplantation) are then determined by flow cytometry analysis for surface phenotype markers or by molecular analysis (FISH for male DNA). These determinations are made on whole-blood and/or marrow cells and this is presumed to mirror the percentage of stem cells in the originally injected competitor population and the percentage of Correspondence: GA Colvin, Department of Research, Roger Williams Medical Center, 825 Chalkstone Ave, Providence, RI 02908-4735, USA; Fax: þ 1 401 456 5759; E-mail: [email protected] 5 Both authors contributed equally to this work Received 8 September 2003; accepted 24 November 2003; Published online 29 January 2004

persisting marked marrow-derived donor stem cells in host mice. In these animal studies, marrow cells with a discernible marker are transplanted into non- or minimally treated hosts. The infused cells compete against resident, residual host marrow. In previous studies we, as well as others, have shown high levels of engraftment into nontreated hosts and into minimally treated hosts.1–17 The key to high levels of engraftment in these models has generally been infusion of high numbers of donor cells to compete with resident or residual host marrow cells. Data from these studies have suggested that marrow space might not need to be ‘opened’ by cytotoxic host treatment for stem cell engraftment to occur. Rather, the critical determinant in the final percentage of chimerism in syngeneic or histocompatible chimerism is the competition between the infused donor and the resident host stem cells. In the past, it has been thought necessary to eradicate existing host marrow stem cells prior to cell infusion, opening up the extracellular sites of marrow hematopoiesis; the so-called niches.18 Given that the production of marrow or blood cells from nontreated mice appear to reflect stem cell content of marrow in these mice, ratios of theses cell types can be used to estimate stem cell content. The hypothesis that marrow engraftment is determined by stem cell competition indicates that final donor chimerism after transplant is directly related to the ratio of donor to residual host stem cells. To evaluate this accurately, knowledge of both distribution and numbers of host marrow cells are critical. This is addressed in this work.

Materials and methods

Mice Male and female BALB/c were purchased from Taconic Farms (Georgetown, NY, USA). Congenic male B6.SJL-PtprcaPep3b/ Boy.J (B6.SJL) (CD45.1), male C57BL/6J-Thy 1.2, Ly5.2-Pep (C57BL/6J) (CD45.2) and male DBA/2J were obtained from the Jackson Lab (Bar Harbor, ME, USA). All animals were housed in a conventional clean facility for at least a week before experimental use, certified to be specific pathogen free and given ad libitum access to autoclaved mouse chow and acidified water. The animals were maintained in accordance with the guidelines of the Institutional Animal Care and Use Committee at the University of Massachusetts Medical School and Roger Williams Medical Center.

Total murine cellularity determination Mice anesthetized with halothane were killed by cervical dislocation. The entire skeletons of mice age 2–16 months were

Total murine cellularity GA Colvin et al

576 manually dissected. The pelvic and thoracic limbs were prepared for removal by cutting away surrounding muscle. The pelvic limbs were then dislocated and removed; the knee joint cut and the bones cleaned. The pelvic bones were carefully dissected away intact. Next, the tail and all bones below the sacrum were removed. The tail was cutoff at the level of the third caudal vertebrae. The lower spine was removed below the 13th thoracic vertebrae. The rib cage, thoracic and cervical spine was then removed. The skull was then removed and cleaned. The bones were then crushed in a mortar with a pestle using cold HBSS (Life Technologies/Gibco/BRL) buffer. This was carried out for individual bones or groups of bones. The various marrow compartment solutions were filtered through a 40 mm strainer, and then washed once. For marrow compartment comparison experiments, marrow from the iliac crests, femurs and tibiae (six bones) were flushed using cold HBSS buffer in a syringe with a 22-gauge needle, filtered through a 40 mm strainer and then washed. The cells were then counted with a Neubauer hemocytometer (VWR Scientific Products, San Dimas, CA, USA). The cells were stained with crystal violet with 0.15%w/v acetic acid or trypan blue. Cell viability after whole marrow harvest was equivalent to that normally obtained with six bones crushed; viability was consistently above 90%.

In vitro progenitor assays Fresh B6.SJL marrow from various marrow spaces was harvested. A total of 20 000 marrow cells were analyzed via an in vitro colony-forming unit culture (CFU-c) and high proliferative potential colony-forming cell (HPP-CFC) assay using a double-layer soft-agar nutrient medium technique in 35 mm plastic culture dishes (BD Biosciences Falcont Discovery Labware, Bedford, MA, USA). The bottom layer consisted of: 1.0 ml underlay of 0.5% agar (Difco, Detroit, MI, USA) with a-MEM (Life Technologies/Gibco/BRL), 20% fetal calf serum (FCS; HyClone Laboratories, Logan, UT, USA), 1% penicillin/streptomycin (Life Technologies/Gibco/BRL) and 1% L-glutamine (Life Technologies/Gibco/BRL). This substrate was inoculated with a ‘seven-factor’ cytokine mixture with recombinant murine (rm) colony-stimulating factor-1 7500 U/dish (R&D Systems, Minneapolis, MN, USA), murine granulocyte– macrophage colony-stimulating factor-3.75 ng/dish (R&D), murine interleukin (IL)-1 alpha 375 U/dish (R&D), rm IL-3 150 U/ dish (Collaborative Research, Bedford, MA, USA), rm stem cell factor 150 ng/dish (R&D), murine granulocyte colony-stimulating factor 7.5 ng/dish (R&D) and rm basic-fibroblast growth factor 7.5 ng/dish (Collaborative). The cells were plated on a 0.5 ml overlay of 0.3% agar in a-MEM with 20% FCS. The dishes were incubated in a humidified 5% O2, 10% CO2, 85% N2 371C water-jacketed incubator (Forma Scientific Inc., Marietta, OH, USA). After 2 weeks, the colonies were scored. Highly dense colonies X0.5 mm in diameter or moderately dense colonies X1.0 mm were considered HPP-CFC. CFU-C were counted as non-HPP colonies with 450 cells. Total progenitors were combination HPP-CFC and CFU-C counts.19,20

tibiae and iliac crests with HBSS using a 22-gauge needle. Engraftment was determined by FISH.

