Bone Marrow Transplantation, (1997) 20, 465–471  1997 Stockton Press All rights reserved 0268–3369/97 $12.00

Effect of chemotherapy for acute myelogenous leukemia on hematopoietic and fibroblast marrow progenitors

C Carlo-Stella1, A Tabilio2, E Regazzi1, D Garau1, R La Tagliata3, S Trasarti3, C Andrizzi3, M Vignetti3 and G Meloni3

1Department of Hematology, University of Parma; 2Department of Hematology, University of Perugia; and 3Department of Cellular Biotechnologies and Hematology, ‘La Sapienza’ University, Rome, Italy

Summary: Keywords: hematopoietic progenitors; microenviron- mental progenitors; acute myelogenous leukemia; chemo- Since reduced marrow cellularity and prolonged pancy- therapy; LTC-IC; hematopoietic engraftment topenia following autologous bone marrow transplan- tation (ABMT) have been frequently observed in patients with acute myelogenous leukemia (AML) included in the AML10 GIMEMA/EORTC trial, the The structural integrity of the hematopoietic system is question was raised to what extent hematopoietic and maintained by a relatively small population of self- microenvironmental progenitor cells were involved in renewing stem cells which can differentiate to produce pro- these patients. Marrow hematopoietic progenitors were genitors committed to terminal maturation.1 The develop- investigated by a short-term methylcellulose assay ment of hematopoietic cells in vivo occurs in intimate quantitating multipotent CFU-Mix, erythroid BFU-E association with a heterogeneous population of mesenchy- and granulocyte–macrophage CFU-GM, as well as a mal, connective tissue type cells and their associated long-term assay quantitating long-term culture- biosynthetic products, which constitute the stromal tissue initiating cells (LTC-IC). The marrow microenviron- of the bone marrow.2 Stromal cells of the hematopoietic ment was studied by evaluating the incidence of fibro- microenvironment include fibroblasts, endothelial cells, blastoid progenitors (CFU-F) and the capacity of stro- adipocytes, and macrophages.2 Based on a number of stud- mal layers to support allogeneic hematopoietic ies,3 the existence of self-renewing stromal stem cells with progenitors. As compared to normal controls (n = 57), multilineage differentiation capacity and capable of gener- AML patients (n = 26) showed a statistically significant ating progenitors with restricted development potential, reduction of the mean (± s.e.m.) number of CFU-Mix including fibroblast, osteoblast and chondrocyte progeni- (5.3 ± 0.6 vs 0.8 ± 0.2, P р 0.0001), BFU-E (68 ± 5 vs tors, has been hypothesized.4–6 20 ± 4, P р 0.0001), CFU-GM (198 ± 11 vs 144 ± 15, P Standard- and high-dose therapies currently used for the р 0.008), and LTC-IC (302 ± 46 vs 50 ± 8, P р 0.001). treatment of hematological and nonhematological malig- The mean (± s.e.m.) incidence of marrow CFU-F was nancies induce transient or permanent damage of hemato- not significantly reduced as compared to normal con- poietic and stromal progenitor cell compartments.7,8 trols (48 ± 6 vs 52 ± 7, P р 0.73). Seventeen AML stro- Despite such chemotherapy-induced defective progenitor mal layers were tested for their capacity to support the cell growth, the reinfusion of autologous marrow can recon- growth of allogeneic hematopoietic progenitors. Seven stitute the hematopoietic system, even in acute myelogen- samples failed to support any progenitor cell growth, ous leukemia (AML) patients treated with remission induc- seven had a significantly lower supportive activity as tion regimens exerting a significant marrow toxicity.9,10 compared to normal stromal layers (13 ± 5 vs 249 ± 56, Recently, therapeutic trials have been conducted in AML P р 0.002), whereas three cultures could not be ana- patients which were aimed at evaluating the efficacy of lyzed due to contamination. In conclusion, induction induction and consolidation regimens containing new anti- and consolidation regimens used in AML patients of the leukemic drugs, such as idarubicin or mitoxantrone.11 AML AML10 protocol induce a markedly defective in vitro patients included in the AML10 GIMEMA/EORTC pilot growth of primitive hematopoietic progenitors and a study receiving idarubicin, cytosine arabinoside, and etopo- severe functional defect of marrow stroma. The associ- side (ICE) as induction therapy and mitoxantrone and cyto- ation of hematopoietic with microenvironmental dam- sine arabinoside (NOVIA) as consolidation therapy, have age might play a key role in the delayed hematopoietic been reported to have reduced marrow cellularity, often regeneration observed following ABMT in patients of preventing marrow harvest and dramatically reducing the the AML10 trial. feasibility of autologous bone marrow transplantation (ABMT).12,13 In addition, as compared to historical controls treated with the conventional daunorubicin and cytosine 14 Correspondence: Dr C Carlo-Stella, Cattedra