Ppm1d Clonal Hematopoiesis Does Not Promote Atherosclerosis in Murine Models

Abstract

Background:

Clonal hematopoiesis (CH) is an age-associated condition common in the elderly that arises when hematopoietic stem cells acquire somatic mutations in an assortment of genes, most commonly DNMT3A, TET2, ASXL1, PPM1D, and JAK2. CH is associated with increased coronary artery disease (CAD) and all-cause mortality. Epidemiological studies have revealed that different CH driver mutations are associated with unique outcomes. PPM1D-CH is more prevalent in patients following radiation and cytotoxic therapy. PPM1D CH has been strongly associated with CAD, peripheral artery disease, and all-cause mortality. However, it is unclear if this relationship is causative.

Methods:

We tested the ability of Ppm1d mutations to promote atherosclerosis in mice. To model Ppm1d CH, we transplanted a mixture of 20% bone marrow from mice expressing Ppm1d T476* truncation mutation under the control of an Scl-Cre and 80% wildtype bone marrow into Ldlr^−/− recipient mice and subjected them to atherosclerosis studies. In parallel, Scl-Cre driven Ppm1d T476* mutations were introduced into Apoe^−/− mice for atherosclerosis studies. To examine the interaction between DNA damage and Ppm1d, we transplanted bone marrow with Ppm1d mutations restricted to monocyte/macrophage or pan-hematopoietic expression of Ppm1d T476* into Ldlr^−/− mice. Cisplatin was administered after establishment of atherosclerosis, and lesion burden and inflammasome activation quantified.

Results:

We found that in vitro Ppm1d mutant macrophages have increased AIM2 inflammasome activation and increased inflammasome activation in response to cisplatin. However, in vivo we found no evidence of elevated inflammasome activation in plaques. Across both Ldlr and Apoe knockout models, hematopoietic Ppm1d mutations did not promote atherosclerosis. Furthermore, Ppm1d mutations did not exacerbate atherosclerosis following administration of cisplatin.

Conclusions:

These findings indicate that in murine models Ppm1d mutations are not sufficient to promote atherosclerosis, suggesting that the epidemiologic association between Ppm1d-CH and CAD may not be due to changes in plaque development.

Graphical Abstract

Introduction

Clonal hematopoiesis (CH) is associated with increased atherosclerotic cardiovascular disease (ASCVD)^1,2. CH arises when hematopoietic stem cells (HSCs) acquire somatic mutations typically in DNMT3A, TET2, ASXL1, PPM1D, or JAK2 which instill a survival advantage^3–5. Through hematopoiesis, mutations in HSCs are packaged into immune cells which can circulate and home to tissue. These mutated immune cells comprise a subset of circulating cells which when associated with a variant allele fraction (VAF) ≥2% in the absence of clinical abnormalities are known as clonal hematopoiesis of indeterminate potential (CHIP). Analyses of large human cohorts have shown outcomes are dependent on the mutated driver gene. For instance, DNMT3A often exhibits weaker associations with ASCVD, while TET2, JAK2, and others consistently show strong links to cardiovascular risk⁶. These varied outcomes may not be too surprising due to the diversity of driver genes. Mechanistic studies in mice and humans have identified unique mechanisms underlying the relationship of Tet2^1,7,8, Jak2^9–12, and Asxll^12,13 with atherosclerosis, implicating a causal relationship between CH and ASCVD.

Protein Phosphatase Magnesium-Dependent 1D (PPM1D) is among the more frequently mutated genes in CH. In 50,122 participants from the UK Biobank and Mass General Brigham Biobank, PPM1D mutations were associated with increased incident pan-arterial atherosclerosis, coronary artery disease (CAD), and peripheral artery disease¹⁴. However, a larger UK Biobank analysis did not detect a significant association between PPM1D-CH and incident ASCVD², leaving the relationship uncertain.

PPM1D-CH typically arise from truncation mutations in the terminal exon that removes the proteasomal degradation signal and preserves the catalytic domain, resulting in reduced turnover and increased protein abundance¹⁵. PPM1D participates in the DNA Damage Response (DDR) by de-phosphorylating P53, Ataxia–Telangiectasia Mutated (ATM), Phosphorylated Histone H2A.X (pγH2AX), Checkpoint Kinase (CHK)1/2, Mouse Double Minute 2 Homolog (MDM2)and other related proteins^16–19. Truncation mutations typically increase PPM1D activity resulting in increased de-phosphorylation of DDR proteins and suppressed DDR signaling. In patients, PPM1D mutations are more frequently found following cancer therapy²⁰. Murine models suggest this enrichment is due to increased HSC survival due to suppressed DDR^21,22. Thus, PPM1D-CH may mark prior exposure to DNA-damaging agents, potentially explaining epidemiologic associations with disease. Nevertheless, the cellular and molecular consequences of PPM1D mutations remain incompletely defined.

Here, we tested whether Ppm1d mutations in hematopoietic cells were sufficient to drive atherosclerosis. Across Ldlr^−/− and ApoE^−/− models, Ppm1d-driven CH did not increase atherogenesis. Moreover, modeling prior genotoxic exposure with cisplatin failed to reveal PPM1D-dependent effects on atherosclerosis. These data argue against a direct role for PPM1D-CH in promoting atherosclerosis and suggest that observed human associations may not reflect changes in lesion development.

Methods

All data will be made available from the corresponding author upon reasonable request.

