Expression and modulation of specific, high affinity binding sites for erythropoietin on the human erythroleukemic cell line K562

Erythroid differentiation is mediated by several interacting factors which include the glycoprotein hormone erythropoietin (Epo), interleukin-3 (IL-3) in the mouse, and erythroid-potentiating activity (EPA) in humans. Each of these factors binds to specific cell surface receptors on responsive target cells, but the way in which these factors interact to modulate erythropoiesis is unknown. In the present study, we used the human erythroleukemic cell line K562 to examine expression and regulation of the receptor for Epo using 125I-labeled, bioactive recombinant human Epo. K562 cells expressed low numbers of a single class of high-affinity Epo receptors corresponding to 4 to 6 receptors per K562 cell (KD = 270 to 290 pmol/L). Treatment of K562 cell cultures with medium conditioned by the EPA-secreting cell line U937 (U937CM) increased receptor expression 2.6 to 3.5-fold to 13 to 17 receptors/cell (KD = 260 to 300 pmol/L). That all of the Epo receptor-potentiating activity in U937CM was accounted for by EPA was shown by a similar increase in Epo receptor expression on K562 cells with recombinant EPA. The effect of U937CM on Epo receptors was reversed by culturing cells in inducer-free medium for 3 days. Medium conditioned by the 5637 cell line had no effect on Epo receptors on K562 cells. In methylcellulose culture, U937CM and Epo acted synergistically to increase erythroid differentiation of K562. Similarly, U937CM stimulated human cord blood CFU-E growth under conditions in which Epo was limiting or in excess. Increases in Epo receptor expression on K562 cells and on CFU-E in response to EPA may mediate the effects of Epo on these cells.     

Loss of suppression of normal bone marrow colony formation by leukemic cell lines after differentiation is induced by chemical agents

The human leukemic cell lines K562 and HL-60 were cocultured with normal bone marrow (BM) cells. Coculture with 10(4) K562 or HL-60 cells results in 50% inhibition of normal CFU-E and BFU-E colony formation. However, when the same number of K562 and HL-60 cells is first treated for two to five days with agents that induce their differentiation, a gradual loss in their capacity to inhibit CFU-E and BFU-E colony formation is observed. The inhibitory material in K562 cells is soluble and present in conditioned medium from cultures of these cells. The degree to which leukemic cell suppression of CFU-E and BFU-E growth is reversed is correlated with the time of exposure to the inducing agent. Suppression is no longer evident after five days of prior treatment with inducers. In fact, up to a 90% stimulation of CFU-E growth is observed in cocultures with K562 cells that have been pretreated with 30 to 70 mumol/L hemin for five days. K562 cells treated with concentrations of hemin as low as 30 mumol/L demonstrate increased hemoglobin synthesis and grow normally, but no longer have an inhibitory effect on CFU-E growth. Hence, reversal of normal BM growth inhibition must be caused by the more differentiated state of the K562 cells and not by a decrease in the number of these cells with treatment. Thus, induction of differentiation in cultured leukemic cells not only alters the malignant cell phenotype but also permits improved growth of accompanying normal marrow progenitor cells. Both are desired effects of chemotherapy.     

Growth hormone polypeptides stimulate proliferation of K562 human erythroleukemia cells

The growth-promoting effect of growth hormones and related polypeptides was examined in vitro with a clonogenic assay using human erythroleukemic cells (K562). The erythroleukemia cells were grown in a serum-substituted methylcellulose culture system and colonies counted after 4 days' incubation. Human growth hormone (hGH) was a potent stimulant for K562 cell growth (60% augmentation). The cys(Cam)53-hGH(1-134) fragment of hGH and human chorionic somatomammotropin had less than half the activity of the intact hGH molecule. Bovine growth hormone was inactive in enhancing K562 colony formation. K562 cell proliferation was stimulated by hGH at concentrations as low as 0.1 ng/ml. Insulin stimulated K562 cell proliferation most effectively at a concentration of 1 ng/ml. The K562 culture system may conveniently be employed for determining the in vitro bioactivity of hGH.     