di Ematologia, Universita` di arabinoside (D3A7) protocol, patients autografted with Parma, Via Gramsci, 14, I-43100 Parma, Italy marrow harvested following the ICE/NOVIA regimens, Received 21 February 1997; accepted 26 May 1997 have experienced a significantly prolonged pancytopenia Bone marrow damage after chemotherapy for AML C Carlo-Stella et al 466 associated with increased transplant-related mortality. apy or at marrow harvest, respectively. Normal values for Reduced marrow cellularity and poor graft function were bone marrow progenitor cell growth were provided from also observed in patients included in the AML10 57 healthy subjects (38 males, 19 females) with a median randomized trial who received the same intercalating drug age of 42 years (range, 20–67). (idarubicin, mitoxantrone or daunorubicin) during remission induction and consolidation therapy.13 These Cell separation procedures clinical findings prompted us to hypothesize that adminis- tration of chemotherapy regimens of the AML10 protocols After informed consent was given, bone marrow was could result in severe toxic effects involving both hemato- obtained by aspiration from the posterior iliac crest and poietic and microenvironmental progenitor cell compart- mononuclear cells (MNC) were separated by centrifugation ments. (400 g, 30 min, 4°C) on a Ficoll–Hypaque gradient (density It was, therefore, the aim of the present study to investi- = 1.077 g/ml). Interface cells were washed and suspended gate marrow hematopoietic and microenvironmental pro- in RPMI-1640 (Gibco, Grand Island, NY, USA) sup- genitors in AML patients of the AML10 pilot or ran- plemented with 20% fetal bovine serum (FBS; Hyclone, domized trials. Hematopoietic progenitors were studied by Logan, UT, USA). means of a short-term methylcellulose assay allowing us to quantitate multipotent (CFU-Mix), erythroid (BFU-E) and CFU-Mix, BFU-E and CFU-GM assay granulocyte–macrophage (CFU-GM) progenitors as well as a long-term assay which allows us to quantitate primitive The assay for multilineage colony-forming units (CFU- progenitors capable of initiating and sustaining hematopo- Mix), erythroid bursts (BFU-E), and granulocyte– iesis in long-term culture (long-term culture initiating cells, macrophage colony-forming units (CFU-GM) was carried LTC-IC).15 The marrow microenvironment was studied by out as described elsewhere.18 Briefly, 5 × 104 MNC were quantitating marrow fibroblast colony-forming cells (CFU- plated in 35-mm Petri dishes in 1-ml aliquots of Iscove’s F) and by evaluating the capacity of stromal layers to modified Dulbecco’s medium (IMDM; Seromed, Berlin, support the growth of allogeneic normal hematopoietic Germany) containing: 30% FBS (Stem Cell Technologies, progenitors.16 Vancouver, Canada); 10−4 m 2-mercaptoethanol (Gibco), and 1.1% (w/v) methylcellulose. Cultures were stimulated with interleukin-3 (10 ng/ml; Sandoz, Basel, Switzerland), Materials and methods granulocyte colony-stimulating factor (10 ng/ml, Amgen, Thousand Oaks, CA, USA), granulocyte–macrophage Patients colony-stimulating factor (10 ng/ml; Sandoz) and erythro- poietin (3 U/ml; Amgen). Progenitor cell growth was evalu- ° Twenty-six patients (12 males, 14 females) with a median ated after incubation (37 C, 5% CO2) for 14–18 days in a age of 45 years (range, 14 to 59) and a diagnosis of AML humidified atmosphere. Four dishes were set up for each were included in this study. The diagnosis of AML was individual data point per experiment. CFU-Mix defined as based on the criteria of the French–American–British containing at least erythroid and granulocytic cells, BFU- (FAB) Cooperative Group.17 The study population included E with у500 cells and CFU-GM with у40 cells were patients with different FAB subtypes (M1 (n = 3), M2 (n scored from the same dish. = 9), M4 (n = 8), M5 (n = 5), and M6 (n = 1)). All patients had been treated at the Department of Hematology, ‘La Long-term culture-initiating cell (LTC-IC) assay Sapienza’ University, Rome, according to different remission induction and consolidation protocols. Nine The long-term culture-initiating cell (LTC-IC) assay was patients enrolled in the AML10 GIMEMA/EORTC pilot performed according to a previously described tech- study received the ICE (idarubicin, cytosine arabinoside, nique.15,19 Briefly, MNC were resuspended in complete etoposide) regimen as remission induction therapy and the medium consisting of alpha-medium (Gibco) supplemented NOVIA (mitoxantrone, cytosine arabinoside) regimen as with fetal bovine serum (12.5%), horse serum (12.5%), l- consolidation therapy, whereas the remaining 17 patients glutamine (4 mm), 2-mercaptoethanol (10−4 m), inositol (0.2 were randomized to receive three different regimens mm), folic acid (20 ␮m) and freshly dissolved hydrocorti- (DCE/DIA (daunorubicin, cytosine arabinoside, eto- sone (10−6 m). MNC (3–5 × 106) were seeded into cultures