Mice

All mice were maintained on a C57BL/6J background. Mice were housed in a specific pathogen-free facility under standards conditions of temperature with a 12-h light-dark cycle and ad libitum access to food. Mouse experiments were approved and conducted in accordance with the guidelines of the University of California San Francisco Institutional Animal Care and Use Committee. Ppm1d^T476 transgenic mice were generated as previously described²¹. Studies utilizing ApoE^−/−-Scl-Ppm1d^T476 mice were conducted in both male and female animals. Studies of Mx1-Cre, Scl-Cre clonal hematopoiesis, and Cx3cr1-Cre were conducted in female animals. For specific expression models Mx1-Cre (The Jackson Laboratory, B6.Cg-Tg(Mx1-Cre)1Cgn/J (003556)), Scl-Cre (The Jackson Laboratory, C57BL/6-Tg(Tal1-cre/ERT)42-056Jrg/J (037466)), and Cx3cr1-Cre (The Jackson Laboratory, B6J.B6N(Cg)-Cx3cr1^{tm1.1(cre)Jung}/J (025524)) mice were crossed with Ppm1d^T476 transgenic mice²¹ to which to generate Mx1- Ppm1d^T476, Scl-Ppm1d^T476, and Cx3cr1- Ppm1d^T476 mice. ApoE^−/− mice (The Jackson Laboratory, B6.129P2-Apoe^tm1Unc/J (002052) were crossed with Scl-Ppm1d^T476 transgenic mice to generate ApoE^−/−-Scl-Ppm1d^T476. Bone marrow transplantations were conducted into Ldlr^−/− (The Jackson Laboratory, B6.129S7-Ldlr^tm1Her/J (002207). Littermate controls lacking Cre were used except for studies utilizing Cx3cr1-Cre mice in which case Cx3cr1-Cre littermate controls were used.

Association of PPM1D CHIP with cardiovascular risk and its interaction with inflammatory gene expression

Analysis was conducted using UK Biobank data, included 438,289 unrelated participants from the UK Biobank study who had undergone whole-exome sequencing of blood-derived DNA and were free of hematologic cancer at enrollment²³. The methodologies for whole-exome sequencing and general detection of clonal hematopoiesis of indeterminate potential (CHIP) were performed as previously described²⁴. This study focused specifically on CHIP driver mutations within the PPM1D gene, and a participant was classified as having PPM1D CHIP if a somatic mutation was detected at a variant allele fraction (VAF) greater than 2%.

The five primary outcomes were atrial fibrillation, bradyarrhythmia, coronary artery disease, heart failure, and all-cause mortality. Each outcome was ascertained using International Classification of Diseases (ICD) codes from inpatient hospital records and national death registries, with a complete list of codes available in the Appendix.

To investigate potential biological mechanisms, we generated genetically predicted expression scores for inflammatory genes involved in the AIM2 inflammasome pathway. The scores were calculated using weights from the eQTLGen Consortium and developed using Pruning + Thresholding and Bayesian PRS-CS methods^25,26. Final scores were optimized against measured RNA-Seq data from the Framingham Heart Study and the Multi-Ethnic Study of Atherosclerosis cohorts, with further technical details described in our prior work²⁷. We then performed stratified analyses to assess whether the associations between these gene scores and cardiovascular outcomes were modified by PPM1D mutation status.

Bone Marrow Transplant

Bone marrow transplantation (BMT) procedures were performed as previously described^9,11,28. Recipient mice aged 8-12 weeks (C57BL/6J Ldlr^−/−) were lethally irradiated once with 10 GY using a PXi Precision X-Ray X-RAD 320. Within 24 hours of irradiation, bone marrow was harvested from sex matched donor mice (6-15 weeks old, indicated genotypes), and total cell numbers were quantified. Irradiated mice were randomized to treatment groups and anesthetized with isoflurane before receiving 4x10⁶ bone marrow cells via retro-orbital injection (100 μL final volume). For studies utilizing Mx1-Cre, mice were injected intraperitoneally (i.p.) with 200 μg/mouse/day polyinosinic:polycytidylic acid (pIpC) (Sigma #P1530) twice, allowing one day between treatments three weeks after BMT.

Atherosclerosis Studies

For atherosclerosis studies, sample sizes were powered based on an assumption of a coefficient of variation about 25-30% that will enable an 80% probability of detecting a >33% difference in atherosclerosis lesion size, necrotic core formation, and macrophage proliferation with P<0.05. 4-5 weeks following BMT, Ldlr^−/− recipient mice fed with Western-type diet (WTD) (Envigo #TD.88137) or WTD containing tamoxifen at 0.5g/kg by weight (Envigo #TD. 130889), as indicated. ApoE^−/− (9-14 weeks old) mice were fed WTD containing tamoxifen for 3 weeks followed by 10 weeks standard WTD. For cisplatin treated mice, mice were fed high fat diet for 10 weeks, injected with Cisplatin (4mg/kg/week) for three weeks, followed by 10 weeks high fat diet.

Mice were anesthetized with isoflurane and bled serially by cheek puncture for analysis of blood parameters at indicated timepoints. For terminal collection, mice were perfused with 20mL PBS (Corning #21-040-CM) and aortic roots were fixed for 48 h in 4% paraformaldehyde, then embedded in paraffin and sectioned for histological analysis. Complete blood count (CBC) was performed utilizing a Heska Elements HT5 blood analyzer.

Oil Red O Staining

Atherosclerotic lesion area was quantified on the intimal surface of the aorta using an en face Oil Red O (ORO) staining method in accordance with established guidelines. Briefly, aortas were dissected, fixed in neutral buffered formalin, and cleared of adventitial and perivascular adipose tissue under a stereomicroscope. Oil Red O stock solution (Thermo Fisher, #A12989.22) was diluted in dH₂O and filtered through 0.45 μm sterile syringe filters prior to use. Aortas were stained for 60 min at room temperature, differentiated in 60% isopropanol for 20 min, and washed three times for 5 min each. The vessels were opened longitudinally using micro-dissecting spring scissors and pinned lumen side up in a wax dish containing 1× PBS (Corning, #21-040-CV) to prevent dehydration. En face images were acquired using a Leica FDC7000T digital microscope at 0.8× magnification. Lesion areas were quantified using ImageJ software within defined regions of the aorta, including the ascending aorta, aortic arch, and the descending Aorta. Atherosclerotic burden was expressed as the percentage of Oil Red O–positive area relative to the total aortic intimal surface area.

