Hypercholesterolemia Increases the Risk of Colorectal Cancer by a Tet-1-Dependent HSC-Autonomous Mechanism

Abstract

The annual, age-standardized colorectal cancer (CRC) incidence rate has decreased by 46% from its peak in 1985. However, this long-standing decline in cases of CRC slowed due mainly to an increase in incidence in individuals younger than 50 years of age. For those less than 50 years of age, CRC is the leading cause of cancer deaths in men and the second in women. At least half of all cases of young-onset CRC are linked to lifestyle risk factors, including obesity. Hypercholesterolemia, a common metabolic disorder in obese people, has been shown to increase the risk of colorectal cancer, but the mechanism is unknown. We will show that hypercholesterolemia increases the incidence and pathological severity of colorectal cancer by inducing an oxidant stress-dependent hematopoietic stem cell-autonomous mechanism. The oxidized-LDL increase in HSC oxidant stress initiates a signaling pathway that culminates in the increased expression of miR101c that downregulates Tet1. This downregulation of Tet1 reduces the expression of the genes critical to the production and cytotoxicity of natural killer T cells and T cells, thereby impairing cancer immunosurveillance against colorectal cancer. This reveals a novel mechanism where a metabolic disorder induces epigenetic reprogramming of natural killer T cells and γδT gene expression within hematopoietic stem cells.

1. Introduction

In 2024, colorectal cancer (CRC), the second most common cause of cancer death in the United States, is anticipated to cause 152,810 new cases of CRC and 53,010 deaths [1]. The annual, age-standardized CRC incidence rate has decreased by 46%, from 66.2 per 100,000 at its peak in 1985 to 35.7 per 100,000 persons in 2019. This decline is similar for men and women and occurred in large part due to the rapid dissemination of colonoscopy, reduced smoking, and more effective treatments [1].

However, this long-standing decline in cases of CRC has abated, driven largely by an increase in individuals younger than 55 years of age developing early-onset CRC at a more advanced stage and in the left colon/rectum [2]. For those less than 50 years of age, CRC is the leading cause of cancer death in men and the second leading cause of cancer death in women. Their presentation at a more advanced stage is attributed to the lack of screenings for that age group [1]. According to an analysis of the Surveillance, Epidemiology, and End Results (SEER) registry data, colon cancer incidence in Americans 20–34 years of age is anticipated to increase by 90% by 2030 and for rectosigmoid and rectal cancers by 124% [2].

The reasons for this devastating young onset of CRC have not yet been elucidated. Most do not have a genetic predisposition. For this reason, at least half of all cases of young-onset CRC are linked to lifestyle risk factors such as obesity, antibiotic use, and low physical activity. Diet may also be important in the development of young-onset CRC [3, 4, 5]. Identifying the pathogenetic mechanisms responsible for the young onset of CRC will permit the development of strategies to improve the clinical outcomes for this group of young patients.

Obesity will soon surpass smoking as the most preventable cause of cancer [6, 7]. Central or abdominal obesity doubles the risk of developing CRC [8]. The studies to determine which metabolic disorder in obese people is responsible for increasing their cancer risk have been unconvincing [9]. A common metabolic disorder in obese people, hypercholesterolemia, has been shown to increase cancer risk, including a significant increase in the risk of colorectal cancer [10]. Abdominal obesity and hypercholesterolemia synergically doubled the risk of CRC [8]. In addition, epidemiological investigations have found that the occurrence of CRC is positively correlated with a high-cholesterol diet [11, 12].

How hypercholesterolemia increases the risk of colorectal cancer has not been established. The mechanism linking hypercholesterolemia to CRC is usually attributed to its luminal effects by increased toxic bile acids and carcinogenic heterocyclic amines. Recently, a persuasive hypothesis has been advanced that hypercholesterolemia exerted a systemic and conditional influence that impaired the cellular components of tumor immunosurveillance [5, 8].

Hypercholesterolemia has also been shown to regulate the inflammatory response and innate immune response [13]. Additional support for the deleterious effects of hypercholesterolemia on the incidence of CRC comes from studies that show statin use can reduce the risk of overall colorectal cancer [14].

While hematopoietic stem cells maintain a quiescent undifferentiated state, they are subject to activating epigenetic enzymes and expression of lineage-associated genes, a process of lineage priming that maintains HSCs responsive to the physiologic and pathological demands for immune cell responses [15, 16, 17, 18]. We have established that hypercholesterolemia profoundly affects HSC’s most fundamental functions. We have shown that hypercholesterolemia induces oxidant stress in HSCs, accelerating their aging and impairing their repopulation capacity [19].

The primary effect of hypercholesterolemia on HSCs is to induce a receptor-mediated uptake of oxidized-low-density lipoprotein (ox-LDL) that increases the HSC oxidant stress that, in turn, activates the p38 MAPK pathway, which increases Notch1 [19]. This increased Notch expression induced a loss of quiescence, reducing the number of “long-term HSCs,” undifferentiated pluripotent stem cells, and increased “short-term HSC,” stem cells committed to various lineage specification pathways. In addition, hypercholesterolemia shortened telomeres and reduced their repopulation capacity. Together, these results show that hypercholesterolemia accelerates the aging of HSCs.

