Aggressive cholesterol lowering normalizes atherosclerosis regression in Jak2^V617F mice

Abstract

The Jak2^V617F (Jak2^VF) mutation is an important cause of both clonal hematopoiesis of indeterminate potential (CHIP) and myeloproliferative neoplasms (MPNs). Mouse models of Jak2^VF CHIP and MPN show accelerated atherosclerosis progression, driven by macrophage inflammasome activation. We undertook the present study to assess the hypothesis that ongoing inflammation would impede atherosclerosis resolution in Jak2^VF mice. Chimeric Jak2^VF/WT or control WT/WT bone marrow was transplanted into Ldlr^−/− mice and, following 13–16 weeks of western diet-induced atherosclerosis progression, cholesterol was lowered either moderately (to 200–300 mg/dl) or markedly (to 100 mg/dl). With moderate cholesterol lowering, there was impaired resolution of lesions in Jak2^VF MPN mice compared to controls. However, with marked cholesterol lowering, progression of lesions was halted in both Jak2^VF MPN and control mice while macrophage burden was decreased and lesional collagen was increased similarly in Jak2^VF MPN and control mice. Two mechanisms of low-density lipoprotein (LDL) lowering-induced suppression of inflammation in plaques were implicated: 1) reversal of increased proliferation, DNA damage and absent in melanoma 2 (AIM2) inflammasome activation specifically in Jak2^VF macrophages and 2) markedly increased macrophage triggering receptor expressed on myeloid cells 2 (TREM2), c-Myc expressing macrophages in both Jak2^VF and control mice. In summary, aggressive LDL lowering reverses inflammasome activation and induces pro-resolving changes in macrophages in Jak2^VF MPN, halting atherosclerosis progression and promoting features of plaque stabilization. These findings suggest that aggressive LDL cholesterol lowering could reverse atherosclerotic cardiovascular disease risk in individuals with JAK2^VF CHIP or MPN.

Graphical abstract

Despite therapeutic advances, atherosclerotic cardiovascular disease (ACVD) remains the leading cause of death in the developed world (1). Although inflammation plays a key role in atherosclerosis progression, cholesterol accumulation remains a prerequisite for lesion development, and low-density lipoprotein (LDL) lowering has consistently shown a benefit in clinical trials (2, 3, 4).

The role of inflammation in increasing ACVD risk has also been well established. The LoDoCo 2 trial found that the risk of cardiovascular events was reduced in patients with chronic coronary disease treated with colchicine (5). Based on these and other results, the US Food and Drug Administration approved the use of low dose colchicine as the first anti-inflammatory drug for heart disease (6). The CANTOS trial involving the use of interleukin-1β (IL-1β) antibodies showed reduced cardiovascular events in high-risk patients with mean LDL cholesterol levels of 81 mg/dl-84 mg/dl and a high-sensitivity C-reactive protein level of at least 2 mg/L. However, there was a small increase in fatal infections, and the Food and Drug Administration (FDA) did not approve this treatment for marketing (7). Many observations suggest that hypercholesterolemia is linked to inflammatory risk. For example, statins lower high-sensitivity C-reactive protein levels (8) and preclinical studies have shown that cholesterol accumulation in plaque macrophages promotes inflammasome activation (9, 10, 11, 12). However, the mechanisms linking hypercholesterolemia, macrophage cholesterol accumulation, and inflammation remain incompletely understood, especially in the clinically important context of atherosclerosis regression induced by LDL lowering (13, 14, 15, 16, 17, 18, 19).

Clonal hematopoiesis (CH) arises from leukemogenic mutations in hematopoietic stem cells that confer a selective advantage and lead to clonal expansion of blood cells. CH commonly arises from mutations in genes that mediate epigenetic modifications or cytokine signaling (20, 21). Clonal hematopoiesis of indeterminate potential (CHIP) is defined as having a mutant allele burden >2% in blood cells in the absence of changes associated with hematopoietic malignancy (20). Among CHIP mutations, JAK2^VF, which increases signaling by hematopoietic cytokines, confers the greatest risk of ACVD (22, 23, 24). Moreover, JAK2^VF has been detected in 3%–4% of a general European population (25). JAK2^VF CHIP and myeloproliferative neoplasms (MPNs) appear to form a disease spectrum, with both increasing atherosclerotic and thrombotic risk (26, 27). Similarly, both Jak2^VF MPN and low allele burden Jak2^VF CHIP in mice that lack changes in blood cell counts or splenomegaly increase thrombosis and atherosclerosis progression involving leukocyte-mediated inflammatory mechanisms (28, 29). In addition, we showed in a hypercholesterolemic mouse model that Jak2^VF CHIP or MPN exacerbates atherosclerosis via inflammasome activation (30) and genetic studies suggest a similar role in humans (31). Based on these findings, we hypothesized that LDL lowering-induced changes in plaques such as halted progression and beneficial remodeling would be impaired in Jak2^VF mice due to persistence of the underlying mechanisms promoting inflammasome activation. Contrary to our hypothesis, we found that cholesterol lowering reversed inflammasome associated changes in Jak2^VF macrophages and that marked cholesterol lowering induced beneficial lesional changes similarly in Jak2^VF MPN mice and controls. These studies suggest the importance of rigorous control of LDL cholesterol levels in individuals with JAK2^VF CHIP or MPN.

Materials and methods

Anesthesia and euthanasia

Mice were anesthetized using 5% isoflurane via inhalation at a flow rate of approximately 3 L/min oxygen using a calibrated vaporizer. Anesthetic depth was assessed using respiratory rate monitoring and the toe-pinch reflex. Isoflurane was administered acutely during procedures requiring anesthesia, including blood collection and euthanasia. Blood collection was performed via retro-orbital sinus for plasma, complete blood counts, and flow cytometric analysis at the indicated time points. For euthanasia, mice were first rendered deeply unconscious with isoflurane, after which cervical dislocation was performed as a secondary method to ensure death, in accordance with institutional IACUC approval, NIH guidelines, and the AVMA Guidelines for the Euthanasia of Animals (2020 Edition). Animals meeting humane end point criteria (e.g., hind limb paralysis/moribund condition) were excluded from end point histological analyses.

Mice

All mice used for these studies were male and on a C57BL/6J background. Mice were housed in a pathogen-free facility under standard conditions of temperature (23°C) with a 12-h light-dark with ad libitum food and water access. Cages and water were changed every 7–14 days. All mouse experiments were approved by the Institutional Animal Care and Use Committee of Columbia University under protocol number AABG0561 and were conducted in accordance with the NIH guidelines and the Guide for the Care and Use of Laboratory Animals. Mice heterozygous for the Jak2^VF conditional knock-in allele were generated as previously described (32). Jak2^VF mice were crossed to Mx1-Cre (The Jackson Laboratory, B6.Cg-Tg(Mx1-Cre)1Cgn/J (003556) or Scl-Cre (The Jackson Laboratory, C57BL/6-Tg(Tal1-cre/ERT)42-056Jrg/J (037466) to generate Mx1-CreJak2^VF and Scl-CreJak2^VF mice, respectively. CD45.1Mx1-CreJak2^VF mice were generated by crossing Mx1-CreJak2^VF mice to CD45.1 (The Jackson Laboratory B6.SJL-Ptprca Pepcb/BoyJ (002014)) mice. CD45.1/CD45.2Scl-CreJak2^VF mice were generated by crossing Scl-CreJak2^VF mice to CD45.1 (The Jackson Laboratory B6.SJL-Ptprca Pepcb/BoyJ (002014)) mice. CD45.1/CD45.2Scl-Cre-Jak2^VF mice were crossed with ZsGreen reporter mice (The Jackson Laboratory, B6.Cg-Gt(ROSA)26Sor^{tm6(CAG-ZsGreen1)Hze}/J (Ai6(RCL-ZsGreen) (007906) to generate CD45.1/CD45.2Scl-CreJak2^VFZsGreen mice. Ldlr^−/− recipient mice for the Mx1-CreJak2^VF moderate cholesterol lowering model that included Myh11-Cre^ERT2 and ZsGreen alleles from a previous lineage-tracing study were obtained from a collaborator and were previously described (33). However, these alleles were not relevant to the experimental outcomes and were not analyzed. Half of recipients were Ldlr^−/− Myh-11-CreZsGreen while the other half either lacked the ZsGreen allele or both the ZsGreen allele and the Myh11-Cre. For the Mx1-CreJak2^VF and Scl-CreJak2^VF aggressive cholesterol lowering studies, Ldlr^−/− recipient mice were purchased from The Jackson Laboratory (B6.129S7-Ldlrtm1Her/J (002207)). Littermate control mice for studies involving Mx1-CreJak2^VF mice contained either an allele with Mx1-Cre or the Jak2^VF transgene, but not both. Littermate control mice for studies involving Scl-CreJak2^VFZsGreen mice contained alleles for Scl-Cre and ZsGreen but not the Jak2^VF transgene. Jak2^VFAim2^−/− mice were generated by crossing Mx1-CreJak2^VF mice to Aim2^−/− mice (The Jackson Laboratory, B6.129P2-Aim2Gt(CSG445)Byg/J) (13144)).