Histological Analysis

Aortic root lesions were serially sectioned at 6-μm thickness. Sections were stained with Hematoxylin (Epredia #7221) and Eosin (Epredia #71304) and lesion area and necrotic core area were determined in the aortic root over a 300-μm span for a total of six sections 60μm apart. Slides were imaged with a Nikon Eclipse Ci microscope and NIS-Elements software (Version 5.42.04). Lesion and necrotic core area were quantified in a blinded manner, and the average of six slides was used to determine area. Necrotic core area was presented as a percentage with the total lesion area being used as the denominator.

Aortic root sections (one per mouse) removed from the thickest sections (mouse to mouse) were stained with picrosirius red kit (Polysciences #24901-500). Cap thickness was measured in the largest lesion of each section at 70-μm intervals. The average cap thickness was reported in length units. Collagen content was measured in each lesion of the same section utilizing the Color Deconvolution tool in FIJI Software²⁹. All lesion analysis was conducted with FIJI software²⁹.

Immunohistochemistry

Aortic root slides from similar regions (mouse to mouse) were baked at 65°C for 45 mins, deparaffinized and rehydrated. Antigen retrieval was performed using a citric acid-based solution (Vector Labs #H-3300) and heated under pressure for 15 mins. Slides were cooled, washed using PBS and PBS containing 0.1% Tween 20 (PBST) (Thermo Scientific #J20605-AP), blocked with 10% goat serum (Biolegend #927503) for 1 hr at room temperature (RT), then incubated with primary antibodies overnight at 4°C: MAC2 (Cedarlane #CL8942AP 1:500), ACTA2 (Sigma Aldrich #A2547 1:500), Cleaved Gasdermin D (Cell Signaling Technology #10137 1:200), and Phospho-Histone H2A.X (Cell Signaling Technology #9718 1:200). Rabbit IgG (Invitrogen #31235), Rat IgG (Invitrogen #31933), and Mouse IgG (Invitrogen #31903) isotype controls were used as negative controls at concentrations matched to their corresponding primary antibodies. Slides were then washed with PBS and PBST, incubated with secondary antibodies (Invitrogen) for 1hr at RT, washed with PBS and PBST, then mounted with Prolonged Gold Anti-Fade Reagent with DAPI (Thermo Scientific #P36931). Slides were imaged on a Nikon Crest LFOV Spinning Disk/C2 Confocal wide-field microscope. All immunofluorescence image analysis was performed in a blinded manner utilizing FIJI software. Signal thresholds in FIJI were established based on isotype controls.

Bone Marrow Derived Macrophage Cultures

Primary bone marrow derived macrophage cultures were generated by flushing bone marrow from hindlimbs with Hanks balanced salt solution (HBSS) (Gibco #14170-112) and filtered in 40-μm cell filters on ice. Cells were centrifuged at 800xg for 10 min at 4°C and resuspended in DMEM (Corning #10-013-CM) with 10% FBS (Gibco #A5669501) containing M-CSF (Thermo Fisher #315-02-10UG). Bone marrow cells were incubated in 100mm x 15mm non-tissue culture treated petri dishes for 7 days and plated into new dishes overnight for downstream assays.

Inflammasome Activation Studies

For NLRP3 Inflammasome assays, differentiated BMDMs were incubated with LPS (20ng/mL) Invitrogen # tlrl-peklps) for 1 h followed by NLPR3 inflammasome agonist ATP (1mM, 2.5mM, or 5mM) for 1 h or AIM2 inflammasome agonist poly(dA:dT, 0.5ug or 2ug/mL) (InvivoGen, #tlrl-patn-1) and Lipofectamine 2000 (Invitrogen, #11668019) diluted in OptiMEM (Gibco, #31985-062) for 6 h. Media was collected for IL-1β (R&D Systems #DY401) with ELISA assay per manufacturer’s instructions. LDH activity was quantified with CyQUANT^™ LDH Cytotoxicity Assay (Invitrogen #CD20301). Absorbance values were measured with a Molecular Devices Spectramax 190 Microplate Reader and SoftMax Pro 7.0 software.

For studies utilizing Cisplatin, differentiated BMDMs were incubated with LPS (5ng/mL) and Cisplatin at indicated concentrations for 24 h. Media was collected for IL-1β (R&D Systems #DY401) with ELISA assay per manufacturer’s instructions. LDH activity was quantified with CyQUANT^™ LDH Cytotoxicity Assay (Invitrogen #CD20301). Absorbance values were measured with a Molecular Devices Spectramax 190 Microplate Reader and SoftMax Pro 7.0 software.

FACS Sorting of Blood Monocytes

Whole blood samples from cisplatin-treated mice were collected and lysed with RBC Lysis Buffer (BioLegend #420302). Samples were centrifuged at 800xg for 5 mins followed by resuspension in 1x PBS (Corning #21-040-CV). Cell suspensions were filtered utilizing a 40μm cell strainer then centrifuged at 800xg for 10 mins then resuspended in 1mL FACS buffer containing anti-mouse/human CD11b- Pacific Blue^™ (BioLegend #101224 (1:100), APC anti-mouse CD115 (CSF-1R) (BioLegend #135510 1:100), or PE anti-mouse Ly-6G/Ly-6C (Gr-1) (BioLegend #108408 1:100) and incubated for 20 mins on ice. Cell suspensions were washed with FACS buffer and centrifuged at 800xg for 5 mins and resuspended in 200μL FACS buffer. Cells were analyzed and sorted on a BD FACS Aria Fusion 1 using a 100μm nozzle.

Monocytes were defined based on forward and side scatter properties, CD11b positive, CD115 positive, and Gr-1 negative. Sorted cells were collected directly into TRI Reagent (Zymo Research #R2050-1-200).

RNA Isolation and cDNA Generation

RNA was isolated from sorted monocytes utilizing Direct-zol RNA MiniPrep (Zymo Research #R2050) according to manufacturer’s instructions. cDNA was generated using Maxima First Strand cDNA Synthesis Kit for RT-qPCR (Thermo Scientific #K1641). RT-qPCR was performed utilizing Fast SYBR^™ Green Master Mix (Applied Biosystems^™ #4385612) according to manufacturer’s instructions on a QuantStudio 5 Real-Time PCR system.