Human leukemia cell line K562 responds to erythroid-potentiating activity

We report that erythroid-potentiating activity (EPA), known to stimulate the proliferation of normal human erythroid precursors in vitro, has a growth-promoting effect on human K562 erythroleukemia cells and Friend mouse erythroleukemia cells. Detailed studies were carried out using an EPA produced by a human T-lymphoblast line (Mo). Although EPA has not been purified to homogeneity, several observations indicate that the factor elaborated by Mo cells that stimulates erythroleukemia cell growth is the EPA molecule. The erythroleukemia growth factor cofractionates with EPA using gel exclusion chromatography, isoelectric focusing, and ion exchange chromatography. In addition, the activities exhibit similar kinetics of heat inactivation. A granulocyte-macrophage colony-stimulating factor also elaborated by Mo cells had no effect on the growth of the erythroleukemia cells. Other sources of EPA, such as peripheral blood leukocyte-conditioned medium, preparations from urine of anemic patients, and medium conditioned by a human monocyte-like cell line, stimulated erythroleukemia cell growth. Mouse sources of EPA (termed "burst-promoting activity") stimulated mouse but not human erythroleukemia cells. The availability of cell lines apparently responsive to EPA should prove useful for examining the mode of action of this regulator of erythropoiesis.     

Hormonal effects on cell proliferation in a human erythroleukemia cell line (K562)

We have investigated the hormonal responsiveness of K562 cells using a serum-substituted in vitro clonogenic assay. Dexamethasone inhibited colony formation by the K562 cells, and the inhibitory effect could be reversed by progesterone (10(-6) M). Fluoxymesterone caused a prominent enhancement of K562 colony growth, whereas estriol had no effect. Stimulation by triiodothyronine was maximal at 10(-7) M, and the thyroid effect could be abrogated by the beta 2-adrenergic antagonist butoxamine in equimolar concentrations. Using standard tissue culture conditions, the beta-adrenergic agent isoproterenol, but not the alpha catecholamine phenylephrine, enhanced the proliferation of K562 cells. When K562 cells were grown under hormone-depleted conditions, they developed responsiveness to phenylephrine and were no longer stimulated by isoproterenol. DbcAMP and prostaglandins of the E series also caused K562 colony enhancement. Prostaglandin F2 alpha had no effect on cell proliferation. Insulin was an effective stimulant of colony formation of K562 cells, as were human growth hormone and ovine prolacin. Bovine growth hormone had no effect. Our results are consistent with the identificaiton of K562 as an erythroid line, and they indicate that K562 cells respond to endocrine hormones in a manner analogous to normal erythroid progenitors.     

Production of erythroid-potentiating activity by a human T-lymphoblast cell line.

We derived a human T-lymphoblast cell line (Mo) that constitutively elaborates certain lymphokines. The Mo cells produce a colony-stimulating factor necessary for the growth of human granulocyte-monocyte precursors in vitro as well as an erythroid-potentiating activity (EPA) that enhances the proliferation of human erythroid progenitors in vitro. In the presence of serum, the EPA in Mo-conditioned medium stimulated the growth of small and large erythroid colonies almost 2-fold. EPA was also produced in serum-free medium, and, when assayed in serum-free cultures of human erythroid progenitors, it stimulated colony growth about 3-fold. The EPA produced by the Mo cell line did not stimulate normal murine erythroid progenitors (CFU-E) or Friend erythroleukemia cell growth in vitro. EPA was inactivated by protease treatment but was remarkably heat stable, with most of the activity recovered after boiling for 15 min. Preliminary biochemical characterization suggests that EPA is an acidic glycoprotein with molecular weight approximately 45,000. EPA is clearly separable from colony-stimulating factor on the basis of heat stability and gel-filtration chromatography. The present observations provide strong support for the concept that activated T cells produce humoral factors important in the regulation of erythropoiesis. The availability of a cell line producing human EPA should facilitate the characterization of the protein and permit definitive studies of its biologic effects.     