poside/daunorubicin, cytosine arabinoside) (n = 6), containing a feeder layer of irradiated (8000 cGy) murine ICE/IDIA (idarubicin, cytosine arabinoside, etoposide/ M2–10B4 cells (3 × 104/cm2, kindly provided by Dr C idarubicin, cytosine arabinoside) (n = 6), MICE/NOVIA Eaves, Terry Fox Laboratories, Vancouver, Canada) engi- (mitoxantrone, cytosine arabinoside, etoposide/ neered by retroviral gene transfer to produce human IL-3 mitoxantrone, cytosine arabinoside) (n = 5)), containing the and human G-CSF.20 Cultures were fed weekly by replace- same intercalating drug during induction and consolidation ment of half the growth medium containing half the nonad- phase.13 Bone marrow cultures were performed before (n herent cells with fresh complete medium. After 5 weeks in = 4) or after (n = 7) consolidation therapy or at marrow culture, nonadherent cells and adherent cells harvested by harvest (n = 15). The median time intervals from diagnosis trypsinization were pooled, washed and assayed together to study were 122 days (range, 34–140), 142 days (range, for clonogenic cells in standard methylcellulose cultures at 90–198) and 150 days (range, 114–248) for patients evalu- an appropriate concentration (usually 5 × 104/ml). The total ated before consolidation therapy, after consolidation ther- number of clonogenic cells (ie CFU-Mix plus BFU-E plus Bone marrow damage after chemotherapy for AML C Carlo-Stella et al 467 CFU-GM) present in 5-week-old LTC provides a relative data (two-tail) was used to test the probability of significant measure of the number of LTC-IC originally present in the differences between samples. test suspension.21 Absolute LTC-IC values were calculated by dividing the total number of clonogenic cells by four, which is the average output of clonogenic cells per LTC- Results IC, according to limiting dilution analysis studies reported by others.21 Hematopoietic progenitors Hematopoietic progenitor cell growth was evaluated in the Assays of the hematopoietic microenvironment bone marrow of 26 AML patients who had achieved com- plete hematological remission with four different protocols In order to evaluate the effect of remission induction and of remission induction and consolidation.13 At the time of consolidation therapy on marrow microenvironment, the the study, 22 patients had received remission induction and incidence of CFU-F and the capacity of marrow cells to consolidation therapy, whereas four cases (Nos 16, 18, 19, generate stromal layers were analyzed. The assay for CFU- 24) had only received remission induction therapy. Blood F was performed according to Castro–Malaspina et al.16 values and marrow morphology were normal and no patient Briefly, 4 ml of marrow MNC (5 × 104/ml) resuspended in had marrow fibrosis. For the entire group of AML patients complete alpha-medium were plated in 60-mm Petri dishes (n = 26), the median interval between diagnosis and study and fibroblast colony growth was evaluated after incubation was 145 days (range, 34–248). ° (37 C, 5% CO2) for 14 days in a humidified atmosphere. Table 1 shows that, as compared to normal controls (n Four dishes were set up for each individual data point per = 57), AML patients showed a statistically significant experiment. For scoring CFU-F, the dishes were stained reduction of the mean (±s.e.m.) number of multipotent with crystal violet and examined under an inverted micro- CFU-Mix (5.3 ± 0.6 vs 0.8 ± 0.2, P р 0.0001), erythroid scope. Fibroblastoid cell aggregates of у50 cells were BFU-E (68 ± 5 vs 20 ± 4, P р 0.0001), and granulocyte– scored as CFU-F. To generate marrow stromal layers, the macrophage CFU-GM (198 ± 11 vs 144 ± 15, P р 0.008). method of Dexter et al1 was used with slight modifications. Briefly, 6-ml aliquots of marrow MNC (1 × 106/ml) resus- pended in complete alpha-medium were inoculated into 25- Table 1 Hematopoietic colony formation cm2 tissue culture flasks and incubated at 37°C in a humidi- Case CFU-Mixa BFU-Ea CFU-GMa LTC-ICb fied atmosphere supplemented with 5% CO2. On a weekly basis, the cultures were fed by complete replacement of 1 0 14 170 15 complete alpha-medium and analyzed for stromal conflu- 2 0 17 311 14 ence and adipocyte formation.22 When a confluent (defined 3 0.5 17 203 38 as Ͼ75% of the flask surface covered by stromal cells) or 4 0.8 36 244 16 subconfluent (defined as 50 to 60% of the flask surface 5 0.3 25 191 24 6 0.5 18 220 3 covered by stromal cells) stromal layer was observed after 7 0 0 100 44 5 weeks, the stromal supportive function was evaluated by 8015671 irradiating (1500 cGy) the flasks and inoculating into each 9 1 40 162 134 flask allogeneic adherent cell-depleted peripheral blood 10 0 0 216 145 × 6 11 2 47 169 94 cells (3 10 ) obtained from G-CSF-mobilized, consenting 12 1 25 153 76 normal donors undergoing harvesting for transplantation. 