Analysis of free cholesterol and cholesteryl esters in BMDMs

BMDMs were cultured in 6-well plates at 1 million cells/well. After 24 hours, cells were washed twice with ice-cold PBS and lysed in the wells with 10% NP-40. Cholesterol extraction was performed by adding the cell lysis solution to 450μL of chloroform-isopropanol solution to give a final ratio of chloroform:isopropanol:NP-40 (7:11:0.1). The extract was then centrifuged at 15,000g for 10 minutes. Avoiding the pellet, the liquid (organic phase) was transferred to a new tube and heated at 50°C for 15 minutes to remove the chloroform, then spun in a speedvac until all organic solvent was removed. Dried lipids were resuspended in the Amplex Red Cholesterol Assay reaction buffer (Invitrogen A12216). Total cellular cholesterol was measured following the kit’s instructions.

Serum Analysis

Mouse serum was isolated using centrifugation of blood at 10,000xg for 10 min at 4°C. Serum cholesterol was quantified using a Cholesterol E assay (FUJIFILM #999-02601). Serum IL-1ß levels were quantified using Mouse IL-1beta/IL-1F2 Quantikine® HS ELISA Kit (R&D Systems #MHSLB00) according to manufacturer’s instructions.

Statistics

Sample sizes are indicated in the figure legends, with n values representing individual mice. Statistical analyses were performed using GraphPad Prism 9. Data normality was assessed using the Kolmogorov–Smirnov and Shapiro-Wilk test. For comparisons between two groups, unpaired two-sided Student’s t-tests were used when data met assumptions of normality; otherwise, the nonparametric Mann–Whitney test was applied. For comparisons involving more than two groups, one-way or two-way ANOVA was used as appropriate. When multiple comparisons were performed, P values were adjusted using the Bonferroni correction. Sex was considered as a biological variable in analyses of the ApoE^−/−-Scl-Ppm1d^T476 cohort. All tests were two-sided unless otherwise specified.

Results

Ppm1d-mutant macrophages have increased AIM2 inflammasome activation

CH mutations can occur in genes restricted to broad cellular functions like epigenetic regulation, signal transduction, and DNA damage response. Within multiple functional pathways CH mutations have been associated with elevated inflammasome activation^1,9,12,30 and increased atherosclerosis³¹. Utilization of CRISPR-Cas9 to impart murine macrophages with Ppm1d mutations led to increased NLRP3 inflammasome activation³². To examine if mice with conditional expression of Ppm1d mutations have increased inflammasome activation, we generated mice with a conditional mutation of Ppm1d at T476 in exon 6²¹, which, under the control of Cre-Recombinase, leads to the expression of a truncated Ppm1d protein similar to mutations found in humans^15,20,22. Mx1-Cre mice were crossed to Ppm1d ^T476* transgenic mice²¹ to generate Mx1-Ppm1d^T476*, which express Ppm1d^T476* in all hematopoietic cells following administration of Polyinosinic-polycytidylic acid (pIpC)⁹. Mx1-Ppm1d^T476* bone marrow derived macrophages (BMDMs) were subjected to NLRP3 and AIM2 inflammasome stimulation. While low concentrations of ATP showed a trend toward decreased NLRP3 inflammasome activation, overall, we found no changes in NLRP3-induced inflammasome activation marked by IL-1β production (Figure 1a). Since PPM1D is involved in DNA damage response and repair we considered the possibility that Ppm1d mutant macrophages may have altered AIM2 inflammasome activation; AIM2 is primarily activated by dsDNA³³. Stimulation of Mx1-Ppm1d^T476* BMDMs with dsDNA fragments (pdAdT) resulted in increased IL-1β secretion relative to control BMDMs, suggesting increased AIM2 inflammasome activation (Figure 1b). Our previous work identified prominent AIM2 inflammasome activation in Jak2 and Asxl1 mutant macrophages^9,12. In humans both JAK2 and ASXL1 CH showed increased coronary artery disease (CAD) risk when AIM2 expression was predicted to be elevated based on a polygenic prediction model for expression in the UK Biobank¹². Therefore, we examined risk in patients with PPM1D mutations in UK Biobank. Unlike JAK2 and ASXL1, PPM1D CH did not show increased risk of CAD based on predicted AIM2 expression (Figure S1, Table S1). Interestingly, patients with PPM1D CH did show an increased risk of bradyarrhythmia when they have higher predicted NLRP3 expression (Figure S1), and a recent report showed a strong association of PPM1D CH with arrhythmia³⁴.

Figure 1. Ppm1d CH does not promote atherosclerosis.

IL-1β quantification in the media from bone marrow derived macrophages (BMDMs) treated with LPS (20ng/mL) for 1 hour followed by a. NLRP3 Inflammasome agonist ATP for 1hour or b. AIM2 Inflammasome agonist pdAdT for 6 hours (n=6). c. Scheme of experimental design for Ppm1d-CH atherosclerosis study. d. Percent Scl-Ppm1d^T476* CD45.2⁺cells in blood after high fat diet feeding. e. Representative hematoxylin and eosin (H&E) images of aortic root lesions, dashed line indicate lesion area (black) and necrotic core (yellow) scale bar, 100μm. f. Quantification of lesion volume (n=15 20% control, n=15 20% Scl-Ppm1d^T476*) and g. percent Necrotic Core area (n=15 20% control, n=15 20% Scl-Ppm1d^T476*). h. Representative immunohistochemical (IHC) analysis of pγH2AX in lesions, yellow dashed lines indicate lesions, yellow arrows indicate pγH2A.X events, scale bar is 100μm. i. Quantification of pγH2AX events in lesions (n=15). j. Representative IHC image of lesions stained of cleaved Gasdermin D (GSDMD), yellow arrow indicates cleaved GSDMD, white line indicates scale 20μm. k. Quantification of GSDMD⁺ cells in lesions (n=15 20% control, n=14 20% Scl-Ppm1d^T476*). Mean ± s.e.m.; two-tailed t-test (f, g, i, & k), two-way ANOVA followed by Bonferroni multiple comparison post hoc test (a, b, & d).