Bone marrow transplantation

Bone marrow transplantations (BMTs) were conducted as previously described with the following exceptions (10). For the Mx1-CreJak2^VF moderate cholesterol lowering study, donor mice were 8–12 weeks old while recipients were 11–14 weeks old. For the Mx1-CreJak2^VF aggressive cholesterol lowering study, donor mice were 7–16 weeks old while recipients were 7 weeks old. For the Scl-CreJak2^VF study, donor mice were 6–12 weeks old while recipients were 8 weeks old. For the Mx1-CreJak2^VF moderate cholesterol lowering study, recipients were lethally irradiated once with 10.5 Gy from a cesium gamma source. Due to decommissioning of the cesium gamma source, recipients from the Mx1-CreJak2^VF and Scl-CreJak2^VF aggressive cholesterol lowering studies were irradiated with 10.5 Gy using a Multirad 350 X-ray Irradiator from Precision X-Ray. Within 24 h of irradiation, bone marrow was isolated from donors with the indicated genotypes. Total bone marrow cell number was quantified using an INCYTO C-Chip hemocytometer (DHC-N01) according to the manufacturer's instructions. Irradiated mice were randomized to treatment groups, anesthetized with isoflurane before receiving 3 x 10⁶ total bone marrow cells via intravenous (i.v.) injection at a total final volume of 100 μl. For the Mx1-CreJak2^VF studies, recipients received either 6 x 10⁵ CD45.1Mx1-CreJak2^VF/WT or 6 x 10⁵ CD45.1 control bone marrow cells (20%) combined with 2.4 x 10⁶ CD45.2Jak2^WT/WT cells (80%). For the Scl-CreJak2^VF study, recipients received either 6 x 10⁵ CD45.1CD45.2Scl-CreJak2^VF/WTZsGreen or 6 x 10⁵ CD45.1CD45.2Scl-CreJak2^WT/WTZsGreen bone marrow cells (20%) combined with 2.4 x 10⁶ CD45.1Jak2^WT/WT cells (80%). For the Mx1-CreJak2^VF studies, mice were allowed to recover for four weeks after BMT before being injected intraperitonially (i.p.) with 50 μg/mouse/day polyinosinic:polycytidylic acid (pIpC) two times, 48 h apart.

Atherosclerosis studies

For the Mx1-CreJak2^VF studies, power analysis using GPower3.1 indicated that with a total sample size of 80 (n = 20 per group), the study had 80% power to detect an effect size of f = 0.32 at α = 0.05, using two-way analysis of variance (ANOVA) across two genotypes and two treatments. For the Scl-CreJak2^VF study, power analysis using GPower3.1 indicated that with a total sample size of 80 (n = 16 per group), the study had 80% power to detect an effect size of f = 0.40 at α = 0.05, using one-way ANOVA across five groups. Four weeks after BMT, Ldlr^−/− recipient mice were fed a western diet (ENVIGO, cat. no. TD.88137) for the indicated times. For the final 10 days of the western diet in the Scl-CreJak2^VF study, mice were switched to a western diet containing tamoxifen at 500 mg/kg (ENVIGO, cat. no. TD130889). To induce LDL lowering in the Mx1-CreJak2^VF study of Fig. 1A, mice were administered 2 x 10¹¹ viral particles per mouse in 200 μl of 1X PBS via tail vein injection of a helper-dependent adenoviral vector (HDAd)-hLDLR (Gene Vector Core Laboratory at the Baylor College of Medicine) and switched to chow diet. To induce LDL lowering in the Mx1-CreJak2^VF study of Fig. 4A and the Scl-CreJak2^VF study of Fig. 6A, mice were administered 2.5 x 10¹¹ genome copies per mouse in 100 μl of 1X PBS via tail vein injection of an Adeno-Associated Virus serotype 8 vector (AAV8)-TBG-mLdlr vector (AAV-263355, Vector Biolabs) and switched to chow diet. Following euthanasia, aortic roots were collected and fixed with 4% paraformaldehyde (PFA) (Electron Microscopy Sciences 15710; 16% paraformaldehyde diluted 1:4 in 1.33X PBS) for 24 h at 4°C and then embedded in paraffin.

Figure 1.

Atherosclerosis regression is incomplete in Jak2^VF MPN mice with moderate cholesterol lowering. A: Study design created with BioRender.com. B: Plasma cholesterol (n = 34, 55, 55, 54, 16, 16 for Ctrl mice, n = 31, 55, 49, 39, 17, 16 for Jak2^VF mice, for weeks 0, 5, 11, 15, 17.5, 21 respectively). P = 0.028, <0.0001, <0.0001, <0.0001, 0.0005, <0.0001 (Ctrl vs. Jak2^VF at weeks 0, 5, 11, 15, 17.5, and 21, respectively). C: H&E images of aortic root lesions. Black lines, necrotic core. Scale bar, 200 μm. D: Lesion area, n = 15–23. P < 0.0001 (Ctrl Baseline vs. Jak2^VF Baseline; Ctrl LDL Lowering vs. Jak2^VF LDL Lowering), P = 0.052 (Jak2^VF Baseline vs. LDL Lowering). E: Necrotic core area, n = 15–23. P = 0.0079 (Ctrl Baseline vs. Jak2^VF Baseline), P = 0.0002 (Ctrl LDL Lowering vs. Jak2^VF LDL Lowering), P = 0.0003 (Jak2^VF Baseline vs. LDL Lowering). F: Picrosirius red-stained aortic root lesions. Scale bar, 200 μm. G: Collagen area as a percentage of lesion area, n = 15–24. P = 0.02 (Ctrl Baseline vs. LDL Lowering). H: Images of aortic root lesions stained for MAC2 (Green) and DAPI (Blue). Scale bar, 200 μm. I: Macrophage area, n = 15–24. P = 0.006 (Ctrl Baseline vs. LDL Lowering), P = 0.16 (Jak2^VF Baseline vs. LDL Lowering). All quantifications shown as mean ± s.e.m. Two-way ANOVA with the Geisser-Greenhouse correction for sphericity and Tukey’s multiple comparisons test (B). Two-way ANOVA with Tukey’s multiple comparisons test (D, E, G, and I). DAPI, 4′,6-diamidino-2-phenylindole; LDL, low-density lipoprotein; MPN, myeloproliferative neoplasm.

Figure 4.

Aggressive cholesterol lowering normalizes regression in Jak2^VF MPN lesions. A: Study design, created with BioRender.com. B: Plasma cholesterol (n = 5, 38, 19, 18, and 18 for Ctrl mice, n = 5, 37, 15, 16, and 15 for Jak2^VF mice, for weeks 0, 4, 13, 15, and 18, respectively). P = 0.0041 for genotype effect by two-way ANOVA with Geisser-Greenhouse correction, P = 0.011, 0.048, 0.056, and 0.0008 (Ctrl vs. Jak2^VF at weeks 0, 4, 13, and 15, respectively). C: H&E images of aortic root lesions. Black lines, necrotic core. Scale bar, 200 μm. D: Lesion area, n = 13–20. P < 0.0001 (Ctrl Baseline vs. Jak2^VF Baseline; Ctrl LDL Lowering vs. Jak2^VF LDL Lowering). Figures 1 and 4 represent independent regression cohorts performed separately; therefore, absolute lesion area should be compared within each cohort rather than between figures. E: Necrotic core area, n = 13–20. P = 0.0003 (Ctrl Baseline vs. Jak2^VF Baseline), P = 0.0006 (Ctrl LDL Lowering vs. Jak2^VF LDL Lowering), P = 0.90 (Jak2^VF Baseline vs. LDL Lowering). F: Picrosirius red-stained aortic root lesions. Scale bar, 200 μm. G: Collagen area as a percentage of lesion area, n = 13–20. P < 0.0001 (Ctrl Baseline vs. LDL Lowering), P = 0.0002 (Jak2^VF Baseline vs. LDL Lowering), P = 0.02 (Ctrl LDL Lowering vs. Jak2^VF LDL Lowering). H: Images of aortic root lesions stained for MAC2 (Green) and DAPI (Blue). Scale bar, 200 μm. I: Macrophage area, n = 13–20. P < 0.0001 (Ctrl and Jak2^VF Baseline vs. LDL Lowering). All quantifications shown as mean ± s.e.m. Two-way ANOVA with Tukey’s multiple comparisons test (D, G, and I). Two-way ANOVA with the Geisser-Greenhouse correction for sphericity and Tukey’s multiple comparisons test (B). Kruskal-Wallis test with Dunn’s multiple comparisons test (E). DAPI, 4′,6-diamidino-2-phenylindole; LDL, low-density lipoprotein; MPN, myeloproliferative neoplasm.