In vivo engraftment assay – minimal-myeloablated model Fresh, 6- to 8-week male B6.SJL marrow was harvested into five groups: (1) crushed spine marrow, (2) flushed iliac, tibiae and femur marrow, (3) crushed iliac, tibiae and femur marrow, (4) crushed femur marrow and (5) crushed whole skeleton marrow. In all, 40 million cells from each group were infused separately via tail vein into C57BL/6J recipients after they received 2 Gy of whole-body irradiation given at a rate of 0.85 Gy/min 3 h prior to bone marrow transplant. A 137cesium source irradiator was utilized (Gammacell 40; Norton International, Nordian, Kanata, Ontario, Canada). The flushed iliac, tibiae and femur marrow cells were obtained via flushing with cold HBSS using a 22gauge needle. The bones from the other groups were crushed in a mortar with a pestle using cold HBSS buffer. The various marrow compartment cell suspensions were filtered through a 40 mm strainer, and then washed once. Peripheral blood chimerism was evaluated at 3, 6, 9, 12 and 15 weeks. Marrow spleen and thymus engraftment was analyzed at 15 weeks time. Flow cytometry was used for analyses.

In vivo engraftment assay – fully-myeloablated model A 6- to 8-week male B6.SJL marrow was injected with an equal amount of fresh male C57BL/6J unseparated bone marrow into lethally irradiated C57BL mice exposed to 9 or 10 Gy in a single dose given at a rate of 0.85 Gy/min. In total, 2 million fresh cells from each animal were competed together and injected via tail vein into each animal in a total volume of 0.5 ml/mouse. Fresh, unseparated 6- to 8-week male B6.SJL marrow was harvested into five groups: (1) crushed spine marrow, (2) flushed iliac, tibiae and femur marrow, (3) crushed iliac, tibiae and femur marrow, (4) crushed femur marrow and (5) flushed whole skeleton marrow. In total, 2 million cells from each group were competed against with 2 106 flushed six-bone (iliac crests, femurs and tibiae) cells. Cells from each group were pooled with equivalent numbers of competing cells (2 106) and infused via tail vein into C57BL/6J recipients after they received 10 Gy of whole-body irradiation. Peripheral blood chimerism was obtained at 3, 6, 9, 12 and 15 weeks. Marrow spleen and thymus engraftment and lineage analysis was performed at 15 weeks time. Flow cytometry was utilized for analyses. In separate experiments, fresh, unseparated 6- to 8-week male B6.SJL marrow was harvested into two groups: (1) crushed spine marrow and (2) crushed iliac, tibiae and femur marrow. Two million cells from each group were competed against 2 106 cells from crushed six bones (iliac crests, femurs and tibiae). Cells from each group were pooled with equivalent number of competing cells (2 106) and infused via tail vein into C57BL/6J recipients after they received 9 Gy of radiation. Analysis was as described above.

Fluorescent in situ Hybridization (FISH) In vivo engraftment assay – nonmyeloablated model Female BALB/c mice 6- to 8-week old were infused via tail vein injection with 40 106 fresh, unseparated male whole bone marrow cells. The cells were obtained via flushing from femurs, Leukemia

The FISH technique used identifies individual cells that contain the Y-chromosome.8 Briefly, cells were fixed, permeabilized with 0.1 N HCl/0.05% Triton X-100; washed with 2 SSC; denatured in 70% formamide in 2 SSC at 701C; dehydrated in


577 ice-cold ethanol and hybridized with a digoxigenin-labeled murine Y-chromosome painting probe at 371C overnight. Unbound probe was removed. The slides were blocked and detection of digoxigenin was carried out using antidigoxigenin fluorescein isothiocyanate-labeled Fab fragments (Boehringer Mannheim, Indianapolis, IN, USA). Finally, slides were mounted in the anti-fade media vectashield (Vector, Newcastle, United Kingdom) with 0.4 pM DAPI (40 6-diamidino-2-phenylindole 2HCl) (Sigma) to label nuclei (DNA). At least 200 cells from three different fields were analyzed per sample.

Flow cytometry engraftment and lineage analysis For the percent blood chimerism (B6.SJL competed with C57BL/ 6J), we incubated 50 ml of blood for 20 min at room temperature (RT) with anti-mouse fluorescein isothiocyanate (FITC)CD45.1 mAb (A-20 line, PharMingen), lysed the samples for 10 min at room temperature with 1.3 ml erythrocyte-lysing solution (150 mmol/l NH4Cl, 10 mmol/l NaHCO3, 1 mmol/l EDTA, pH 7.4), washed the samples with phosphate-buffered saline (PBS) and resuspended in PBS and 1% paraformaldehyde (Sigma). We measured the donor cell percentage by dividing FITC-positive cells by total nucleated cells. For bone marrow, spleen and thymus evaluation, we incubated 2 106 cells in PBS–5% HI–FCS with FITCCD45.2 mAb and biotin-CD45.1 mAb (PharMingen) for 30 min at 41C, washed and secondarily labeled with conjugated streptavidin allophycocyanin (SAv-APC) (Molecular Probes, Eugene, OR, USA). Donor chimerism was calculated as CD45.1/(CD45.1 þ CD45.2) as nucleated red blood cells do not label with CD45.1 or CD45.2 For lineage analysis, we labeled bone marrow with rat-antimouse Ly6G/GR-1 (8C5 hybridoma), CD45R/B-220 (RA3 hybridoma) or CD3 (T cells) (PharMingen) mAbs. Donor specificity was detected using a mouse-anti-mouse biotin-CD 45.1 mAb (PharMingen). We then labeled with an FITCpolyclonal goat anti-rat mAb (Southern Biotechnology Associates, Birmingham, AL, USA) and SAv-APC (Molecular Probes). For organ and lineage analysis, propidium iodide (Sigma) was used immediately preceding flow cytometry to gate out dead cells.21