Ppm1d CH does not promote atherosclerosis in mice.

To determine if Ppm1d mutations promote atherosclerosis, we generated mice that express Ppm1d^T476* in all hematopoietic cells following tamoxifen administration by crossing Ppm1d^T476* mice to Scl-ERT2-Cre mice (Scl-Ppm1d^T476*)^11,35. We then modeled Ppm1d CH by transplanting mixtures of 80% CD45.1⁺ WT bone marrow, with either 20% CD45.2⁺ Scl-Ppm1d^T476* or 20% CD45.2⁺ littermate control bone marrow into lethally irradiated atherosclerosis-prone Ldlr^−/− mice (Figure 1c). After recovery, Ppm1d^T476* expression was induced by tamoxifen administration followed by a 16-week high fat diet (HFD) to promote atherosclerosis (Figure 1c). Following 16 weeks of feeding, Ppm1d mutant cells comprised ~10% of monocytes, neutrophils and T-cells in blood (Figure 1d), equivalent to a VAF of 5% in PPM1D patients. Ppm1d mutations in bone marrow did not alter blood cells counts or spleen weight, consistent with previous reports^21,22 (Figure S2a–g), suggesting no major alteration in hematopoiesis. In addition, body weight and serum cholesterol were unchanged (Figure S2h&i).

Examination of atherosclerosis in proximal aortas revealed no change in lesion size (Figure 1e&1f). Because PPM1D mutations increase cellular survival following DNA damaging therapies^21,22 we examined necrotic core area in lesions as a proxy of altered macrophage survival. Ppm1d mutations did not alter necrotic core formation (Figure 1e&1g). Hematopoietic cells interact with stromal cells to support matrix production and dynamics. Therefore, we examined collagen content in plaques. Ppm1d mutant mice displayed no Ppm1d-dependent alteration in collagen content or cap thickness (Figure S2j–l). In addition, immunohistochemical (IHC) analysis of smooth muscle cell (SMC) and macrophage density (Figure S2m–o) both indicated no Ppm1d-dependent change.

Intracellular DNA damage can activate the AIM2 inflammasome, leading to Caspase 1 activation and proteolytic cleavage of IL-1β, IL-18, and Gasdermin D (GSDMD)³⁶. To determine whether Ppm1d CH increases DNA damage within lesions, we stained for 8-hydroxy-2’deoxyguanosine (8-OHdG) a marker of oxidized DNA damage and pγH2Ax, a marker of DNA double strand breaks. We found a significant increase in DNA double strand breaks, marked by elevated pγH2Ax (Figure 1h&i), with no change in 8-OHdG (Figure S2p&q). Despite this increase in DNA damage, we detected no evidence of enhanced inflammasome activation, as assessed by cleaved GSDMD³⁷ staining (Figure 1j&k).

Together, these findings suggest that in mice, Ppm1d-CH does not alter atherosclerosis progression, or plaque composition.

Hematopoietic expression of Ppm1d mutations in Apoe^−/− mice does not alter atherosclerosis.

We next considered the possibility that a more severely altered lipid state may be required to identify the impact of Ppm1d on atherosclerosis. Prior studies demonstrated that complete deletion of Ppm1d in Apoe^−/− mice resulted in decreased atherosclerosis through altered autophagy³⁸ and decreased cholesterol accumulation in macropahges³⁸. Therefore, we examined if Ppm1d mutant macrophages have altered cholesterol levels. Unlike Ppm1d knockout macrophages³⁸, Ppm1d mutant macrophages had no alteration in cholesterol accumulation (Figure S3a).

To further assess if Ppm1d mutations may promote atherosclerosis in Apoe^−/− mice we generated ApoE^−/−Scl-Ppm1d^T476* mice, by crossing Scl-Ppm1d^T476* mice onto an ApoE^−/− background. 8–12-week-old male and female ApoE^−/−Scl-Ppm1d^T476* were fed a HFD containing tamoxifen for three weeks to induce Ppm1d^T476* expression, followed by an additional 10 weeks of high fat diet feeding to promote atherosclerosis (Figure 2a). Similar to Scl-Ppm1d^T476* on a WT background ApoE^−/−Scl-Ppm1d^T476* mice did not have altered circulating blood cell counts or spleen weight (Figure S3b–h), suggesting no major alteration in hematopoiesis. No significant change in body weight was observed; however, serum cholesterol was significantly increased in female ApoE^−/−Scl-Ppm1d^T476* mice (Figure S3i&j).

Figure 2. Ppm1d mutations do not promote atherosclerosis in ApoE^−/− mice.

a. Scheme of experimental design for ApoE^−/−Scl-Ppm1d^T476* atherosclerosis study. b. Representative Oil red O (ORO) staining of aorta. c. quantification of ORO-stained aortas, (n=6 males, female n=8 control, n=6 males, female n=6 ApoE^−/−Scl-Ppm1d^T476*). d. Representative H&E images of aortic root lesions, dashed line indicate lesion area (black) and necrotic core (yellow), scale bar 100μm. Quantification of e. lesion volume (n=8 male control, n=9 female control, n=14 male ApoE^−/−Scl-Ppm1d^T476*, n=8 female ApoE^−/−Scl-Ppm1d^T476*) and f. percent necrotic core area (n=8 male control, n=9 female control, n=14 male ApoE^−/−Scl-Ppm1d^T476*, n=8 female ApoE^−/−Scl-Ppm1d^T476*). g. Representative IHC images of pγH2Ax in lesions, scale bar 100μm. h. quantification of pγH2A.X+ cells in lesions (n=8 male control, n=9 female control, n=14 male ApoE^−/−Scl-Ppm1d^T476*, n=7 female ApoE^−/−Scl-Ppm1d^T476*). i. Representative IHC images of GSDMD in lesions, scale bar 100μm. j. quantification of cleaved Gasdermin D (GSDMD)⁺ cells in lesions (n=8 male control, n=9 female control, n=14 male ApoE^−/−Scl-Ppm1d^T476*, n=8 female ApoE^−/−Scl-Ppm1d^T476*) k. Terminal serum quantification of IL-1β. (n=8 male control, n=9 female control, n=14 male ApoE^−/−Scl-Ppm1d^T476*, n=8 female ApoE^−/−Scl-Ppm1d^T476*) Mean ± s.e.m.; two-way ANOVA followed by Bonferroni multiple comparison post hoc test (c, e, f, h, j, & k).

