- Email: [email protected]
Alterations in macrophage phenotypes in experimental venous thrombosis
FROM BENCH TO BEDSIDE From the American Venous Forum
Alterations in macrophage phenotypes in experimental venous thrombosis Katherine A. Gallagher, MD, Andrea T. Obi, MD, Megan A. Elfline, MS, Emily Hogikyan, Catherine E. Luke, LVT, Samuel Henke, Dawn Coleman, MD, and Peter K. Henke, MD, Ann Arbor, Mich Objective: Macrophages are involved in venous thrombus (VT) resolution and vein wall remodeling. This study was undertaken to identify variations in macrophage phenotypes in thrombi and vein wall in multiple models of VT to clarify the natural history of macrophage polarization in clearance of VT. We also sought to demonstrate the feasibility of macrophage phenotyping in human VT. Methods: Established murine models of VT were used to mimic the clinical spectrum of human VT (stasis and nonstasis models). Vein wall and thrombi were isolated at acute (2 days) or chronic (6-21 days) time points and analyzed by Bio-Plex assay (Bio-Rad, Carlsbad, Calif) for cytokines (interleukin [IL]-1b, IL-6, IL-10, IL-12), by immunohistochemistry for “M1-like” (IL-12) or “M2-like” (arginase 1 [Arg-1]) markers, and by histology for intimal thickness and collagen content (Sirius red staining). Bone marrow was harvested from animals 2 days after undergoing sham, stasis, or nonstasis surgery. Macrophages were skewed toward M1 using lipopolysaccharide, and RNA analysis was done for inflammatory cytokine genes (IL-1b, IL-12). Human blood samples were similarly analyzed with reverse transcription polymerase chain reaction for macrophage polarization markers (CD206, inducible nitric oxide synthase, CCR2) and thrombi with immunohistochemistry (inducible nitric oxide synthase, Arg-1). Results: Stasis (chronic) and nonstasis (acute and chronic) thrombi were characterized by a predominance in antiinflammatory (M2) macrophages (n [ 4-5/group; P < .05). Larger thrombi were found in the stasis model at both time points (n [ 3; P < .01), correlating with decreased intrathrombus inflammatory (M1) cytokines (IL-1b, P [ .03; IL-12, P [ .17; n [ 4) and diminished inflammatory response From the Section of Vascular Surgery, Department of Vascular Surgery, University of Michigan Medical School. This work was supported by HL K08112897 and HL R01HL092129, Vascular Cures Foundation, University of Michigan Cardiovascular Center McKay and Heart of Champion Funds, and the Taubman Scholars Foundation. Author conflict of interest: none. Presented in part at the Twenty-sixth Annual Meeting of the American Venous Forum, New Orleans, La, February 19-21, 2014. Correspondence: Peter K. Henke, MD, 1500 E Medical Center Dr, Rm 5463, Cardiovascular Center, Ann Arbor, MI 48109-5867 (e-mail: [email protected]). The editors and reviewers of this article have no relevant financial relationships to disclose per the Journal policy that requires reviewers to decline review of any manuscript for which they may have a conflict of interest. 2213-333X Copyright Ó 2016 by the Society for Vascular Surgery. Published by Elsevier Inc. http://dx.doi.org/10.1016/j.jvsv.2016.03.005
of bone marrow-derived macrophages to lipopolysaccharide (IL-1b, P [ .03; IL-12, P [ .04; n [ 4) compared with nonstasis model. Anti-inflammatory (M2 [Arg-1]) macrophage cell counts were elevated in the post-thrombotic vein wall of stasis mice compared with nonstasis mice (acute: n [ 4, P < .05; chronic: n [ 5, P < .01), consistent with increased intimal thickness (P < .01; n [ 4-6) and collagen deposition chronically (P [ .005; n [ 12). M2-like thrombi (Arg-1, P < .05; n [ 4-7) and circulating markers (CD206, P < .05; n [ 917) decreased over time in human VT. Conclusions: Experimental VT is characterized by an antiinflammatory predominant macrophage phenotype, possibly impairing thrombus resolution, and is model dependent. Altering the M1/M2 macrophage balance may accelerate thrombus resolution and allow the development of translatable novel therapies to treat VT and to prevent post-thrombotic syndrome. (J Vasc Surg: Venous and Lym Dis 2016;4:463-71.)
Clinical Relevance: In humans, the longer time that the thrombus is in contact with the vein wall, the greater the damage. Therefore, treatment strategies designed to minimize thrombus burden at early time points are imperative. Direct modulation of inflammation has not been explored in venous thrombus, and we speculate that altering the M1/M2 balance may accelerate thrombus resolution and allow the development of translatable novel therapies to treat venous thrombus. These new therapies may allow the thrombus to be more lysable, which may represent an alternative treatment avenue for reduction of post-thrombotic syndrome and avoidance of complications related to standard therapies.
Venous thromboembolism (VTE) is a substantial health problem, resulting in significant morbidity and mortality.1,2 Even among patients undergoing adequate anticoagulation, the current mainstay therapy, complications associated with VTE are prevalent.3 Postthrombotic syndrome (PTS), a late complication associated with VTE, is associated with significant morbidity and results in added health care costs.4,5 Little progress has been made in the treatment of PTS in recent years, particularly with the recent data that active compression does not prevent PTS. Currently, no pharmacologic agent exists for the treatment or prevention of PTS. Experimentally, venous thrombus (VT) resolution and vein wall healing are mediated by leukocytes and their 463
464 Gallagher et al
JOURNAL OF VASCULAR SURGERY: VENOUS AND LYMPHATIC DISORDERS October 2016
associated chemokines, plasminogen activators, matrix metalloproteinases, and proinflammatory cytokines.6-10 The process of thrombus resolution resembles wound healing with distinct phases of neutrophil and monocyte influx, followed by resolution and fibrosis.11-13 The vein wall response to thrombus bears many similarities to the process of sterile inflammation, and leukocytes are primary drivers of this process.14 Macrophages are key mediators of the generation and resolution of inflammation and are critical for thrombus resolution as they clear cellular debris and have profibrinolytic activity.8 Macrophages are phenotypically defined by the expression of specific surface markers and gene expression. Classically activated macrophages (“M1-like”) express a defined set of proinflammatory mediators (interleukin [IL]-1b, IL-6, IL-12, inducible nitric oxide synthase [iNOS], CCR2), whereas alternatively activated macrophages (“M2-like”) with profibrinolytic activity display an anti-inflammatory phenotype (IL-10, arginase 1 [Arg-1], CD206).15 Currently, the role of macrophage functional phenotypes in VT resolution is unknown. The objective of this study was to identify variations in macrophage phenotypes and hence functions in two common models of VT at both acute and chronic time points after VT to provide a broader understanding of the human pathophysiology of this disease process. Further, we have examined thrombus from a small group of human patients after VT to demonstrate feasibility of such an approach. We hypothesized that macrophages display a decreased M1-like, proinflammatory activity and increased M2-like, profibrotic function during the venous thrombosis process, ultimately resulting in alterations in vein wall fibrosis. METHODS Venous thrombosis models. Male mice, aged 6 to 8 weeks on a C57BL6 background, were purchased from Jackson Laboratory (Bar Harbor, Me) and were used for all studies. For all procedures, mice underwent general anesthesia with isoflurane and oxygen with continuous monitoring. All animal studies were done with University of Michigan Animal Care and Use committee approval. Stasis model of VT. Mice underwent surgical ligation of the inferior vena cava (IVC) and visible contributing vessels below the renal veins, producing complete blood stasis and subsequent VT formation. This model is well characterized and consistently (>97%) produces a VT.10,11,14,16,17 IVC, thrombus, and bone marrow were harvested at acute (2 days) and chronic (6, 8, 14, or 21 days) time points after induction of stasis for analysis. Nonstasis model of VT. An endothelial injury model of VT18 was used to establish thrombus in the setting of continued blood flow. In brief, mice underwent intraluminal electrolytic injury to the IVC. A 25-gauge needle was inserted into the IVC and put in contact with the anterior IVC wall between the renal veins and iliac bifurcation. A current of 25 mAmp was applied for 15 minutes, consistently producing nonocclusive VT. IVC, thrombus, and bone marrow were harvested as listed previously.