Transplantation quantification In order to facilitate the presentation of transplantation data, especially quantitatively on a per organ or marrow compartment

(per bone) basis, we have arbitrarily designated 1% donor chimerism (in each model) to represent one transplantation unit. Therefore, the estimate of engraftable stem cells per organ would involve total cells per organs transplant units: resulting in transplant units per organ.

Differential cell count Cells from culture samples were centrifuged onto slides (350 rpm for 5 min) by cytospin (Shandon, Pittsburgh, PA, USA). Slides were stained with Wright–Giemsa (Sigma) staining. At least 100 mononuclear cells were counted.

Statistics Quantitative data are summarized by mean and standard error of the mean. For comparative analysis, the Wilcoxon’s rank-sum and the Kruskal–Wallis test were used. Kendall’s correlation coefficient, Cramer’s V (f) and Cuzick’s nonparametric trend test were also utilized.22–24

Results

Marrow cellularity Calculations of the total murine cellularity of male and female BALB/c were performed. Mice were 2–16 months of age. The mice were killed and the bones manually cleaned. Sample marrow distribution percentage in different marrow spaces are presented in Figure 1. Male and female BALB/c mice were analyzed separately for their total murine marrow cellularity. Table 1 shows mean cellularity of BALB/c and other strains at 8 weeks of age. Male and female BALB/c data from 21 male and 12 female mice ages 8–52 weeks of age were combined (Kendall’s correlation coefficient t ¼ 0.9045; Cramer’s V (f) ¼ 1.0). The mean BALB/ c marrow cellularity of mice 8–52 weeks old was 530720 million cells/animal (n ¼ 33). Marrow cellularity was equivalent from 2 to 12 months of age with a decrease to 246715 million cells at 16 months (n ¼ 2) (Figure 2). Comparing several commonly used murine strains, male and female BALB/c mice contained 503767 106 marrow cells at 8 weeks of age, C57/BL6 males had 466748 106 (n ¼ 5) and DBA/2 males had 284734 106 marrow cells (n ¼ 5) at 8 weeks of age (Table 1). DBA/2J mice contained significantly

Figure 1 Percentage of total marrow cellularity in separate marrow spaces. The entire skeleton of a 2-month-old BALB/c mouse was cleaned manually and the individual bones or groups of bones were crushed via mortar and pestle. The percent of marrow in each space is outlined in the legend with the spine housing approximately half of all hematopoietic cells in this representative animal. Leukemia


578 Table 1 strains

Total bone marrow cellularity among several 8-week-old

Table 2

Total number of progenitors per B6.SJL marrow space

Marrow space Strain

Sex

BALB/c BALB/c C57/BLK DBA/2

Female Male Male Male

Age (weeks)

N

Cellularity

s.e.m.

8 8 8 8

5 5 5 5

556 106 450 106 466 106 284 106

73 106 58 106 48 106 34 106

Tail Mandible Radii Tibiae Humeri Ribs Femurs Iliac crests Skull Lower spine Upper spine

HPP-CFC

CFU-c

Total progenitors

14587261 41087406 55657494 16437143 31807257 48237282 24127222 45857439 69967477 61227893 18 12671869 24 24872493 10 97171703 33 23172908 44 20272618 12 87971918 25 75872304 38 63772799 14 31072623 56 28677899 70 59679824 15 21171591 43 96474526 59 17574539 18 65672432 39 93673550 58 59273188 42 02974116 133 666710 837 175 695711 264 57 18777659 113 68576954 170 87279970

Total per mouse 182 87772142 476 52373813 Frequency 0.00034505 0.0008991

659 40074359 0.00124415

The mean and s.e. of the mean are presented.

Figure 2 Total bone marrow cellularity of BALB/c according to age. BALB/c mice were analyzed for their total marrow cellularity. The numbers of animals analyzed per time point are within parenthesis. Mean cellularity was found to be stable from 2 months to 1 year (8–52 weeks) at 530720 million with some drop-off in cellularity in the two mice analyzed at 16 months (64 weeks). The mean and s.e.m. are presented.

fewer marrow cells compared with the other two groups (Po0.0001). Boggs had performed similar studies in the female B6 DBA/2 using erythroid cell 59Fe radioisotope uptake and had obtained a remarkable close number of 260–280 106 as the total marrow cellularity for that strain.25