Oil Red O (ORO) staining of aortas revealed no alteration in lipid accumulation (Figure 2b&c). Consistently histological analysis of proximal aortas showed no change in lesions area in mice with Ppm1d mutations in all hematopoietic cells (Figure 2d&e). Because Ppm1d mutations may alter features commonly associated with plaque stability like necrotic core formation, cap thickness, and macrophage density³⁹ we examined plaque composition. We found no alteration in necrotic core area (Figure 2d&f). Similarly, picrosirius red staining of plaques revealed no modification in collagen content or cap thickness (Figure S3k–m). IHC analysis of ACTA2 and MAC2 showed no alteration in SMCs or macrophages in atheromas (Figure S3n–p). Examination of DNA damage markers revealed no alteration in pγH2Ax (Figure 2g&h) or 8-OHdG (Figure S3q&r).

Furthermore, we found no evidence of altered inflammasome activation marked by cleaved GSDMD in lesions (Figure 2i&j) or circulating IL-1β concentrations (Figure 2k). Together, these findings indicate that in Apoe^−/− mice Ppm1d mutations in hematopoietic cell are not sufficient to promote inflammasome activation or exacerbate atherosclerosis.

Cisplatin does not selectively aggravate atherosclerosis in mice with monocyte/macrophage restricted expression of Ppm1d mutations.

PPM1D CH occurs at a higher rate following radiation and platinum-based chemotherapy²⁰. This enrichment may be due to increased PPM1D-mediated suppression of DNA damage response and cell death. In macrophages, DNA damage can promote an inflammatory response⁴⁰ and we have previously shown that enhanced DNA damage promotes AIM2 inflammasome activation and atherosclerosis in JAK2-CH⁹. Since Ppm1d mutant macrophages had increased AIM2 inflammasome activation in response to pdAdT (Figure 1b) we hypothesized that Ppm1d mutant macrophages treated with DNA damaging agents like cisplatin, will have elevated AIM2 inflammasome activation. Administration of cisplatin to Ppm1d mutant macrophages resulted in elevated IL-1β secretion compared to control (Figure 3a), suggesting increased inflammasome activation. Although LDH release increased with increasing doses of cisplatin, no genotype-dependent alteration was observed, suggesting that increased IL-1β secretion from Ppm1d mutant macrophages was not caused by increased cisplatin-induced cell death (Figure 3b).

Figure 3. Cisplatin Does Not Promote Atherosclerosis in Mice Selectively Expressing Ppm1d Mutations in Monocytes/Macrophages.

BMDMs were treated with 5ng/mL LPS and Cisplatin at the indicated concentrations for 24 hours and a. IL-1β and b. LDH activity was quantified in media (n=6-8). c. Scheme of experimental design for Cx3cr1-Ppm1^dT476* atherosclerosis studies. d. Representative H&E images of aortic root lesions, dashed line indicate lesion area (black) and necrotic core (yellow), scale bar 100μm. Quantification of e. lesion area (n=7 control, n=7 Cx3cr1-Ppm1^dT476*) and f. percent Necrotic Core area (n=7 control, n=7 Cx3cr1-Ppm1^dT476*). g. IHC staining of aortic root plaques, yellow dashed lined indicate plaque, scale bar 100μm. h. Percent of SMC area (n=6 control, n=7 Cx3cr1-Ppm1^dT476*) and i. percent macrophage area (n=6 control, n=7 Cx3cr1-Ppm1^dT476*).j. Representative IHC images of pγH2Ax in lesions, white line indicate 20μm. k. quantification of pγH2Ax+ cells in lesions (n=5 control, n=7 Cx3cr1-Ppm1^dT476*). l. Representative IHC images of GSDMD in lesions, scale bar 100μm. m. quantification of cleaved Gasdermin D (GSDMD)⁺ cells in lesions (n=5 control, n=6 Cx3cr1-Ppm1^dT476*). qPCR analysis of CD11b⁺ cells isolated from lesions; n. Aim2, o. Il1b, and p. Il6 (n=15 Control, n=14 Cx3cr1-Ppm1^dT476*). Mean ± s.e.m.; two-tailed t-test (e, f, h, i, k, m-p), two-way ANOVA followed by Bonferroni multiple comparison post hoc test (a & b).

Based on these findings, we hypothesized, that in established lesions harboring macrophages with Ppm1d mutations, enhanced DNA damage will promote AIM2 inflammasome activation and atherosclerosis. To test this, we generated monocyte/macrophage restricted Ppm1d mutant mice by crossing Ppm1d^T476* with Cx3cr1-Cre mice (Cx3cr1-Ppm1d^T476*). Cx3cr1-Cre⁺ mice have suppressed Cx3cr1 expression⁴¹, therefore Cx3cr1-Cre littermates were used as controls. Bone marrow was transplanted into lethally irradiated Ldlr^−/− mice. Mice were subjected to a 10-week high fat diet to establish atherosclerosis and then were treated with cisplatin for three weeks, followed by an additional 10 weeks of high fat diet feeding (Figure 3c). This approach seeks to model individuals receiving therapy with preexisting atherosclerosis. No changes in blood cell counts, body weight, or serum cholesterol were found following cisplatin treatment (Figure S4a–h). H&E analysis of lesions revealed no alteration in plaque size or necrotic core area (Figure 3d–f). Picrosirius red staining showed no difference in collagen content or cap thickness (Figure S4i–k). IHC analysis of ACTA2 and MAC2 showed no change in SMC or macrophage area in lesions (Figure 3g–i). Cisplatin administration did not lead to genotype dependent changes in DNA damage marked by 8-OHdG (Figure S4l&m) or pγH2Ax (Figure 3j&k) in Ppm1d lesions. Similarly, Ppm1d mutations did not alter cleaved GSDMD in lesions (Figure 3l&m). To further determine if Ppm1d mutant macrophages were intrinsically altered in plaques we isolated CD11b⁺ cells from digested aortas by fluorescence-activated cell sorting (FACs). qPCR analysis revealed no alteration in genes related to inflammation (Aim2, Il1b, or Il6) (Figure 3n–p) or DNA proliferation (Pcna and Mki67), DNA damage repair (Neil1 and Ogg1) (Figure S4n–q). Together these data indicate that in mice, even under cytotoxic stress Ppm1d mutations in macrophages do not promote inflammasome activation or atherosclerosis.