13 0 5 104 76 Nonadherent cells were obtained by removing monocyte– 14 2 16 137 0 15 1 5 36 8 macrophages by a 1 h (37°C, 5% CO2) plastic-adherence procedure. To quantitate the stromal supportive function, 16 0 6 80 73 17 0 13 93 26 cultures were processed as for the LTC-IC assay (see 18 0.8 19 194 94 above) and the supportive function of AML stroma was 19 0 1 21 20 assessed as number of progenitors detected after 5 weeks 20 0.3 23 80 0 in LTC. In preliminary experiments, counting of confluent 21 1.8 10 94 NE 22 1.3 19 69 76 stromal layers following trypsinization did not reveal sig- 23 1.5 33 206 NE nificant quantitative variation from patient to patient. The 24 0 0 0 NE function of AML stroma was compared in parallel experi- 25 2.5 28 149 NE ments with that of normal marrow stroma. 26 3.7 90 287 NE Normal controls (n = 57) Mean 5.3c 68c 198d 302e Statistical analysis s.e.m. 0.6 5 11 46 Range 1.0–19 8–208 33–483 13–1859 Four plates were scored for each data point per experiment and the results were expressed as the mean ±1 s.e.m. Stat- aMean values per 5 × 105 mononuclear cells. istical analysis was performed with the statistical package bMean values per 2 × 106 mononuclear cells inoculated on day 0. cCompared with AML-derived CFU-Mix or BFU-E, P р 0.0001. Statview (BrainPower, Calabasas, CA, USA) run on a Mac- dCompared with AML-derived CFU-GM, P р 0.008. intosh Performa 6300 personal computer (Apple Computer, eCompared with AML-derived LTC-IC, P р 0.001. Cupertino, CA, USA). The Student’s t-test for unpaired NE = not evaluable for culture contamination. Bone marrow damage after chemotherapy for AML C Carlo-Stella et al 468 A statistically significant growth reduction was also dose chemotherapy regimens used in AML patients were observed for the primitive LTC-IC progenitors which were not able to quantitatively damage marrow CFU-F. reduced on average by six-fold as compared to normal con- To characterize the marrow microenvironent at a func- trols (302 ± 46 vs 50 ± 8, P р 0.001). tional level, the formation of stromal layers and their capa- As compared to controls, AML patients studied at the bility to support the growth of allogeneic hematopoietic time of marrow harvest (n = 15) showed a persistent progenitors were analyzed. As shown in Table 2, 62% of reduction of LTC-IC (302 ± 46 vs 55 ± 11, P р 0.008) AML patients generated a confluent stroma whereas 19% associated with normalization of CFU-GM growth generated a subconfluent layer and 19% failed to produce (198 ± 11 vs 167 ± 16, P р 0.2). No substantial difference any kind of adherent stromal layer. In contrast, nearly all in the incidence of committed and primitive progenitors (95%) normal marrow samples (n = 28) used as control was observed when AML patients were stratified according produced a confluent marrow stroma with only two samples to the different induction and consolidation regimens, or resulting in subconfluent growth and no sample showing inclusion in the pilot or randomized AML10 trial. absence of stromal layer. Adipocyte formation was evident in 62% of AML patients and 100% of normal samples (Table 2). No difference between normal and AML was Microenvironment progenitors observed in terms of time to reach the confluence and time In contrast to hematopoietic progenitors, the mean (± of appearance of adipocytes. s.e.m.) incidence of fibroblastic progenitors was not sig- Thirteen confluent and four subconfluent AML stromal nificantly reduced as compared to normal controls (48 ± 6 layers could be tested for their capacity to support the vs 52 ± 7, P р 0.73) (Table 2). No difference in CFU-F growth of allogeneic peripheral blood LTC-IC. Seven AML growth could be detected when AML patients were strati- stromal layers, including five confluent and two subconflu- fied according to time of the study (preconsolidation vs ent layers, failed to support any progenitor cell growth, post-consolidation vs harvest), different induction–consoli- whereas seven AML samples, including six confluent and dation regimens, inclusion in the pilot or randomized one subconfluent stromal cultures, had a markedly defective AML10 trial. Overall, these results suggest that standard- supportive capacity which was significantly lower as com- pared to normal marrow-derived stromal layers (13 ± 5 vs 249 ± 56, P р 0.002) (Figure 1). Table 2 Fibroblastoid progenitor cell growth and stroma character- istics