Murine histology and immunohistologic analysis. Tissues were formalin fixed, paraffin embedded, and cut into 5-mm sections. Nonspecific sites were blocked with serum, and sections were incubated with primary antibodies to IL-12 (Abcam, Cambridge, UK; catalog #ab131039) and Arg-1 (Novus Biologicals, Littleton, Colo; catalog #NBP1-54621). A species-specific avidinbiotin complex peroxidase kit for rabbit or rat (Vector Laboratories Inc, Burlingame, Calif) was used according to the manufacturer’s instructions. Color development was performed with diaminobenzidine (Vector Laboratories). The slides were counterstained with hematoxylin (Vector Laboratories). In a blinded fashion, positive cells in five high-power fields (400) embedded in the IVC thrombus were counted and totaled. Collagen content was quantified with picrosirius red stain with Zeiss Axio M1 scope and Zeiss AxioVision software (Carl Zeiss Microimaging GmbH, Göttingen, Germany).19-21 Two images were analyzed in crossedplane polarized light from a monochromatic light source at 0 and 90 degrees to the plane of polarization to capture the birefringence of fibers extinguished in one direction. The IVC wall was outlined as a region of interest, and then the image underwent threshold segmentation to differentiate collagen from other (cells or empty space) components of the vein wall using Image J software (National Institutes of Health, Bethesda, Md). Vein wall collagen score was calculated as (% birefringent area) (measured vein wall area)/(total specimen area). To account for noncollagen vein wall changes, intimal thickness assessed from hematoxylin and eosin-stained sections was scored by a board-certified veterinary pathologist in a blinded manner.21,22 Briefly, a midsection of the thrombosed IVC was blindly analyzed on a scale of 0 to 4, with 0 representing intima as space occupied by only endothelial cells and 4 representing a greatly thickened intima containing fibroblasts, white blood cells, or hemorrhage at its widest point.22 Cell culture and real-time quantitative polymerase chain reaction. Bone marrow cells were collected by flushing mouse femur and tibia bones with RPMI medium at 2 days after the stasis or nonstasis model of the IVC. Bone marrow-derived macrophages were cultured as previously detailed. On day 6, the cells were replated, and after resting for 24 hours, they were incubated with or without interferon-g (100 ng/mL; Invitrogen, Carlsbad, Calif) and lipopolysaccharide (100 ng/mL) for 6 hours. Total RNA extraction was then performed using TRIzol (Invitrogen) per the manufacturer’s instructions. Total RNA was reverse transcribed to complementary DNA with M-MLV (Invitrogen). Reverse transcription polymerase chain reaction (PCR) was performed with TaqMan PCR mix (2) using the 7500 Real-Time PCR System (Applied Biosystems, Foster City, Calif) with primers for IL-1b and IL-12 (Applied Biosystems). Glyceraldehyde 3-phosphate dehydrogenase was used as the internal control. All standards and samples were assayed in triplicate. The threshold cycle values were used to plot a standard curve.
JOURNAL OF VASCULAR SURGERY: VENOUS AND LYMPHATIC DISORDERS Volume 4, Number 4
Gallagher et al 465
Fig 1. Thrombus size was measured and weighed after both stasis and nonstasis formation in early (2 days) and chronic (8 days) venous thrombus (VT; all n ¼ 3; *P < .01; A). Immunohistochemistry was performed using arginase 1 (Arg-1 [M1]) and interleukin-12 (IL-12 [M2]) antibodies, and cell counts of thrombus were undertaken at 2 days (n ¼ 4-5; *P < .05; B, E, G) and 8 days (n ¼ 4-5; *P < .05; C, E, G). Hematoxylin and eosin-stained sections of midpoint of thrombi were obtained for nonstasis (D) and stasis (F) models at both time points (10 representative photographs shown). Representative thrombi of nonstasis (E) and stasis (G) stained for Arg-1 at 8 days are shown at 20. All data are represented as mean 6 standard error of the mean. hpf, High-power field.
Bio-Plex. Levels of IL-1b, IL-6, IL-10, and IL-12 present in thrombus from the stasis and nonstasis VT murine models were determined using the Luminex/Bio-Plex assay 200 system (Bio-Rad, Carlsbad, Calif) per the manufacturer’s protocol. The limit of detection for each cytokine was 0.5 rg/mL. The cytokine levels were normalized to protein present in a cell-free preparation of each sample measured by Bradford assay (Bio-Rad). Human sample collection. Tissue. Human tissue was obtained from patients undergoing endophlebectomy of common femoral veins followed by recanalization of iliocaval obstruction for the treatment of debilitating PTS.23,24 All patients were treated with anticoagulation alone (no thrombolysis or procedural intervention for treatment of VT). VT occurring <1 year before acquisition of the specimen was classified as acute, whereas that occurring at later time points was considered chronic. Specimens were
formalin fixed and paraffin embedded. This protocol had been previously approved by the Institutional Review Board of the ProMedica Health System and adhered to the principles of the Declaration of Helsinki. Blood. Blood from patients presenting to our diagnostic vascular laboratory with acute VT in the popliteal, femoral, or iliac system was obtained at enrollment and 6 months after diagnosis, which is the subject of a separate published study.25 A sample of the collected blood was placed in PAXgene Blood RNA System collecting tubes (BD Biosciences, Becton Lakes, NJ) according to the manufacturer’s instructions. The messenger RNA obtained was subject to reverse transcription by incubation with oligo-dT primer and M-MLV reverse transcriptase (Life Technologies, Carlsbad, Calif) at 94 C for 3 minutes and then incubated at 40 C for 70 minutes. The complementary DNA obtained was amplified using TaqPolymerase
466 Gallagher et al
JOURNAL OF VASCULAR SURGERY: VENOUS AND LYMPHATIC DISORDERS October 2016
Fig 2. Thrombi at 2 days underwent Bio-Plex assessment for cytokine markers indicative of M1 macrophage phenotype, interleukin-1 (IL-1; n ¼ 4; *P ¼ .03; A) and interleukin-12 (IL-12; n ¼ 4; P ¼ .17; B). Bone marrow-derived macrophages from animals undergoing stasis or nonstasis thrombosis or sham operation were obtained at 2 days after thrombosis and were cultured for 6 days in standard conditions and treated with lipopolysaccharide (100 mg/mL) for 6 hours. Reverse transcription polymerase chain reaction (PCR) was performed to determine transcription levels of key M1 genes IL-1 (C; n ¼ 10; *P ¼ .03) and IL-12 (D; n ¼ 10; *P ¼ .04). Similarly, thrombi at 8 days underwent Bio-Plex assessment for M1 phenotype cytokine markers IL-1 (n ¼ 4; *P ¼ .04; E) and IL-6 (n ¼ 4; P ¼ .06; F). M2 cytokine IL-10 (n ¼ 4; * P ¼ .002; G) was additionally assessed. Top panel, All results at 2 days. Bottom panel, All results at 8 days.