To determine if cisplatin interacts with Ppm1d mutations broadly in hematopoietic cells to promote atherosclerosis, we expressed Ppm1d^T476* in all hematopoietic cells using an Mx1-Cre (Mx1-Ppm1d^T476*). Using a similar approach as with Cx3cr1-Ppm1d^T476* mice, we transplanted Mx1-Ppm1d^T476* bone marrow into Ldlr^−/− mice and fed the mice a HFD to promote atherosclerosis (Figure 4a). Once plaques were established (after 10 weeks of HFD feeding) mice were subjected to three weeks of cisplatin treatments followed by and addition 10 weeks HFD feeding. Ppm1d mutant mice showed no changes in blood cell counts (Figure S5a–f). In addition, mouse weight and serum cholesterol were unchanged (Figure S5g&h). Oil Red O staining of aortas showed no change in lipid accumulation (Figure 4b&c). Consistently proximal aorta lesion area was unchanged in Mx1-Ppm1d^T476* mice (Figure 4d&e). Necrotic core area was similarly unaltered (Figure 4d–f), as well as collagen content and fibrous cap thickness (Figure S5i–k). Similar to other models we found no change in SMC or macrophage density, (Figure 4g–i). Together, these findings suggest that in mice, Ppm1d mutations do not alter atherosclerosis in response to cisplatin.

Figure 4. Cisplatin does Not Selectively Promote Atherosclerosis in Mice with Ppm1d Mutations in all Hematopoietic Cells.

a. Scheme of experimental design for Mx1-Ppm1d^T476* atherosclerosis studies. b. Representative Oil red O (ORO) staining of aorta. c. quantification of ORO-stained aortas, (n=16 control, n=16 Mx1-Ppm1d^T476*). d. H&E images of aortic root lesions, dashed line indicate lesion area (black) and necrotic core (yellow), scale bar 100μm. Quantification of e. lesion area (n=14 control, n=18 Mx1-Ppm1d^T476*) and f. percent Necrotic Core area (n=14 control, n=18 Mx1-Ppm1d^T476*). g IHC staining of aortic root plaques, yellow dashed lined indicate plaque, scale bar 100μm. h. Percent of SMC area (n=14 control, n=18 Mx1-Ppm1d^T476*) and i. percent macrophage area (n=14 control, n=18 Mx1-Ppm1d^T476*). j. Representative IHC images of pγH2Ax in lesions, scale bar 100μm. k. quantification of pγH2Ax⁺ cells in lesions (n=14 control, n=18 Mx1-Ppm1d^T476*). l. Representative IHC images of GSDMD in lesions. m. quantification of cleaved Gasdermin D (GSDMD)⁺ cells in lesions (n=12 control, n=18 Mx1-Ppm1d^T476*). n. Ferric chloride (FeCl₃) induced arterial thrombosis (n=10 control, n=9 Mx1-Ppm1d^T476*). Mean ± s.e.m.; two-tailed t-test (c, e, f, h, I, k, m & n).

Our macrophage culture system found that Ppm1d mutant macrophages had increased cisplatin induced inflammasome activation. Therefore, we examined markers of DNA damage in lesions of mice treated with cisplatin. We found no change in 8-OHdG (Figure S5l&m) or pγH2Ax (Figure 4j&k). Furthermore, inflammasome activation marked by cleaved GSDMD was unaltered (Figure 4l&m). Suggesting even with cisplatin administration Ppm1d mutations do not exacerbate inflammasome activation within lesions.

We next investigated whether cisplatin induces systemic inflammasome activation in vivo. Mice were injected with cisplatin, and serum IL-1β concentrations were quantified 24 hours later. Cisplatin increased serum IL-1β levels in both control and Ppm1d mutant mice; however, Ppm1d mutations did not further elevate IL-1β levels (Figure S5n). Neither cisplatin or Ppm1d mutations altered the proportion of Ly6c^High monocytes (Figure S5o). Furthermore, FACS isolated monocytes showed no change in Il1b, Nlrp3, Aim2, and Il10 expression (Figure S5p–s). Together, these data indicate that while acute cisplatin administration can induce inflammasome activation in vivo, Ppm1d-mutations does not amplify this response in mice.

Ppm1d mutations do not promote arterial thrombosis.

Atherosclerotic complications can result from plaque rupture leading to thrombotic events⁴². Therefore, we subjected mice with Mx1-Ppm1d^T476*bone marrow to a well-established model of arterial thrombosis to determine if Ppm1d mutations alter thrombotic responses. Using a FeCl₃-induced arterial thrombosis model,⁴³ we found that Ppm1d mutations in hematopoietic cells does not alter arterial thrombosis (Figure 4n).

Discussion

Human population genetics identified an association between PPM1D CH and atherosclerosis, however the cellular and molecular mechanisms underlying this association are not well understood. Here we subjected mice with Ppm1d mutations in hematopoietic cells to four unique atherosclerosis models, spanning 13-23 weeks of HFD feeding and found that Ppm1d mutations were not sufficient to worsen disease. Together these findings suggests that the association between PPM1D and ASCVD may not be causative, however studies in humans will be required to rigorously rule out a causal role.

