Ex Vivo Expansion of Bone Marrow from Breast Cancer Patients: Reduction in Tumor Cell Content Through Passive Purging

Bone Marrow Transplantation, (1998) 22, 153–159  1998 Stockton Press All rights reserved 0268–3369/98 $12.00 http://www.stockton-press.co.uk/bmt Ex vivo expansion of bone marrow from breast cancer patients: reduction in tumor cell content through passive purging

BI Lundell1, JJ Vredenburgh2, C Tyer2, K DeSombre2 and AK Smith1

1Aastrom Biosciences, Ann Arbor, MI; and 2Duke University Bone Marrow Transplant Program, Duke University Medical Center, NC, USA

Summary: a trend of higher relapse rates in breast cancer patients receiving tumor cell positive PBPC collections has been High-dose chemotherapy (HDC) with hematopoietic reported.7 support appears promising in the treatment of breast High-dose chemotherapy with hematopoietic rescue cancer, although reinfusion of contaminating tumor appears promising in the treatment of breast cancer,10,11 but cells may contribute to disease relapse. Ex vivo expan- the presence of contaminating tumor cells in the reinfused sion may reduce tumor cell content through use of a hematopoietic cell product remains a concern. Retrovirally small inoculum volume and by passive purging during mediated gene marking techniques and retrospective clini- culture. We assessed the ex vivo expansion potential of cal trials have demonstrated in patients with acute myelo- tumor cell positive bone marrow (BM) from breast can- genous leukemia, neuroblastoma, B cell lymphoma, chronic cer patients and the effect of ex vivo expansion on tumor lymphocytic leukemia, or chronic myelogenous leukemia cell content. Cryopreserved/thawed mononuclear cell that reinfusion of contaminating tumor cells present in the (C/T MNC) BM harvests with known tumor cell con- BM harvest can contribute to disease relapse in the BM tamination (n = 7) were assessed for tumor cells pre- itself or in extramedullary sites.12–16 The contribution of and post-expansion using immunocytochemical (ICC) reinfused tumor cells to the relapse of breast cancer is cur- staining. Pre-expansion inoculum samples contained a rently unclear. Studies monitoring the reinfusion of tumor × 6 range of 6–2128 tumor cells per 5.0 10 nucleated cells. positive hematopoietic cell products in breast cancer Ex vivo expansion resulted in fold expansions of 6.67 patients were unable to detect a correlation between tumor and 11.37 for total cells and CFU-GM, respectively. cell contamination and site of relapse.6,8 Tumor cells were undetectable in four of the seven post- In vitro clonogenic assays suggest, however, that tumor expansion samples and were reduced in the remaining cells remaining in BM harvests or PBPC collections can three samples. The data demonstrate passive purging of be viable and clonogenic12,13,17,18 potentially contributing to breast cancer cells during ex vivo expansion, with hema- disease relapse. The homing of reinfused tumor cells to topoietic progenitor cell expansion comparable to that sites of previous disease may prevent the distinction of normal BM. Reduction in tumor cell number con- between relapse due to residual tumor cells or reinfused tained in the small volume culture inoculum combined tumor cells. Altered adhesion receptor expression or func- with passive purging during the ex vivo expansion pro- tion, similar to that previously demonstrated in breast can- cess suggest a potential 2–4+ log reduction in tumor cell cer tumor cells,19 may contribute to the homing of reinfused content in the reinfused cell product. tumor cells from the circulation to a microenvironment cap- Keywords: bone marrow transplantation; ex vivo expan- able of supporting tumor cell growth.20,21 Many HDC trials sion; tumor cell contamination are currently conducted without some form of tumor cell purging. Removal of tumor cells from the reinfused hematopoietic stem cell product may result in improved disease- Immunologic techniques have revealed tumor cell infil- free and overall patient survival. tration of the BM in as many as 48% of early stage breast Previous ex vivo expansion studies using CD34+ PBPC cancer patients at the time of diagnosis.1–4 The presence of derived from breast cancer patients have reported median tumor cells in the BM has been shown to be predictive of progenitor expansions of 10- to 46-fold,22 with a relative relapse in early stage5–7 and advanced stage8,9 breast cancer disadvantage of tumor cell growth compared to hematopo- patients. In addition, the number of tumor cells identified ietic cell growth in cultures seeded with tumor cell lines.23 in the BM has been correlated with shorter disease-free and The ex