Case CFU-F Time to Adipocyte Stromal Discussion per 106 confluence formation supportive MNCa (weeks) (weeks)b function Stem cell transplantation (SCT) is an increasingly used therapeutic approach for the treatment of hematological 1164c 4 Absent and non-hematological diseases of neoplastic and non-neo- 2 0 4 4 Absent 23 3 48 No growth — — plastic origin. Successful SCT depends on engraftment of 4 28 No growth — — pluripotent hematopoietic stem cells in the marrow 5 48 2 Absent Defective microenvironment and the capability of stem cells of hom- 6 68 4 4 Defective ing to the marrow and docking at specific sites.24 Therefore, 7NDNDNDND the spatial organization of stem cells in the marrow, 8 108 3 Absent Defective 9 124 3 Absent Defective mediated by the hematopoietic microenvironment and 10 24 No growth — — extracellular matrix, is crucial for hematopoietic regener- 11 64 ND ND ND ation following SCT.2 12 76 Subconfluenced 4 Defective Since reduced marrow cellularity and prolonged pancyto- 13 16 No growth — — 14 88 ND ND ND penia following ABMT have been frequently observed in 15 24 5 Absent Absent AML patients of the AML10 protocols, the question was 16 36 Subconfluence 3 Absent raised as to what extent hematopoietic and microenviron- 17 36 3 3 Defective 18 68 3 3 Absent 19 8 4 3 Defective 300

20 44 3 2 Absent cells 21 52 Subconfluence 3 Contamination 6 10 200

22 104 ND ND ND × 23 44 4 3 Contamination 24 28 3 3 Contamination 100 25 4 Subconfluence 3 Absent P р 0.002 26 ND ND ND ND 0 LTC-IC per 2 LTC-IC Normal Stroma AML Stroma aMean (± s.e.m.) CFU-F generated by normal controls (n = 28) were 52 ± 7 (P р 0.73). Figure 1 Marrow stromal function in normal subjects and AML patients. bAfter first adipocytic area observed. All stromas were recharged with allogeneic adherent cell-depleted peri- cConfluence defined as у75% of the flask surface covered by stromal cells. pheral blood cells (3 × 106). Cultures were processed as for the LTC-IC dSubconfluence defined as 50 to 60% of the flask surface covered by assay (see Materials and methods) and the supportive function was stromal cells. expressed as number of progenitors produced by normal LTC-IC after 5 ND = not done due to low cell number. weeks in long-term culture. Bone marrow damage after chemotherapy for AML C Carlo-Stella et al 469 mental progenitor cells were damaged in these patients. In stantial contribution of CFU-GM in the early phase of agreement with previously published papers, data reported engraftment. Indeed, using a homogeneous population of herein clearly demonstrate that patients with AML treated primitive human hematopoietic cells, in vitro evidence has with anthracycline- or mitoxantrone-containing regimens been provided suggesting the existence of a mechanism that have a significant reduction of primitive (LTC-IC) and can provide both short-term and long-term hematopoietic committed (CFU-Mix, BFU-E, CFU-GM) marrow progeni- repopulation.37 Therefore, the marked LTC-IC reduction tors.25–27 In our patients, progenitor cell damage was detected in our AML patients might be viewed as one detected at the time they achieved complete remission fol- biologically relevant factor conditioning the prolonged lowing induction chemotherapy and was still evident at the pancytopenia associated with increased transplant-related time of marrow harvest when, despite a substantial evi- mortality observed following ABMT. dence of CFU-GM regeneration, as suggested by the nor- The CFU-F assay identifies a class of clonogenic fibro- mal incidence of these progenitors, primitive LTC-IC blast progenitors38 which belong to the osteogenic stromal remained markedly reduced. It has been shown that follow- lineage39 and play an important role in establishing the ing a cytotoxic insult, self-renewal of stem cells takes pre- hematopoietic microenvironment both in vitro and in cedence over differentiation usually allowing the stem cell vivo.3,40 Ectopic transplantation of individual fibroblastic system to recover to a normal level.28 In our study popu- clones grown in vitro from mouse marrow beneath the renal lation, we could demonstrate CFU-GM, but not LTC-IC capsule of syngeneic hosts demonstrated that approximately regeneration. The persistent derangement of the LTC-IC 15% produced a marrow organ containing the full spectrum compartment associated with a normal CFU-GM incidence of stromal cell types of the hematopoietic microenviron- suggests a permanent defect, due to an irreversible toxic ment.4,5 A proportion of marrow CFU-F cultured in the effect of chemotherapy, rather than a transient phenomenon presence of hydrocortisone is induced to adipogenesis.6 related to the short time interval between chemotherapy In vivo as well as in vitro chemotherapy has been shown administration and time of the study. to damage marrow stroma.7,25,41,42 We demonstrate that The association of a decreased number of hematopoietic AML marrow has a maintained CFU-F incidence but fails progenitors with normal peripheral blood counts indicates to generate confluent stromal layers in a significant percent- that the hematopoietic system is capable of compensating age of cases. This suggests that fibroblastic progenitors, for the deficiency of early hematopoietic progenitor cells, although quantitatively normal, are functionally altered in as previously shown in patients who have undergone highly that they have lost their proliferative and differentiative aggressive chemotherapy or allogeneic