Figure 6.

Cholesterol lowering suppresses Jak2^VF macrophage proliferation and DNA damage. A: Study design created with BioRender.com. B: Plasma cholesterol (n = 79, 78, and 78 for Baseline mice for weeks 3, 11, and 15, respectively; n = 13, 14 for Ctrl Progression mice, n = 16, 16 for Ctrl LDL Lowering mice, n = 15, 15 for Jak2^VF Progression mice, n = 16, 16 for Jak2^VF LDL Lowering mice, for weeks 18 and 23, respectively). C: Images of aortic root lesions for MAC2 (Green), Cleaved GasD (Red), and DAPI (Blue). Scale bar, 60 μm. D: Log₁₀ transformed cleaved GasD mean fluorescence intensity (MFI) in lesions with the addition of constant 1, n = 14–16. P = 0.01 (Baseline vs. Jak2^VF Progression), P = 0.051 (Jak2^VF Progression vs. Jak2^VF LDL Lowering). E: Log₁₀ transformed cleaved GasD MFI in the necrotic core with the addition of constant 1, n = 14–16. P = 0.04 (Baseline vs. Jak2^VF Progression), P = 0.013 (Jak2^VF Progression vs. Jak2^VF LDL Lowering). F: Images of aortic root lesions stained for MAC2 (Green), Cleaved GasD (Red), and DAPI (Blue). Scale bar, 60 μm. White arrows, cleaved GasD + macrophages. G: Log₁₀ transformed cleaved GasD positive macrophages per section with the addition of constant 1, n = 14–16. H: Images of aortic root lesions for MAC2 (Red), Ki67 (White), and ZsGreen (Green). Scale bar, 60 μm. White arrows, macrophages double positive for Ki67 and ZsGreen. I: Macrophages positive for both Ki67 and ZsGreen per section, n = 13–15. P = 0.0019 (Baseline vs. Jak2^VF Progression), P = 0.0006 (Ctrl Progression vs. Jak2^VF Progression), P = 0.0003 (Jak2^VF Progression vs. Jak2^VF LDL Lowering). J: Macrophages positive for Ki67 but negative for ZsGreen per section, n = 13–16. K: Images of aortic root lesions stained for pγH2AX (Red), ZsGreen (Green), and DAPI (Blue). Scale bar, 60 μm. White arrows, pγH2AX positive cells. L: Log₁₀ transformed cells double positive for pγH2AX and ZsGreen in lesions with the addition of constant 1, n = 13–16. P = 0.0024 (Baseline vs. Jak2^VF Progression), P = 0.0019 (Jak2^VF Progression vs. Jak2^VF LDL Lowering). M: Log₁₀ transformed cells positive for pγH2AX but negative for ZsGreen in lesions with the addition of constant 1, n = 13–16. All quantifications shown as mean ± s.e.m. One-way ANOVA with Holm–Sidak’s multiple comparisons test (D and E). Kruskal–Wallis test with Dunn’s multiple comparison’s test (G, I, J, L, and M). DAPI, 4′,6-diamidino-2-phenylindole; GasD, gasdermin D; LDL, low-density lipoprotein; pγH2AX, phosphorylated histone H2A.X.

Histological analysis

Lesion area and necrotic core area were quantified as previously described (22, 30). Blinded researchers sectioned paraffin-embedded aortic root lesions at similar points between mice over a 150 μm span for a total of 50 sections spaced 3 μm apart with 2 sections per slide on 25 slides. Six total slides spaced 30 μm apart were then stained with hematoxylin & eosin (H&E) (Sigma-Aldrich MHS32-1L; Abcam ab246824) and imaged with an Olympus DP27 camera attached to an Olympus CX43 biological microscope set to the 10X objective using Olympus cellSens Entry 3.1 software. Lesion area was quantified in a blinded fashion, and the average of the six slides was used to determine lesion area. The Necrotic core area was similarly quantified in a blinded fashion using the same six H&E sections. Collagen was stained using the Polysciences Picrosirius Red Stain Kit (cat. no. 24901) according to the manufacturer’s instructions. The same slide between mice was stained and the collagen quantified by applying a consistent color threshold to calculate the collagen-positive area as a percentage of total lesion area. Picrosirius red was imaged via the same system used for H&E imaging. All lesion analysis was conducted using FIJI software.

Immunofluorescence

Identical slides of aortic roots between mice were baked in an incubator at 60°C for 30 min and deparaffinized in Histo-clear (National Diagnostics, HS-200/50-899-90147) three times for 10 min. Histo-clear was removed by 20 shakes in 100% ethanol twice followed by 20 shakes in HistoPrep Reagent Alcohol (Fisher Scientific, HC-600-1GAL) before being washed in tap water seven times. Antigen retrieval was performed by placing slides either in citric acid-based antigen retrieval solution (Vector Laboratories, H3300) or tris-based antigen retrieval solution (Vector Laboratories, H3301) and pressure cooking for 15 min with a 5-min natural release in an Instant Pot IP-LUX pressure cooker. Citric acid-based antigen retrieval was performed for all antibodies except for the antibody against cleaved gasdermin D (GasD). After pressure cooking, slides were cooled for 10 min and then washed three times for 5 min in 1X phosphate buffered saline (PBS) (Boston Bioproducts BM-220; diluted 1:10). Hydrophobic circles were drawn using a liquid blocker pen (Sigma-Aldrich, Z377821). For antibodies sourced from mice, mouse-on-mouse blocking was performed by incubating sections in 10% goat serum (Thermo Fisher Scientific, 10000C) in 1X PBS (Corning 21-040-CM) with 1 drop of mouse-on-mouse block per 1.25 ml of solution (Vector Laboratories, NC9290646) for 1 h at room temperature in a humidified chamber. Slides were then washed in 1X PBS containing Tween 20 (Thermo Fisher Scientific, BP337-500) at a concentration of 0.01% (PBS-T) twice for 5 min and then in 1X PBS once for 5 min. For other antibodies, sections were blocked by incubating in Normal Serum Block (BioLegend, 927503) for 1 h at room temperature in a humidified chamber. Following blocking, sections were incubated with the following primary antibodies at the indicated concentrations overnight at 4 °C in a humidified chamber: Absent in melanoma 2 (AIM2) (Abcam, ab119791, 1:250), Cleaved GasD (Cell Signaling, 10137, 10 μg/ml), c-Myc (Cell Signaling, 5605, 5.76 μg/ml), Ki67 (Abcam, ab15580, 9 μg/ml), MAC2 (Cedarlane, CL8942AP 1 μg/ml), MAC2 conjugated to Alexa Fluor 488 (Cedarlane, CL8942AF4, 1 μg/ml), mer proto-oncogene tyrosine kinase (MerTK) (R&D, BAF591, 2 μg/ml), pγH2AX (Cell Signaling, 9718, 0.74 μg/ml), ZsGreen (Thermo Fisher Scientific, TA180002, 10 μg/ml), triggering receptor expressed on myeloid cells 2 (TREM2) (Denali, 4D9 DC1847, 1:100). Anti-AIM2 and anti-Ki67 (Fig. 2J only) were biotinylated to allow co-staining with antibodies sourced from the same species using the Abcam Biotinylation Kit/Biotin Conjugation Kit (Fast, Type A)-Lightning-Link® (ab201795). The following IgG antibodies were used as negative controls for primary antibodies at the same concentration: Mouse IgG (ChromPure, 015-000-003), Rabbit IgG (Abcam, ab172730), Rabbit IgG (Novus Biologicals, NBP2-24891), Rabbit IgG (ChromPure 011-000-003), Rat IgG (ChromPure, 012-000-003), Rat IgG (BioLegend, 400431), Rat IgG conjugated to Alexa Fluor 488 (Invitrogen, 53-4321-80). Slides were then washed twice with PBS-T for 5 min and once with 1X PBS for 5 min. The following secondary antibodies were used at a dilution of 1:200 for 1 h at room temperature: Rat conjugated to Alexa Fluor 488 (Invitrogen A11006), Rat conjugated to Alexa Fluor 568 (Invitrogen, A11077), Rat conjugated to Alexa Fluor 647 (Invitrogen A-21247), Rabbit conjugated to Alexa Fluor 568 (Invitrogen, A11011), and Rabbit conjugated to Alexa Fluor 647 (A27040). The following secondary antibodies were used at a dilution of 1:800 for 1 h at room temperature: Mouse conjugated to Alexa Fluor 647 (Invitrogen, A21237). Streptavidin conjugated to Alexa Fluor 568 (Invitrogen, S11226) or Streptavidin conjugated to Alexa Fluor 647 (Invitrogen, S32357) were used at a dilution of 1:800 for 1 h at room temperature to detect biotinylated antibodies. DAPI (4′,6-diamidino-2-phenylindole, Dilactate) (Biolegend, 422801) was included during incubation with secondary antibodies at a concentration of 5 μg/ml. Slides were then washed twice with PBS-T for 5 min and once with 1X PBS for 5 min. After washing, slides were mounted using Prolong Gold Antifade Reagent with DAPI (Invitrogen, P36935). TUNEL staining was performed using the In Situ Cell Death Detection Kit TMR Red (Roche, 12156792910) according to the manufacturer’s instructions. Slides were imaged on a Leica Fluorescent DMI 6000B wide-field microscope or a Nikon Ti Eclipse inverted microscope with high-sensitivity confocal imaging at the Confocal and Specialized Microscopy Shared Resource (CSMSR) Core at Columbia University using identical exposure and gain settings. Except for TUNEL, quantitative analyses were performed on raw images using fixed thresholds/parameters applied uniformly across all samples; brightness/contrast adjustments were used for figure visualization only. For TUNEL, display ranges (brightness/contrast) of the TUNEL channel were adjusted per section so that nuclei showing only baseline background fluorescence became faintly visible but not completely absent; a nucleus was scored as TUNEL-positive when, under those settings, it retained clearly stronger nuclear-localized signal relative to neighboring background nuclei within the lesion ROI. Scoring was restricted to the intimal region and performed on original 16 bit data. Brightness and contrast adjustments were linear, applied to the entire image, and used only for visualization. Negative controls for all stains were processed in parallel with identical antigen retrieval and imaged using the same acquisition parameters as their respective stains listed in the legend. All images were analyzed in a blinded fashion using FIJI software.