vivo expansion potential of primary tumor cells and overall survival.6 Although the prognostic significance of their impact on hematopoietic progenitor cell expansion in tumor cells in PBPC collections requires further evaluation, a stromal-based culture system has not been evaluated. The intent of this study was (1) to evaluate the ex vivo expansion potential of BM samples obtained from stage IV breast Correspondence: Dr BI Lundell, Aastrom Biosciences, Inc., 24 Frank cancer patients which were known to be tumor cell positive Lloyd Wright Drive, Domino’s Farms, Lobby L, Ann Arbor, MI 48105, USA using a small scale culture system, and (2) to determine the Received 21 January 1998; accepted 18 March 1998 effect of ex vivo expansion on tumor cell content. Passive tumor cell purging during ex vivo expansion BI Lundell et al 154 Materials and methods Two BM aspirates from each normal donor were pooled and the nucleated cell concentration was determined. MNC Bone marrow samples were separated by density gradient centrifugation (Ficoll– Paque Plus; Pharmacia Biotech, Uppsala, Sweden; The Duke University Bone Marrow Transplant Program s.g.1.077) at 300 g for 20 min at 25°C. Fresh BM MNC has performed a series of clinical protocols for patients util- were washed twice with LTBMC medium prior to culture. izing a variety of purging techniques. Seven back-up BM samples collected from stage IV breast cancer patients having a median age of 46 years (range 34–51 years) were Small scale cell culture expansion system available for ex vivo expansion. All seven patients exhibited Fresh and C/T MNC expansion cultures were initiated in tumor cell infiltration of the BM by routine histologic 24-well tissue culture plates (Costar, Cambridge, MA, examination of BM biopsies. The BM samples used USA) at a viable nucleated cell density of 0.5 × 106 cells for these studies were obtained using standard BM per well in a final volume of 0.6 ml. Ex vivo expansion harvesting techniques. cultures were performed using LTBMC medium sup- For a controlled comparison, small volume BM aspirates plemented with recombinant human (rh) erythropoietin = (n 7) were collected after obtaining informed consent (Epoetin Alfa; Amgen, Thousand Oaks, CA, USA) at 0.1 from normal volunteers having a median age of 37 years U/ml, rh PIXY321 (Immunex Corporation, Seattle, WA, (range 21–45 years). Briefly, two aspirates were collected USA) at 5 ng/ml, and rh flt-3 ligand at 25 ng/ml (Immunex from adjacent sites on the posterior iliac crest. For each Corporation). Ex vivo expansion cultures were incubated at aspiration, 3 ml of BM was collected from each of four 37°C in a humidified atmosphere with 5% CO and fed ° 2 quadrants, rotating the BM needle bevel 90 between quad- with a 50% medium exchange on days 4, 7, 9, 10 and 11. rants, for a total volume of 12 ml. Cells were anticoagulated To prevent cell loss during feeding, medium removed from with 20 U/ml preservative-free heparin (Apothecon, Prince- each well was centrifuged at 300 g for 5 min. The resulting ton, NJ, USA) and processed within 5 h of collection. cell pellet was resuspended in fresh expansion medium and returned to culture. Expansion cultures were harvested on Sample processing day 12 using trypsin/EDTA (Gibco BRL) and Dulbecco’s PBS (Gibco BRL). The expanded cell product was enumer- An unpurged aliquot of patient BM MNC was control rate ated on a cell counter as described above. cryopreserved and stored for a median period of 33 months (range 12–42 months). Cryopreserved patient samples were Hematopoietic cell assays shipped on dry ice to Aastrom Biosciences, where they were stored in the vapor phase of liquid nitrogen for up CFU-GM: Cells were inoculated in methylcellulose (final to 1 month prior to thawing for initial tumor cell content concentration 1.0%; Sigma) prepared with IMDM sup- assessment and ex vivo expansion. Cryopreserved samples plemented with 2 mm glutamine (Gibco BRL), 30% FBS, were thawed rapidly in a 37°C circulating waterbath, 1.0% BSA, 1% monothioglycerol (Sigma), 0.1% penicillin- diluted with approximately two volumes of long-term bone streptomycin (Gibco BRL), 5 ng/ml rh PIXY321, 5 ng/ml marrow culture medium ((LTBMC; Iscove’s modified Dul- rh G-CSF (Amgen) and 10 U/ml rh erythropoietin. Tripli- becco’s medium (IMDM) with 4 mml-glutamine (Gibco cate 1 ml aliquots containing 1.0–2.0 × 104 cells per ml BRL, Grand Island, NY, USA), 10% fetal bovine serum (fresh and C/T MNC expansion culture inoculums) or 3.0– (FBS; BioWhittaker, Walkersville, MD, USA), 10% horse 6.0 × 103 cells per ml (expanded cell products) were plated serum (Gibco BRL), 20 ␮g/ml vancomycin (Vancocin HCl, in gridded 