Stem cell/progenitor distribution We next evaluated stem cell/progenitor cell distribution content within different marrow spaces. This was carried out in part because of previous studies by Va´cha et al,26 who have shown that there is marked variation of the percent of erythropoiesis both between murine strains and location. We performed an in vitro HPP-CFC progenitor assay to complement the engraftment data (Figure 3). On a per cell basis, total in vitro progenitor content (HPP-CFC and CFU-c) for the spine was significantly higher than marrow from six bones crushed (Po0.02). Marrow from the femur had more progenitors than six bones crushed (Po0.02). Marrow from the whole skeleton had more progenitor expression than flushed and crushed six bones (Pp0.002). Looking just at HPP-CFC content, marrow from crushed femurs and whole skeleton had more progenitor expression than marrow from flushed six bones (Po0.005). When in vitro HPP-CFC progenitor assay results from individual murine bones were applied to the cellularity of that particular marrow space, we were able to calculate the total number of HPP-CFC, CFC-c and total progenitors contained within each marrow space (Table 2). The total number of HPPLeukemia

CFC per animal was 182 877 colonies with a frequency of 0.00034/cell in whole bone marrow using 530 million cells as the mean cellularity. Table 2 demonstrated that the upper and lower spine contained equivalent amounts of HPP-CFC, CFC-c and total progenitors. The spine collectively contains the vast majority of HPP-CFC progenitors in the mouse (Po0.0009). Equivalent numbers of HPP-CFC were observed between humerii, ribs, femurs, iliac crests and the skull. This group collectively had more HPP-CFC progenitors than the tail and mandible (Po0.03). The Kruskal–Wallis test was performed to compare all groups. To evaluate engraftable stem cells, we used a competitive in vivo assay in three separate models, a nonmyeloablated, a minimally myeloablated and lethally ablative model. In the last two models, we utilized B6.SJL (CD45.1) to C57BL/6 (CD45.2) transplants. Figure 4 shows peripheral blood engraftment from 3 to 15 weeks after transplant of 40 106 unseparated marrow cells into 2 Gy hosts. Engraftment into spleen and thymus at 15 weeks is shown in Figure 5. Figure 6 shows peripheral blood data from the same experimental model in 10 Gy hosts. Analysis of donor marrow from the lethally ablated model showed equivalent multilineage donor lineage expression of CD3, B220 and GR-1 antibodies from spine and leg engrafted marrow at 15 weeks with 2.370.2, 23.871.4 and 52.572.4% from leg marrow and 2.570.3 20.572.7 and 49.974.4% from engrafted spine marrow, respectively (P ¼ NS). Thus, these assays, whether measuring peripheral blood engraftment or marrow engraftment, showed significant increases in total spine engraftable units.

Stem cell competition An accurate tally of the total murine cellularity of BALB/c mice allowed us to test our hypothesis that in the syngeneic setting, the final host chimerism results from competition between donor and resident host stem cells. Previous dogma has held that marrow space had to be created by cytotoxic treatment opening special stem cell niches in order to make space for marrow stem cells to engraft. The critical issue is whether these niches are limiting for stem cell homing and engraftment. A number of groups, including our own have shown substantial engraftment into marrow of nonablated mice, but have not specifically


579

Transplant units per organ (%)

a

Transplant units (%)

3

6

9

12

Spleen

Thymus

40 30 20 10 0 6Bone Flush

6Bone Crush

Femur Crush

Spine Crush

All Crush

180 160 140 120 100 80 60 40 20 0 6Bone Flush

a 50

Bone Marrow

50

b Absolute # transplant units per marrow space (millions)

Figure 3 In vitro HPP-CFC analysis of fresh bone marrow from various marrow spaces. In vitro HPP-CFC analysis of fresh B6.SJL marrow from various marrow spaces of cells harvested. On a per cell basis, total in vitro progenitor content (HPP-CFC and CFU-c) for the spine was significantly higher than marrow from crushed legs (iliac, femur and tibia); Po0.02. Marrow from the femur had more progenitors than crushed legs Po0.02. Marrow from the whole skeleton had more progenitor expression than flushed and crushed legs Pp0.002. Marrow from the femur and whole skeleton had a higher concentration of HPP-CFC progenitor expression than marrow from flushed leg bones (Po0.005). The Kruskal–Wallis test was performed to compare all groups. The mean and s.e.m. are presented.

70 60

6Bone Crush

Femur Crush

Spine Crush

All Crush

15

Figure 5 Final donor chimerism and calculated donor cell volume after transplant of 40 106 marrow cells into 2 Gy irradiated hosts from various marrow spaces. (a) Marrow and organ engraftment analysis after transplant of 40 106 unseparated marrow cells into 2 Gy hosts from various marrow spaces performed 15 weeks posttransplant. (b) Final calculated donor cell volume distribution per marrow space showing that most of the resident donor cells reside in the spine as compared to six bones flush, six bones crush and femur crush at 15 weeks after transplant (Po0.0001). The Kruskal–Wallis test was performed to compare all groups. The mean and s.e.m. are presented.

40 30 20 10 0 6Bone Flush

6Bone Crush

Femur Crush

Spine Crush

All Crush

Spine Crush

All Crush

Absolute # transplant units per marrow space (millions)

b 250

3

6

9

12

15

200 150 100 50 0 6Bone Flush

6Bone Crush

Femur Crush

Figure 4 Blood chimerism and calculated donor cell volume after transplant of 40 106 marrow cells into 2 Gy irradiated hosts from various marrow spaces. (a) Peripheral blood engraftment after transplant of 40 106 unseparated B6.SJL marrow cells into 2 Gy C57BL/6 hosts in various marrow spaces 3, 6, 9, 12 and 15 weeks after transplant. (b) Calculated donor cell distribution per marrow space showing that most of the resident donor cells reside in the spine (Po0.0001). There were three to six animals per group transplanted with donor marrow from a total of 18 mice. The Kruskal–Wallis test was performed to compare all groups. The mean and s.e.m. are presented.