Based on these findings, we advanced the hypothesis that hypercholesterolemia induces an oxidized-LDL-dependent increase in HSC oxidant stress that reduces the production of innate immune cells that, in turn, impairs immunosurveillance against CRC.

2. The mechanisms by which hypercholesterolemia increases the incidence and pathological severity of colorectal neoplasia

2.1 Hypercholesterolemia increases the incidence and histopathologic severity of colorectal neoplasia by an HSC-autonomous mechanism

Colorectal neoplasia was induced with azoxymethane (AOM) in two mouse models of hypercholesterolemia, the ApoE−/− mouse and the C57BL/6 mouse fed a high cholesterol diet (HCD). The average tumor number was more than two-fold higher in hypercholesterolemic mice than in WT mice (Figure 1A). The histopathological stage of the tumors in hypercholesterolemic mice was also more advanced than in WT mice. The tumors in the hypercholesterolemic mice progressed to the late stages of tumorigenesis, including adenoma+++ and carcinoma stages. Meanwhile, the tumors at the early stages of tumorigenesis, including hyperplasia and adenoma+, were dramatically reduced in hypercholesterolemic mice (Figure 1B). Together, these results show that hypercholesterolemia increases the number and pathological severity of colorectal tumors [13].

Figure 1.

Hypercholesterolemia increases the average tumor number and histopathologic stage of colorectal neoplasia through a hematopoietic stem cell-autonomous manner. A. Average tumor number per mouse from WT, ApoE-/-, and HCD mice. B. Histopathologic stages of the tumors from WT, ApoE-/-, and HCD mice. n=12, *, p<0.05; **, p<0.01, vs. WT. C. Average tumor number from WT recipient mice reconstituted with HSCs from WT or ApoE-/- mice. D. Histopathologic stages of the tumors from WT recipient mice reconstituted with HSCs from WT or ApoE-/- mice. n=12, *, p<0.05; **, p<0.01, vs. WT [13].

To determine if hypercholesterolemia increases the incidence and pathologic severity of colorectal cancer by an HSC-autonomous mechanism, we transplanted HSCs from WT (CD45.2) and hypercholesterolemic mice into lethally irradiated WT mice (CD45.1). The serum cholesterol and white blood cell counts of the recipient WT mice were normal (Figure 2D and E). Remarkably, we found that the WT mice reconstituted with HSCs from hypercholesterolemic mice had nearly two-fold more tumors, and the histopathological severity was worse than that in WT mice reconstituted with WT HSCs (Figure 1C and D). These results show that hypercholesterolemia increases the incidence and pathologic severity of colorectal tumors by an HSC-autonomous mechanism.

Figure 2.

WT recipient mice reconstituted with HSCs from hypercholesterolemic ApoE-/- mice have normal serum cholesterol levels and white blood cell counts. A. Serum cholesterol levels of WT recipient mice reconstituted with HSCs from WT or ApoE-/- mice. B. White blood cell counts of WT recipient mice reconstituted with HSCs from WT or ApoE-/- mice [13].

2.2 Hypercholesterolemia specifically reduces the differentiation of HSCs toward NKT and γδ T cells

Given that hypercholesterolemia increases the incidence of colorectal neoplasia by an HSC-autonomous mechanism, we hypothesized that hypercholesterolemia impairs HSC production of immune cells that weakens cancer immunosurveillance against colorectal neoplasia. We found that NKT and γδ T cells were significantly reduced in number in the circulation. While these cells are rare in circulation and secondary lymphoid tissue, they are enriched in many peripheral tissues, such as the skin, intestines, and lungs. They are critical in the colon submucosa’s response to infection and tumors. NKT and γδ T cells are unique because they exhibit innate and adaptive immune responses. NKT and γδ T cells secrete various cytokines critical for the antitumor functions of cytotoxic T cells. NKT and γδ T cells also interact with antigen-presenting cells to induce them to secrete cytokines that recruit and stimulate the antitumor functions of cytotoxic T cells, boosting innate and adaptive antitumor responses [19, 20, 21].

NKT and γδ T cell concentrations were also reduced in the thymus of hypercholesterolemic mice (Figure 3A and B). At phases 1 CD44⁻NK1.1⁻] and 2 [CD44⁺NK1.1⁻; intrathymic NKT-cell development in hypercholesterolemic mice was identical to WT (Figure 4C) as was the CD4+ subsets of NKT cells (Figure 4D). In all of the mouse models, the T-cell developmental intermediates were similar (Figure 4E and F) FACS analysis did not show any significant change in CDe3+, CD4+, and CD8+ T-cell populations in the peripheral blood of hypercholesterolemic mice apart from a slight decrease in B cells and a slight increase in NK cells (Figure 4G). In lethally irradiated WT recipient mice that were reconstituted with HSCs from hypercholesterolemic mice, we observed in the thymus an almost identical decrease in the HSC differentiation toward γδ T cells and NKT cells as that seen in hypercholesterolemic mice (Figure 3C and D). Thus, hypercholesterolemia induces an HSC-autonomous reduction in the production of HSCs toward NKT and γδ T cells [13].

Figure 3.