35-mm dishes (Nunc, Naperville, IL, USA) and Lilly, Indianapolis, IN, USA), 5 ␮g/ml gentamicin incubated at 37°C in a humidified atmosphere with 5% ␮ (Fujisawa USA, Deerfield, IL, USA), and 5 m hydrocorti- CO2. At day 14 of culture, CFU-GM (colony forming units sone (Solu-Cortef, Upjohn, Kalamazoo, MI, USA) plus 200 granulocyte/macrophage) were enumerated as previously ␮g/ml DNAse (Sigma, St Louis, MO, USA) and incubated described.24 at room temperature for 20 min. C/T MNC were pelleted by centrifugation at 300 g for 7 min and resuspended in LTC-IC: To determine the frequency of long-term culture- fresh LTBMC medium. Aliquots of thawed cells were initiating cells (LTC-IC), fresh and C/T MNC and their assayed to determine nucleated cell concentration and cell subsequent expanded cell products were plated in limiting viability. Cells were diluted in isotonic-buffered saline dilutions on irradiated stromal layers (20 replicates, inocu- (Diluent 2; Stephens Scientific, Riverdale, NJ, USA) and lum cell number = 2.5, 5.0, 10.0 and 20.0 × 103 cells per nucleated cells were enumerated on a Coulter ZM cell well; 20 replicates, expanded cell number = 7.5, 15.0, 30.0, counter (Coulter Electronics, Hialeah, FL, USA) after and 60.0 × 103 cells per well) as previously described.24 erythrocyte lysis with Manual Lyse (Stephens Scientific). Cultures were maintained at 33°C in a humidified atmos- Cell viability was determined using propidium iodide phere with 5% CO2, and fed at weekly intervals with a 50% (Molecular Probes, Eugene, OR, USA) staining. After 10 medium exchange (100 ␮l). After 5 weeks, the wells were min, cells were washed and resuspended in 0.5 ml phos- harvested using trypsin/EDTA and plated in non-tissue cul- phate-buffered saline (PBS; Gibco BRL) containing 1% ture-treated 24-well plates (Falcon, Lincoln Park, NJ, USA) bovine serum albumin (BSA; Leptab-7, Intergen, Purchase, containing 0.25 ml methylcellulose medium supplemented NY, USA) for analysis on a FACS Vantage flow cytometer as described above. Wells were scored for the presence or (Becton Dickinson, San Jose, CA, USA). absence of secondary colony-forming cells (CFC) on day Passive tumor cell purging during ex vivo expansion BI Lundell et al 155 14. The absolute number of LTC-IC was calculated based GM per well (range 8231–31 792 CFU-GM per well, n = on the maximum likelihood method.25,26 7 patient samples in triplicate). When expressed relative to productivity per surface area, ex vivo expanded tumor cell positive C/T MNC samples produced an average of Immunocytochemical staining 11 195 Ϯ 3876 CFU-GM per cm2 (range 4243–16 388 Quantitation of the tumor cell population present in the C/T CFU-GM per cm2, n = 7 patient samples in triplicate). Due MNC samples pre- and post-expansion was determined by to variability in the CFU-GM content of the C/T MNC ICC staining, as previously described.27 Ten coded slide inocula, the corresponding CFU-GM-fold expansion aver- preparations containing 5.0 × 105 nucleated cells per slide aged 11.37 Ϯ 3.76 (range 5.68–16.75, n = 7 patient samples were prepared for each sample. The slides were air dried in triplicate). LTC-IC, as determined by limiting dilution and stored for up to 1 month prior to staining. Tumor cell analysis, were maintained or expanded in six of seven antigens were identified using a murine monoclonal anti- tumor cell positive C/T MNC samples. body cocktail containing 260F9, 317G5 and 520C9, anti- Ex vivo expansion potential of tumor cell positive C/T bodies previously shown to have reactivity against breast MNC harvested from stage IV breast cancer patients was cancer cells, with no reactivity against BM cells.28 A coded not significantly different from that of fresh BM MNC set of cytospin preparations containing normal BM MNC obtained from normal donors. As shown in Table 1, spiked with CAMA-1 or SKBr3, breast cancer cell lines, nucleated cell expansion potential, expanded nucleated cell at a frequency of 1:104, 1:105 or 1:106 was included with yield, CFU-GM expansion potential, expanded CFU-GM each batch of slides stained. Tumor cells and normal BM yield, LTC-IC expansion potential, and expanded LTC-IC cells were distinguished on the basis of ICC staining and yield for tumor cell positive C/T MNC obtained from BM cell morphology. A complete battery of quality control harvests were not significantly different from the respective cytospin preparations was included with each staining pro- values obtained from the ex vivo expansion of fresh MNC cedure. This included the following negative and positive collected from normal donors by small volume BM control samples: negative controls included (1) fresh