5.1 Independent growth of extramedullary tumors

Another clinical situation that illustrates the autonomous growth and resistance of hematologic tumors is extramedullary relapse after stem cell transplants given for systemic disease. High percentages of reported post-transplant extramedullary relapses in organs occur without systemic disease relapse.51, 52 Relapse in breast without marrow relapse was reported in 2 cases of transplant for myeloma53, 54 and 82 transplanted for leukemia (32 ALL and 50 AML).51 In published experience, most cases of post-transplant relapse as tumors initially treated with systemic agents, without excision or surgery, did not achieve eradication of tumors or prevention of spread, and many cases of fatal aplasia of normal donor marrow occurred.51 Tumor resistance to chemotherapy is also documented in cases where relapse after chemotherapy or transplant occurs simultaneously in extramedullary sites and marrow in both leukemia and myeloma. In many of these reports, asynchronous response is noted: systemic therapy achieves remission in the marrow but not in tumor.1, 55-57

5.2 Adipose tissue in extramedullary sites

In leukemia and myeloma, the marrow microenvironment has long been thought to permit growth of malignant cells protected from chemotherapy and to direct metastasis to bone in myeloma and some solid cancers. A similar influence of microenvironment of extramedullary organs has been proposed to influence the growth, metastasis, and chemoresistance of tumors in them. Adipose tissue is a major component of the microenvironment of bone marrow (BMAT). In some extramedullary organs, it is white adipose tissue (WAT), particularly in breast, omentum, and subcutaneous tissue. Other settings where hematopoiesis occurs in close proximity to adipose cells include most embryonic organs that produce blood cells before there is functioning bone marrow58 and in diseases where marrow is fibrotic or dysfunctional, as in myelofibrosis and thalassemia, where tumors of reactive hematopoiesis may occur in non-marrow organs.59 In solid cancers, the presumed reciprocal interactions of adipose and cancer cells, enabled by their proximity, is postulated to facilitate the remodeling and transformation of both cells that results in chemoresistance.

Differences in gene expression between marrow and adipose tissue were shown by Fideles et al., with higher expression of genes related to osteoblast differentiation and osteogenesis noted in marrow mesenchymal stem cells (MSC). By contrast, subcutaneous adipose tissue MSC showed higher expression of genes related to extracellular matrix organization, lipid metabolism, and adipocyte differentiation.60 It is possible that the metastatic tropism to soft tissue sites, seen clinically in the three hematologic breast tumors in preference to bone, could reflect these differences.

In 1962, Trubowitz and Sims speculated that "fat may be more often involved in leukemia or lymphoma than is generally appreciated”. This was based on their post-mortem biopsies of apparently uninvolved abdominal skin of patients with either leukemia or lymphoma that found infiltration of subcutaneous tissue in 30%.61 It is interesting that three decades earlier, cancer pathologist James Ewing had observed "plasma cell changes occur mainly in fat areas".62

Among the cells that compose adipose tissue are adipocytes, pre-adipocytes, fibroblasts, macrophages, stromal cells, endothelial cells and monocytes. In addition to storing lipid that can be transferred to cells as an energy source, WAT, like BMAT, secretes myriad factors that could influence tumor growth. They include leptin, inflammatory cytokines and growth factors, including IL-6, IL-1β, collagen 6, chemokines, and binding proteins, and FABP4, CD36, and lipoprotein lipase involved in fatty acid transfer. These and other lipid metabolism genes were found to be among the most significantly deregulated in our pilot study of 11 AML breast tumor samples by RNA sequencing, only part of which has been published.4 Several groups have suggested adipokines as potential treatable targets in both breast cancer and hematologic malignancies.63-65

There are in vitro studies that support the role of adipose tissue in the transformation of cells. Co-culture of breast cancer cells with adipocytes resulted in the cancer cells adopting an invasive phenotype and tumor cells becoming “less compact…more scattered colonies”, with downregulation of E-cadherin.66 These changes are consistent with the observation of discohesive cells in samples of all 3 hematologic breast tumors and invasive lobular breast cancer referenced above. Five-fold increases in IL-6 and IL-1β mRNA were also noted, and higher levels of IL-6 expression were seen in samples from larger tumors and those with lymph node involvement.66 Co-culture studies of adipose cells with AML blasts resulted in upregulation of FABP4 mRNA in both cell types and increase in AML survival.67 Co-culture experiments of leukemic and myeloma cells with adipocytes have shown decreased responsiveness to commonly used therapeutic agents.65, 68 In a transplant model study using a murine leukemia line and oncogenic KRAS, the development of leukemic tumors in adipose tissues was demonstrated.69

WAT may also play a role in the development of desmoplastic fibrosis, another factor in chemoresistance seen in breast tumors of all three hematologic malignancies and in solid cancers. Studies have suggested that adipocytes may be reprogrammed by breast cancer cells into fibroblast-like cells that overexpress extracellular matrix components, particularly collagen 6, a known growth factor for breast cancer, with creation of dense desmoplastic reaction.70