Figure 2.

Moderate cholesterol lowering reverses macrophage AIM2 inflammasome activation, DNA damage, and proliferation in Jak2^VF MPN lesions. A: Images of aortic root lesions stained for MAC2 (Green), Cleaved GasD (Red), and DAPI (Blue). Scale bar, 60 μm. B: Log₁₀ transformed cleaved GasD mean fluorescence intensity (MFI) in lesions with the addition of constant 1, n = 15–24. P = 0.021 (Ctrl Baseline vs. Jak2^VF Baseline). C: Log₁₀ transformed cleaved GasD mean fluorescence intensity (MFI) in necrotic cores with the addition of constant 1, n = 15–24. P = 0.0008 (Ctrl Baseline vs. Jak2^VF Baseline), P = 0.032 (Ctrl LDL Lowering vs. Jak2^VF LDL Lowering). D: Images of aortic root lesions stained for MAC2 (Green), Cleaved GasD (Red), and DAPI (Blue). Scale bar, 60 μm. White arrows, cleaved GasD + macrophages. E: Log₁₀ transformed cleaved GasD positive macrophages per section with the addition of constant 1, n = 15–24. P = 0.024 (Jak2^VF Baseline vs. LDL Lowering). F: Images of aortic root lesions stained for MAC2 (Green), AIM2 (Red), and DAPI (Blue). Scale bar, 60 μm. White arrows, AIM2 positive macrophages. G: Log₁₀ transformed AIM2 positive macrophages per section with the addition of constant 1, n = 15–23. P = 0.0001 (Ctrl Baseline vs. Jak2^VF Baseline), P = 0.03 (Jak2^VF Baseline vs. LDL Lowering). H: Images of aortic root lesions stained for MAC2 (Green), pγH2AX (Red), and DAPI (Blue). Scale bar, 60 μm. White arrows, pγH2AX positive cells. I: Log₁₀ transformed pγH2AX positive cells per section with the addition of constant 1, n = 15–23. P = 0.0019 (Ctrl Baseline vs. Jak2^VF Baseline), P = 0.045 (Jak2^VF Baseline vs. LDL Lowering). J: Images of aortic root lesions stained for MAC2 (Green), Ki67 (Red), and DAPI (Blue). Scale bar, 60 μm. White arrows, Ki67 positive macrophages. K: Log₁₀ transformed Ki67 positive macrophages per section with the addition of constant 1, n = 15–23. P = 0.0019 (Jak2^VF Baseline vs. LDL Lowering). MAC2 intensity differences reflect independent staining and imaging sessions across panels. All quantifications shown as mean ± s.e.m. Two-way ANOVA with Tukey’s multiple comparisons test (B and C). Kruskal–Wallis test with Dunn’s multiple comparisons test (E, G, I, and K). AIM2, absent in melanoma 2; DAPI, 4′,6-diamidino-2-phenylindole; GasD, gasdermin D; LDL, low-density lipoprotein; MPN, myeloproliferative neoplasm; pγH2AX, phosphorylated histone H2A.X.

Blood isolation and analysis

Blood was collected at the indicated time points via retro-orbital sinus using heparinized microhematocrit capillary tubes (Fisherbrand 22–362566) into EDTA-coated microvette tubes (Sarstedt 20.1278.100). Complete blood cell counts were quantified using a VetScan HM5 Hematology system (Zoetis/Abaxis). Plasma was collected by centrifuging blood at 13,000 g for 10 min at 4°C. Plasma cholesterol was quantified using the Cholesterol E kit (Fujifilm Wako Pure Chemical Corporation, #999–02601/NC9138103) according to the manufacturer’s instructions. Flow cytometric analysis for blood leukocytes was performed by lysing in RBC lysis buffer (Biolegend 420302) for 2 min at room temperature before centrifuging at 3,000 g for 2 min at 4°C. Cells were washed in ice cold MACS Buffer (0.5% BSA and 2 mM EDTA in 1X PBS) (Thermo Fisher Scientific BP9706-100, Invitrogen 15575-038) and centrifuged at 800 g for 10 min at 4°C before staining with antibodies at a 1:100 dilution in ice cold MACS buffer for 20 min. Cells were washed once more in ice cold MACS Buffer, centrifuged at 800 g for 10 min at 4°C before being resuspended in 250 μl of MACS Buffer. Antibodies used in each experiment are listed in Supplemental Tables S1–S3. Cells were captured using a BD Fortessa Flow Cytometer running BD FACSDiva software at the Flow Cytometry Shared Core of the Columbia Center for Translational Immunology (CCTI) and Herbert Irving Comprehensive Cancer Center (HICCC). Flow cytometric data were analyzed using FCS Express 7 Research Edition.

Bone marrow-derived macrophage cultures

Hindlimbs were isolated and stored in ice cold DMEM. The ends of the bones were cut and the bones placed into 600 μl microcentrifuge tubes with a hole punctured at the bottom and sitting inside of 1.5 ml microcentrifuge tubes containing 100 μl of ice cold DMEM. The bones were pulse centrifuged at 4°C for 8 s. The pellet was then suspended in 1.5 ml of ice cold DMEM before being filtered through a 40 μm cell strainer (Thermo Fisher Scientific, 22-363-547) into 50 ml of ice cold DMEM. Suspensions were centrifuged at 800 g for 10 min at 4°C and the supernatant aspirated. The pellet was resuspended in M-CSF complete media for the Jak2^VFAim2^−/− experiment (DMEM + 20 ng/ml M-CSF + 10% FBS + 1% PENSTREP) (M-CSF from Cell Signaling, 33444S) or LCM complete media (all other experiments) (DMEM + 20% LCM + 10% FBS + 1% PENSTREP) and plated into tissue culture treated dishes (Corning CLS430167). Cells were differentiated into macrophages by incubating for 5 days at 37°C and 5% CO₂. After 5 days, the media were aspirated and cells washed in 20 ml of 1X PBS before being stripped by incubation in 5 ml of Corning Cell Stripper (Corning 25-056-CL) for 5 min before being collected using 5 ml of complete media. Cells were then centrifuged at 800 g for 10 min at 4°C and the supernatant aspirated. Cells were counted using an INCYTO C-Chip hemocytometer (DHC-N01) according to the manufacturer's instructions. Cells were plated into well plates of various sizes depending on experiment at a density of 333,333 cells/ml (3 ml of suspension for 6-well plates, 2 ml of suspension for 12-well plates). Cells were allowed to reattach overnight before proceeding with further experiments.

acLDL cell culture experiments

Preparation of acetylated LDL (acLDL) was performed as previously described (34). Bone marrow-derived macrophages (BMDMs) were incubated in M-CSF-free media containing either vehicle or 25 μg/ml acLDL overnight at 37°C and 5% CO₂.

oxPAPC cell culture experiments

BMDMs were incubated in M-CSF-free media containing either vehicle or 50 μg/ml oxidized 1-palmitoyl-2-arachidonoyl-sn-glycero-3-phosphocholine (oxPAPC) overnight at 37°C and 5% CO₂.