addressed the question as to whether niches were a limiting factor for engraftment.4,13–17 As will be demonstrated below, considering accurate total murine cellularity numbers, which equate with total stem cell numbers, and calculating theoretical engraftment when 40 million male BALB/c marrow cells are infused into nontreated female BALB/c hosts, and comparing these theoretical engraftment figures with those actually observed in 72 independent determinations, we arrive at virtually identical values. This would indicate that essentially all marrow engraftable stem cells engraft in marrow and that the final percent chimerism is determined solely by stem cell competition. In these models, we have assumed that: (1) all marrow stem cells home and engraft in marrow only; (2) that total marrow cells in a nonmanipulated cell population correlated directly with the number of stem cells in that population and (3) that at different times after marrow cell infusion in a transplant, the numbers of donor and host cells will correlate directly with the number of donor and host stem cells. We then either assumed that infused marrow physically displaced host marrow (replacement model) or added to host marrow (augmentation model). Given these conditions, the replacement and augmentation models gave theoretic engraftments (chimerism) of 7.5 and 7.0%, respectively. As controls for previously carried out experiments, we have studied 72 nontreated female BALB/c mice infused with 40 million male Leukemia


580

a 100

3

6

9

12

Transplant units (%)

80 60 40 20 0 6Bone Flush

6Bone Crush

Femur Crush

Spine Crush

All Crush

Absolute # transplant units per marrow space (millions)

b 300

3

6

9

12

250

Figure 7 Theoretical and experimental syngeneic nonmyeloablative transplant. Theoretical engraftment with replacement or augmentation models compared to observed engraftment. A total of 72 unperturbed 6- to 8-week-old female BALB/c mice were injected with 40 106 male whole marrow cells, as controls for other experiments. The mean engraftment by FISH was found to be 6.9%, which fits with the stem cell competition hypothesis.

200 150 100 50 0 6Bone Flush

6Bone Crush

Femur Crush

Spine Crush

All Crush

Figure 6 Blood chimerism after competitive transplant of marrow cells from various marrow spaces with flushed six-bone marrow into 10 Gy irradiated hosts. (a) Peripheral blood engraftment after transplant of 2 106 unseparated B6.SJL marrow cells into 10 Gy C57BL/6 hosts from various marrow spaces competed against 2 106 unseparated flushed iliac, femur and tibia marrow 3, 6, 9 and 12 weeks after transplant. (b) Calculated donor cell volume distribution per marrow space showing that most of the resident donor cells reside in the spine as compared to six bones flush, six bones crush and femur crush (Po0.0001). There were three to six animals per group transplanted with donor marrow from a total of 18 mice. The Kruskal–Wallis test was performed to compare all groups. The mean and s.e.m. are presented.

BALB/c marrow cells and evaluated engraftment results 8 weeks after cell infusion. The observed engraftment in these mice was 6.9170.4% (Figure 7). These results indicate an extraordinarily high percent chimerism and show that chimerism in a syngeneic setting is determined by host to donor stem cell ratios. Additional evidence for support of this hypothesis comes from experiments performed by Ramshaw et al,27 in which varying doses and schedules of male BALB/c marrow were infused into nontreated female BALB/c hosts showing that engraftment was 0.17% per million cells infused. Therefore if 40 106 cells were infused, the resulting chimerism would be 6.8%. This same report27 indicated that neither marrow cellularity nor content of progenitors was increased in mice injected with 200 million marrow cells over 5 days. In further work Bloomberg et al28 injected a total of 800 million marrow cells over time into nonmyeloablated BALB/c mice and showed that there was no increase in marrow or splenic cellularity. There was also no increase in marrow progenitors. These data indicate that the replacement model is probably the correct theoretical model. Using this model, we have also analyzed previously reported data from BALB/c female mice, which had been transplanted with 40 million male BALB/c cells for 5 conLeukemia

secutive days; a total of 200 million marrow cells.6 These mice were analyzed out to 25 months, close to the lifespan of a BALB/ c mouse. Theoretical engraftment using the replacement model is 37.7% while the actually observed engraftment for three mice at 25 months was 36.774.4% and for a total of seven mice at from 21 to 25 months, the observed engraftment was 31.373.8%. These data, utilizing a different total engrafted cell dose, represent strong additional verification of the stem cell competition hypothesis. In addition, the observations of longterm (25 month) multilineage (granulocyte, B- and T-cell) engraftment, as well as secondary and tertiary serial engraftment, indicate that true stem cells are engrafting in this nonmyeloablated transplant model.6,8 This model can be applied to minimally treated hosts. To evaluate the effect of 100 cGy on whole marrow, irradiated whole marrow was infused into nontreated hosts, which showed a reduction of expected engraftment by 91.4%. This would mean that a BALB/c mouse irradiated with 100 cGy would have the functional equivalent of 45 million residual whole marrow cells (530 106 8.6%). We can then use this number to interpret the results when 40 106 fresh male marrow cells are infused into 100 cGy irradiated female hosts. Average engraftment was 67.7711.5% for the nonirradiated donor cells.11 A theoretical engraftment level in 100 cGy mice with marrow replacement would be 40/45 106 or 88%. A model in which host marrow cells are augmented to host marrow, theoretical engraftment would be 40/85 106 or 47%. These data are consistent with the stem cell competition hypothesis with theoretically predicted engraftment in the replacement model exceeding experimental value by 21%. In addition, experiments performed by Cronkite et al29 using whole marrow transplants of 200 106 CBA/2 marrow into female recipients or same sex C57BL/6J transplants showed 16– 40% engraftment and increased progenitor expression (CFU-S) for the first 2 months after transplant. This was interpreted as stem cell augmentation. Assuming no alloimmunoreactivity,30


581 applying these data to a theoretical stem cell replacement model would yield chimerism of 200 106/280 106 ¼ 71% and an augmentation model would yield chimerism of 200/ (200 þ 280) ¼ 42%.