Hypercholesterolemia significantly impairs the differentiation of HSCs towards NKT and γδT cells which are critical components of innate immunity against colorectal neoplasia. A. Hypercholesterolemia significantly impairs the differentiation of HSCs toward NKT and γδT cells, which are critical components of innate immunity against colorectal neoplasia. A, Frequency and total number of NKT cells in thymus of WT and ApoE−/− mice. n = 8. B. Frequency and total number of γδT cells in thymus of WT and ApoE−/− mice. n = 8. C. Frequency and total number of NKT cells in thymus of lethally irradiated WT recipients reconstituted with HSCs from WT or ApoE−/− mice. n = 8. D. Frequency and total number of γδT cells in thymus of lethally irradiated WT recipients reconstituted with HSCs from WT or ApoE−/− mice. n = 8. E. Frequency of submucosal NKT cells in colon of WT and ApoE−/− mice. n = 8. F. Frequency of submucosal γδT cells in colon of WT and ApoE−/− mice. n = 8. G. Average tumor number and histopathologic stage of AOM-induced colorectal neoplasia in CD1d−/− mice. n = 10. H. Average tumor number and histopathologic stages of AOM-induced colorectal neoplasia in Tcrd−/− mice. n = 10. I. Frequency of NKT cells in the tumors from WT and ApoE−/− mice. n = 8. J, Frequency of γδT cells in the tumors from WT and ApoE−/− mice. n = 8 [13].

Figure 4.

Hypercholesterolemia reduces the frequency and alters specific subsets and maturation of NKT and γδT cells in the thymus. A. Frequency and number of NKT cells in thymus of WT and HCD mice. n=6, *, p<0.05, vs. WT. B. Frequency and number of γδT cells in thymus of WT and HCD mice. n=6, *, p<0.05, vs. WT. C. DN1 (CD44+CD25–DN), DN2 (CD44+CD25+DN) and DN3 (CD44–CD25+DN) populations in thymus of WT, ApoE-/- and HCD mice. n=8, *, p<0.05, vs. WT. D. DP (CD4+CD8+), DN (CD4-CD8-), CD4+ and CD8+ populations in thymus of WT, ApoE-/- and HCD mice. n=8, *, p<0.05, vs. WT. E. Stage 0 (CD44-NK1.1-), Stage 1 (CD44+NK1.1-) and Stage 2 (CD44+NK1.1+) of NKT cells in thymus of WT and ApoE-/- mice. F, CD4+ and CD4-CD8- subsets of NKT cells in thymus of WT and ApoE-/- mice. n = 5, *, p<0.05, vs. WT. G. Frequency of B cells, NK cells, CD3e+, CD4+ and CD8+ cells in peripheral blood of WT, ApoE-/- and HCD mice. n = 5, *, p<0.05, vs. WT. H. NK cells (CD45+CD3e-NKp46+), CD11b- dendritic cells (CD11c+CD11b-CD103+F4/80-), CD11b+ dendritic cells (CD11c+CD11b+CD103+F4/80-), CD11c- macrophages (CD11c-CD11b+CD103-F4/80+) and CD11c+ macrophages (CD11c+CD11b+CD103-F4/80+) in the colon of WT, ApoE-/- and HCD mice. n = 5, *, p<0.05, vs. WT [13].

In harmony with these findings in the thymus, in the colon submucosa of hypercholesterolemic mice, we found a substantial reduction of NKT cells [up to 6-fold] and γδ T cells [3-fold] (Figure 3E and F). In the meantime, we found in the colon of hypercholesterolemic mice that the other major cellular components of cancer immunosurveillance, including NK cells, CD11b⁻ dendritic cells, CD11b⁺ dendritic cells, CD11c⁻ macrophages, and CD11c⁺ macrophages did not show any meaningful changes (Figure 4H).

2.3 NKT and γδ T cells are critical components of immunosurveillance against colorectal neoplasia induced by azoxymethane

To provide additional support for our hypothesis that NKT and γδ T cells play a critical role in immunosurveillance against colorectal neoplasia, colorectal neoplasia was induced with azoxymethane in mice which lack γδ T cells (Tcrd−/− mice), and in mice which lack NKT cells (CD1d−/−mice). Both mouse strains showed a much higher incidence and greater histopathologic severity of colorectal neoplasia than that documented in their control background strains (Figure 3G and H). We also showed significantly lower concentrations of NKT and γδ T cells in the early but not later stages of tumorigenesis in hypercholesterolemic mice than that in WT mice (Figure 3I and J). These findings show that hypercholesterolemia reduced the concentrations of NKT and γδ T cells, which impaired tumor immunosurveillance against colorectal neoplasia [13].