(Promega, Madison, Wisc) in the Rotor-Gene quantitative real-time PCR system (Qiagen, Hilden, Germany). SYBR green intercalating dye was used to monitor levels of DNA amplification for each gene. Commercially available primers were purchased for the following genes of interest: b-actin (RefSeq# NM_001101.3), CD206(RefSeq# NM_002438), iNOS (RefSeq# NM_000625), and CCR2 (RefSeq# NM_001123396, Qiagen). The study was previously approved by the University of Michigan Institutional Review Board for Human Subjects research, adhered to the principles of the Declaration of Helsinki, and included only patients who gave informed consent. Statistical analysis. All data are shown as a mean 6 standard error of the mean. Student t-test or analysis of
variance, as appropriate, was used for comparison between the stasis and nonstasis models (GraphPad Prism, San Diego, Calif); P values # .05 were considered significant. RESULTS VT resolution is characterized by a predominantly M2 response. The natural history of macrophage phenotype in VT resolution was examined by microscopy of thrombus and vein wall at various time points after thrombosis (Fig 1). Initial VT formation in stasis and nonstasis models is characterized by smaller thrombus in the nonstasis model at both 2 and 8 days after thrombosis, consistent with perithrombus blood flow (Fig 1, A, D, and F). The VT histologic appearance between the two
JOURNAL OF VASCULAR SURGERY: VENOUS AND LYMPHATIC DISORDERS Volume 4, Number 4
Gallagher et al 467
Fig 3. Immunohistochemistry of the vein wall was performed using arginase 1 (Arg-1 [M1]; E and G) and interleukin12 (IL-12 [M2]) antibodies at 2 days after thrombosis (n ¼ 4; *P < .05; A) and 8 days after thrombosis (n ¼ 5; *P < .001; B, E, G). Changes in vein wall were assessed at chronic time points after thrombosis (6-8 days “mid” or 1421 days “late”) to assess for late changes in vein wall phenotype. Intimal thickness was scored in a blinded fashion by a veterinary pathologist at five points along each vein wall (n ¼ 4-6; *P < .01; C). At day 21, collagen content was analyzed using Sirius red staining measured with Image J software (n ¼ 12; *P ¼ .005; D, F, H). Photomicrographs of Arg-1-positive cells shown at 100 in nonstasis (E) and stasis (G) models. Sirius red staining shown in black and white, for nonstasis (F) and stasis (H) models; T marks thrombus and arrows denote vein wall border.
468 Gallagher et al
JOURNAL OF VASCULAR SURGERY: VENOUS AND LYMPHATIC DISORDERS October 2016
Fig 4. M1 (inducible nitric oxide synthase [iNOS]) and M2 (arginase 1 [Arg-1]) balance was assessed in human acute (<1 year) and chronic ($1 year) deep venous thrombosis (DVT) samples obtained from endophlebectomy by immunohistochemistry (n ¼ 4-7; *P < .05; A, C, D, G, H). Gene expression in humans with history of DVT was assessed at time of diagnosis (“enrollment”; n ¼ 17) and 6 months (n ¼ 9) for M2 marker CD206 (P ¼ .03; B). M1 markers iNOS (P ¼ .08; E) and CCR2 (P ¼ .04; F) were similarly assessed. All data are represented as mean 6 standard error of the mean. Representative histologic sections of iNOS (C and D) and Arg-1 (G and H) at 40 in an early (9-month) and late (25-year) human thrombus.
models is similar but demonstrates maintenance of a flow lumen in the nonstasis model (Fig 1, D) compared with an occlusive thrombus in the stasis model (Fig 1, F). Inflammatory monocytes-macrophages recruited to thrombus underwent immunohistochemical analysis for class M1-like (IL-12) or M2-like (Arg-1) proteins at both time points. Macrophages expressing an M2-like marker were more prevalent than cells expressing M1-associated IL-12 protein in the stasis model chronically (Fig 1, B and G) and in the nonstasis model at both acute and chronic time points (Fig 1, C and E). Thrombus and bone marrow-derived cells have decreased M1 markers under conditions of complete blood stasis. To better define M1 and M2 phenotypic difference between stasis and nonstasis thrombosis, we examined the thrombus for inflammatory (M1) and anti-inflammatory (M2) cytokines using Luminex/BioPlex (Fig 2). We found that the stasis model had significantly decreased levels of M1 marker IL-1b (P < .01) and moderately decreased IL-12 at day 2 (P ¼ .17) compared with nonstasis thrombi (Fig 2, A and B). We then examined gene expression from cells derived at the same acute time point (day 2) from the bone marrow and cultured ex vivo into macrophages. IL1-b and IL-12 transcript levels in bone marrow macrophages after lipopolysaccharide stimulation were significantly decreased in stasis models (but not in nonstasis models) compared with sham controls (Fig 2, C and D). Chronically, at day 8, stasis thrombi retained decreased M1-like macrophages (Fig 2, E and F) but significantly increased anti-
inflammatory (M2; Fig 2, G) cytokine profile compared with nonstasis thrombi. Divergent responses in vein wall at early and mid time points after thrombosis result in altered vein wall phenotype chronically. Changes in the vein wall are characteristic of the post-thrombotic fibrotic response26,27 and were examined in conjunction with macrophage phenotype (Fig 3). We found that macrophages expressing an M2-like marker were more prevalent than cells expressing the M1associated IL-12 protein in the stasis model compared with the nonstasis model (Fig 3, A and B). The macrophage response did not heavily favor M1 or M2 phenotype in the nonstasis model but was predominantly M2 when the thrombi were formed under conditions of complete blood stasis (Fig 3, A, B, E, and G). Consistent with this observation, the intimal thickness in the stasis model was found to be greater compared with the nonstasis model at mid and late time points (Fig 3, C). Furthermore, vein wall fibrosis, as measured by collagen content, was elevated in the stasis VT model (Fig 3, D, F, and H), characteristic of an overly robust M2 response. Circulating and tissue anti-inflammatory macrophages decrease over time in human VT. Human tissue from acute (<1 year) or chronic ($1 year) VT was analyzed to determine if such an approach could be similarly applied across species (Fig 4). Thrombi samples were stained for M1-like marker iNOS (Fig 4, A, C, and D) and M2-like marker Arg-1 (Fig 4, A, G, and H). Levels of both M1and M2 macrophages decreased over time (Fig 4, A), without considerable predominance of one
JOURNAL OF VASCULAR SURGERY: VENOUS AND LYMPHATIC DISORDERS Volume 4, Number 4
Gallagher et al 469
Fig 5. Schematic illustrating results.