bone marrow trans- potential which substantially contribute to the generation of plantation.29 The reduced number of hematopoietic pro- a confluent marrow stroma.4,5 Four patients (cases 3, 4, 10 genitors coexisting with normal white blood cell counts and 13) exhibited growth of CFU-F but failed to generate suggests an inverse relationship between the differentiative stromal cell layers in LTC culture conditions. Differences capacity and potential for stem cell self-renewal.29 This in cell concentration, frequency of medium change, scoring implies that in our AML patients an increased demand on time and culture duration, which distinguish the culture sys- stem cells for differentiation takes precedence over the tems we used to grow CFU-F or generate stromal layers, demand for self-renewal, thus compensating for main- might only partially explain the discrepancy between taining the supply of end cells to the peripheral blood. Such growth of CFU-F and failure to generate stromal cell layers. a mechanism might explain the LTC-IC damage we Since stromal layers are generated either by a stromal stem detected in AML patients evaluated at the time of mar- cell(s) or different classes of progenitor(s) which include, row harvest. but are not limited to CFU-F,3 we hypothesize that in some Despite recent evidence demonstrating that the LTC-IC patients chemotherapy-induced stromal toxicity might pref- assay detects a class of primitive stem cells with phenotypic erentially affect stromal progenitors other than CFU-F and characteristics of transplantable murine in vivo repopulating this might translate into failure of stromal layer generation cells30 and self-renewal potential,31 the role of primitive in 5-week-old LTC associated with growth of day-14 LTC-IC in promoting hematopoietic recovery following CFU-F. SCT as well as the dose of these progenitors which is Either a complete failure or a markedly defective required for a prompt and durable engraftment is still a capacity of AML stroma to support the growth of allo- matter of controversy. Distinct subsets of stem cells may geneic normal hematopoietic progenitors was observed. be responsible for different phases of engraftment after The number of stromal cells in confluent cultures was not stem cell transplantation.32,33 Murine studies in which iso- assessed on a routine basis. However, in preliminary enzyme analysis and retroviral gene marking of hematopo- experiments, counting of confluent stromal layers following ietic cells have been used to track the fate of stem cells trypsinization did not reveal significant quantitative vari- support the existence of short-term and long-term reconsti- ation from patient to patient. Since the evidence of a defec- tuting stem cell populations.34–36 In human subjects there tive supportive function was derived from six confluent, is no clear evidence for the separation of distinct subsets and only one subconfluent, stromal cell layers, we assume of cells with differing repopulating potentials. A functional that the defective supportive ability of AML stroma is due heterogeneity of transplantation potential of human stem to qualitative differences in stromal cell function rather cells can be argued from recipients of mobilized peripheral than quantitative differences in the number of stromal cells. blood SCT who have a rapid hematopoietic reconstitution Under physiological conditions, stromal cells of the hem- suggesting either that mobilized peripheral blood is atopoietic microenvironment include fibroblasts, endo- enriched for short-term repopulating stem cells or a sub- thelial cells, adipocytes and macrophages.2 These different Bone marrow damage after chemotherapy for AML C Carlo-Stella et al 470 cell types provide the physical framework within which opment: experimental and clinical studies. Leukemia 1989; 3: hematopoiesis occurs, play a role in directing the processes 469–474. by synthesizing, sequestering or presenting growth- 3 Gronthos S, Simmons PJ. The biology and application of stimulatory and growth-inhibitory factors, and also express human bone marrow stromal cell precursors. J Hematother a broad repertoire of adhesion molecules that serve to 1996; 5: 15–23. 4 Owen ME, Cave J, Joyner CJ. Clonal analysis in vitro of mediate specific interactions with hematopoietic stem/ 2,43 osteogenic differentiation of CFU-F. J Cell Sci 1987; 87: progenitor cells. Based on our data, it cannot be ruled 731–739. out that differences in cell composition between AML and 5 Owen ME, Friedenstein AJ. Stromal stem cells: marrow- normal stromal layers might be an additional factor respon- derived osteogenic precursors. CIBA Found Symp 1988; 136: sible for the lack of supportive function by AML stroma. 