2.4 The incidence of colorectal neoplasia is a linear function of hypercholesterolemia-induced HSC oxidant stress

Hypercholesterolemia induced a receptor-mediated uptake of oxidized-LDL that induced HSC oxidant stress that profoundly affected normal HSC function, accelerating their aging and reducing their repopulation capacity [19]. These effects were reduced by the administration of N-acetylcysteine [NAC]. Therefore, we hypothesized that NAC would also reverse the effects of hypercholesterolemia on NKT and γδ T cell number and cancer immunosurveillance. In pursuit of this possibility, we confirmed in hypercholesterolemic mice that NAC also rescued the impaired differentiation of HSCs toward NKT and γδ T cells (Figure 5A and B). In addition, NAC significantly decreased the average tumor number in both ApoE−/− and HCD mice. While NAC reduced the histopathologic severity of tumors in HCD mice, the reduction in the histopathologic severity of tumors in ApoE−/− mice did not reach statistical significance (Figure 5C and D). In both ApoE−/− and HCD mice, NAC also increased the infiltration of NKT and γδ T cells in the early stages of tumorigenesis (Figure 5E and F). Linear regression analysis between the level of HSC oxidant stress and the number of tumors per mouse revealed an incredible linear correlation between these variables [r² = 0.87] (Figure 5G). These findings reveal that hypercholesterolemia-induced hematopoietic stem cell oxidant stress directly mediates the reduction of HSC production of NKT and γδ T cells, and this reduction in NKT and γδ T cells in tumor number and their histopathological severity [13].

Figure 5.

Hypercholesterolemia-induced oxidant stress in HSCs strongly correlates with the incidence of AOM-induced colorectal neoplasia. A. Frequency and total number of NKT cells in thymus of ApoE-/- and N-Acetyl Cysteine (NAC) treated ApoE-/- mice. n=8, *, p<0.05, vs. ApoE-/-. NAC was given in drinking water for 8 weeks (150mg/kg/day). B. Frequency and total number of γδT cells in thymus of ApoE-/- and NAC treated ApoE-/- mice. n=8, *, p<0.05, vs. ApoE-/-. C. Average tumor number and histopathologic stages of tumors isolated from ApoE-/- and NAC treated ApoE-/- mice. n=12, *, p<0.05, vs. ApoE-/-. D. Average tumor number and histopathologic stages of tumors isolated from High Cholesterol Diet (HCD) and NAC treated HCD mice. n=12, *, p<0.05, vs. HCD. E. NKT cell infiltration in the early stages of tumorigenesis in ApoE-/- and NAC treated ApoE-/- mice. n=6, *, p<0.05, vs. ApoE-/- . F. γδ T cell infiltration in the early stages of tumorigenesis in ApoE-/- and NAC treated ApoE-/- mice. n=12, *, p<0.05, vs. ApoE-/-. G. Regression analysis between oxidant stress in HSCs and tumor number [13].

2.5 Hypercholesterolemia-induced downregulation of Tet1 in HSCs impairs their differentiation toward NKT and γδ T cells

Since the effects of hypercholesterolemia on NKT and γδ T cells are cell-autonomous and sustained after transplantation into WT mice who have normal cholesterol levels, there are at least two possible molecular mechanisms: oxidative stress-induced DNA mutations or oxidative stress-induced changes in epigenetic enzyme expression. We found oxidant stress-dependent reduction in the expression of Tet1 in HSCs from hypercholesterolemic mice (Figure 6A and B). Tet2 has been shown to regulate key HSC functions, including self-renewal, proliferation, and hematopoiesis [22, 23, 24]. However, the role of tet1 in these key HSC functions is unknown. In order to determine if Tet1 plays a direct role in the reduced production of NKT and γδ T cells and the consequent impairment of tumor immunosurveillance, we determined the proportion and number of NKT and γδ T cells in the thymus of WT and Tet1−/− mice. As we found in hypercholesterolemic mice, the proportion and number of NKT and γδ T cells in the thymus of Tet1−/− mice was significantly lower than that of WT mice (Figure 6C and D). Also consistent with our findings in hypercholesterolemic mice, the proportion and number of NKT and γδ T cells in the colon submucosa of Tet1−/− mice was substantially lower than that in WT mice (Figure 6E and F). We did not find any changes in the peripheral blood of Tet1−/− mice CD3e+. CD4+, or CD8+ cells. We found a slight decrease in B cells and NK cells in Tet1−/− mice (Figure 7A and B).

Figure 6.

Hypercholesterolemia-induced oxidant stress in HSCs positively correlates with AOM-induced colorectal neoplasia. A. Expression of Tet1, Tet2, and Tet3 in HSCs from WT and ApoE−/− mice; n = 6, **, P < 0.01, versus WT. B. Downregulation of Tet1 expression in HSCs from ApoE−/− is oxidant stress dependent in mice; n = 6, *P < 0.05; **, P < 0.01, versus ApoE−/−. C. Frequency and number of NKT cells in thymus of WT and Tet1−/− mice; n = 5. *, P < 0.05, versus WT. D. Frequency and number of γδT cells in thymus of WT and Tet1−/− mice; n = 5. *, P < 0.05, versus WT. E. Frequency of submucosal NKT cells in colon of WT and Tet1−/− mice; n = 5, *, P < 0.05, versus WT. F. Frequency of submucosal γδT cells in colon of WT and Tet1−/− mice. n = 5; *, P < 0.05, versus WT. G. Tet1-relative expression following its overexpression in WT and Apoe−/− HSCs. n = 6; *, P < 0.05, versus WT; ##, P < 0.01, versus ApoE−/−. H. Frequency of NKT cells in thymus of recipient mice transplanted with WT HSCs, ApoE−/− HSCs, Tet1-overexpressing WT HSCs, or Tet1-overexpressing ApoE−/− HSCs. n = 6; *, P < 0.05; **, P < 0.01, versus WT+Mock; ##, P < 0.01, versus ApoE−/−+Mock. I. Frequency of γδT cells in thymus of recipient mice transplanted with WT HSCs, ApoE−/− HSCs, Tet1-overexpressing WT HSCs, or Tet1-overexpressing ApoE−/− HSCs. n = 6; *, P < 0.05; **, P < 0.01, versus WT+Mock; ##, P < 0.01, versus ApoE−/−+Mock [13].