MβCD-cholesterol cell culture experiments

BMDMs were primed in complete media containing 20 ng/ml recombinant M-CSF (Cell Signaling 33444S) and 100 ng/ml lipopolysaccharide (Cell Signaling 14011) for 3 h at 37°C and 5% CO₂ before treating with complete media containing vehicle, (2-(2,2,6,6-tetramethylpiperidin-1-oxyl-4-ylamino)-2-oxoethyl)triphenylphosphonium chloride (mitoTEMPO) (MedChemExpress, HY-112879) (10 μM), methyl-β-cyclodextrin-cholesterol (MβCD-Cholesterol) (Sigma-Aldrich, C4951) (30 μg/ml) or a combination of mitoTEMPO (10 μM) and MβCD -Cholesterol (30 μg/ml) for 4 h at 37°C and 5% CO₂. Lipopolysaccharides remained in the media for the duration of these treatments at 100 ng/ml.

Bone marrow-derived macrophage proliferation quantification

BMDM proliferation was measured using the Invitrogen Click-iT™ Plus EdU Alexa Fluor™ 488 Flow Cytometry Assay Kit according to the manufacturer’s instructions. After incubation with EdU, cells were washed in 1X PBS before being stripped by incubation in 500 μl of Corning Cell Stripper (Corning 25-056-CL) for 5 min. Cells were captured using a BD Fortessa Flow Cytometer running BD FACSDiva software at the Flow Cytometry Shared Core of the Columbia Center for Translational Immunology (CCTI) and Herbert Irving Comprehensive Cancer Center (HICCC). Flow cytometric data were analyzed using FCS Express 7 Research Edition.

Immunoblotting

Cells were lysed in radioimmunoprecipitation assay (RIPA) buffer containing protease and phosphatase inhibitors (Invitrogen 87786, Thermo Fisher Scientific, 78420) for 30 min at 4°C on a rocker. Lysates were then frozen at −80°C to enhance lysis. Lysates were then thawed at 4°C, centrifuged at max speed for 5 min at 4°C before being sonicated at 50% power for 30 s. Total protein concentration was determined by bicinchoninic acid (BCA) analysis (Thermo Fisher Scientific, 23225) according to the manufacturer’s instructions. Equal protein amounts were then loaded into polyacrylamide gels (Bio-Rad, 5671094) using 1X Running Buffer (Boston Bioproducts BP-150) and transferred to nitrocellulose membranes (Bio-Rad, 1620115) using transfer buffer (Boston Bioproducts BP-190) with 20% ethanol (Fisher 2701). Membranes were washed in Tris-buffered saline with Tween 20 (TBS-T) (Boston Bioproducts IBB-180X) for 5 min at room temperature before blocking for 1 h at room temperature in 3% bovine serum albumin (BSA) (Fisher, BP9706-100) in TBS-T while gently rocking. After blocking, membranes were incubated overnight at 4°C in 1:1000 primary antibody solution in 3% BSA in TBS-T with the following antibodies: pγH2AX (Cell Signaling, 9718) and total GasD (Cell Signaling, 38754) or for 20 min at room temperature in 1:80,000 primary antibody solution with the following antibodies: β-Actin (Cell Signaling, 12262). After incubation with non-HRP-conjugated primary antibodies, membranes were washed in TBS-T 5 times for 5 min before incubation in a horseradish peroxidase (HRP) conjugated anti-rabbit secondary antibody (7074 Cell Signaling Technology) in 3% BSA in TBS-T for 1 h at room temperature while gently rocking. After HRP-conjugated primary or secondary antibody incubation, membranes were washed five times for 5 min before incubation for 30 s with a chemiluminescent substrate (Thermo Fisher Scientific, 34578). Membranes were then exposed in a dark room using film (Thermo Fisher Scientific, XAR ALF 2025). Densitometric quantification was performed using FIJI software.

IL-1β ELISA detection

IL-1β detection was quantified using the Mouse IL-1 beta/IL-1F2 DuoSet enzyme-linked immunosorbent assay (ELISA) (R&D Systems DY401) according to the manufacturer’s instructions. The plate was detected using a SpectraMax M2 (Molecular Devices).

Statistics

Statistical analyses were conducted using GraphPad Prism v10.5. Data are presented as the mean ± s.e.m. All statistical tests were performed using two-sided analyses. Normality was tested using the D’Agostino-Pearson omnibus and visual inspection of Q–Q plots. Where appropriate, log transformation was applied to approximate normality. Repeated measures data were analyzed using two-way ANOVA with the Geisser-Greenhouse correction for sphericity and Tukey’s multiple comparisons test with individual variance estimation. For datasets involving two independent variables, two-way ANOVA with Tukey’s multiple comparisons test or Kruskal–Wallis with Dunn’s multiple comparisons test were used based on distribution. For datasets involving five or more experimental groups, one-way ANOVA with Holm–Sidak’s multiple comparisons test or Kruskal–Wallis with Dunn’s multiple comparisons test was used based on distribution. Formal outliers were removed using a two-tailed Grubbs’ test (α = 0.05). For experiments with technical replicates from a single mouse, 3–6 wells were plated per condition.

Results

Atherosclerosis regression is impaired with moderate cholesterol lowering in Jak2^VF mice

To assess the effect of Jak2^VF MPN on atherosclerosis regression, we placed Jak2^VF MPN male mice and controls on a western diet for 16 weeks and then induced cholesterol lowering using an HDAd vector containing the human LDL receptor gene (HDAd-hLDLR) combined with a switch to chow diet for an additional 5 weeks (Fig. 1A). Jak2^VF MPN mice exhibited splenomegaly, lower body weight, and increased RBCs, hematocrit, red blood cell distribution width (RDWc), and neutrophils, mirroring previous findings (Supplemental Fig. 1A–I) (30). In this model, CD45.1+ cells are Jak2^VF in experimental mice and WT in controls. CD45.1+ neutrophils and monocytes were elevated while CD45.1+ lymphocytes were reduced (Supplemental Fig. 1J–L). Plasma cholesterol was reduced in Jak2^VF MPN mice on western diet (Fig. 1B) similar to previous findings in both mice and humans (23, 24, 30). Plasma cholesterol remained moderately elevated for 10 days post-virus treatment but returned to pre-diet levels (∼200–300 mg/dl) by the end of the LDL lowering period (Fig. 1B). At baseline, lesion area and necrotic core area were significantly increased in the aortic root of male Jak2^VF MPN mice compared to controls, confirming that Jak2^VF MPN worsens atherosclerosis in both male and female mice (Fig. 1C–E) (30). Moderate cholesterol lowering halted progression of lesion and necrotic core area in control mice whereas Jak2^VF MPN mice developed significantly larger necrotic cores and displayed a trend toward increased lesion area (P = 0.052) (Fig. 1C–E). Consistent with previous studies (35, 36), lesional collagen was increased in control mice after cholesterol lowering with no significant change in Jak2^VF MPN mice, further indicating incomplete resolution in the Jak2^VF MPN mice (Fig. 1F, G). The most consistent phenotype found in atherosclerosis regression studies has been a decrease in macrophage burden (13, 17, 18, 19, 36). Macrophages were significantly decreased in control mice but not significantly in Jak2^VF MPN mice (P = 0.16) after cholesterol lowering (Fig. 1H, I; IgG control antibody, Supplemental Fig. 2A). These results indicate that atherosclerosis resolution is incomplete in Jak2^VF MPN with moderate cholesterol lowering.

We next assessed inflammatory features in plaques. Necrotic, inflammatory cell death is mediated by activated GasD downstream of inflammasome activation (37). We used an antibody to the active N-terminal fragment of GasD to assess pyroptosis in plaques (29). Total cleaved GasD in plaques as well as cleaved GasD in the necrotic core were increased in Jak2^VF MPN lesions at baseline and were not reduced by cholesterol lowering (Fig. 2A–C; IgG control antibody, Supplemental Fig. 2B). However, macrophage-specific cleaved GasD was decreased by cholesterol lowering in both groups, with a significant decrease in Jak2^VF MPN lesions, suggesting that macrophage pyroptosis had been suppressed by the time of harvest (Fig. 2D, E). Consistent with earlier findings, AIM2 + macrophages and phosphorylated histone H2A.X (pγH2AX), an indicator of double-stranded DNA break formation, were increased in Jak2^VF MPN lesions at baseline and were decreased by cholesterol lowering, suggesting reversal of AIM2 inflammasome activation (Fig. 2F–I, streptavidin only control, Supplemental Fig. 2C) (30). In addition, macrophage proliferation trended higher at baseline in Jak2^VF MPN mice and was significantly reduced by cholesterol lowering (Fig. 2J, K, streptavidin only control, Supplemental Fig. 2D). These data suggest reversal of macrophage AIM2 inflammasome activation and pyroptosis by moderate cholesterol lowering in Jak2^VF MPN mice, correlating with decreased macrophage proliferation and decreased DNA damage response. The persistence of cleaved GasD in the necrotic core area may reflect ongoing pyroptosis during the early phase of LDL lowering when cholesterol levels were persistently elevated.