42–60. We suggest that the functional damage of the microen- 6 Bennet JH, Joyner CJ, Triffitt JT, Owen ME. Adipocyte cells vironment we detected in AML patients exerts a primary cultured from marrow have osteogenic potential. J Cell Sci role in conditioning the poor graft function observed in 1991; 99: 131–139. AML10 patients following ABMT. This hypothesis is 7 Greenberger JS. Toxic effects on the hematopoietic microen- indirectly supported by the evidence that chemotherapy- vironment. Exp Hematol 1991; 19: 1101–1109. 8 Betticher DC, Huxol H, Muller R et al. Colony growth in related hematopoietic progenitor cell damage is a well- cultures from bone marrow and peripheral blood after curative known consequence of antineoplastic therapy, but usually treatment for leukemia and severe aplastic anemia. Exp Hema- does not prevent autografting procedures. Thus, the tol 1993; 21: 1517–1521. microenvironmental injury might play a primary role in 9 Burnett AK. Autologous transplantation in acute leukemias, supporting a poor graft function observed in AML patients including purging. Curr Opin Oncol 1990; 2: 263–273. of the AML10 trials. 10 Meloni G, De Fabritiis P, Petti MC, Mandelli F. BAVC regi- The observation that hematopoietic regeneration occurs men and autologous bone marrow transplantation in patients within the expected length of time in AML10 patients with acute myelogenous leukemia in second remission. Blood undergoing allogeneic transplantation (GM, unpublished 1990; 75: 2282–2285. data) suggests that engraftment may take place even in the 11 Berman E, Heller G, Santorsa J et al. Results of a randomized trial comparing idarubicin and cytosine arabinoside with dau- presence of a damaged microenvironment. However, norubicin and cytosine arabinoside in adult patients with patients undergoing allogeneic transplant are reinfused not newly diagnosed acute myelogenous leukemia. Blood 1991; only with normal hematopoietic cells but also with 77: 1666–1674. ‘engraftment facilitating cells’, which are donor-derived 12 Crump M, Keating A. Acute leukemia in adults. Curr Opin marrow accessory cells separate from hematopoietic stem Hematol 1995; 2: 247–254. cells.44 13 Muus P, Zittoun R, Mandelli F et al. Randomized phase III In conclusion, induction–consolidation regimens used in study of induction (ICE vs MICE vs DCE) and intensive con- AML patients of the AML10 protocols induce a markedly solidation (IDIA vs NOVIA vs DIA) followed by stem cell defective in vitro growth of primitive hematopoietic pro- transplantation in AML: EORTC/GIMEMA AML10 study. genitors and a severe functional defect of marrow stroma. Eighth Int Symp Autologous Marrow and Blood Transplant 1996; 4 (Abstr.). The association of hematopoietic with microenvironmental 14 Zittoun RA, Mandelli F, Willemze R et al. Autologous or allo- damage might play a key role in the delayed hematopoietic geneic bone marrow transplantation compared with intensive regeneration observed following ABMT in patients of the chemotherapy in acute myelogenous leukemia. New Engl J AML10 trial. Despite there being no consistent evidence Med 1995; 332: 217–223. for the transplantability of human marrow stroma,45 it can 15 Sutherland HJ, Eaves CJ, Eaves AC et al. Characterization be hypothesized that harvesting, ex vivo expansion and rein- and partial purification of human marrow cells capable of initi- fusion of autologous stromal progenitors at later stages in ating long-term hematopoiesis in vitro. Blood 1990; 74: the course of the treatment might overcome poor graft func- 1563–1570. tion in these patients. 16 Castro-Malaspina H, Gay RE, Resnick G et al. Characteri- zation of human bone marrow fibroblast colony-forming cells (CFU-F) and their progeny. Blood 1980; 56: 289–301. 17 Bennett JM, Catovsky D, Daniel MT et al. Proposed revised Acknowledgements criteria for the classification of acute myeloid leukemia: a report of the French–American–British Cooperative Group. This work was supported in part by grants from Consiglio Nazion- Ann Intern Med 1985; 103: 620–625. ale delle Ricerche (Progetto Finalizzato ACRO), Ministero dell’U- 18 Carlo-Stella C, Mangoni L, Piovani G et al. Identification of niversita` e della Ricerca Scientifica e Tecnologica (MURST, 40% Philadelphia-negative granulocyte–macrophage colony- and 60%), and Associazione Italiana per la Ricerca sul Cancro forming units generated by stroma-adherent cells from chronic (AIRC, Milano). myelogenous leukemia patients. Blood 1994; 83: 1373–1380. 19 Carlo-Stella C, Regazzi E, Garau D et al. Effect of the protein tyrosine kinase inhibitor genistein on normal and leukemic hematopoietic progenitor cell growth. Br J Haematol 1996; References 93: 551–557. 20 Sutherland HJ, Eaves CJ, Lansdorp PM et al. Differential 1 Dexter TM, Allen TD, Lajtha LG. Conditions controlling the regulation of primitive human hematopoietic cells in long- proliferation of haemopoietic stem cells in vitro. J Cell Physiol term cultures maintained on genetically engineered murine 1977; 91: 335–345. stromal cells. Blood 1993; 78: 666–672. 