CLL and associated stromal cells are exposed to lower oxygen tension in lymph nodes and other lymphoid organs; the impact of this on gene expression profiles (GEPs) and cellular functions is not understood. In this study the effects of low oxygen tension (1%; “hypoxia”) on stromal and CLL cell GEPs and interactions were explored using a co-culture system employing a bone marrow derived human stromal cell line (HS-5) and CLL B cells. HS-5 was grown at 1% O2 for 4 weeks (n ≥ 3 experiments) and at weekly intervals GEPs were analyzed using whole genome expression beadchips (HumanHT-12 v 4, Illumina). Differential GEPs were determined using GenomeStudio and Partek programs, and gene sets/pathways were identified using Gene Set Enrichment Analysis (GSEA) and Ingenuity Pathway Analysis (IPA) with significance defined as p<0.01, FDR <0.25. The proportion of live, apoptotic, and metabolically active stromal cells was quantified using Neutral Red, APOpercentage apoptosis and MTT assays, respectively. Receptors on CLL cells were investigated by flow cytometry.
GEPs obtained under hypoxic vs. normoxic (21% O2) conditions showed that at low O2 tension HS-5 cells underexpress (fold difference ≥2 and p<0.01) a discrete set of proinflammatory chemokines and genes which promote adhesion and overexpress genes associated with cell movement and cytoskeletal (re)organization. Similar gene sets were identified using IPA and GSEA. Specifically, CCL2, CCL5, CCL7and CCL8 were differentially underexpressed under low O2 tensions over the whole course of the study and CCL3, CCL3L1 and CCL3L3 decreased progressively at weekly intervals. In sharp, contrast, the same ligands were overexpressed under normoxia because levels were higher at the outset. Interestingly, co-culture of HS-5 with CLL cells under both hypoxic and normoxic conditions led to enhanced production of CCL3 and CCL3-like mRNAs and a decreased production of CCL2, CCL5, CCL7 and CCL8, although the latter was more striking under normoxia. Similarly, when HS-5 was cultured alone at 1% O2, DLC1, IL24, NQO1, TIMP1 and NRG, genes which promote adhesion, were downregulated; when co-cultured with CLL cells these were upregulated. Interestingly over the course of the entire study, hypoxic stroma upregulated ANGPTL4, ANXA2, BCAR1, BHLHE40, LOX, and SLC2A1 genes associated with cytoskeletal organization and migration. These data suggest hypoxia induced migration for the stroma and CLL induced capacity to retain leukemic cells on the stroma.
Stromal-CLL cell interactions were studied by analyzing chemokine and SLAM family receptors on CLL cells. CLL cells from IGHV mutated (M-CLL, n=2) and unmutated (U-CLL, n=2) patients showed downregulation of CCR2 (receptor for CCL2, CCL7, CCL8) and CCR5 (receptor for CCL3, CCL5, CCL8) at one week under hypoxia and/or when co-cultured on stroma. Interestingly, M-CLL clones had higher percentage of CCR1 and CCR5 expressing cells compared to U-CLL samples under normoxia but with minute or no differences under hypoxia or with co-culture, suggesting M-CLL cells display these receptors similar to (bad prognosis) U-CLL cells under hypoxia/co-culture conditions. Moreover, U-CLL cells upregulated SLAM family markers CD48 and CD150 on their surface at hypoxic/co-culture conditions compared to M-CLL cells, implying that hypoxia downregulates lymphocyte recruiting chemokine receptors and upregulates SLAM family receptors on U-CLL cells. Moreover, under hypoxia stromal cells showed a significant decrease in Apoptosis dye staining, increase in neutral red staining for live cells, and decrease in MTT reduction at respective 1, 2 or 3 weeks indicating controlled cell growth under hypoxia.
In summary, despite promoting controlled growth characteristics, these findings suggest that, in isolation, hypoxic stroma do not promote inflammatory processes and thus negatively regulate cell recruitment and interactions. However in the presence of CLL cells, stromal cells upregulate selective chemokines and genes controlling adhesion to the stroma; these changes are CLL B-cell induced and enhance the likelihood the leukemic cells and macrophages and potentially other immune/inflammatory cells will be attracted to and retained on the stroma. The latter effects would favor interactions with and survival of CLL cells.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.