nor- aspiration. mal PBPC, (2) thawed normal PBPC and (3) normal expanded BM MNC; positive controls included (1) the Reduction in tumor cell content during ex vivo expansion breast cancer cell lines CAMA-1 and SKBr3, (2) known The effect of ex vivo expansion on tumor cell content was tumor cell positive pleural effusion and (3) known tumor evaluated using ICC staining. Immunostained cells were cell positive BM aspirate. The slides were independently considered positive for tumor cell identification when stain- scored by at least two observers. The tumor cell frequencies ing was observed on at least 75% of the cell membrane and in inoculum and expanded samples were calculated based cytoplasm, and the cell morphology was consistent with an on the number of tumor cells counted on 10 slides, each epithelial tumor cell. This technique can detect one tumor containing 5.0 × 105 nucleated cells. cell in 106 normal BM cells.27 Ex vivo expansion of tumor cell positive C/T BM MNC Statistical analysis obtained from stage IV breast cancer patients resulted in passive purging in all samples examined. Tumor cell fre- Results of experimental points obtained from multiple Ϯ quency decreased over the course of ex vivo expansion in experiments were reported as means s.d. Significance all cases. The evaluation of 5 × 106 nucleated cells levels were determined by two sided Student’s t-test increased the likelihood of identifying tumor cells present analysis. at low frequency. Pre-expansion tumor cell positive C/T MNC inoculum samples contained a range of 6–2128 tumor cells per 5.0 × 106 nucleated cells. No tumor cells were Results detectable in four of seven C/T MNC expanded cell products. In the remaining three C/T MNC expanded cell pro- Ex vivo expansion potential of tumor cell positive bone ducts, the number of tumor cells decreased from 44 to 2, marrow 40 to 2, and 2128 to 4 per 5.0 × 106 nucleated cells, respect- Inoculum samples for ex vivo expansion were obtained ively. Viability testing using propidium iodide indicated from unpurged C/T MNC harvests collected from stage IV that more than 92% of cells were viable after ex vivo expan- breast cancer patients. After 12 days of ex vivo culture, the sion. As shown in Table 2, the amount of tumor cell con- total nucleated cell fold expansion averaged 6.67 Ϯ 1.98 tamination in the culture inocula did not correlate with the (range 3.14–9.13, n = 7 patient samples in triplicate) with subsequent ex vivo expansion potential. The total numbers a corresponding mean nucleated cell yield of of tumor cells remaining in the expanded cell products were 3.33 Ϯ 0.99 × 106 cells per well (range 1.57–4.57 × 106 significantly less than the initial numbers present in the pre- cells per well, n = 7 patient samples in triplicate). expansion culture inocula. A tumor cell reduction of 1 to To determine progenitor cell ex vivo expansion potential, 4 logs occurred during the ex vivo expansion process. the tumor cell positive C/T MNC expansion culture inocula and the resulting post-expansion cell products were assayed Discussion for CFU-GM and LTC-IC. After 12 days in culture, the ex vivo expanded cell products derived from tumor cell posi- In contrast to the therapeutic use of BM harvests or apher- tive C/T MNC yielded an average of 21 718 Ϯ 7520 CFU- esis of PBPC, ex vivo expansion of small volume BM aspir- Passive tumor cell purging during ex vivo expansion BI Lundell et al 156 Table 1 Expansion potential of tumor cell positive cryopreserved/thawed BM MNC compared to normal fresh BM MNC

Normal Tumor cell positive Fresh MNC C/T MNC

Expanded nucleated cells per well Mean Ϯ s.d. 3.43 Ϯ 1.03 × 106 3.33 Ϯ 0.99 × 106 Range 1.84–5.01 × 106 1.57–4.57 × 106 Fold expansion, mean Ϯ s.d. 6.86 Ϯ 2.06 6.67 Ϯ 1.98 Expanded CFU-GM per wella Mean Ϯ s.d. 28 959 Ϯ 11 845 21 718 Ϯ 7520 Range 12 333–45 392 8231–31 792 Fold expansion, mean Ϯ s.d. 13.44 Ϯ 6.31b 11.37 Ϯ 3.76 Expanded LTC-IC per wella Mean Ϯ s.d. 42 Ϯ 36 42 Ϯ 24 Range 2–119 13–85 Fold expansion, mean Ϯ s.d. 1.48 Ϯ 1.14b 0.70–0.32

Seven tumor cell positive C/T MNC samples and seven normal fresh MNC samples were cultured for 12 days in triplicate wells of 24-well tissue culture plates with an inoculating density of 0.5 × 106 viable cells per well. aExpanded CFU-GM and LTC-IC per well were calculated by converting the number of colonies scored on an aliquot of the expanded cell product to reﬂect the total nucleated cell yield after ex vivo expansion. For the CFU-GM assay the number of colonies scored per 35-mm dish ranged from 4 to 40. For the LTC-IC assay by limiting dilution analysis, the number of secondary CFC scored per well ranged from 0 to 25. bn = 6.