phenotype. Circulating human leukocytes were isolated and examined for gene expression of inflammatory markers (CCR2, iNOS) and anti-inflammatory scavenger receptor (CD206), using samples from a prior published study.25 Levels were compared between controls and patients with acute VT or at 6 months after VT. CD206 (M2 marker) was significantly decreased at 6 months after VT, consistent with histologic observations (Fig 4, B). M1-like markers (iNOS, P ¼ .08; CCR2, P ¼ .04) were decreased (Fig 4, E and F). DISCUSSION Venous thrombogenesis and resolution are dependent on leukocytes and their interactions with other cellular and noncellular elements within the local environment.9 Macrophages are essential for thrombus resolution and significantly contribute to vein wall remodeling and fibrosis by mediating the inflammatory response.26 In this study, we show for the first time in experimental VT that (1) thrombus resolution is characterized by M2-like macrophage predominance, (2) M1/M2 balance in the thrombus correlates with changes in bone marrowderived cells, and (3) divergent macrophage phenotypes between thrombosis models correspond to differing fibrotic response of the vein wall. We also demonstrate feasibility of characterizing human macrophage phenotypes in thrombus and blood samples and present data suggesting that phenotypic changes in thrombi are similar to changes in circulating monocytes over time. A schematic of our findings is shown in Fig 5.
Thrombi. Macrophages are essential to thrombus resolution. Previous studies have shown that direct injection of intraperitoneal macrophages into experimental caval thrombi dramatically decreases thrombus size and increases recanalization.15 Among murine monocytes, presence of CCR2 is characteristic of a “proinflammatory” (M1) phenotype and absence is indicative of an “anti-inflammatory” (M2) phenotype.28,29 Experimentally, gene deletion of CCR2 results in decreased macrophage recruitment and impaired thrombus resolution.11,15 Exogenous interferong (Th1 cytokine) can reverse impaired VT resolution in a CCR2/ mouse, perhaps by influencing macrophage phenotypes because interferon-g is well known to skew macrophages toward an M1-like phenotype.11 Therefore, the consequence of an early M1-like phenotype would be expected to lead to increased clearance of thrombosis. However, in our study, the thrombus was characterized by a predominantly M2-like phenotype, particularly during chronic resolution, suggesting that the presence of residual venous thrombi may be related to macrophage polarization. This may indicate that the process of thrombus formation in the venous system induces changes in the bone marrow progenitor cells that influence their plasticity as mature monocyte-macrophages in the periphery. The intrathrombus M2 cytokine phenotype was more pronounced and M1 diminished in our stasis model, corresponding to larger residual thrombus burden. Interestingly, this phenotype corresponded to that of bone marrowderived cells, suggesting that targeting of bone marrowderived progenitor cells might represent a method to alter macrophage phenotypes involved in thrombus clearance.
470 Gallagher et al
JOURNAL OF VASCULAR SURGERY: VENOUS AND LYMPHATIC DISORDERS October 2016
Others have shown feasibility of such an approach in a rat VT model; stimulation of bone marrow with recombinant human granulocyte-stimulating factor increases circulating CCR2 and macrophage recruitment to venous thrombi and improves thrombus clearance.30 The discovery of ways to modulate thrombus resolution is beneficial because observations suggest that rapid VT resolution may be associated with a lower risk of PTS31 and may not confer the associated bleeding risks of anticoagulation or thrombolysis. Vein wall. Changes in macrophage phenotypes were also evident in vein wall over time, with a heavy predilection toward M2 phenotype in the stasis model at acute and chronic time points, correlating with increased vein wall thickness and fibrosis. This is consistent with observations in other organ systems, in which M2 polarization has been associated with pulmonary32 and renal fibrosis.33 Previous data have shown that the mechanism of VT, stasis vs nonstasis, contributes to altered vein wall phenotype.10 We propose that a contributing factor might be differences in macrophage polarization, with a predominantly M2 phenotype predisposing the vein wall to a more intense fibrotic response. Whether the M2 macrophages are set at the bone marrow level, as suggested here indirectly from the systemic response to venous thrombosis, or become M2 from an initial M1 influx is not clear from our studies. For example, recent work in a sterile inflammatory model of liver disease suggests an active change due to the local environment in macrophage function rather than separate recruitments.34 From a translational aspect, early interventions designed to alter macrophage phenotype toward an inflammatory state would be most important in altering the fibrotic response. Human. Our human data demonstrate for the first time in man that venous thrombi show variation in M1 and M2 markers. Although our study was not powered to detect differences among such a heterogeneous group of patients and our results should be verified with a prospectively identified, larger population, we did identify that tissue macrophages expressing either phenotype seemed to decrease over time, correlating with circulating macrophage phenotypes. For the first time, we are able to identify that variations in macrophage polarization exist in human VT. As human trials targeting macrophage response are already ongoing for inflammatory diseases such as nephritis, atherosclerosis, psoriasis, and cancer,35 these data represent a potentially new avenue of investigation for the treatment of PTS. Limitations. Sterile inflammation is a complex process, and although our data suggest that macrophages play a key role, we cannot exclude the effects of other cells and signaling pathways shown to play a role in inflammation. In addition to Toll-like receptor stimulation, nuclear factor kB or TRIF signaling pathways may play critical roles in macrophage activation during VT.14,36 Further, because iron has recently been shown to play a role in macrophage polarization in venous ulcers, and iron has been shown to accumulate during 8 days in VT, it is unclear what role this may play in influencing macrophage polarization in VT resolution.37
Although previous research supports that neutrophils and macrophages are among the most critical cells early in VT,11,38 we have not accounted for the role that platelets play in this process. It has been shown recently that platelets may promote necrotic inflammation.39-41 CONCLUSIONS In humans, the longer time that the thrombus is in contact with the vein wall, the greater the damage. Therefore, treatment strategies designed to minimize thrombus burden at early time points are imperative. Direct modulation of inflammation has not been explored in VT, and we speculate that altering the M1/M2 balance may accelerate thrombus resolution and allow the development of translatable novel therapies to treat VT. These new therapies may allow the thrombus to be more lysable, which may represent an alternative treatment avenue for reduction of PTS and avoidance of complications related to standard therapies. AUTHOR CONTRIBUTIONS Conception and design: KG, PH Analysis and interpretation: KG, AO, DC, PH Data collection: ME, EH, CL, SH Writing the article: KG, AO, PH Critical revision of the article: KG, AO, ME, CL, PH Final approval of the article: KG, AO, ME, EH, CL, SH, DC, PH Statistical analysis: KG, AO, CL, SH, PH Obtained funding: KG, PH Overall responsibility: PH REFERENCES 1. Coon WW, Willis PW 3rd, Keller JB. Venous thromboembolism and other venous disease in the Tecumseh community health study. Circulation 1973;48:839-46. 2. Anderson FA Jr, Wheeler HB, Goldberg RJ, Hosmer DW, Patwardhan NA, Jovanovic B, et al. A population-based perspective of the hospital incidence and case-fatality rates of deep vein thrombosis and pulmonary embolism. The Worcester DVT Study. Arch Intern Med 1991;151:933-8. 3. Weitz JI, Bates SM. New anticoagulants. J Thromb Haemost 2005;3: 1843-53. 4. Heit JA, Silverstein MD, Mohr DN, Petterson TM, Lohse CM, O’Fallon WM, et al. The epidemiology of venous thromboembolism in the community. Thromb Haemost 2001;86:452-63. 5. Henke PK, Comerota AJ. An update on etiology, prevention, and therapy of postthrombotic syndrome. J Vasc Surg 2011;53:500-9. 6. Deatrick KB, Luke CE, Elfline MA, Sood V, Baldwin J, Upchurch GR Jr, et al. The effect of matrix metalloproteinase 2 and matrix metalloproteinase 2/9 deletion in experimental post-thrombotic vein wall remodeling. J Vasc Surg 2013;58:1375-1384.e2. 7. Engelmann B, Massberg S. Thrombosis as an intravascular effector of innate immunity. Nat Rev Immunol 2013;13:34-45. 8. Saha P, Humphries J, Modarai B, Mattock K, Waltham M, Evans CE, et al. Leukocytes and the natural history of deep vein thrombosis: current concepts and future directions. Arterioscler Thromb Vasc Biol 2011;31:506-12. 9. Wakefield TW, Myers DD, Henke PK. Mechanisms of venous thrombosis and resolution. Arterioscler Thromb Vasc Biol 2008;28: 387-91.
