AUTOLOGOUS PERIPHERAL BLOOD STEM CELL TRANSPLANTATION: GRAFT ENGINEERING

Autologous peripheral blood stem cell transplantation (aPBSCT), also known as high-dose chemotherapy with autologous PBSC support, is effective treatment for certain patients with cancer. PBSC transplantation has proven to be useful in treatment of a wide variety of malignancies including breast cancer, leukemia, and lymphoma (1, 2). aPBSCT has been ongoing for over two years at YNHH. Over 75 patients have undergone this procedure at Yale in order to treat various resistant or aggressive forms of cancer. In aPBSCT, cells capable of regenerating all of the hematopoietic cells (so called "stem cells") which naturally circulate at low levels in blood, are collected from cancer patients while they are in remission or undergoing therapy. Then, after a large dose of chemotherapy that destroys the bone marrow, these stem cells are reinfused and engraft to re-form the patient's own hematopoietic cell compartment. While a PBSCT allows a patient to receive high dose chemotherapy with prompt regeneration of hematopoiesis and without the need for an allogeneic bone marrow donor, an obvious problem is that the reinfused cells may contain tumor cells. This is because the pheresis procedure collects all peripheral blood mononuclear cells (PBMC), of which "stem cells" are actually a very small proportion (<1%). Thus, tumor cells can be included among the stem cells, macrophages and lymphocytes that constitute PBMC. Recently, to address this problem, a new experimental technology called "CD34 selection" has been introduced at YNHH. This article will describe in more detail the rationale for removing tumor cells from PBSC products, the methods by which this is done in the Blood Bank, and the future directions for this technology.

To minimize the risk that autologous PBSC contain clonogenic tumor cells that could contribute to cancer relapse, PBSC are usually collected from patients who are in remission. Despite this precaution, tumor cells can be detected in both bone marrow and peripheral blood stem cell products. When a patient's PBSC are going to be collected by pheresis, the number of stem cells in the product can be increased by first treating the patient with chemotherapy and stem cell growth hormones. This is referred to as "mobilization" of PBSC. However, both tumor cells and hematopoietic stem cells (PBSC) are mobilized into the peripheral bloodstream by the mobilization regimens used prior to collection.

Although it has not yet been proven whether tumor cells reinfused into the patient can cause tumor relapse, proof that tumor cells present in the stem cell product can contribute to relapse of disease has been shown using gene marking studies. The cells infused into the patients (both PBSC and any contaminating tumor cells) were "marked" with a neomycin resistance gene prior to reinfusion into the patient, and these "marked" cells were later found in the tumors of the patients who relapsed (3). However, it is still not known whether tumor in the PBSC product leads to a higher incidence of relapse, shorter disease-free survival, or shorter overall survival. In one study designed to look at this, there was no correlation between the detection of contaminating tumor cells in the PBSC product and disease-free survival or overall survival (4).To evaluate this question definitively, a large randomized controlled study of patients with equivalent prognoses being treated with "purged" vs. "unpurged" PBSC needs to be performed. The "purging" technique used will need to be evaluated based on its ability to remove detectable cancer cells.

Detection of Tumor Cells in PBSC Products

Very sensitive techniques have been developed to detect tumor in PBSC products. For breast cancer, several antibody-based techniques have been published. Using antibodies against cytokeratin, breast cancer cells have been detected in PBSC using FACS analysis. This technique has a limit of detection of 1/200,000 cells. This would mean that a PBSC transplant that has "undetectable" tumor by this method could still have as many as 20,000 contaminating tumor cells in the PBSC product. A higher sensitivity for breast cancer detection (1/1,000,000 cells) has been reported using immunohistochemistry.

Contaminating lymphoma cells are best detected using molecular biology techniques. DNA-based probes are used to identify tumor cells based on specific DNA sequences unique to the tumor cells. For example, the reciprocal translocation t(14;18)(q32;q21) involving juxtaposition of the bcl2 gene with the immunoglobulin heavy-chain joining locus is a characteristic marker found in approximately 85% of follicular lymphomas and up to 40% of diffuse large cell lymphomas. In addition to this identifiable marker, the majority of remaining non-Hodgkin's lymphomas have an identifiable molecular marker due to clonal rearrangements in the JH locus (B cell lymphomas) or TCR gene (T cell lymphomas). This JH clonal marker approach can also be used in multiple myeloma. Such molecular markers are detected via PCR of the specifically rearranged gene or specific translocation. These techniques have a limit of sensitivity of between 1/300,000 to 1/1,000,000 cells.