Table 2 Reduction in tumor cell content during ex vivo expansion

Sample Total cell Tumor cells per Tumor cells per Minimal log identiﬁcation fold expansion 5 × 106 cells 5 × 106 cells reduction pre-expansion post-expansion

1 8.16 42 0a у1.0 2 3.15 44 2 1.8 3 6.46 40 2 2.1 4 9.13 32 0a у1.0 5 7.60 6 0a у0.8 6 7.68 2128 4 3.6 7 4.49 10 0a у1.0

aBreast cancer tumor cells were detected by ICC staining. The limit of tumor cell detection was one tumor cell per 1 × 106 viable nucleated cells.

ates may provide a hematopoietic cell product with reduced This study confirms our ability to successfully expand levels of tumor cell contamination. Previous studies from BM MNC harvested from tumor cell positive stage IV several groups have documented the safety of an ex breast cancer patients, with total and progenitor cell expan- vivo expanded BM or PBPC stem cell product for sions equivalent to normal BM. Expansion of BM MNC transplant.29–32 We recently demonstrated the ability to from tumor cell positive stage IV breast cancer patients in expand small volume BM MNC separated from WBM aspi- the small scale culture system was similar to expansion of rates (р40 ml) using a stromal-based perfusion culture to BM MNC from tumor cell negative advanced stage breast obtain sufficient cells for successful hematopoietic cancer patients in the Aastrom Cell Production System engraftment following myeloablative chemotherapy in using the same cytokine conditions as listed in Materials advanced stage breast cancer patients.33 While sustained and methods. Mean total cell fold and CFU-GM expansions clinical engraftment requires long-term follow-up, we are in the small scale vs the clinical scale system were 6.7 vs encouraged by these findings. All patients enrolled in our 6.4 and 11.37 vs 12.2, respectively.33 LTC-IC were main- clinical studies had histologically normal BM biopsies. An taned during ex vivo expansion in both the small scale and earlier report of successful hematopoietic engraftment using clinical scale systems. Although the number of LTC-IC an ex vivo expanded cell product derived from PBPC did does not equate with the presence of long-term repopulating not include a myeloablative chemotherapy regimen.29 A cells,30,34 infusion of our expanded cell product provided more recent report by Holyoake et al30 indicated that unlike sustained engraftment following a myeloablative chemo- ex vivo expansion of BM MNC in a stromal-based per- therapy regimen. This suggests that our ex vivo expanded fusion culture system, the ex vivo expansion of CD34+ cell product derived from BM MNC contains the necessary PBPC in gas permeable bags did not provide a hematopo- cell population(s) in sufficient number(s) for the reconsti- ietic cell product capable of providing short- or long-term tution of hematopoiesis. Sustained long-term engraftment engraftment following a myeloablative chemoradiotherapy with our expanded cell product can only be confirmed as regimen. these patients are followed over time. Passive tumor cell purging during ex vivo expansion BI Lundell et al 157 In addition to successful expansion of hematopoietic than aliquots of BM harvests, are required to determine if cells, passive purging during ex vivo expansion in a small the reinfused tumor cell content can be reduced below scale culture system decreased the tumor cell frequency and detectable limits in a higher percentage of cases. Passive the total tumor cell content contained in the resulting cell purging in our ex vivo culture system offers a simpler product. The sensitivity of ICC staining can be increased method of tumor cell reduction, as the large number of cells by evaluating a higher number of cells. Our evaluation of provided by multiple PBPC collections or BM harvests 5 × 106 nucleated cells exceeded the 95% confidence limit make purging and positive CD34+ cell selection costly and of sensitivity for ICC staining, which is about one tumor difficult to perform. cell in 106 nucleated cells for 3 × 106 cells examined.35 In vitro clonogenic assays suggest that tumor cells Alternate methods of epithelial tumor cell detection, such remaining in BM harvests or PBPC collections can be as the more sensitive reverse transcriptase polymerase chain viable and clonogenic.12,13,17,18 The clonogenic potential of reaction based on cytokeratin 19 and clonogenic assays are residual tumor cells