JOURNAL OF VASCULAR SURGERY: VENOUS AND LYMPHATIC DISORDERS Volume 4, Number 4
10. Henke PK, Varma MR, Moaveni DK, Dewyer NA, Moore AJ, Lynch EM, et al. Fibrotic injury after experimental deep vein thrombosis is determined by the mechanism of thrombogenesis. Thromb Haemost 2007;98:1045-55. 11. Henke PK, Pearce CG, Moaveni DM, Moore AJ, Lynch EM, Longo C, et al. Targeted deletion of CCR2 impairs deep vein thombosis resolution in a mouse model. J Immunol 2006;177:3388-97. 12. Soehnlein O, Lindbom L. Phagocyte partnership during the onset and resolution of inflammation. Nat Rev Immunol 2010;10:427-39. 13. Nathan C, Ding A. Nonresolving inflammation. Cell 2010;140: 871-82. 14. Henke PK, Mitsuya M, Luke CE, Elfline MA, Baldwin JF, Deatrick KB, et al. Toll-like receptor 9 signaling is critical for early experimental deep vein thrombosis resolution. Arterioscler Thromb Vasc Biol 2011;31:43-9. 15. Ali T, Humphries J, Burnand K, Sawyer B, Bursill C, Channon K, et al. Monocyte recruitment in venous thrombus resolution. J Vasc Surg 2006;43:601-8. 16. Deatrick KB, Eliason JL, Lynch EM, Moore AJ, Dewyer NA, Varma MR, et al. Vein wall remodeling after deep vein thrombosis involves matrix metalloproteinases and late fibrosis in a mouse model. J Vasc Surg 2005;42:140-8. 17. Henke PK, Varga A, De S, Deatrick CB, Eliason J, Arenberg DA, et al. Deep vein thrombosis resolution is modulated by monocyte CXCR2mediated activity in a mouse model. Arterioscler Thromb Vasc Biol 2004;24:1130-7. 18. Diaz JA, Hawley AE, Alvarado CM, Berguer AM, Baker NK, Wrobleski SK, et al. Thrombogenesis with continuous blood flow in the inferior vena cava. A novel mouse model. Thromb Haemost 2010;104:366-75. 19. Junqueira LC, Bignolas G, Brentani RR. Picrosirius staining plus polarization microscopy, a specific method for collagen detection in tissue sections. Histochem J 1979;11:447-55. 20. Cuttle L, Nataatmadja M, Fraser JF, Kempf M, Kimble RM, Hayes MT. Collagen in the scarless fetal skin wound: detection with picrosirius-polarization. Wound Repair Regen 2005;13:198-204. 21. Baldwin JF, Sood V, Elfline MA, Luke CE, Dewyer NA, Diaz JA, et al. The role of urokinase plasminogen activator and plasmin activator inhibitor-1 on vein wall remodeling in experimental deep vein thrombosis. J Vasc Surg 2012;56:1089-97. 22. Wojcik BM, Wrobleski SK, Hawley AE, Wakefield TW, Myers DD Jr, Diaz JA. Interleukin-6: a potential target for post-thrombotic syndrome. Ann Vasc Surg 2011;25:229-39. 23. Laser A, Elfline M, Luke C, Slack D, Shah A, Sood V, et al. Deletion of cysteine-cysteine receptor 7 promotes fibrotic injury in experimental post-thrombotic vein wall remodeling. Arterioscler Thromb Vasc Biol 2014;34:377-85. 24. Comerota AJ, Oostra C, Fayad Z, Gunning W, Henke P, Luke C, et al. A histological and functional description of the tissue causing chronic postthrombotic venous obstruction. Thromb Res 2015;135: 882-7. 25. Deatrick KB, Elfline M, Baker N, Luke CE, Blackburn S, Stabler C, et al. Postthrombotic vein wall remodeling: preliminary observations. J Vasc Surg 2011;53:139-46. 26. Obi AT, Diaz JA, Ballard-Lipka NL, Roelofs KJ, Farris DM, Lawrence DA, et al. Plasminogen activator-1 overexpression decreases experimental postthrombotic vein wall fibrosis by a non-
27.
28.
29.
30.
31.
32.
33.
34.
35. 36.
37.
38.
39.
40. 41.