2 Dexter TM. Regulation of hemopoietic cell growth and devel- 21 Udomsakdi C, Lansdorp PM, Hogge DE et al. Characteri- Bone marrow damage after chemotherapy for AML C Carlo-Stella et al 471 zation of primitive hematopoietic cells in normal human peri- 34 Lemischka IR, Raulet DH, Mulligan RC. Developmental pheral blood. Blood 1992; 80: 2513–2521. potential and dynamic behaviour of hematopoietic stem cells. 22 Perkins S, Fleischman RA. Hematopoietic microenvironment: Cell 1986; 45: 917–927. origin, lineage, and transplantability of the stromal cells in 35 Jordan CT, Lemischka IR. Clonal and systemic analysis of long-term bone marrow cultures from chimeric mice. J Clin long-term hematopoiesis in the mouse. Genes Dev 1990; 4: Invest 1988; 81: 1072–1077. 220–232. 23 Appelbaum FR. Marrow transplantation for hematologic 36 Abkowitz JL, Linenberger ML, Newton MA et al. Evidence malignancies: a brief review of current status and future pros- for the maintenance of hematopoiesis in a large animal by pects. Semin Hematol 1988; 25 (Suppl. 3): 16–22. sequential activation of stem cell clones. Proc Natl Acad Sci 24 Gordon MY, Greaves MF. Physiological mechanisms of stem USA 1990; 87: 9062–9066. cell regulation in bone marrow transplantation and haemopo- 37 Gordon MY, Blackett NM, Lewis JL, Goldman JM. Evidence iesis. Bone Marrow Transplant 1989; 4: 335–338. for a mechanism that can provide both short-term and long- 25 Chang J, Geary CG, Testa NG. Long-term bone marrow dam- term haemopoietic repopulation by a seemingly uniform popu- age after chemotherapy for acute myeloid leukaemia does not lation of primitive human haemopoietic precursor cells. Leu- improve with time. Br J Haematol 1990; 75: 68–72. kemia 1995; 9: 1252–1256. 26 Minderman H, Linssen P, van der Lely N et al. Toxicity of 38 Friedenstein AJ, Chailakhjan RK, Lalykina KS. The develop- idarubicin and doxorubicin towards normal and leukemic ment of fibroblast colonies in monolayer cultures of guinea- human bone marrow progenitors in relation to their proliferat- pig bone marrow and spleen cells. Cell Tiss Kin 1970; 3: ive state. Leukemia 1994; 8: 382–387. 393–403. 27 Damon LE, Rug HS, Ries CA, Linker CA. Delayed 39 Ashton BA, Allen TD, Howlett CR et al. Formation of bone engraftment of 4-hydroperoxycyclophosphamide-purged auto- and cartilage by marrow stromal cells in diffusion chambers logous bone marrow after induction treatment containing in vivo. Clin Orthop 1980; 151: 294–307. mitoxantrone for acute myelogenous leukemia. Bone Marrow 40 Friedenstein AJ, Chailakhyan RK, Latsinik NV et al. Stromal Transplant 1996; 17: 93–99. cells responsible for transferring the microenvironment of the 28 Ross EAM, Anderson N, Micklem HS. Serial depletion and hemopoietic tissues. Cloning in vitro and retransplantation in regeneration of the murine hematopoietic system: implications vivo. Transplantation 1974; 17: 331–340. for hematopoietic organization and the study of aging. J Exp 41 Siena S, Castro-Malaspina H, Gulati SC et al. Effects of in Med 1982; 155: 432–444. vitro purging with 4-hydroperoxycyclophosphamide on the 29 Li S, Champlin R, Fitchen JH, Gale RP. Abnormalities of hematopoietic and microenvironmental elements of human myeloid progenitor cells after ‘successful’ bone marrow trans- bone marrow. Blood 1985; 65: 655–662. plantation. J Clin Invest 1984; 75: 234–241. 42 Greenberger JS, Palaszynsky EW, Pierce JH. Biological 30 Lemieux ME, Rebel VI, Lansdorp PM, Eaves CJ. Characteriz- effects of prolonged l-phenylalanine mustard treatment of ation and purification of a primitive hematopoietic cell type murine long-term bone marrow cultures and IL-3-dependent in adult mouse marrow capable of lymphomyeloid differen- haemopoietic progenitor cell lines. J Natl Cancer Inst 1985; tiation in long-term marrow ‘switch’ cultures. Blood 1995; 86: 74: 247–253. 1339–1347. 43 Gordon MY, Riley GP, Watt SM, Greaves MF. Compart- 31 Petzer AL, Hogge DE, Lansdorp PM et al. Self-renewal of mentalization of a haemopoietic growth factor (GM-CSF) by primitive human hematopoietic cells (long-term culture-initiat- glycosaminoglycans in the bone marrow microenvironment. ing cells) in vitro and their expansion in defined medium. Proc Nature 1987; 326: 403–405. Natl Acad Sci USA 1996; 93: 1470–1474. 44 Kaufman CL, Colson YL, Wren SM et al. Phenotypic charac- 32 Jones RJ, Wagner JE, Celano P et al. Separation of pluripotent terization of a novel bone marrow-derived cell that facilitates haematopoietic stem cells from spleen colony-forming cells. engraftment of allogeneic bone marrow stem cells. Blood Nature 1990; 347: 188–189. 1994; 84: 2436–2446. 33 Fraser CC, Hoffman R. Hematopoietic stem cell behavior: 45 Simmons PJ, Przepiorka D, Thomas ED, Torok-Storb B. Host potential implications for gene therapy. J Lab Clin Med 1995; origin of marrow stromal cells following allogeneic bone mar- 125: 692–702. row transplantation. Nature 1987; 328: 429–432.