As a composite, CH is associated with a ~40% higher all-cause mortality⁴. This risk has been attributed to increased cardiovascular disease¹ as well as other diseases including cancer². For CAD, emerging studies suggest that DNMT3A-driven CH may exert only modest effects, whereas TET2 and JAK2 consistently associate with higher risk, underlying the importance of examining each variant independently. Because CH frequency is relatively rare, occurring in ~20% of the population aged 70 years⁵, less common variants like PPM1D are more difficult to study. Thus, larger populations may be required to precisely identify drivers of PPM1D- risk.

Although CAD patients with PPM1D-CH have increased all-cause mortality⁴⁴, factors underlying this association are unclear. Our data suggest, at least in mice, that the underlying pathogenesis of atherosclerosis is not altered by Ppm1d mutations. However, secondary complications including stroke, arrythmia and heart failure may be increased in the setting of established atherosclerosis. Notably, stroke⁴⁵, heart failure³² and arrythmia³⁴ have all been associated with PPM1D CH, Our murine models were not designed to detect these outcomes, and future studies may be required to determine whether PPM1D mutations influence cardiovascular complications independent of plaque burden.

Unlike other common CH variants like DNMT3A, TET2, ASXL1, and JAK2 which are directly linked to hematological malignancy, the role of PPM1D mutations on cellular function and malignancy is not well defined⁴⁶. Human and mouse studies show that PPM1D promotes cell survival and proliferation following genotoxic stress^21,22. This is likely due to the suppression of the DDR, which may lead to genomic instability, but in the short-term aids in cell survival. In the absence of DNA damage, Ppm1d-mutant mice show no overt malignant behavior (consistent with our findings). This suggest that the association of PPM1D CH with atherosclerosis may merely be a marker of exposure to DNA damage, a process that itself can promote atherosclerosis^47,48.

Prior work using Ppm1d-deficient mice demonstrated protection from atherosclerosis attributed to enhanced autophagy, improved cholesterol efflux, and altered systemic metabolism³⁸. In contrast, our mice harbor a conditional Ppm1d truncation mutation (T476*) similar to mutations frequently found in patients with PPM1D CH, which can enhance PPM1D activity. Unlike deletion models, we observed no changes in cholesterol accumulation, body weight, or aortic lipid deposition. Collectively, these findings suggest that complete loss of PPM1D and CH-associated gain-of-function truncations exert distinct biological effects; CH-associated truncating mutations do not exert opposing effects on atherosclerosis through altered lipid accumulation.

Ppm1d mutations have also been shown to promote heart failure in mice³². However, this approach utilized CRISPR editing of hematopoietic cells to model Ppm1d CH and Ppm1d macrophages had extensive genomic instability along with prominent inflammation. Our data similarly found an increase in DNA-damage associated inflammation in Ppm1d mutant macrophages in vitro. However, in vivo we found minimal evidence of sustained DNA damage within lesions, which was likely insufficient to promote inflammasome activation and atherosclerosis, even when mice were treated with cisplatin. In contrast, Jak2-driven CH increased atherosclerosis through AIM2 inflammasome activation likely due to DNA damage associated with mitochondrial ROS generation or increased replicative stress^9,37,49. It remains possible that in humans, decades of cumulative DNA damage could interact with PPM1D mutations to promote genotype-dependent increases in IL-1β production.

Consistent with our findings, analysis of TOPMed data did not identify associations between PPM1D-CH and plasma lipid profiles or common inflammatory markers⁵. However, it is possible that enhanced inflammatory signaling is particularly relevant under specific contexts or at specific sites of injury. For instance, PPM1D CH was strongly associated with arrythmia, even when controlled for previous cancer therapy³⁴ as well as chronic obstructive pulmonary dysfunction⁵⁰ and stroke⁴⁵. Together these data suggest that specific pathological contexts may modify PPM1D risk.

A key limitation of our study is reliance on mouse models. Although they recapitulate key Ppm1d-driven changes in hematopoiesis^21,22, they may not fully capture the chronic DNA damage and inflammatory milieu characteristic of human atherosclerosis⁵¹, which unfolds over decades. Despite examining both short- and long-term disease models and introducing exogenous DNA damage with cisplatin, we did not detect Ppm1d-dependent changes in plaque development. However, prolonged exposure in humans may produce cumulative effects not evident in mice.

Overall, our data indicate that PPM1D mutations do not promote atherosclerosis in mice. In humans, the observed, although inconsistent, epidemiologic association may reflect confounding by prior genotoxic exposure, which both drives expansion of PPM1D-mutant clones and independently increases ASCVD risk. Future large-scale genetic and mechanistic studies will be necessary to further disentangle these relationships.

Clinical Implications.

Clonal hematopoiesis (CH), a condition common in older adults, arises when hematopoietic stem cells acquire somatic mutations in genes such as DNMT3A, TET2, ASXL1, PPM1D, and JAK2. CH has been associated with increased coronary artery disease (CAD) and all-cause mortality. Among CH subtypes, PPM1D mutations are more prevalent in patients exposed to radiation or cisplatin, and epidemiologic studies have linked PPM1D-CH to CAD.

Here, we demonstrate that Ppm1d mutations in hematopoietic cells are not sufficient to promote atherosclerosis in murine models. These findings suggest that the observed association between PPM1D-CH and CAD may reflect confounding by prior genotoxic exposure, which both drives clonal expansion and independently increases cardiovascular risk. Alternatively, these associations may be explained by an increased risk of secondary cardiovascular complications such as stroke, arrhythmia, and heart failure in the setting of established atherosclerosis.

Non-standard Abbreviations and Acronyms

8-OHdG

8-hydroxy-2’deoxyguanosine

ASCVD

Atherosclerotic Cardiovascular Disease

CAD

Coronary Artery Disease

ASCVD

Atherosclerotic Cardiovascular Disease

CH

Clonal Hematopoiesis

CHIP

Clonal Hematopoiesis of Indeterminate Potential

DDR

DNA Damage Response

GSDMD

Gasdermin D

HFD

High Fat Diet

HSCs

Hematopoietic Stem Cells

IHC

Immunohistochemistry

pγH2AX

Phosphorylated Histone H2A.X

SMC

Smooth Muscle Cell

VAF

Variant Allele Fraction