detected after ex vivo expansion in our required to verify the reduction in tumor cell content. culture system has not been determined. Although the con- The patient MNC samples analyzed in this study were tribution of reinfused tumor cells to disease relapse remains aliquots of processed BM harvests obtained by bilateral controversial, several recent studies suggest that reduction aspiration from two to four skin sites per side, with 10 to in the number of tumor cells reinfused may delay and 20 bone punctures per skin site. We expect the initial tumor potentially reduce the frequency of disease relapse in breast cell number contained in a small volume BM aspirate to cancer patients.6,9 These studies employed immunologic be less than the amount present in the BM harvests reported techniques which increase the sensitivity of tumor cell here. It is likely that reductions in the total volume of BM detection. Analysis of 246 patients indicate a decreased collected (0.5–1.0 l/BM harvest vs 0.04 l/small volume BM overall survival in stage IV breast cancer patients who aspirate) and the number of bone sites sampled received a transplant with BM or PBPC containing detect- (bilateral/20–80 sites vs unilateral/two sites) would contrib- able contaminating tumor cells.9 Additional studies suggest ute to lower tumor cell number in the small volume BM BM metastasis and higher numbers of tumor cells identified inoculum, which may further reduce the tumor cell content in the reinfused cell product correlate with decreased dis- in the expanded cell product. ease-free and overall patient survival in patients with multi- Initially, apheresis products were considered to have less node positive breast cancer treated with chemotherapy.6 tumor cell contamination than BM harvests.17 Subsequent The data reported in this study demonstrate that ex vivo studies indicate that tumor cell contamination in mobilized expansion of tumor cell positive C/T BM MNC from stage PBPC collections remains frequent,23,36,37 indicating that IV breast cancer patients will yield total and progenitor cell apheresis is not a reliable method for eliminating tumor numbers equivalent to that of normal expanded BM MNC cells from the hematopoietic cell product. Due to the with concomitant reduction in tumor cell content by passive infusion of greater numbers of hematopoietic cells in PBPC purging during the culture period. We expect the use of collections, the actual numbers of tumor cells reinfused MNC from a small volume BM aspirate as inoculum for may be similar to that in BM harvests. Apheresis of breast ex vivo expansion would further decrease the level of tumor cancer patients may be further complicated by the less uni- cell contamination in the reinfused cell product, as com- form, thus less predictable, mobilization pattern in patients pared to the use of a full BM harvest or apheresis products who have previously received myelotoxic chemotherapy. for stem cell transplantation. Reduction in initial BM sam- In addition, tumor cell contamination of PBPC can occur ple volume, reduction in the number of BM aspiration sites regardless of the patient’s clinical status at the time of required for sample collection, and passive purging during apheresis.38 Mobilization for PBPC collection also has the ex vivo expansion could result in a potential 2 to 4+ log potential of mobilizing tumor cells from sites other than reduction in tumor cell content in the reinfused cell product. the BM, or from areas of BM metastases not sampled in a Further studies evaluating the clonogenic potential of tumor small volume aspiration or harvest. These negative factors cells will be performed to accurately reflect the viable associated with the collection of mobilized PBPC in breast tumor cell content remaining in the ex vivo expanded cell cancer patients would be eliminated with the infusion of a product. Clinical trials using the Aastrom Cell Production cell product obtained by ex vivo expansion of MNC from System are underway to determine the therapeutic benefits a small volume BM aspirate. of the expanded cell product in breast cancer patients and Although active tumor cell purging using chemical further investigate passive tumor cell purging during ex agents or antibodies can be employed to decrease the num- vivo expansion. Long-term patient follow-up will aid in ber of tumor cells reinfused with the