Gallagher et al 471
vitronectin-dependent mechanism. J Thromb Haemost 2014;12: 1353-63. Obi AT, Diaz JA, Ballard-Lipka NL, Roelofs KJ, Farris DM, Lawrence DA, et al. Low-molecular-weight heparin modulates vein wall fibrotic response in a plasminogen activator inhibitor 1dependent manner. J Vasc Surg Venous Lymphat Disord 2014;2: 441-450.e1. Willenborg S, Lucas T, van Loo G, Knipper JA, Krieg T, Haase I, et al. CCR2 recruits an inflammatory macrophage subpopulation critical for angiogenesis in tissue repair. Blood 2012;120:613-25. Geissmann F, Jung S, Littman DR. Blood monocytes consist of two principal subsets with distinct migratory properties. Immunity 2003;19: 71-82. Chen YK, Jiang XM, Gong JP. Recombinant human granulocyte colony-stimulating factor enhanced the resolution of venous thrombi. J Vasc Surg 2008;47:1058-65. Enden T, Haig Y, Klow NE, Slagsvold CE, Sandvik L, Ghanima W, et al. Long-term outcome after additional catheter-directed thrombolysis versus standard treatment for acute iliofemoral deep vein thrombosis (the CaVenT study): a randomised controlled trial. Lancet 2012;379:31-8. Sun L, Louie MC, Vannella KM, Wilke CA, LeVine AM, Moore BB, et al. New concepts of IL-10-induced lung fibrosis: fibrocyte recruitment and M2 activation in a CCL2/CCR2 axis. Am J Physiol Lung Cell Mol Physiol 2011;300:L341-53. Anders HJ, Ryu M. Renal microenvironments and macrophage phenotypes determine progression or resolution of renal inflammation and fibrosis. Kidney Int 2011;80:915-25. Dal-Secco D, Wang J, Zeng Z, Kolaczkowska E, Wong CH, Petri B, et al. A dynamic spectrum of monocytes arising from the in situ reprogramming of CCR2þ monocytes at a site of sterile injury. J Exp Med 2015;212:447-56. Hamilton JA, Achuthan A. Colony stimulating factors and myeloid cell biology in health and disease. Trends Immunol 2013;34:81-9. Dewyer NA, El-Sayed OM, Luke CE, Elfline M, Kittan N, Allen R, et al. Divergent effects of Tlr9 deletion in experimental late venous thrombosis resolution and vein wall injury. Thromb Haemost 2015;114:1028-37. Sindrilaru A, Peters T, Wieschalka S, Baican C, Baican A, Peter H, et al. An unrestrained proinflammatory M1 macrophage population induced by iron impairs wound healing in humans and mice. J Clin Invest 2011;121:985-97. Henke PK, Varma MR, Deatrick KB, Drewyer NA, Lynch EM, Moore AJ, et al. Neutrophils modulate post-thrombotic vein wall remodeling but not thrombus neovascularization. Thromb Haemost 2006;95:272-81. Brill A, Fuchs TA, Chauhan AK, Yang JJ, De Meyer SF, Kollnberger M, et al. von Willebrand factor-mediated platelet adhesion is critical for deep vein thrombosis in mouse models. Blood 2011;117:1400-7. Jackson SP, Schoenwaelder SM. Procoagulant platelets: are they necrotic? Blood 2010;116:2011-8. Panigrahi S, Ma Y, Hong L, Gao D, West XZ, Salomon RG, et al. Engagement of platelet toll-like receptor 9 by novel endogenous ligands promotes platelet hyperreactivity and thrombosis. Circ Res 2013;112:103-12.
Submitted Nov 17, 2015; accepted Mar 12, 2016.
Alterations in macrophage phenotypes in experimental venous thrombosis Katherine A. Gallagher, MD, Andrea T. Obi, MD, Megan A. Elfline, MS, Emily Hogikyan, Catherine E. Luke, LVT, Samuel Henke, Dawn Coleman, MD, and Peter K. Henke, MD, Ann Arbor, Mich Objective: Macrophages are involved in venous thrombus (VT) resolution and vein wall remodeling. This study was undertaken to identify variations in macrophage phenotypes in thrombi and vein wall in multiple models of VT to clarify the natural history of macrophage polarization in clearance of VT. We also sought to demonstrate the feasibility of macrophage phenotyping in human VT. Methods: Established murine models of VT were used to mimic the clinical spectrum of human VT (stasis and nonstasis models). Vein wall and thrombi were isolated at acute (2 days) or chronic (6-21 days) time points and analyzed by Bio-Plex assay (Bio-Rad, Carlsbad, Calif) for cytokines (interleukin [IL]-1b, IL-6, IL-10, IL-12), by immunohistochemistry for “M1-like” (IL-12) or “M2-like” (arginase 1 [Arg-1]) markers, and by histology for intimal thickness and collagen content (Sirius red staining). Bone marrow was harvested from animals 2 days after undergoing sham, stasis, or nonstasis surgery. Macrophages were skewed toward M1 using lipopolysaccharide, and RNA analysis was done for inflammatory cytokine genes (IL-1b, IL-12). Human blood samples were similarly analyzed with reverse transcription polymerase chain reaction for macrophage polarization markers (CD206, inducible nitric oxide synthase, CCR2) and thrombi with immunohistochemistry (inducible nitric oxide synthase, Arg-1). Results: Stasis (chronic) and nonstasis (acute and chronic) thrombi were characterized by a predominance in antiinflammatory (M2) macrophages (n [ 4-5/group; P < .05). Larger thrombi were found in the stasis model at both time points (n [ 3; P < .01), correlating with decreased intrathrombus inflammatory (M1) cytokines (IL-1b, P [ .03; IL-12, P [ .17; n [ 4) and diminished inflammatory response From the Section of Vascular Surgery, Department of Vascular Surgery, University of Michigan Medical School. This work was supported by HL K08112897 and HL R01HL092129, Vascular Cures Foundation, University of Michigan Cardiovascular Center McKay and Heart of Champion Funds, and the Taubman Scholars Foundation. Author conflict of interest: none. Presented in part at the Twenty-sixth Annual Meeting of the American Venous Forum, New Orleans, La, February 19-21, 2014. Correspondence: Peter K. Henke, MD, 1500 E Medical Center Dr, Rm 5463, Cardiovascular Center, Ann Arbor, MI 48109-5867 (e-mail: [email protected]). The editors and reviewers of this article have no relevant financial relationships to disclose per the Journal policy that requires reviewers to decline review of any manuscript for which they may have a conflict of interest. 2213-333X Copyright Ó 2016 by the Society for Vascular Surgery. Published by Elsevier Inc. http://dx.doi.org/10.1016/j.jvsv.2016.03.005
of bone marrow-derived macrophages to lipopolysaccharide (IL-1b, P [ .03; IL-12, P [ .04; n [ 4) compared with nonstasis model. Anti-inflammatory (M2 [Arg-1]) macrophage cell counts were elevated in the post-thrombotic vein wall of stasis mice compared with nonstasis mice (acute: n [ 4, P < .05; chronic: n [ 5, P < .01), consistent with increased intimal thickness (P < .01; n [ 4-6) and collagen deposition chronically (P [ .005; n [ 12). M2-like thrombi (Arg-1, P < .05; n [ 4-7) and circulating markers (CD206, P < .05; n [ 917) decreased over time in human VT. Conclusions: Experimental VT is characterized by an antiinflammatory predominant macrophage phenotype, possibly impairing thrombus resolution, and is model dependent. Altering the M1/M2 macrophage balance may accelerate thrombus resolution and allow the development of translatable novel therapies to treat VT and to prevent post-thrombotic syndrome. (J Vasc Surg: Venous and Lym Dis 2016;4:463-71.)
Clinical Relevance: In humans, the longer time that the thrombus is in contact with the vein wall, the greater the damage. Therefore, treatment strategies designed to minimize thrombus burden at early time points are imperative. Direct modulation of inflammation has not been explored in venous thrombus, and we speculate that altering the M1/M2 balance may accelerate thrombus resolution and allow the development of translatable novel therapies to treat venous thrombus. These new therapies may allow the thrombus to be more lysable, which may represent an alternative treatment avenue for reduction of post-thrombotic syndrome and avoidance of complications related to standard therapies.