Cost–benefit analysis Table 3 depicts cost of obtaining 3.6 billion whole bone marrow cells (our usual starting number of cells) for stem cell purification when we use BALB/c mice. Five harvesting techniques were compared with animal use, technician use and time to complete procedure using data from our laboratory. We estimate that we save approximately $200 dollars/euros per harvest when obtaining marrow from the whole skeleton compared with crushing six bones as well as using 23 less animals per stem cell purification (at $10.50/mouse). This may be important in cases where one is attempting stem cell purification with a rare, expensive or hard to breed mouse or in general when availability of marrow cell numbers is limiting. For example, we are breeding GFP þ transgenic mice and estimate the cost of each mouse based on breeding and housing costs at approximately $234. We estimate the total marrow approach on these mice saves $179.23 per mouse or $5377 for a 30-mouse kill when we compare six bones crush with whole skeleton crush. This value increases to $203.27/mouse or $6098 when we compare it to six bones flush.

Discussion These results extend our knowledge of the numbers and distribution of hematopoietic cells in adult mouse bone marrow. Previous estimates of murine marrow cellularity were 260–280 million in female B6D2f1 (C57BL/6j female DBA2/J male) mice.25 Several different approaches have been taken to obtain these results, most include indirect methods such as uptake or turnover of iron using radioactive 59Fe labeling with iron citrate given intraperitoneally, and then extrapolating total body marrow cellularity after manual counting of a bone such as the humerus or femur. In Boggs initial studies, individual bones were analyzed after removal of all overlying soft tissue from the bones by baking the carcass in foil. Alternatively, the carcass was cut up into pieces with scissors and the pieces weighed and analyzed via scintillation counting. Accurate results were obtained this way, with some increased variability. Additional experiments by Boggs and Patrene31 used the hide beetle (Dermestes maculates) to clean the skeleton prior to scintillation. This technique required that the skeleton needed to be cleaned of all residual soft tissue as only B40% of the injected 59 Fe was recovered in the marrow with the remainder found in Table 3

various tissues. We selected a method involving mechanical cleaning of bones followed by mortar and pestle crushing. Our data and that of others have indicated that flushing of long bones is less efficient than crushing with B9% of cells remaining near the endosteum, but, in any case, of course flushing is not feasible (or at least would be very awkward) from some of the bony regions, that is, skull, vertebrae, mandible, etc.25 Using this approach, we found that there were distinct strain differences in total marrow cellularity. At 8 weeks of age BALB/c mice had a cellularity of 503750 million cells and C57BL/6 mice a cellularity of 466748 million cells, while male DBA2/J mice had a mean cellularity of 284734 million cells. This latter figure is consistent with that previously reported by Boggs and co-workers of 280 million cells, but is significantly less than that seen in the BALB/c mice (Po0.0001) or C57BL/6 mice (Po0.03). The total marrow cellularity of BALB/c mice was relatively stable from 8 to 56 weeks of age. A fall-off at 64 weeks of age was seen, but was based on only two BALB/c mice. The bulk of marrow cells were found in the spine, femurs and skull, and there were no significant alterations in cell types between these marrow regions. In general, the concentration of engraftable stem cells in different bony regions showed variability and was not consistently different between spine and other regions. The concentration of progenitors on a per cell basis was relatively increased in the spine over the six-bone crush. In contrast, the absolute numbers of progenitors and engraftable units, using multiple different in vivo assays, were predominantly located in the spine, indicating a real potential for sparing the use of mice by harvesting these bones with a crush technique. Harvesting all the bones further increases cell yield, but this is at a cost, involving increased time and technical work. We estimate that an entire skeletal harvest with four experienced technicians takes approximately 2 h, a spine skull harvest 1 h and 30 min, a tibia femur/iliac crest crush 1 h and 15 min, a tibia/femur iliac crest flush 1 h and a simple tibia blow out 45 min. Costs of these approaches are presented in Table 3. It is clear from this evaluation that the harvest of all bones is both mouse and cost-effective. These estimates do not include setup time, which is similar for each approach. Probably, if one is carrying out an experiment with engraftment of whole marrow cells for an estimate of stem/progenitor cell content in multiple mice a simple blowout should be used, while if one were attempting stem cell purification, necessitating relatively large numbers of cells, harvest of all bones would be the best approach. Beyond practical considerations of harvesting mouse marrow cells for study, the estimation of total mouse marrow cellularity is essential for arriving at a quantitative estimate of stem cell engraftment, which in turn is a necessary prerequisite for addressing the role of stem cell competition vs opening space in

Summary and analysis of methods used to harvest marrow for equivalent number of cells

Collection technique

Femurs flushed

The requisites for obtaining 3.6 billion BALB/c marrow cells Average yield per marrow space (million cells) 47.7 # mice for desired cell yield 75 BALB/c cost – $10.50 each $792 Time to obtain cells (min) 180 Technician cost Eight operators @ $20/h $480 Total cost for equivalent yield