Cholesterol lowering restores impaired efferocytosis and increases TREM2+ macrophages in Jak2^VF mice

LDL lowering has been associated with improvements in pro-resolving features of lesions such as increased macrophage efferocytosis, MerTK expression (13, 38, 39), increased numbers of TREM2 + macrophages and limitation of necrotic core formation in plaques (16, 40). During atherosclerosis progression, macrophage MerTK and TREM2 were suppressed in Jak2^VF CH mouse plaques as a result of interleukin-1 (IL-1) signaling from mutant to WT cells (29). Consistent with these studies, we found that macrophage MerTK was decreased in Jak2^VF MPN lesions at baseline compared to controls (Fig. 3A, B, streptavidin only control, Supplemental Fig. 2E). MerTK levels were completely restored by cholesterol lowering (Fig. 3A, B). In addition, we found that cholesterol lowering dramatically increased the percentage of macrophages positive for TREM2 in both control and Jak2^VF MPN lesions (Fig. 3C, D, IgG control antibody, Supplemental Fig. 2F). Consistent with these findings, macrophage-associated TUNEL + nuclei as a percentage of total lesional TUNEL + nuclei, an estimate of in situ efferocytosis (14), were significantly decreased in Jak2^VF MPN lesions at baseline and restored by cholesterol lowering, (Fig. 3E, F, dye only negative control, Supplemental Fig. 2G). Together, these data suggest that cholesterol lowering restores impaired efferocytosis in Jak2^VF MPN lesions and increases the proportion of inflammation-resolving macrophages in lesions. Thus, macrophage inflammasome activation and associated defects in efferocytosis were reversed by 5 weeks of moderate cholesterol lowering in Jak2^VF MPN mice. The increased necrotic core and decreased collagen content in these mice could reflect limited resolution of hypercholesterolemia during the earlier phases of LDL lowering.

Figure 3.

Moderate cholesterol lowering reverses impaired efferocytosis in Jak2^VF MPN lesions while increasing TREM2^Hi macrophages in control and Jak2^VF MPN lesions. A: Images of aortic root lesions stained for MAC2 (Green), MerTK (Red), and DAPI (Blue). Scale bar, 60 μm. B: Percentage of MAC2 positive area double positive for MerTK and MAC2, n = 15–23. P = 0.02 (Ctrl Baseline vs. Jak2^VF Baseline), P < 0.0001 (Jak2^VF Baseline vs. LDL Lowering). C: Images of aortic root lesions stained for MAC2 (Green), TREM2 (Red), and DAPI (Blue). Scale bar, 60 μm. D: Percentage of MAC2 positive area double positive for TREM2 and MAC2, n = 15–24. P < 0.0001 (Ctrl and Jak2^VF Baseline vs. LDL Lowering). E: Images of in situ efferocytosis in aortic root lesions: MAC2 (Green), TUNEL (Red), and DAPI (Blue). Scale bar, 60 μm. White arrows, nuclei double positive for TUNEL and MAC2. White wedges, nuclei positive for TUNEL but negative for MAC2. F: Percentage of TUNEL positive nuclei also positive for MAC2. P = 0.011 (Ctrl Baseline vs. Jak2^VF Baseline), P = 0.021 (Jak2^VF Baseline vs. LDL Lowering). MAC2 intensity differences reflect independent staining and imaging sessions across panels, n = 13–24. All quantifications shown as mean ± s.e.m. Two-way ANOVA with Tukey’s multiple comparisons test (D). Kruskal-Wallis test with Dunn’s multiple comparisons test (B and F). DAPI, 4′,6-diamidino-2-phenylindole; LDL, low-density lipoprotein; MerTK, mer proto-oncogene tyrosine kinase; MPN, myeloproliferative neoplasm; TREM2, triggering receptor expressed on myeloid cells 2.

More effective cholesterol lowering restores atherosclerosis regression

To test this hypothesis, we performed a similar study but with more effective LDL cholesterol lowering. Murine LDL does not bind as strongly to the human low-density lipoprotein receptor (LDLR) (employed in Fig. 1) as to the mouse LDLR (41, 42). Therefore, we employed an AAV8 containing the mouse Ldlr gene which more effectively and more quickly lowered cholesterol levels (Fig. 4A). Blood parameters were similar to the first experiment and not appreciably changed by cholesterol lowering (Supplemental Fig. 3, Fig. 4B). Consistent with the first study, lesion area and necrotic core area were significantly increased in Jak2^VF MPN mice at baseline. However, in contrast to the first study, both groups showed similar nonsignificant upward trends of 1.29-fold (Ctrl) and 1.25-fold (Jak2^VF) in lesion area and 2.0-fold (Ctrl) and 1.67-fold (Jak2^VF) in necrotic core area after aggressive LDL lowering (Fig. 4C–E). Furthermore, aggressive cholesterol lowering significantly increased lesional collagen and decreased macrophage burden in both groups (Fig. 4F–I, IgG control antibody, Supplemental Fig. 4A). Analysis of cleaved GasD showed an increase in macrophages in Jak2^VF mice at baseline that was reversed by cholesterol lowering, and in contrast to the first study a decrease in cleaved GasD in the necrotic core that appeared similar in control and Jak2^VF MPN mice (Fig. 5A–D). Similarly, cleaved GasD was increased in CD11b + splenocytes of Jak2^VF MPN mice at baseline and reversed by cholesterol lowering (Fig. 5E, F). Consistent with the first study, double stranded DNA breaks (pγH2AX) were increased in Jak2^VF MPN lesions at baseline and more strongly decreased by aggressive cholesterol lowering along with a significant decrease in macrophage proliferation (Fig. 5G, H, Supplemental Fig. 5A, B, IgG control antibodies, Supplemental Fig. 5C, D). In addition, macrophage MerTK was increased by aggressive cholesterol lowering (Supplemental Fig. 6A, B). These results indicate that more effective cholesterol lowering over 5 weeks reversed inflammatory features and induced similar levels of atherosclerosis resolution in control and Jak2^VF MPN lesions. However, studies of moderate and intensive cholesterol lowering may not be directly comparable as they were conducted at different times and baseline lesions in control and Jak2^VF mice were somewhat larger in the former study.

Figure 5.

Aggressive cholesterol lowering more strongly decreases macrophage pyroptosis and DNA damage in Jak2^VF MPN lesions. A: Images of aortic root lesions stained for MAC2 (Green), cleaved GasD (Red), and DAPI (Blue). Scale bar, 60 μm. B: Log₁₀ transformed cleaved GasD MFI in the necrotic core with the addition of constant 1, n = 14–20. P = 0.0064 for genotype effect and P = 0.021 for treatment effect by two-way ANOVA. C: Images of aortic root lesions stained for MAC2 (Green), Cleaved GasD (Red), and DAPI (Blue). Scale bar, 60 μm. White arrows, cleaved GasD + macrophages. D: Log₁₀ transformed cleaved GasD positive macrophages per section with the addition of constant 1, n = 13–20. P = 0.023 (Ctrl Baseline vs. Jak2^VF Baseline), P = 0.005 (Jak2^VF Baseline vs. LDL Lowering). E: Representative immunoblot analysis of full-length and cleaved GasD in CD11b + splenocytes. F: Log₁₀ transformed densitometric quantification of the ratio of cleaved GasD to full-length GasD from E with the addition of constant 1. n = 7–8 biological replicates. P = 0.0003 (Ctrl Baseline vs. Jak2^VF Baseline), P = 0.0046 (Jak2^VF Baseline vs. LDL Lowering). G: Images of aortic root lesions stained for MAC2 (Green), pγH2AX (Red), and DAPI (Blue). Scale bar, 60 μm. White arrows, pγH2AX positive cells. H: Log₁₀ transformed pγH2AX positive cells per section with the addition of constant 1, n = 14–20. P = 0.016 (Ctrl Baseline vs. LDL Lowering), P = 0.0016 (Ctrl Baseline vs. Jak2^VF Baseline), P < 0.0001 (Jak2^VF Baseline vs. LDL Lowering). All quantifications shown as mean ± s.e.m. Two-way ANOVA with Tukey’s multiple comparisons test (B, D, F, and H). DAPI, 4′,6-diamidino-2-phenylindole; GasD, gasdermin D; LDL, low-density lipoprotein; MFI, mean fluorescence intensity; MPN, myeloproliferative neoplasm; pγH2AX, phosphorylated histone H2A.X.