hematopoietic cell pro- determining the contribution of reinfused tumor cells to dis- duct, these procedures have many disadvantages including ease relapse in breast cancer patients and the ability of an toxicity to normal cells, stem cell loss, and uncertain effi- ex vivo expanded cell product to provide sustained cacy. Positive CD34+ cell selection of BM harvests or engraftment. Gene marking studies may also elucidate the PBPC collections from breast cancer patients has been significance of contaminating tumor cells. shown to reduce the number of detectable tumor cells in the hematopoietic graft.39 As in the study by Shpall et al,39 References tumor cell content after ex vivo expansion was reduced in all samples examined, but remained above detectable limits 1 Mansi JL, Easton D, Berger U et al. Bone marrow micromet- in some of the samples. Additional ex vivo expansion stud- astases in primary breast cancer: prognostic significance after ies using MNC from small volume BM aspirates, rather 6 years’ follow-up. Eur J Cancer 1991; 27: 1552–1555. Passive tumor cell purging during ex vivo expansion BI Lundell et al 158 2 Kirk SJ, Cooper GG, Hoper M et al. The prognostic signifi- ietic growth factor alone or after cyclophosphamide followed cance of marrow micrometastases in women with early breast by growth factor. J Clin Oncol 1996; 9: 2569–2575. cancer. Eur J Surg Oncol 1990; 16: 481–485. 19 Liapis H, Flath A, Kitazawa S. Integrin alpha V beta 3 3 Diel IJ, Kaufmann M, Goerner R et al. Detection of tumor expression by bone-residing breast cancer metastases. Diagn cells in bone marrow of patients with primary breast cancer: Mol Pathol 1996; 5: 127–135. a prognostic factor for distant metastasis. J Clin Oncol 1992; 20 Urashima M, Chen BP, Chen S et al. The development of a 10: 1534–1539. model for the homing of multiple myeloma cells to human 4 Harbeck N, Untch M, Pache L et al. Tumour cell detection in bone marrow. Blood 1997; 90: 754–765. the bone marrow of breast cancer patients at primary therapy: 21 Messinger Y, Chelstrom L, Gunther R et al. Selective homing results of a 3-year median follow-up. Br J Cancer 1994; 69: of human leukemic B-cell precursors to specific lymphohema- 566–571. topoietic microenvironments in SCID mice: a role for the beta 5 Cote RJ, Rosen PP, Lesser ML et al. Prediction of early 1 integrin family surface adhesion molecules VLA-4 and relapse in patients with operable breast cancer by detection of VLA-5. Leuk Lymphoma 1996; 23: 61–69. occult bone marrow micrometastases. J Clin Oncol 1991; 9: 22 Shapiro F, Yao T-J, Raptis G et al. Optimization of conditions 1749–1756. for ex vivo expansion of CD34+ cells from patients with stage 6 Vredenburgh JJ, Silva O, Broadwater G et al. The significance IV breast cancer. Blood 1994; 84: 3567–3574. of tumor contamination in the bone marrow from high-risk 23 Vogel W, Behringer D, Scheding S et al. Ex vivo expansion primary breast cancer patients treated with high-dose chemo- of CD34+ peripheral blood progenitor cells: implications for therapy and hematopoietic support. Biol Blood Marrow the expansion of contaminating epithelial tumor cells. Blood Transplant 1997; 3: 91–97. 1996; 88: 2707–2713. 7 Schulze R, Schulze M, Wischnik A et al. Tumor cell contami- 24 Koller MR, Palsson MA, Manchel I et al. Long-term culture- nation of peripheral blood stem cell transplants and bone initiating cell expansion is dependent on frequent medium marrow in high-risk breast cancer patients. Bone Marrow exchange combined with stromal and other accessory cell Transplant 1997; 19: 1223–1228. effects. Blood 1995; 86: 1784–1793. 8 Brockstein BE, Ross AA, Moss TJ et al. Tumor cell contami- 25 de St Groth SF. The evaluation of limiting dilution assays. J nation of bone marrow harvest products: clinical consequences Immunol Meth 1982; 49: R11–R23. in a cohort of advanced-stage breast cancer patients undergo- 26 Finney DJ. The estimation of bacterial densities from dilution ing high-dose chemotherapy. J Hematother 1996; 5: 617–624. series. J Hygiene 1951; 49: 26–35. 9 Moss TJ, Cooper B, Kennedy MJ et al. The prognostic value 27 Vredenburgh JJ, Silva O, Tyer C et al. A comparison of of immunocytochemical (ICC) analysis on bone marrow (BM) immunohistochemistry, two-color immunofluorescence, and and stem cell products (PBSC) taken from patients with stage flow cytometry with cell sorting for the detection of IV breast cancer undergoing autologous transplant (ABMT) micrometastatic breast cancer in the bone marrow. J Hema- therapy. Proc Am Soc Clin Oncol 1997; 16: 90a (Abstr. 