Activities derived from established human myeloid cell lines reverse the suppression of cell line colony formation by lactoferrin and transferrin

Myeloid cell lines were evaluated for the release of substances needed for colony formation by their own colony-forming cells (CFC) and by other myeloid cell lines. Dialyzed U937 conditioned medium (CM) had no effect on the cloning efficiency of U937 cells, whether or not U937 CFC had been induced for MHC class-II antigens by preincubation of these cells for 72 h with indomethacin and human gamma interferon (HuIFN gamma). Dialyzed U937 CM, however, restored colony formation of HuIFN gamma-induced U937 cells suppressed by lactoferrin (LF) or transferrin (TF). Dialyzed U937 CM did not restore colony formation of U937 cells suppressed by acidic isoferritins (AIF) or prostaglandin E2 (PGE2). Detection of the growth-restoring effects of U937 CM required that U937 CM be prepared in the presence of indomethacin or that the CM be dialyzed to remove inhibitors of U937 colony formation. Dialyzed U937 CM did not inactivate LF. Dialyzed U937 CM did not stimulate or enhance colony formation of normal human bone marrow granulocyte-macrophage (CFU-GM), erythroid (BFU-E), or multipotential (CFU-GEMM) progenitor cells, but did contain potent inhibitory activity against these progenitor cells. HL-60, EM2, EM3, and K562 cells were also evaluated. HL-60-, EM3-, and K562-CFC that were not preincubated with HuIFN gamma did not express MHC class-II antigens, and colony formation by these cells was not influenced by LF, TF, or AIF. Noninduced EM2-CFC constitutively expressed MHC class-II antigens, and colony formation by these cells was suppressed by LF, TF, and AIF. After induction of MHC class-II antigens on HL-60- and EM3-CFC by HuIFN gamma, colony formation by these cells was suppressed by LF, TF, and AIF. Colony formation by HuIFN gamma-induced EM2 cells was more responsive to inhibition by LF, TF, and AIF than was colony formation by noninduced EM2 cells. K562 cells were not induced into a responsive state to LF, TF, or AIF by HuIFN gamma. Dialyzed CM from HL-60, EM2, and EM3 cells contained activities that restored colony formation by their own LF-suppressed CFC. The activities present in dialyzed CM from U937, HL-60, EM2, and EM3 cells may be similar since they could each restore LF-suppressed colony formation of U937, HL-60, EM2, or EM3 cells.(ABSTRACT TRUNCATED AT 400 WORDS)     

Tumor necrosis factor-alpha and hematopoietic progenitors: effects of tumor necrosis factor on the growth of erythroid progenitors CFU-E and BFU-E and the hematopoietic cell lines K562, HL60, and HEL cells

Macrophages can modulate the growth of hematopoietic progenitors. We have examined the effects of tumor necrosis factor-alpha, a product of activated macrophages, on human erythroid progenitors (CFU-E, BFU-E) and the hematopoietic cell lines K562, HL60, and HEL cells. Tumor necrosis factor (TNF) significantly inhibited CFU-E and BFU-E growth at concentrations as low as 10(-11)-10(-12) M (0.2 U/ml), although erythroid colony and burst formation were not totally ablated. Preincubation of marrow samples with TNF for 15 min was sufficient to suppress erythroid colony and burst formation. Addition of TNF after the start of culture inhibited CFU-E- and BFU-E-derived colony formation if TNF was added within the first 48 h of culture. Additionally, TNF inhibited the growth of highly purified erythroid progenitors harvested from day 5 BFU-E. The colonies which formed in cultures treated with TNF were significantly smaller than those formed in control cultures. TNF (10(-8)-10(-10) M) also suppressed the growth of the hematopoietic cell lines K562, HL60, and HEL cells, with 40%-60% of the cells being sensitive to TNF. Preincubation of HL60 cells with TNF for 15 min significantly inhibited their growth. K562, HL60, and HEL cells expressed high-affinity receptors for TNF in low numbers (6000-10,000 receptors per cell). Fluorescence-activated cell sorter analysis of TNF binding to HEL cells demonstrated that the majority of these cells expressed TNF receptors. These data suggest that: (1) TNF is a rapid irreversible and extremely potent inhibitor of CFU-E, BFU-E, and hematopoietic cell lines K562, HL60, and HEL cells; (2) TNF appears to be acting on a subpopulation of erythroid cells, predominantly CFU-E, BFU-E, and possibly proerythroblasts; (3) TNF appears not to require accessory cells such as lymphocytes or macrophages to inhibit erythroid progenitors; and (4) the presence of TNF receptors on hematopoietic cells is not sufficient to confer sensitivity to TNF since the majority (80%-95%) of HEL cells express TNF receptors while only 40%-60% are inhibited by TNF.     