Figure 7.

Frequency of T cell intermediate populations in thymus and cellular components responsible for cancer immunosurveillance in peripheral blood and colon of WT and Tet-/- mice. A. Frequency of B cells, NK cells, CD3e+, CD4+ and CD8+ cells in peripheral blood of WT and Tet1-/- mice. n = 6, *, p<0.05, vs. WT. B. NK cells (CD45+CD3e-NKp46+), CD11b- dendritic cells (CD11c+CD11b-CD103+F4/80-), CD11b+ dendritic cells (CD11c+CD11b+CD103+F4/80-), CD11c- macrophages (CD11c-CD11b+CD103-F4/80+) and CD11c+ macrophages (CD11c+CD11b+CD103-F4/80+) in the colon of WT and Tet1-/- mice. n = 6, *, p<0.05, vs. WT [13].

In contrast to these findings to these findings in Tet1 knockout mice, we found that overexpression of Tet1in HSCs from WT and hypercholesterolemic mice resulted in a 7-fold increase of NKT cells in WT mice and an almost 20-fold increase in NKT cells hypercholesterolemic mice, in vivo and in vitro. In parallel finding, overexpression of Tet1in HSCs from WT and hypercholesterolemic mice resulted in a 10-fold increase in WT and a 20-fold increase in γδ T cells (Figures 6G–I and 8A, D and E). These results support the novel and specific role of Tet1 in HSCs lineage specification toward NKT and γδ T cells.

Figure 8.

The expression of Tet1 determines the differentiation of HSCs towards NKT and γδT cells in vitro. A Downregulation of Tet1 by shRNA in HSCs . n=6, *, p<0.05; **, p<0.01, vs. WT; #, p<0.05, vs. ApoE-/-. B Differentiation of HSCs towards NKT cells in vitro following Tet1 downregulation C. Differentiation of HSCs towards γδT cells in vitro following Tet1 downregulation. n=6, *, p<0.05; **, p<0.01, vs. WT; #, p<0.05, vs. ApoE-/-. D. Differentiation of HSCs towards NKT cells in vitro following Tet1 overexpression E. Differentiation of HSCs towards γδT cells in vitro following Tet1 overexpression. n=6, *, p<0.05; **, p<0.01, vs. WT; #, p<0.05, vs. ApoE-/- [13].

Given these effects of overexpression of Tet1 on the production of γδ T cells we sought to test the hypothesis that overexpression of Tet1 in hypercholesterolemic mice would restore the production of NKT and γδ T cells and thereby immunosurveillance against colorectal neoplasia. To this end, we reconstituted lethally irradiated WT mice with HSCs from WT and ApoE−/− mice that overexpress Tet1. Unexpectedly, lethally irradiated mice reconstituted HSCs that overexpressed Tet1, and all died. To address this problem, we included 1/3 non-transduced HSCs. Under these conditions, all mice survived, indicating clearly that Tet1 has an important role in HSC engraftment after irradiation. Overexpression of Tet1 in the HSCs of hypercholesterolemic mice restored the concentration of NKT and γδ T cells in the thymus and colon submucosa to that in WT mice (Figure 9A–E). Perhaps our most important finding is, overexpression of Tet1 in HSCs from hypercholesterolemic mice reduced the number of colorectal tumors by 55%, similar to that in WT mice (Figure 9F–G). Overexpression of Tet1 in HSCs of hypercholesterolemic also greatly reduced the histopathological severity of the colorectal neoplasia (Figure 9G). In WT and hypercholesterolemic mice, overexpression of Tet1 eliminated the progression of tumors to the carcinoma stage.

Figure 9.

Hypercholesterolemia induced oxidant stress downregulates the expression of Tet1 in HSCs that impairs their differentiation towards NKT and γδT cells. A. Frequency of cells derived from Tet1-overexpressing HSCs. The transplantation of Tet1-overexpressing WT HSCs was supported with WT HSCs and the transplantation of Tet1-overexpressing ApoE−/− HSCs was supported with ApoE−/− HSCs, both at the ratio of 3:1. n = 8; *, P < 0.05; **, P < 0.01, versus WT+Mock; ##, P < 0.01, versus ApoE−/−+Mock. B. Frequency and total number of NKT cells in thymus of the recipients after transplantation with WT HSCs, ApoE−/− HSCs, Tet1-overexpressing WT HSCs+WT HSCs, or Tet1-overexpressing ApoE−/− HSCs+ApoE−/− HSCs. n = 8; *, P < 0.05, versus WT+Mock→WT; #, P < 0.05, versus ApoE−/−+Mock→WT. C. Frequency and total number of γδT cells in thymus of the recipients. n = 8; *, P < 0.05, versus WT+Mock→WT; #, P < 0.05, versus ApoE−/−+Mock→WT. D. Frequency of NKT cells in colon submucosa of the recipients. n = 8; **, P < 0.01, versus WT+Mock→WT; #, P < 0.05, versus ApoE−/−+Mock→WT. E. Frequency of γδT cells in colon submucosa of the recipients. n = 8; **, P < 0.01, versus WT+Mock→WT; #, P < 0.05, versus ApoE−/−+Mock→WT. F. Average tumor number per mouse in the recipients. n = 12; *, P < 0.05, versus WT+Mock→WT; #, P < 0.05, versus ApoE−/−+Mock→WT. G. Histopathologic stages of tumors. n = 12; *, P < 0.05; **, P < 0.01 versus WT+Mock→WT; #, P < 0.05; ##, P < 0.01, versus ApoE−/−+Mock→WT [13].