$1272

Six bones flushed

Six bones crushed

Spine crushed

All bone crushed

109.2 33 $346 90

120.0 30 $315 80

275.6 13 $137 120

530.0 7 $71 100

$240

$213

$320

$267

$586

$528

$457

$338

Leukemia


582 marrow engraftment. We have previously reported on engraftment into nontreated BALB/c mice, showing relatively high levels of engraftment when ‘adequate numbers’ of marrow cells were infused. In the course of these studies, as noted above, we evaluated a total of 72 female BALB/c mice injected with 40 million BALB/c marrow cells and evaluated for 8-week engraftment. These were the controls for a variety of different experiments and provided a unique opportunity to assess whether engraftment was due to stem cell competition or space opening. The total number of marrow cells in a nonmyelosuppressed recipient mouse presumably is an accurate measure of the stem cell pool size in that mouse. Given this assumption, one can use the percent of donor cells in the recipient after marrow transplantation as a direct measure of the engraftment of donor cells in the transplanted mouse. Thus, knowledge of the total marrow cellularity in the normal host mouse and the final percentage of donor cells in a nontreated host allows one to calculate the expected vs observed engraftments in a transplantation setting. We have made two other minimalist assumptions: (1) that all stem cells home to the marrow and (2) that the infused cells replace or augment host marrow cells. It turns out that the replacement and augmentation models give very similar results in the theoretical models when 40 million marrow cells were infused, although as noted our results with analyzing marrow cellularity in nonmyeloablated mice transfused with from 200 to 800 million marrow cells suggest that the replacement model is the appropriate model. We found that if we assumed marrow replacement, the expected engraftment when 40 106 cells are infused into a nontreated BALB/c host would be 7.5%, while if an augmentation model is used the expected engraftment would be 7.0%. The observed engraftment in 72 male BALB/c to female BALB/c was 6.9170.4%. Both calculated values are remarkably close to the observed values, suggesting that virtually all marrow-engrafting stem cells home to marrow and engraft. These data were further confirmed by the studies of BALB/c female mice transplanted with 200 million BALB/c male marrow cells and analyzed 25 months later.6 These data raise a major question as to why the observed values are so high, since we know that marrow ‘stem cells’ engraft in the spleen and presumably seed other organs as well. In addition, this would be a remarkably efficient process to give a value so close to the predicted values. This does suggest that the marrow homing cell is responsible for ongoing hematopoiesis and that the cells homing elsewhere may, in fact, not be marrow homing cells and may not contribute to ongoing hematopoiesis. Obviously, in the context of these experiments, they are not marrow homing cells. It also raises the question as to whether the homing engrafting cells may be favored over host cells, although we have no data on this point. In addition, if BALB/c mice are exposed to 100 cGy radiation, 8-week engraftable stem cells are reduced to 8.6272.7% of nonirradiated mice. Extrapolating to total murine cellularity of 530 million cells, this would be equivalent to a residual of 45 106 marrow cells. When 40 106cells are infused into 100 cGy treated mice, mean 8-week engraftment of 67.7711.5% was obtained. A theoretical engraftment level in 100 cGy mice with marrow displacement would be 40/45 106 or 88%. A model in which marrow cells are added to host marrow, theoretical engraftment would be 40/85 106 or 47%. This is in the range of engraftment one finds in 100 cGy mice infused with 40 106 cells. All together these data suggest that a specific stem cell homes to marrow, that this process is efficient, and that stem cell competition is the critical determinant of sustained marrow Leukemia

engraftment. It is clear that cytotoxic marrow ablative or partially ablative host treatments markedly increase the percent donor cells in engraftment models, as would be expected from the stem cell depletion associated with these approaches. The observation that increased engraftment is seen in mice exposed to 100 cGy whole-body irradiation out to 1 year after exposure and that this increased host engraftability correlated with continued suppression of engraftable stem cells in the irradiated host marrow, is also consistent with stem cell competition determining engraftment in syngeneic transplantation.8 Whole skeleton BM cellularity in BALB/c mice is higher than previously measured in other strains and averages 530 million cells. Total bone marrow harvest is feasible, economical and provides high numbers of hematopoietic cells rich in progenitors. Depending on the details of the experimental approach, the total bone harvest can be recommended as saving both mouse use and money.

Acknowledgements This work was supported by National Institutes of Health sponsored grants: P01 HL56920-03, P01 DK50222-03, R01 DK27424-18, R01 DK60084, R01 DK60090-01A1, and K08 DK64980-01.