Production of leukemic blast growth factor by a human bladder carcinoma cell line

Circulating blast cells from the peripheral blood of acute myeloblastic leukemia patients include a subpopulation capable of colony formation in the presence of phytohemagglutinin-stimulated leukocyte-conditioned medium (PHA-LCM). We describe the complete replacement in the blast assay of PHA-LCM by conditioned medium from a human bladder carcinoma cell line, HTB9. Both conditioned media contain a stimulator of blast cell growth that elutes as a single peak from a Sephadex G100 column with an apparent molecular weight of 30,000. It is shown that this leukemic blast growth factor is distinct from erythroid-potentiating activity (EPA) and possibly granulocyte macrophage colony-stimulating factor.     

mda-7/IL-24 inhibits the proliferation of hematopoietic malignancies in vitro and in vivo

OBJECTIVE: Melanoma differentiation-associated gene-7/interleukin-24 (mda-7/IL-24) has been consistently shown to exert growth inhibitory effects on various tumor types. However, the majority of these reports were limited to solid tumors. The purpose of this study was to investigate the antitumor activity of mda-7/IL-24 and the underlying mechanism in hematopoietic malignancies. MATERIALS AND METHODS: We determined the expression of mda-7/IL24 and its heterodimeric receptors in hematopoietic tumor cell lines and then stably transfected mda-7/IL-24 into K562 (leukemia) and Namalwa (lymphoma) cell lines to assess the effects of mda-7/IL-24 on cell proliferation, cell cycle, apoptosis, colony-forming ability, and tumor growth in vivo. Microarray analysis was performed to determine the genes that were differentially regulated by mda-7/IL-24 in K562 cells. RESULTS: Expression of mda-7/IL-24 or its intact receptor pairs was not detected in the 11 cell lines tested. Ectopic expression of mda-7/IL-24 induced significant (p < 0.05) inhibition of cell growth and colony formation in both K562 and Namalwa cells, and the growth inhibition in K562 cells was associated with G(0)/G(1) cell-cycle arrest. Results of in vivo studies showed good correlation with in vitro inhibition of tumor cell proliferation in both the cell lines. We also showed that the increase in p21(WAF-1) and BCCIP and decrease in cdk6, smurf2, and phosphorylated pRb, which are regulators of cell-cycle progression, might account for G(0)/G(1) cell-cycle arrest in K562 cells. CONCLUSIONS: The present study demonstrated for the first time the potential antitumor activity of mda-7/IL-24 in chronic myelogenous leukemia and lymphoma.     

Expression and induction of cathepsins B and D in K562 cells

Three distinct lysosomal protease activities have been identified in the human leukemia cell line, K562. These include cathepsin D, the classic protease of the mature red blood cell, as well as two proteases, cathepsins B and H, which have been associated with development and differentiation in a variety of tissues. Each of these three lysosomal proteases was expressed in a specific fashion during hemoglobin induction in K562 cells. Both cathepsin B and cathepsin D activities could be induced by growth of K562 cells in medium containing either hemin or heat-treated serum or by increasing the concentrations of untreated serum in the medium. Cathepsin H activity in the same cells remained unchanged. This is the first report of inducible protease activities in K562 cells. Our identification of specific well-characterized protease activities that change differentially during K562 induction provides a framework for additional studies on the role of proteases in hematopoietic differentiation.     

HSPB8 is methylated in hematopoietic malignancies and overexpression of HSPB8 exhibits antileukemia effect

HSPB8 has been shown to be involved in regulation of cell proliferation and apoptosis, and it has also been found to have divergent properties in solid tumors. The purpose of this study was to investigate the expression and function of HSPB8 in hematopoietic malignancies. Expression and induced expression of HSPB8 was evaluated in hematopoietic tumor cell lines and bone marrow samples from patients with leukemia. Methylation status was investigated by methylation-specific polymerase chain reaction. The role of HSPB8 in hematopoietic malignancies was addressed by reintroducing HSPB8 expression into the K562 (leukemia) and Namalwa (lymphoma) cell lines. Expression of HSPB8 was absent in hematopoietic tumor cell lines and primary patient and normal volunteer samples. Promoter DNA methylation of HSPB8 was observed in these cells. HSPB8 expression could be restored after demethylation treatment with 5-aza-2'-deoxycytidine. Overexpression of HSPB8 reduced colony formation of both K562 and Namalwa cell lines, inhibited the cell growth of Namalwa in?vitro, and suppressed tumor formation of K562 cells in?vivo. The present study demonstrates that HSPB8 is silenced by DNA methylation in hematopoietic malignant and normal cells and its expression can be induced by treatment with 5-aza-2'-deoxycytidine. Overexpression of HSPB8 may have an antitumor activity in chronic myelogenous leukemia and lymphoma.     