317). tother 1996; 5: 57–62. 10 Stemmer SM, Cagnoni PJ, Shpall EJ et al. High-dose paclit- 28 Johnston CS, Shpall EJ, Williams S et al. Detection of mini- axel, cyclophosphamide, and cisplatin with autologous hema- mal residual breast cancer in bone marrow. Prog Clin Biol topoietic progenitor-cell support: a phase I trial. J Clin Oncol Res 1992; 377: 637–642. 1996; 14: 1463–1472. 29 Brugger W, Heimfeld S, Berenson RJ et al. Reconstitution 11 Bezwoda WR, Seymour L, Dansey RD. High-dose chemo- of hematopoiesis after high-dose chemotherapy by autologous therapy with hematopoietic rescue as primary treatment for metastatic breast cancer: a randomized trial. J Clin Oncol progenitor cells generated ex vivo. New Engl J Med 1995; 333: 1995; 13: 2483–2489. 283–287. 12 Rill DR, Santana VM, Roberts WM et al. Direct demon- 30 Holyoake TL, Alcorn MJ, Richmond L et al. CD34 positive stration that autologous bone marrow transplantation for solid PBPC expanded ex vivo may not provide durable engraftment tumors can return a multiplicity of tumorigenic cells. Blood following myeloablative chemoradiotherapy regimens. Bone 1994; 84: 380–383. Marrow Transplant 1997; 19: 1095–1101. 31 Zimmerman TM, Williams SF, Bender JG et al. Clinical use 13 Brenner MK, Rill DR, Moen RC et al. Gene-marking to trace + origin of relapse after autologous bone-marrow transplan- of selected and expanded peripheral blood CD34 cells: a pre- tation. Lancet 1993; 341: 85–86. liminary report of feasibility and safety. J Hematother 1995; 14 Provan D, Bartlett-Pandite L, Zwicky C et al. Eradication of 4: 527–529. polymerase chain reaction-detectable chronic lymphocytic leu- 32 Champlin R, Mehra R, Gajewski J et al. Ex vivo expanded kemia cells is associated with improved outcome after bone progenitor cell transplantation in patients with breast cancer. marrow transplantation. Blood 1996; 88: 2228–2235. Blood 1995; 86: 295a (Abstr. 1166). 15 Gribben JG, Freedman AS, Neuberg D et al. Immunologic 33 Stiff PJ, Oldenberg D, Hsi E et al. Successful hematopoietic purging of marrow assessed by PCR before autologous bone engraftment following high-dose chemotherapy using only ex marrow transplantation for B cell lymphoma. New Engl J Med vivo expanded bone marrow grown in Aastrom (stromal- 1991; 325: 1525–1533. based) bioreactors. Proc Am Soc Clin Oncol 1997; 16: 88a 16 Deisseroth AB, Zu Z, Claxton D et al. Genetic marking shows (Abstr. 309). that Ph+ cells present in autologous transplants of chronic 34 Sutherland HJ, Hogge DE, Lansdorp PM et al. Quantitation, myelogenous leukemia (CML) contribute to relapse after auto- mobilization, and clinical use of long-term culture-initiating logous bone marrow in CML. Blood 1994; 83: 3068–3076. cells in blood cell autografts. J Hematother 1995; 4: 3–10. 17 Ross AA, Cooper BW, Lazarus HM et al. Detection and 35 Sharp JG. Micrometastases and transplantation. J Hematother viability of tumor cells in peripheral blood stem cell collec- 1996; 5: 519–524. tions from breast cancer patients using immunocytochemical 36 Ross AA, Loudovaris M, Hazelton B et al. Immunocytochem- and clonogenic assay techniques. Blood 1993; 82: 2605–2610. ical analysis of tumor cells in pre- and post-culture peripheral 18 Passos-Coelho JL, Ross AA, Kahn DJ et al. Similar breast blood progenitor cell collections from breast cancer patients. cancer cell contamination of single-day peripheral-blood pro- Exp Hematol 1995; 23: 1478–1483. genitor-cell collections obtained after priming with hematopo- 37 Brugger W, Bross KJ, Glatt M et al. Mobilization of tumor Passive tumor cell purging during ex vivo expansion BI Lundell et al 159 cells and hematopoietic progenitor cells into peripheral blood 39 Shpall EJ, Jones RB, Bearman SI et al. Transplantation of of patients with solid tumors. Blood 1994; 83: 636–640. enriched CD34-positive autologous marrow into breast cancer 38 Ross AA, Layton TJ, Ostrander AB et al. Comparative analy- patients following high-dose chemotherapy: influence of sis of breast cancer contamination in mobilized and nonmobil- CD34-positive peripheral-blood progenitors and growth fac- ized hematopoietic grafts. J Hematother 1996; 5: 549–552. tors on engraftment. J Clin Oncol 1994; 12: 28–36.