These results show that the mechanism by which hypercholesterolemia increases the risk of colorectal neoplasia is by inducing a Tet1-dependent HSC-autonomous mechanism that epigenetically reprograms the number and gene expression of NKT and γδ T cells. Given the unexpected simplicity of this finding, it could be leveraged into the creation of a cell immunotherapy for a variety of cancers.

2.6 MiR101c mediates the downregulation of Tet1 in HSCs isolated from hypercholesterolemic mice

We have shown that hypercholesterolemia induces an oxidized-LDL-dependent increase in HSC oxidant stress that initiates a signaling pathway culminating in the reduction to Tet1. How is Tet1 directly regulated? To address this question, we performed miRNA microarray analysis in HSCs isolated from WT and ApoE−/− mice (Figure 10A). MiR101c, predicted to target Tet1 directly, was upregulated significantly in HSCs from ApoE−/− mice. This increased level of miR101c was validated by RT-PCR (Figures 10B and 11A). Application of NAC effectively reduced the overexpression of miR101c in HSCs from ApoE−/− mice (Figures 10B and 11A). Overexpression of miR101c in HSCs from WT and hypercholesterolemic mice greatly reduced Tet1 expression (Figure 11B and C). Transfection of an inhibitor of miR101c greatly increased Tet1 expression (Figure 11D and E). In a luciferase assay of the Tet1 3′-UTR, miR101c reduced luciferase activity, whereas when the Tet1 binding sites were blocked, miR101c failed to increase luciferase activity (Figure 11F) These findings indicate that miR101c directly binds to the 3′-UTR of Tet1.

Figure 10.

Expression of miRNA in HSCs from hypercholesterolemic mice. A. Microarray profiling analysis in HSCs from WT and ApoE-/- mice. n=4. B. Oxidant stress dependent upregulation of miR-101c expression in HSCs from HCD fed mice. n=4, *, p<0.05, vs WT control; #, p<0.05, vs HCD control [13].

Figure 11.

Reconstitution of lethally irradiated WT mice with ApoE-/- HSCs that overexpress Tet1 restores immunosurveillance against colorectal neoplasia. A. The frequency and total number of NKT cells in thymus and blood of the recipients after transplantation with WT HSCs, ApoE-/- HSCs, Tet1-overexpressing WT HSCs+WT HSCs, or Tet1-overexpressing ApoE-/- HSCs+ApoE-/- HSCs. n=8, *, p<0.05, vs. WT→WT; #, p<0.05, vs. ApoE-/-→WT. B. The frequency and total number of γδT cells in thymus and blood of the recipients. n=8, *, p<0.05, vs. WT→WT; #, p<0.05, vs. ApoE-/-→WT. C. The frequency of NKT cells in colon submucosa of the recipients. n=8, **, p<0.01, vs. WT→WT; #, p<0.05, vs. ApoE-/-→WT. D. The frequency of γδT cells in colon submucosa of the recipients. n=8, **, p<0.01, vs. WT→WT; #, p<0.05, vs. ApoE-/-→WT. E. Average tumor number per mouse in the recipients. n=12, *, p<0.05, vs. WT→WT; #, p<0.05, vs. ApoE-/-→WT. F. Histopathologic stages of tumors. n=12, *, p<0.05, **, p<0.01 vs. WT→WT; #, p<0.05, ##, p<0.01, vs. ApoE-/-→WT [13].

2.7 Tet1 directly induces the expression of genes critical for HSC differentiation toward NKT and γδ T cells

While the detailed mechanism by which HSCs produce NKT and γδ T cells remains incompletely described, a group of genes has been shown to be critical to this process [25, 26]. We sought to characterize the effects of Tet1 on the expression and epigenetic regulation of these genes that are required for HSC production of NKT and γδ T cells (Table 1). Of these genes, five, Fyn, Sox13, IL15R, ITK, and SH2D1a, had lower expression in ApoE−/− HSCs than in WT HSCs (Figure 12A). When Tet1 was overexpressed in HSCs from WT and ApoE−/− mice, their expression increased substantially (Figure 12A). Because Tet1 regulates the methylation of genes [23, 27], we proceeded to use pyrosequencing to characterize the effects of Tet1 on the methylation at key regulatory regions of these five genes that were affected by hypercholesterolemia. As anticipated, we found hypermethylation of Fyn, Sox13, Il15R, IKT and SH2D1a in HSCs from hypercholesterolemic mice (Figure 12B).