Killing of K562 cells with conjugates between human transferrin and a ribosome-inactivating protein (SO-6)

Cellular iron uptake is mediated by binding of transferrin with specific surface receptors and internalization of the Fe-transferrin-receptor complex. This has been examined as a possible pathway for carrying into leukaemic cells a ribosome-inactivating protein (RIP), SO-6, derived from Saponaria officinalis. Purified human differic transferrin was conjugated with SO-6 and a pool of proteins was obtained, with variable numbers of SO-6 molecules linked to a single transferrin molecule. Human erythroleukaemic K562 cells were grown in the presence of human transferrin, SO-6 and human transferrin conjugated with SO-6. The conjugate was found to be internalized via binding with transferrin receptor. Whereas the presence of unconjugated human transferrin and SO-6 in the medium did not significantly influence K562 cell growth, the conjugated proteins displayed an inhibitory activity on cell proliferation. This was maximal after 72 h at a transferrin concentration of 10(-9) M, with about 50% of cells being killed. Bovine transferrin, present in fetal calf serum, did not appear to compete with human diferric transferrin in binding to K562 cells in suspension culture. In a clonogenic assay, colony formation by leukaemic cells was not influenced by free SO-6 or transferrin, whereas the conjugated proteins were markedly inhibitory (about 100% at 10(-9) M). Our findings indicate that SO-6 can be efficiently carried into mammalian cells via the transferrin-transferrin receptor cycle and exert its ribosome inactivating activity. This is in keeping with the existence of an alternative pathway of transferrin endocytosis in addition to the classic acidic endosome pathway. From a practical viewpoint, conjugates between transferrin and SO-6 can be useful tools for studying the expression of transferrin receptors, and deserve also to be investigated for a possible use in cancer therapy.     

Human T-lymphocyte products stimulate human hemopoietic progenitor cell proliferation in diffusion chambers in vivo

Human myeloid colony formation in diffusion chambers in mice (CFU-DG) was enhanced following administration of a human T-cell-line-derived conditioned medium (Mo). The Mo cell line also elaborates activities stimulating human myeloid colony formation in vitro in agar (CSF) and potentiating erythroid colony formation in vitro in methylcellulose (EPA). Depletion of CSF from Mo conditioned medium by heat inactivation or gel exclusion chromatography did not affect CFU-DG formation. EPA and CFU-DG stimulating activities are heat stable and have approximately the same molecular weight. Culture of human bone marrow cells in diffusion chambers in mice for 4 days under the influence of Mo conditioned medium resulted in significant increment of BFU-E and CFU-DG as judged by subculture of diffusion chamber contents. No effect on CFU-C could be detected.     

The role of folates in the development of methotrexate resistance in human leukemia cell line K562

The effect of reduced and oxidized folates on the development of methotrexate (MTX) resistance has been examined in human leukemia cell line K562 (K562/S). K562/S cells were made resistant to MTX by soft-agar cloning either in RPMI-1640 medium (K562/MTX-PGA) or in folic-acid-free RPMI-1640 medium containing 10 nM leucovorin (K562/MTX-LV). The optimal concentrations of leucovorin for the growth of K562/S, K562/MTX-PGA and K562/MTX-LV cells were 1 nM, 5 nM and 10 nM respectively. K562/MTX-PGA cells were 24-fold resistant to MTX as noted by impaired MTX transport. In contrast, K562/MTX-LV cells were 26-fold resistant to MTX as noted by gene amplification of dihydrofolate reductase. Furthermore cross-resistance to cytosine arabinoside was only demonstrated in K562/MTX-PGA, while the K562/MTX-LV cells showed no significant cross-resistance to cytosine arabinoside. These results suggest that the type and level of folates used during the development of MTX resistance may play a role in the mechanism for MTX resistance. Leukemia cells that are grown in leucovorin might serve as a model for acquired MTX resistance in vivo.     