Cholesterol lowering suppresses proliferation of Jak2^VF macrophages and increases TREM2+ macrophages

To remove differences in lesion area at baseline and to simulate the acquisition of the Jak2^VF mutation in life after atherosclerosis has already developed, we performed a third study in which Jak2^VF was activated in hematopoietic stem cells at the time of LDL lowering by using a tamoxifen-inducible Scl-driven cre recombinase along with a floxed ZsGreen reporter allele (Jak2^VF Scl CH) to assess cell genotype specific changes (Fig. 6A). In addition to the two LDL lowering groups, we included control and Jak2^VF mice that remained on the western diet (progression groups). As expected, plasma cholesterol remained elevated in the progression groups but was markedly and rapidly decreased in both LDL lowering groups (Fig. 6B). The Jak2^VF MPN phenotype was partially apparent as indicated by increased spleen weights, RDWc, and hematocrit as well as expansion of Jak2^VF (CD45.1+CD45.2+) blood monocytes and neutrophils, but less pronounced than in the earlier studies when Jak2^VF was activated prior to western diet feeding (Supplemental Fig. 7).

Compared to the earlier studies, the impact of Jak2^VF on lesion progression was less pronounced reflecting the shorter period of Jak2^VF activation (Supplemental Fig. 8A–C). However, consistent with the earlier studies, the necrotic core significantly worsened in the Jak2^VF progression mice but not in the control mice compared to baseline, while the necrotic core was not significantly changed in either LDL lowering group compared to baseline (Supplemental Fig. 8C). In addition, macrophage burden was similarly decreased in both LDL lowering groups, (Supplemental Fig. 8D, E). These studies further support that aggressive cholesterol lowering similarly promotes resolution in control and Jak2^VF mice with acquisition of Jak2^VF in the context of established atherosclerosis.

Total and necrotic core Cleaved GasD were increased in Jak2^VF lesions compared to baseline with a significant decrease in the necrotic core and a total trend lower (P = 0.051) after cholesterol lowering, consistent with aggressive cholesterol lowering reversing pyroptosis in lesions (Fig. 6C–E). Cleaved GasD + macrophages were directionally consistent with the Mx1-cre LDL lowering models but did not reach statistical significance (Fig. 6F, G). In the Jak2^VF progression group, proliferation was significantly increased in Jak2^VF (green) macrophages but not in WT (non-green) macrophages indicating a mutant cell-specific effect, as seen previously (30). Importantly, this increased proliferation of Jak2^VF cells was reversed by cholesterol lowering, (Fig. 6H–J). The DNA damage response (pγH2AX) was significantly increased in Jak2^VF cells and was reversed by cholesterol lowering while WT cells showed no significant proliferative changes or changes in the DNA damage response within Jak2^VF lesions (Fig. 6K–M, ZsGreen^−/− control, Supplemental Fig. 9A).

Given that a proliferative efferocytic population with high c-Myc expression has been described with LDL lowering (13, 38), we assessed c-Myc expression along with TREM2. In contrast to the decreases in Ki67, macrophage c-Myc and c-Myc + TREM2^Hi macrophages were increased by aggressive LDL lowering in both control and Jak2^VF mice, suggesting a selective increase of TREM2+, anti-inflammatory macrophages despite a decrease in total macrophages (Supplemental Fig. 10A–D, IgG antibody control, Supplemental Fig. 10E).

Cholesterol loading increases AIM2 inflammasome activation and the DNA damage response in macrophages

The parallel decrease in proliferation and DNA damage response in Jak2^VF macrophages (Fig. 6H–M) suggested that reversal of proliferation could be a key mechanism to explain reduced inflammasome activation in response to cholesterol lowering. To further assess the relationship between proliferation, the DNA damage response and inflammasome activation, we carried out mechanistic studies in BMDMs. We found that cholesterol loading of macrophages with acLDL, which promotes entry of cholesterol via the scavenger receptor A/endo-lysosomal system (43), caused a modest increase in proliferation (Fig. 7A). While Jak2^VF macrophages showed significantly increased pγH2AX compared to controls, there was no effect of acLDL loading on pγH2AX (Fig. 7B, C). We also showed that oxPAPC, a major component of oxidized LDL, stimulated macrophage proliferation, but this was associated with a reduced level of pγH2AX (Fig. 7D–F). We next employed methyl-β-cyclodextrin (MβCD)-cholesterol, a distinctive mode of cholesterol loading that increases mitochondrial cholesterol content (44), promoting mitochondrial reactive oxygen species (ROS) formation and AIM2 inflammasome activation (45). MβCD-cholesterol increased IL-1β secretion into cell culture media especially in Jak2^VF macrophages, and this was reversed by the mitochondrial-specific antioxidant mitochondria-targeted TEMPO (a triphenylphosphonium-conjugated nitroxide) (MitoTEMPO) (Fig. 7G), suggesting that the effect was dependent on mitochondrial ROS production, consistent with prior studies (45). The increase in IL-1β secretion in Jak2^VF macrophages in response to MβCD-cholesterol was reversed by Aim2 deficiency implicating the AIM2 inflammasome (Fig. 7H). MβCD-cholesterol increased pγH2AX in control and Jak2^VF macrophages, with a larger effect in the latter (two-way ANOVA) (Fig. 7I, J). In contrast, MβCD-cholesterol loading did not increase macrophage proliferation (Fig. 7K). Together these findings suggest that increased AIM2 inflammasome activation in response to MβCD-cholesterol loading may largely reflect increased mitochondrial ROS and DNA damage, rather than changes in cell proliferation.

Figure 7.

Cholesterol increases AIM2 inflammasome activation and nuclear dsDNA breaks via increased mitochondrial ROS in Jak2^VF macrophages. A: EdU + BMDMs as a percentage of total BMDMs after overnight incubation with acLDL (25 μg/ml) in base media without LCM or recombinant M-CSF during the proliferation assay. n = 4 biological replicates. P = 0.032 for acLDL effect. B: Representative immunoblot analysis of pγH2AX in BMDMs after overnight incubation with acLDL (25 μg/ml) in base media without LCM or recombinant M-CSF during the proliferation assay. C: Densitometric quantification of pγH2AX normalized to β-Actin from B. Each data point represents a cell-well technical replicate from one mouse (1 biological replicate, 3 technical replicates). P = 0.022 (Ctrl Vehicle vs. Jak2^VF Vehicle), P = 0.038 (Ctrl acLDL vs. Jak2^VF acLDL). Statistical comparisons in this panel were performed across cell-well technical replicates from a single mouse (1 biological replicate; n = 3 wells per condition) to assess within-experiment reproducibility. D: EdU + BMDMs as a percentage of total BMDMs after overnight incubation with oxPAPC (50 μg/ml) in base media without LCM or recombinant M-CSF during the proliferation assay. Each data point represents a cell-well technical replicate from one mouse (1 biological replicate, 3 technical replicates). P = 0.022 (Ctrl Vehicle vs. Ctrl oxPAPC), P < 0.0001 (Jak2^VF Vehicle vs. Jak2^VF oxPAPC), P = 0.0004 (Ctrl oxPAPC vs. Jak2^VF oxPAPC). Statistical comparisons in this panel were performed across cell-well technical replicates from a single mouse (1 biological replicate; n = 3 wells per condition) to assess within-experiment reproducibility. E: Representative immunoblot analysis of pγH2AX in BMDMs after overnight incubation with oxPAPC (50 μg/ml) in base media without LCM or recombinant M-CSF during the proliferation assay. F: Densitometric quantification of pγH2AX normalized to β-Actin from E. Each data point represents a cell-well technical replicate from one mouse (1 biological replicate, 3 technical replicates). P = 0.026 for oxPAPC effect. Statistical comparisons in this panel were performed across cell-well technical replicates from a single mouse (1 biological replicate; n = 3 wells per condition) to assess within-experiment reproducibility. G: ELISA quantification of total IL-1β in cell culture media of BMDMs incubated with MβCD-cholesterol (30 μg/ml) alone, mitoTEMPO (10 μM) alone, or MβCD-cholesterol (30 μg/ml) in combination with mitoTEMPO (10 μM) for 4 h after priming with LPS (100 ng/ml) for 3 h in complete media containing 20 ng/ml recombinant M-CSF. Data are shown as technical replicates from n = 2–3 independent experiments (biological replicates). Wells failing prespecified quality-control criteria in which the coefficient of variation (CV) exceeded 20% were excluded. P = 0.0007 (Jak2^VF Vehicle vs. Jak2^VF MβCD-cholesterol), P = 0.019 (Jak2^VF MβCD-cholesterol vs. Jak2^VF MβCD-cholesterol + mitoT), H: ELISA quantification of total IL-1β in cell culture media of BMDMs incubated with MβCD-cholesterol (30 μg/ml) for 4 h after priming with LPS (100 ng/ml) for 3 h in complete media containing 20 ng/ml recombinant M-CSF. n = 6 cell-well technical replicates from a single mouse (1 biological replicate). No biological replicates were available. P = 0.0017 (Jak2^VF Vehicle vs. Jak2^VF MβCD-cholesterol), P = 0.0026 (Jak2^VF MβCD-cholesterol vs. Jak2^VFAim2^−/− MβCD-cholesterol). Statistical comparisons in this panel were performed across cell-well technical replicates from a single mouse (1 biological replicate; n = 6 wells per condition) to assess within-experiment reproducibility. I: Representative immunoblot analysis of pγH2AX in BMDMs after a 4-h incubation with MβCD-cholesterol (30 μg/ml) in complete media containing 20 ng/ml recombinant M-CSF or LCM. Irrelevant lanes were removed from the blot. Vertical lines indicate where lanes were spliced from the same membrane and exposure. J: Densitometric quantification of pγH2AX normalized to β-Actin from H. n = 8 biological replicates for the vehicle groups and n = 4 biological replicates for the MβCD-cholesterol groups. P = 0.0019 (Ctrl Vehicle vs. Ctrl MβCD-cholesterol), P = 0.0028 (Jak2^VF Vehicle vs. Jak2^VF MβCD-cholesterol). P = 0.019 for genotype effect. P < 0.0001 for MβCD-cholesterol effect. Ctrl Vehicle data include earlier assessments of this end point in untreated cells, prior to the initiation of the treatment arm. All samples were processed and analyzed using the same protocol. K: EdU + BMDMs as a percentage of total BMDMs after a 4-h incubation with MβCD-cholesterol (30 μg/ml) in complete media containing 20 ng/ml recombinant M-CSF. n = 3 biological replicates. P = 0.043 for genotype effect. All quantifications shown as mean ± s.e.m. Two-way ANOVA with Tukey’s multiple comparisons test (A, C, D, F, H, J, and K). Kruskal–Wallis test with Dunn’s multiple comparisons test (G). acLDL, acetylated LDL; AIM2, absent in melanoma 2; BMDM, bone marrow-derived macrophage; IL-1β, interleukin-1β; LPS, lipopolysaccharide; MitoTEMPO, mitochondria-targeted TEMPO (a triphenylphosphonium-conjugated nitroxide); MβCD, methyl-β-cyclodextrin; oxPAPC, oxidized 1-palmitoyl-2-arachidonoyl-sn-glycero-3-phosphocholine; pγH2AX, phosphorylated histone H2A.X; ROS, reactive oxygen species.