Genes related to NKT cell differentiation	Genes related to γ8 T cell differentiation
Interleukin-2 receptor β (IL-2Rb)	B-cell lymphoma/leukemia 11B (BCL11b)
Interleukin-15 receptor (IL-15R)	Early growth response protein 2 (EGR2)
E26 Transformation specific transcription factor 1 (Ets1)	Ets variant 5 (ETV5)
Myeloid Elf-1-like factor (MEF)	Inhibitor of DNA binding protein 2 (ID2)
Interferon regulatory factor 1 (IRF-1)	Inhibitor of DNA binding protein 3 (ID3)
Fyn	Interleukin-2-inducible T-cell kinase (ITK)
Interleukin-2-inducible T-cell kinase (Itk)	Interleukin 7 receptor (IL-7R)
Activator protein-1 (AP-1)	Interleukine-15 receptor (IL-15R)
T cell factor 1 (TCF-1)	PHD finger protein 1 (PHF1)
Nuclear factor κB p50 (NFκb)	SLAM-Associated Protein (SAP, SH2D1a)
RELb	Sry-related HMG box 13 (Sox13)
IκB kinase 2 (IKK2)	T cell factor 12 (TCF12)
Protein kinase C-θ (PKCθ)	Zinc finger and BTB domain-containing protein 16 (ZBTB16)
Signaling lymphocytic activation molecule F1 (SLAMF1)
Signaling lymphocytic activation molecule-associated protein (SAP)
Krüppel-like factor 2 (KLF2)
CCR9

Table 1.

The genes related to the differentiation of NKT and γ8 T cells. The genes highlighted with green color showed significant changes in our study.

Figure 12.

Tet1 regulates the expression of the key regulatory genes in the differentiation of NKT and γδT cells. A. The expression of genes in cells from in vitro co-culture with WT HSCs, ApoE-/- HSCs, Tet1 overexpressing WT HSCs or Tet1 overexpressing ApoE-/- HSCs. n=4, *, p<0.05, **, p<0.01, vs. WT+mock; #, p<0.05, ##, p<0.01, vs. ApoE-/-+mock. B. DNA methylation status of the targeted gene. n=4, *, p<0.05, **, p<0.01, vs. WT+mock; #, p<0.05, ##, p<0.01, vs. ApoE-/-+mock [13].

As expected, the overexpression of Tet1 in HSCs from WT and ApoE−/− mice decreased the methylation and increased the expression of these five genes (Figures 12B and 13B). In addition to these five genes, an additional five genes, ETV5, EGR2, SLAMF1, ZBTB16, and RELB, whose expression was unaffected in HSCs from ApoE−/− mice, nonetheless their expression increased after Tet1 overexpression (Figure 13A). Both NKT and γδ T are being evaluated intensely as cell immunotherapies for a variety of cancers. This opens the possibility of a combined cell therapy directed by Tet1 overexpression in HSCs.

Figure 13.

In HSCs, Tet1 overexpression reduces the expression of the key regulatory genes in their differentiation of NKT and γδT cells. A. Gene expression analysis in WT HSCs, ApoE-/- HSCs, Tet1 overexpressing WT HSCs and Tet1 overexpressing ApoE-/- HSCs. n=4, *, p<0.05, **, p<0.01, vs. WT+mock; #, p<0.05, ##, p<0.01, vs. ApoE-/-+mock. B. DNA methylation status of the genes analyzed in A. n=4, *, p<0.05, **, p<0.01, vs. WT+mock; #, p<0.05, ##, p<0.01, vs. ApoE-/-+mock [13].

The methylation of ETV5, EGR2, and NFKB1was higher in HSCs derived from ApoE−/− mice than in HSCs from WT mice, but overexpression of Tet1 reduced this methylation at or below that in WT HSCs (Figure 13B) Taken together, all of these results show that Tet1regulates the expression of multiple genes that are essential to produce NKT and γδ T cells and this regulation is disrupted by hypercholesterolemia.

2.8 Tet1 regulates histone modifications by the O-linked N-acetylglucosamine transferase (OGT)

O-linked N-acetylglucosamine transferase is an evolutionarily conserved enzyme whose primary function is to catalyze)-linked protein glycosylation. Tet2 and Tet3 have been shown to act as stable partners if OGT in the nucleus [28, 29, 30]. Their interaction with OGT induces GlcNAcylation of Host Cell Factor-1 that enhances the H3K4 methyltransferase SET1/COMPASS complex, indicating that Tet enzymes increase H3K4me3 modifications that result in transcriptional activation [31]. Our immunoprecipitation results establish that OGT interacts with Tet1 in HSCs (Figure 14A). This Tet1-OGT interaction was significantly less in HSCs from hypercholesterolemic mice, consistent with the decrease in Tet1 expression. Overexpression of Tet1 enhanced the Tet1-OGT interaction but had no effect on the expression or interaction of Tet3 and OGT (Figure 14A and B). As anticipated by these findings, overexpression increased H3K4me3-induced methylation at the promoters of all of the genes evaluated except RELb and NFKB1. These findings indicate that Tet1’s interaction with OGT increases H3K4me3 levels that help maintain an active chromatin structure near many of the genes required to produce NKT and γδ T cells by HSCs (Figure 14C). These results show that Tet1 regulates the expression of multiple genes required for NKT and γδ T cells by multiple mechanisms.