Cell interactions influencing murine marrow megakaryocytes: nature of the potentiator cell in bone marrow

Auxiliary bone marrow cells are required for optimal murine megakaryocyte colony formation in addition to progenitor cells and a colony stimulating activity (CSA) present in WEHI-3 cell conditioned medium. These auxiliary cells are adherent, with a sedimentation rate of 5.8 mm hr-1 and buoyant density of 1.065-1.078 gcm-3. The activity from bone marrow cells is loss at irradiation doses above 900 rad. Bone marrow cells with these characteristics, and supernatants from lung, bone shafts, and peritoneal exudate cells were all active in enhancing megakaryocyte colony incidences in mouse bone marrow cultures above those stimulated by an obligatory activity in WEHI-3 cell conditioned medium. Certain macrophage cell lines (J774, P388D1) could elaborate the activity. This study confirms that a potentiation activity enhances CSA stimulation of megakaryocyte colony formation. The potentiator is elaborated by bone marrow cells in limiting amounts requiring either high cell concentrations or an exogenous source of the activity for optimal colony growth.     

Induction of erythropoietic colonies in a human chronic myelogenous leukemia cell line.

The ability of cells derived from the K562 cell line to generate erythropoietic colonies was studied. The K562 cell line was derived from a patient with chronic myelogenous leukemia 8 yr ago by Lozzio and Lozzio. Rare benzidine-positive colonies formed when these cells were cloned in plasma clots (3 +/- 1/10(4) cells), and their number was not substantially increased by the addition of erythropoietin (9.5 +/- 1/10(4) cells). Sodium butyrate was capable of markedly enhancing the number of benzidine-positive colonies (19.5 +/- 1/10(4) cells) formed, while the combination of sodium butyrate plus erythropoietin exerted a synergistic effect on erythropoietic colony formation (57 +/- 4/10(4) cells). The K562 cell line is a long-term culture system that contains human erythropoietic stem cells. This cell line should be useful in future studies on the cellular and molecular events associated with human erythroid cell differentiation.     

Colony growth of normal and neoplastic cells in various concentrations of methylcellulose

To assess the semisolid character of methylcellulose (MC) and its ability to prevent cell migration and aggregation in clonogenic assays, we studied the influence of various concentrations of MC (0.7%-1.26%) on colony growth of neoplastic cell lines, normal bone marrow cells, and hairy cell leukemia (HCL). All cell lines (K562, HL-60, JOK-1, Daudi, and BB3, an IgM-kappa B-cell line) showed a prominent decrease in colony numbers and remarkable changes in colony morphology at rising MC concentrations, whereas no such influence could be demonstrated for HCL, mixed lineage colony-forming units (CFU-GEMM), granulocyte-macrophage CFU (CFU-GM), erythroid burst-forming units (BFU-E), and erythroid CFU (CFU-E). Despite a decrease in colony numbers at high MC concentrations, some cell lines showed a sustained proliferation as measured by growth index calculations and bromodeoxyuridine (BrdUrd) incorporation. This indicates that at certain MC concentrations colony formation is not always a reflection of proliferation. BrdUrd incorporation yielded an extremely low proliferation capacity for HCL. It is likely that HCL cells, which strongly aggregate, formed pseudo-colonies in spite of high MC concentrations.     

Natural killer cell-mediated inhibition of growth of myeloid and lymphoid clonogenic leukemias

We have shown that peripheral blood (PB) lymphocytes of normal donors inhibit colony formation by myeloid (K-562) and lymphoid (Molt-4) leukemia cell lines in a clonogenic assay in vitro. The inhibitor cells were identified as natural killer (NK) cells based on their large granular lymphocyte (LGL) morphology and CD16+CD5- cell surface phenotype. The levels of inhibition were dependent both on the leukemia: effector (L:E) cell ratio, as well as on the time of preincubation of the leukemic and effector cells; maximum inhibition was observed at a 1:20 L:E cell ratio, and required 6-16 h preincubation of the leukemia and effector cells. Colony formation of both K-562 and Molt-4 was also inhibited by a soluble factor derived from coculture of K-562 and PB lymphocytes. The finding that the growth of clonogenic Molt-4 cells was almost completely abolished following treatment with interferon (IFN)-alpha rA in the dose of 10(3) U/ml, whereas growth of K-562 cells was only slightly affected by similar treatment, suggested that NK cells may mediate inhibition of clonogenic leukemias through various mechanisms.