References 1 Lambert JF, Colvin GA, Zhong S, Wang H, D’Hondt L, Abedi M et al. H2-mismatched transplantation with repetitive cell infusions and CD40 ligand antibody infusions without myeloablation. Br J Haematol 2002; 119: 155–163. 2 Bradley MB, Sattler RM, Raftopoulos H, Ward M, Grossman IR, Townes TM et al. Correction of phenotype in a thalassemia mouse model using a nonmyeloablative marrow transplantation regimen. Biol Blood Marrow Transplant 2002; 8: 453–461. 3 Quesenberry PJ, Stewart MF, Peters S, Nillson S, Ramshaw H, Rao S et al. Engraftment of hematopoietic stem cells in nonmyeloablated and myeloablated hosts. Stem Cells 1997; 15 (Suppl 1): 167–169. 4 Stewart FM, Crittenden RB, Lowry PA, Pearson-White S, Quesenberry PJ. Long-term engraftment of normal and post-5-fluorouracil murine marrow into normal nonmyeloablated mice. Blood 1993; 81: 2566–2571. 5 Quesenberry PJ, Ramshaw H, Crittenden RB, Stewart FM, Rao S, Peters S et al. Engraftment of normal murine marrow into nonmyeloablated host mice. Blood Cells 1994; 20: 348–350. 6 Rao SS, Peters SO, Crittenden RB, Stewart FM, Ramshaw HS, Quesenberry PJ. Stem cell transplantation in the normal nonmyeloablated host: relationship between cell dose, schedule, and engraftment. Exp Hematol 1997; 25: 114–121. 7 Ramshaw HS, Crittenden RB, Dooner M, Peters SO, Rao SS, Quesenberry PJ. High levels of engraftment with a single infusion of bone marrow cells into normal unprepared mice. Biol Blood Marrow Transplant 1995; 1: 74–80. 8 Stewart FM, Zhong S, Lambert JF, Colvin GA, Abedi M, Dooner MS et al. Host marrow stem cell potential and engraftability at varying times after low-dose whole-body irradiation. Blood 2001; 98: 1246–1251. 9 Sykes M, Spitzer TR. Non-myeloblative induction of mixed hematopoietic chimerism: application to transplantation tolerance and hematologic malignancies in experimental and clinical studies. Cancer Treat Res 2002; 110: 79–99. 10 van den Bos C, Kieboom D, Wagemaker G. Correction of murine beta-thalassemia by partial bone marrow chimerism: selective advantage of normal erythropoiesis. Bone Marrow Transplant 1993; 12: 9–13. 11 Mardiney III M, Malech HL. Enhanced engraftment of hematopoietic progenitor cells in mice treated with granulocyte colonystimulating factor before low-dose irradiation: implications for gene therapy. Blood 1996; 87: 4049–4056.


583 12 Stewart FM, Zhong S, Wuu J, Hsieh C, Nilsson SK, Quesenberry PJ. Lymphohematopoietic engraftment in minimally myeloablated hosts. Blood 1998; 91: 3681–3687. 13 Micklem HS, Clarke CM, Evans EP, Ford CE. Fate of chromosomemarked mouse bone marrow cells tranfused into normal syngeneic recipients. Transplantation 1968; 6: 299–302. 14 Takada A, Takada Y, Ambrus JL. Proliferation of donor spleen and bone-marrow cells in the spleens and bone marrows of unirradiated and irradiated adult mice. Proc Soc Exp Biol Med 1971; 136: 222–226. 15 Takada Y, Takada A. Proliferation of donor hematopoietic cells in irradiated and unirradiated host mice. Transplantation 1971; 12: 334–338. 16 Brecher G, Ansell JD, Micklem HS, Tjio JH, Cronkite EP. Special proliferative sites are not needed for seeding and proliferation of transfused bone marrow cells in normal syngeneic mice. Proc Natl Acad Sci USA 1982; 79: 5085–5087. 17 Saxe DF, Boggs SS, Boggs DR. Transplantation of chromosomally marked syngeneic marrow cells into mice not subjected to hematopoietic stem cell depletion. Exp Hematol 1984; 12: 277–283. 18 Weiss L. The hematopoietic microenvironment of the bone marrow: an ultrastructural study of the stroma in rats. Anat Rec 1976; 186: 161–184. 19 Kriegler AB, Verschoor SM, Bernardo D, Bertoncello I. The relationship between high proliferative potential colony-forming cells in mouse bone marrow. Exp Hematol 1994; 22: 432–440. 20 McNiece IK, Bertoncello I, Kriegler AB, Quesenberry PJ. Colonyforming cells with high proliferative potential (HPP-CFC). Int J Cell Cloning 1990; 8: 146–160. 21 Riedy MC, Muirhead KA, Jensen CP, Stewart CC. Use of a photolabeling technique to identify nonviable cells in fixed

22 23 24 25 26

27

28 29 30

31

homologous or heterologous cell population. Cytometry 1991; 12: 133–139. Wilcoxon F. Individual comparisons by ranking methods. Biometrics 1945; 1: 80–83. Armitage PE, Berry G. Statistical Methods in Medical Research, 3rd edn. Oxford Blackwell Science Ltd, 1994. Cuzick J. A Wilcoxon-type test for trend. Stat Med 1985; 4: 87–90. Boggs DR. The total marrow mass of the mouse: a simplified method of measurement. Am J Heme 1984; 16: 277–286. Va´cha J, Hola J, Dungel J, Znojil V. The distribution of erythropoiesis over the various anatomical regions of the erythropoietic system in some inbred strains of mice. Exp Heme 1982; 10: 768–773. Ramshaw HS, Crittenden RB, Dooner M, Peters SO, Rao SS, Quesenberry PJ. High levels of engraftment with a single infusion of bone marrow cells into normal unprepared mice. Biol Blood Marrow Transplant 1995; 1: 74–80. Blomberg M, Rao S, Reilly J, Tiarks C, Peters S, Kittler E et al. Repetitive bone marrow transplantation in nonmyeloablated recipients. Exp Hematol 1998; 26: 320–324. Cronkite EP, Bullis JE, Brecher G. Marrow transfusions increase pluripotential stem cells in normal hosts. Exp Hematol 1985; 13: 802–805. van Os R, Sheridan TM, Robinson S, Drukteinis D, Ferrara JL, Mauch PM. Immunogenicity of Ly5 (CD45)-antigens hampers long-term engraftment following minimal conditioning in a murine bone marrow transplantation model. Stem Cells 2001; 19: 80–87. Boggs DR, Patrene KD. Marrow mass and distribution in murine skeletons cleaned by beetles as compared to cut up carcasses and a further simplification of the latter technique. Am J Hematol 1986; 21: 49–55.

Leukemia