Methods to Reduce Tumor Cell Contamination

Several approaches have been developed to reduce or eliminate tumor cells from bone marrow and PBSC products. These approaches can be conceptually divided into "negative" selection (sometimes called "purging") - in which the goal is to kill or select out tumor cells - and "positive" selection, in which the stem cell is selected and other cells (including tumor) are left behind.

A nonspecific negative selection approach has been to treat the stem cell product with chemotherapy (4-hydroperoxycyclophosphamide) before reinfusion of the product into the patient. While a prospective randomized controlled study has not been done to prove the effectiveness of this technique, the patients do appear to have less relapse than historical controls. One of the major drawbacks to this approach has been that the hematopoietic reconstituting cells are also damaged by the chemotherapy, and patients can take 30-50 days to engraft. This puts the patients at very high risk of lethal infection or bleeding, and often requires extended hospitalization.

More specific approaches to tumor purging have involved the use of immunotoxins. Immunotoxins have a toxic chemical linked to an antibody against a molecule that is expressed on the patient's tumor cells, but not on normal nonmalignant cells. Alternatively, complement-mediated lysis of antibody-coated tumor cells has been used to kill tumor cells in the stem cell product prior to reinfusion into the patient. Problems with this technique have included optimizing the toxins used and identifying appropriate antigens for each tumor.

More recently, research has been focused on positive selection strategies. These methods use the CD34 antigen as a target for positive selection. The CD34 antigen is expressed on early hematopoietic progenitors, but not on most malignant cells including lymphoma, multiple myeloma, and breast cancer. CD34 selection is currently performed by first treating PBMC (which contain a few stem cells, many lymphocytes and macrophages, and possibly some tumor cells) with a mouse monoclonal antibody to the CD34 molecule. After incubation and washing, the cells which bound the antibody are "caught" on a solid phase support (usually an array of beads) that is designed to bind the anti-CD34 monoclonal antibody. For example, a bead may be covalently coated with sheep anti-mouse IgG which in turn will bind to the murine-derived anti-CD34. The cells that did not bind antibody and therefore did not stick to the beads are then washed away. Finally, the cells which did bind the beads are released through either an enzymatic or physical process. These cells are further characterized and cryopreserved for later infusion into the patient. Usually, the recovered cells represent less than 1% of the starting material, but seem to contain nearly all of the capacity to reconstitute hematopoeisis in the patient. Two companies - Cellpro, Inc. and Baxter Immunotherapy - have so far developed devices for accomplishing this selection. At Yale, we are currently using the Baxter technology. We have just completed enrollment in a small study on breast cancer patients. At present, patients with various B cell malignancies may be eligible to enter the CD34 selection protocol here at Yale.

Many recent studies including those at YNHH have demonstrated that CD34 positive cells isolated from PBSC products are capable of reconstituting hematopoiesis in patients after they receive high-dose chemotherapy. In addition, in studies ongoing at Yale New Haven Hospital and elsewhere, the depletion of tumor cells from the PBSC product by CD34 selection is being measured. The findings indicate that CD34 selection results in approximately a 3-log reduction in tumor cell contamination; that is 999 out of every 1000 tumor cells is removed from the PBSC product by this technique. Still unknown is whether such reductions in tumor content of the PBSC product will confer a decreased risk of tumor relapse. Long-term follow up on current patients as well as future large randomized controlled studies should eventually address this important issue. To learn more about graft engineering of PBSC products at Yale, call Drs. Diane Krause or Mark Shlomchik at the Blood Bank (5-2441).

References

1. Bezwoda WR, Seymour L, Dansey RD. High-dose chemotherapy with hematopoietic rescue as primary treatment for metastatic breast cancer: a randomized trial. J Clin Oncology 1995;13:2483-2489.
2. Philip T, et al. Autologous bone marrow transplantation as compared with salvage chemotherapy in relapses of chemotherapy-sensitive non-Hodgkin's lymphoma. NEJM 1995;333:1540-1545.
3. Rill DR. et al. Direct demonstration that autologous bone marrow transplantation for solid tumors can return a multiplicity of tumorigenic cells. Blood 1994;84:380-83.
4. Hardingham JE. et al. Significance of molecular marker-positive cells after autologous peripheral-blood stem-cell transplantation for non-Hodgkin's lymphoma. J Clin Oncol 1995;13:1073-79.

Diane Krause, M.D., Ph.D. Mark Shlomchik, M.D., Ph.D.