Venous thromboembolism (VTE) is a substantial health problem, resulting in significant morbidity and mortality.1,2 Even among patients undergoing adequate anticoagulation, the current mainstay therapy, complications associated with VTE are prevalent.3 Postthrombotic syndrome (PTS), a late complication associated with VTE, is associated with significant morbidity and results in added health care costs.4,5 Little progress has been made in the treatment of PTS in recent years, particularly with the recent data that active compression does not prevent PTS. Currently, no pharmacologic agent exists for the treatment or prevention of PTS. Experimentally, venous thrombus (VT) resolution and vein wall healing are mediated by leukocytes and their 463
464 Gallagher et al
JOURNAL OF VASCULAR SURGERY: VENOUS AND LYMPHATIC DISORDERS October 2016
associated chemokines, plasminogen activators, matrix metalloproteinases, and proinflammatory cytokines.6-10 The process of thrombus resolution resembles wound healing with distinct phases of neutrophil and monocyte influx, followed by resolution and fibrosis.11-13 The vein wall response to thrombus bears many similarities to the process of sterile inflammation, and leukocytes are primary drivers of this process.14 Macrophages are key mediators of the generation and resolution of inflammation and are critical for thrombus resolution as they clear cellular debris and have profibrinolytic activity.8 Macrophages are phenotypically defined by the expression of specific surface markers and gene expression. Classically activated macrophages (“M1-like”) express a defined set of proinflammatory mediators (interleukin [IL]-1b, IL-6, IL-12, inducible nitric oxide synthase [iNOS], CCR2), whereas alternatively activated macrophages (“M2-like”) with profibrinolytic activity display an anti-inflammatory phenotype (IL-10, arginase 1 [Arg-1], CD206).15 Currently, the role of macrophage functional phenotypes in VT resolution is unknown. The objective of this study was to identify variations in macrophage phenotypes and hence functions in two common models of VT at both acute and chronic time points after VT to provide a broader understanding of the human pathophysiology of this disease process. Further, we have examined thrombus from a small group of human patients after VT to demonstrate feasibility of such an approach. We hypothesized that macrophages display a decreased M1-like, proinflammatory activity and increased M2-like, profibrotic function during the venous thrombosis process, ultimately resulting in alterations in vein wall fibrosis. METHODS Venous thrombosis models. Male mice, aged 6 to 8 weeks on a C57BL6 background, were purchased from Jackson Laboratory (Bar Harbor, Me) and were used for all studies. For all procedures, mice underwent general anesthesia with isoflurane and oxygen with continuous monitoring. All animal studies were done with University of Michigan Animal Care and Use committee approval. Stasis model of VT. Mice underwent surgical ligation of the inferior vena cava (IVC) and visible contributing vessels below the renal veins, producing complete blood stasis and subsequent VT formation. This model is well characterized and consistently (>97%) produces a VT.10,11,14,16,17 IVC, thrombus, and bone marrow were harvested at acute (2 days) and chronic (6, 8, 14, or 21 days) time points after induction of stasis for analysis. Nonstasis model of VT. An endothelial injury model of VT18 was used to establish thrombus in the setting of continued blood flow. In brief, mice underwent intraluminal electrolytic injury to the IVC. A 25-gauge needle was inserted into the IVC and put in contact with the anterior IVC wall between the renal veins and iliac bifurcation. A current of 25 mAmp was applied for 15 minutes, consistently producing nonocclusive VT. IVC, thrombus, and bone marrow were harvested as listed previously.
Murine histology and immunohistologic analysis. Tissues were formalin fixed, paraffin embedded, and cut into 5-mm sections. Nonspecific sites were blocked with serum, and sections were incubated with primary antibodies to IL-12 (Abcam, Cambridge, UK; catalog #ab131039) and Arg-1 (Novus Biologicals, Littleton, Colo; catalog #NBP1-54621). A species-specific avidinbiotin complex peroxidase kit for rabbit or rat (Vector Laboratories Inc, Burlingame, Calif) was used according to the manufacturer’s instructions. Color development was performed with diaminobenzidine (Vector Laboratories). The slides were counterstained with hematoxylin (Vector Laboratories). In a blinded fashion, positive cells in five high-power fields (400) embedded in the IVC thrombus were counted and totaled. Collagen content was quantified with picrosirius red stain with Zeiss Axio M1 scope and Zeiss AxioVision software (Carl Zeiss Microimaging GmbH, Göttingen, Germany).19-21 Two images were analyzed in crossedplane polarized light from a monochromatic light source at 0 and 90 degrees to the plane of polarization to capture the birefringence of fibers extinguished in one direction. The IVC wall was outlined as a region of interest, and then the image underwent threshold segmentation to differentiate collagen from other (cells or empty space) components of the vein wall using Image J software (National Institutes of Health, Bethesda, Md). Vein wall collagen score was calculated as (% birefringent area) (measured vein wall area)/(total specimen area). To account for noncollagen vein wall changes, intimal thickness assessed from hematoxylin and eosin-stained sections was scored by a board-certified veterinary pathologist in a blinded manner.21,22 Briefly, a midsection of the thrombosed IVC was blindly analyzed on a scale of 0 to 4, with 0 representing intima as space occupied by only endothelial cells and 4 representing a greatly thickened intima containing fibroblasts, white blood cells, or hemorrhage at its widest point.22 Cell culture and real-time quantitative polymerase chain reaction. Bone marrow cells were collected by flushing mouse femur and tibia bones with RPMI medium at 2 days after the stasis or nonstasis model of the IVC. Bone marrow-derived macrophages were cultured as previously detailed. On day 6, the cells were replated, and after resting for 24 hours, they were incubated with or without interferon-g (100 ng/mL; Invitrogen, Carlsbad, Calif) and lipopolysaccharide (100 ng/mL) for 6 hours. Total RNA extraction was then performed using TRIzol (Invitrogen) per the manufacturer’s instructions. Total RNA was reverse transcribed to complementary DNA with M-MLV (Invitrogen). Reverse transcription polymerase chain reaction (PCR) was performed with TaqMan PCR mix (2) using the 7500 Real-Time PCR System (Applied Biosystems, Foster City, Calif) with primers for IL-1b and IL-12 (Applied Biosystems). Glyceraldehyde 3-phosphate dehydrogenase was used as the internal control. All standards and samples were assayed in triplicate. The threshold cycle values were used to plot a standard curve.
JOURNAL OF VASCULAR SURGERY: VENOUS AND LYMPHATIC DISORDERS Volume 4, Number 4
Gallagher et al 465
Fig 1. Thrombus size was measured and weighed after both stasis and nonstasis formation in early (2 days) and chronic (8 days) venous thrombus (VT; all n ¼ 3; *P < .01; A). Immunohistochemistry was performed using arginase 1 (Arg-1 [M1]) and interleukin-12 (IL-12 [M2]) antibodies, and cell counts of thrombus were undertaken at 2 days (n ¼ 4-5; *P < .05; B, E, G) and 8 days (n ¼ 4-5; *P < .05; C, E, G). Hematoxylin and eosin-stained sections of midpoint of thrombi were obtained for nonstasis (D) and stasis (F) models at both time points (10 representative photographs shown). Representative thrombi of nonstasis (E) and stasis (G) stained for Arg-1 at 8 days are shown at 20. All data are represented as mean 6 standard error of the mean. hpf, High-power field.