Discussion

Based on prior studies (30), we hypothesized that LDL lowering alone would not be sufficient to fully prevent plaque progression in Jak2^VF mice. Although this proved to be the case with moderate LDL lowering, aggressive cholesterol lowering halted progression and induced beneficial remodeling changes, including decreased macrophage burden and increased lesional collagen content similarly in controls and Jak2^VF MPN mice. Surprisingly, moderate or intensive LDL lowering reversed the major cellular changes in Jak2^VF macrophages that have been associated with athero-progression (30). This included reduced AIM2 expression and pyroptosis and increased MerTK, TREM2 and in situ efferocytosis that have all been mechanistically linked in prior studies of Jak2^VF mice (29). Our prior studies have shown that inflammasome activation leads to reduced levels of MerTK and TREM2 (29), suggesting that increases in these factors during cholesterol lowering may be secondary to reduced inflammasome activation.

Prior studies have shown that LDL lowering in mice with established atherosclerosis may induce atherosclerosis regression, limitation of progression, or beneficial remodeling involving multiple mechanisms (46) including macrophage emigration (47), decreased monocyte entry (48) and a role of acquired immunity and T-regs (17). Our studies using a cleaved GasD antibody to monitor pyroptosis in lesions, indicated that reversal of inflammasome activation may also be a key mechanism underlying beneficial remodeling of plaques to a more stable phenotype, especially in settings where there is underlying increased inflammasome activation as occurs in Tet2 and Jak2VF CH. Similar to macrophage proliferation being a major factor accounting for increased macrophage number during advanced atherosclerosis progression (49), suppression of proliferation may be an important mechanism decreasing macrophage burden with LDL lowering (35). Our studies suggest that reversal of the proliferative tendency of Jak2^VF macrophages is also a key mechanism decreasing macrophage burden with LDL lowering in Jak2^VF mice. The mechanism of proliferation could involve uptake of modified LDL by scavenger receptors into the endolysosomal system (43) consistent with our observations. The decreased proliferation of Jak2VF macrophages signifies that there are fewer inflammatory macrophages in lesions, consistent with the virtual disappearance of Jak2^VF macrophages in lesions with aggressive cholesterol lowering.

The parallel decrease in macrophage proliferation and DNA damage response suggested that decreased replication induced DNA damage could be a mechanism to explain decreased AIM2 inflammasome activation and pyroptosis. However, mechanistic studies in Jak2^VF BMDMs indicated that macrophage proliferation could be clearly dissociated from DNA damage responses and AIM2 inflammasome activation that were rather dependent on mitochondrial ROS generation and DNA damage as suggested by earlier studies (45, 50). The increased DNA damage and inflammasome activation in Jak2^VF macrophages compared to controls likely reflects increased glycolysis, mitochondrial ROS formation, and oxidative mitochondrial damage (30, 51). However, the decreased Jak2^VF macrophage burden, as well as decreased AIM2 expression, and decreased cholesterol-induced AIM2 inflammasome activation together explain why effective LDL lowering was sufficient to reverse the adverse effects of Jak2^VF on atherosclerosis.

Paradoxically, despite the overall reduction in macrophage burden, during resolution there may also be a relative increase in a proliferative set of efferocytic macrophages that have increased c-Myc expression and contribute to plaque stability (13, 14). In addition to a potential role in cell proliferation, c-Myc increases Akt signaling and cell survival (52, 53). Previously, we observed a 4-fold increase in TREM2^Hi macrophages induced by both LDL lowering and Jak2^VF inactivation in Jak2^VF CH mice (16). In the present study, we found that LDL lowering alone led to a marked increase in TREM2 staining in macrophages, as well as elevated c-Myc expression specifically within TREM2^Hi macrophages, in both control and mutant mice. In addition, c-Myc expression is enhanced by efferocytosis in WT macrophages (13, 38), suggesting that this effect is not Jak2^VF-specific. This contrasts with the decreased proliferation and DNA damage response found specifically in the Jak2^VF macrophages. Our findings together suggest that there is an increase in TREM2^Hi, c-Myc, and MerTK expressing macrophages induced by LDL lowering either through increased proliferation or increased survival of these subsets that then have beneficial effects on efferocytosis and plaque stability. Given that inflammasome activation and IL-1 release have been linked to reduced MerTK and TREM2 as a result of a disintegrin and metalloproteinase 17 (ADAM17)-mediated cleavage (29, 54), the increase in this proresolving macrophage population may in part be secondary to reduced cholesterol-induced inflammasome activation.

Our study has several limitations. Moderate versus aggressive LDL lowering studies were conducted at different times and lesion area at baseline in both control and Jak2^VF mice was higher in the moderate group which could have impaired regression. It is also possible that a longer period of cholesterol lowering at a moderate level could have achieved the same beneficial results. Human JAK2-mutant clones may persist at low allele burden for years prior to detection of CHIP or development of overt MPN (55, 56, 57). Accordingly, the degree of benefit from intensive LDL-C lowering may vary across patient trajectories and clonal exposure duration. In response to LDL lowering, regression of lesion and necrotic core area was only moderate or nonexistent. Both control and Jak2^VF mice showed similar upward trends in necrotic core area following cholesterol lowering, although only significant in the Jak2^VF moderate group. This could suggest that the necrotic core is slower to resolve than other features such as inflammatory macrophages.

LDL lowering guidelines have recommended progressively lower targets based on clinical trial evidence. Current recommendations suggest an LDL cholesterol target below 70 mg/dl for patients with established ACVD or in certain high-risk groups (58). Statin use has been associated with improved outcomes in human MPN patients and was associated with a 37% reduction in the risk of thrombosis in patients who had polycythemia vera or essential thrombocythemia as well as a 22% reduction in all-cause mortality (59). Although data from mice should be extrapolated to humans with caution, the AIM2 inflammasome seems to increase ACVD risk in both mice and humans carrying the Jak2^VF mutation (31). Nevertheless, the present preclinical study adds mechanistic findings to support the idea that intensive LDL-C lowering in Jak2^VF MPN patients would reduce ACVD risk.

Data availability

The datasets generated for this study are available from the corresponding author upon reasonable request.

Supplemental data

This article contains supplemental data.

Conflict of interests

A. R. T. is a consultant for CSL Behring and is on the scientific advisory board of Beren Therapeutics. All other authors declare that they have no conflicts of interest with the contents of this article.

Supplemental data

Supplemental

Material