Figure 14.

Tet1 regulates H3K4me3 modification of the key regulatory genes in the differentiation of HSCs towards NKT and γδT cells in vitro. A. Detection of the expression of Tet1 and its association with OGT. Immunoprecipitation was performed with WT HSCs, ApoE-/- HSCs, Tet1 overexpressing WT HSCs or Tet1 overexpressing ApoE-/- HSCs. B. Detection of the expression of Tet3 and its association with OGT. C. H3K4me3 modification of the key regulatory genes in the differentiation of HSCs towards NKT and γδT cells in vitro. n=4, *, p<0.05, **, p<0.01, vs. WT+mock; #, p<0.05, ##, p<0.01, vs. ApoE-/-+mock [13].

2.9 Hypercholesterolemia reduces Tet1 expression in human HSCs that impairs their production of NKT and γδ T cells

To determine if the effects of hypercholesterolemia on murine HSCs can be extrapolated to human HSCs, we exposed human HSCs to oxidized-LDL, the source of oxidant stress in HSCs from hypercholesterolemic mice [19]. Oxidized-LDL induced a concentration-dependent decrease in the expression of Tet1 (Figure 15C). This reduction in Tet1 expression reduced the differentiation of human HSCs toward NKT and γδ T cells (Figure 15A and B). These preliminary results suggest that the extensive findings from these studies are generalizable to humans.

Figure 15.

Ox-LDL impairs the differentiation of human HSCs towards NKT and γδT cells in vitro. A. Differentiation of human HSCs towards iNKT cells in vitro. B. Differentiation of human HSCs towards γδT cells in vitro. C. Relative expression of Tet1 in human HSCs treated with oxLDL. n=3, *, p<0.05, vs. control [13].

3. Conclusions

For those under 50 years of age, CRC is the leading cause of cancer death in men and the second leading cause of death in women. More than 50% of cases are due to lifestyle risk factors. Obesity will soon replace tobacco abuse as the most preventable cause of cancer and is a significant risk for CRC. However, obesity is associated with multiple metabolic abnormalities, including hypercholesterolemia, which is a well-defined risk factor for CRC. We have shown that hypercholesterolemia impairs immunosurveillance against colorectal neoplasia by reducing the number and function of NKT and γδ T cells. In support of this role for NKT and γδ T cells in CRC immunosurveillance, mice lacking NKT T cells have very substantial increases in colorectal neoplasia.

At a molecular level, hypercholesterolemia induces an oxidized-LDL increase in HSC oxidant stress that initiates a signaling pathway culminating in the downregulation of Tet1whihc is directly regulated by miR101c. These effects of hypercholesterolemia on CRC are not due to a direct effect on NKT and γδ T cells, as is universally held. Rather, hypercholesterolemia increases the incidence of CRC by a Tet-1dependent HSC-autonomous mechanism that epigenetically reprograms their gene expression within the HSC. This reduction in gene expression is achieved by bivalent regulation comprised of a gain of repressive DNA methylation and loss of activating H3K4me3. Thus, we show for the first time that HSCs not only produce immune cells, but they can also regulate their gene expression by oxidant stress-dependent epigenetic reprogramming.

Tet1 is a critical tumor suppressor in multiple human cancers, including colorectal cancer [32]. The analysis of tumor methylomes of tumor cell lines and primary tumors of multiple carcinomas and lymphomas, including gastric and colorectal carcinomas, showed that Tet1 is frequently methylated and consequently downregulated [33, 34, 35]. The overexpression of the Tet1 catalytic domain can significantly reduce the methylation of tumor suppressor genes and thereby restore their expression [34]. Reduced Tet1 expression can repress the expression of the DKK gene and the activation of the WNT pathway, which increases tumorigenesis in the colon. In contrast, restoring Tet1 expression in colon carcinoma cells inhibits their proliferation as well as tumor xenografts [35]. All of these studies strongly indicate that Tet1 plays a critical role in the transformation of colon cells. So, in addition to these findings, we show that Tet1 regulates the production of NKT and γδ T cells that is central to immunosurveillance against colorectal neoplasia.

Are these findings generalizable? We have also shown that type 2 diabetes mellitus (T2DM) impairs wound healing by an HSC-autonomous mechanism [36]. In parallel to our findings with hypercholesterolemia, T2DM induced a Nox-2-dependent increase in HSC oxidant stress that initiated a signaling pathway that culminated in the increased expression of Dnmt1 by reduced expression of let-7-3p. These changes reduced the expression of Klf4, Notch-1and Pu.1, genes critical to the production of monocytes and macrophages. Consequently, across the first two stages of wound healing, T2DM reduced the number of macrophages and increased their polarization toward the M-1 phenotype. Reconstituting T2DM mice with HSCs from T2DM in which an shRNA against Dnmt1 restored a normal rate of wound healing. Thus, this may be a framework to identify the mechanism of the other risk factors for young-onset CRC may be evaluated. Each risk factor induces a risk factor-specific HSC oxidant stress that initiates a signaling pathway culminating in the changed expression of an epigenetic enzyme(s) that is directly regulated by a miRNA. In this way, risk factors affect the gene expression by inducing redox-dependent epigenetic reprogramming of their gene expression within the hematopoietic stem cell.