Bio-Plex. Levels of IL-1b, IL-6, IL-10, and IL-12 present in thrombus from the stasis and nonstasis VT murine models were determined using the Luminex/Bio-Plex assay 200 system (Bio-Rad, Carlsbad, Calif) per the manufacturer’s protocol. The limit of detection for each cytokine was 0.5 rg/mL. The cytokine levels were normalized to protein present in a cell-free preparation of each sample measured by Bradford assay (Bio-Rad). Human sample collection. Tissue. Human tissue was obtained from patients undergoing endophlebectomy of common femoral veins followed by recanalization of iliocaval obstruction for the treatment of debilitating PTS.23,24 All patients were treated with anticoagulation alone (no thrombolysis or procedural intervention for treatment of VT). VT occurring <1 year before acquisition of the specimen was classified as acute, whereas that occurring at later time points was considered chronic. Specimens were
formalin fixed and paraffin embedded. This protocol had been previously approved by the Institutional Review Board of the ProMedica Health System and adhered to the principles of the Declaration of Helsinki. Blood. Blood from patients presenting to our diagnostic vascular laboratory with acute VT in the popliteal, femoral, or iliac system was obtained at enrollment and 6 months after diagnosis, which is the subject of a separate published study.25 A sample of the collected blood was placed in PAXgene Blood RNA System collecting tubes (BD Biosciences, Becton Lakes, NJ) according to the manufacturer’s instructions. The messenger RNA obtained was subject to reverse transcription by incubation with oligo-dT primer and M-MLV reverse transcriptase (Life Technologies, Carlsbad, Calif) at 94 C for 3 minutes and then incubated at 40 C for 70 minutes. The complementary DNA obtained was amplified using TaqPolymerase
466 Gallagher et al
JOURNAL OF VASCULAR SURGERY: VENOUS AND LYMPHATIC DISORDERS October 2016
Fig 2. Thrombi at 2 days underwent Bio-Plex assessment for cytokine markers indicative of M1 macrophage phenotype, interleukin-1 (IL-1; n ¼ 4; *P ¼ .03; A) and interleukin-12 (IL-12; n ¼ 4; P ¼ .17; B). Bone marrow-derived macrophages from animals undergoing stasis or nonstasis thrombosis or sham operation were obtained at 2 days after thrombosis and were cultured for 6 days in standard conditions and treated with lipopolysaccharide (100 mg/mL) for 6 hours. Reverse transcription polymerase chain reaction (PCR) was performed to determine transcription levels of key M1 genes IL-1 (C; n ¼ 10; *P ¼ .03) and IL-12 (D; n ¼ 10; *P ¼ .04). Similarly, thrombi at 8 days underwent Bio-Plex assessment for M1 phenotype cytokine markers IL-1 (n ¼ 4; *P ¼ .04; E) and IL-6 (n ¼ 4; P ¼ .06; F). M2 cytokine IL-10 (n ¼ 4; * P ¼ .002; G) was additionally assessed. Top panel, All results at 2 days. Bottom panel, All results at 8 days.
(Promega, Madison, Wisc) in the Rotor-Gene quantitative real-time PCR system (Qiagen, Hilden, Germany). SYBR green intercalating dye was used to monitor levels of DNA amplification for each gene. Commercially available primers were purchased for the following genes of interest: b-actin (RefSeq# NM_001101.3), CD206(RefSeq# NM_002438), iNOS (RefSeq# NM_000625), and CCR2 (RefSeq# NM_001123396, Qiagen). The study was previously approved by the University of Michigan Institutional Review Board for Human Subjects research, adhered to the principles of the Declaration of Helsinki, and included only patients who gave informed consent. Statistical analysis. All data are shown as a mean 6 standard error of the mean. Student t-test or analysis of
variance, as appropriate, was used for comparison between the stasis and nonstasis models (GraphPad Prism, San Diego, Calif); P values # .05 were considered significant. RESULTS VT resolution is characterized by a predominantly M2 response. The natural history of macrophage phenotype in VT resolution was examined by microscopy of thrombus and vein wall at various time points after thrombosis (Fig 1). Initial VT formation in stasis and nonstasis models is characterized by smaller thrombus in the nonstasis model at both 2 and 8 days after thrombosis, consistent with perithrombus blood flow (Fig 1, A, D, and F). The VT histologic appearance between the two
JOURNAL OF VASCULAR SURGERY: VENOUS AND LYMPHATIC DISORDERS Volume 4, Number 4
Gallagher et al 467
Fig 3. Immunohistochemistry of the vein wall was performed using arginase 1 (Arg-1 [M1]; E and G) and interleukin12 (IL-12 [M2]) antibodies at 2 days after thrombosis (n ¼ 4; *P < .05; A) and 8 days after thrombosis (n ¼ 5; *P < .001; B, E, G). Changes in vein wall were assessed at chronic time points after thrombosis (6-8 days “mid” or 1421 days “late”) to assess for late changes in vein wall phenotype. Intimal thickness was scored in a blinded fashion by a veterinary pathologist at five points along each vein wall (n ¼ 4-6; *P < .01; C). At day 21, collagen content was analyzed using Sirius red staining measured with Image J software (n ¼ 12; *P ¼ .005; D, F, H). Photomicrographs of Arg-1-positive cells shown at 100 in nonstasis (E) and stasis (G) models. Sirius red staining shown in black and white, for nonstasis (F) and stasis (H) models; T marks thrombus and arrows denote vein wall border.
468 Gallagher et al
JOURNAL OF VASCULAR SURGERY: VENOUS AND LYMPHATIC DISORDERS October 2016
Fig 4. M1 (inducible nitric oxide synthase [iNOS]) and M2 (arginase 1 [Arg-1]) balance was assessed in human acute (<1 year) and chronic ($1 year) deep venous thrombosis (DVT) samples obtained from endophlebectomy by immunohistochemistry (n ¼ 4-7; *P < .05; A, C, D, G, H). Gene expression in humans with history of DVT was assessed at time of diagnosis (“enrollment”; n ¼ 17) and 6 months (n ¼ 9) for M2 marker CD206 (P ¼ .03; B). M1 markers iNOS (P ¼ .08; E) and CCR2 (P ¼ .04; F) were similarly assessed. All data are represented as mean 6 standard error of the mean. Representative histologic sections of iNOS (C and D) and Arg-1 (G and H) at 40 in an early (9-month) and late (25-year) human thrombus.
models is similar but demonstrates maintenance of a flow lumen in the nonstasis model (Fig 1, D) compared with an occlusive thrombus in the stasis model (Fig 1, F). Inflammatory monocytes-macrophages recruited to thrombus underwent immunohistochemical analysis for class M1-like (IL-12) or M2-like (Arg-1) proteins at both time points. Macrophages expressing an M2-like marker were more prevalent than cells expressing M1-associated IL-12 protein in the stasis model chronically (Fig 1, B and G) and in the nonstasis model at both acute and chronic time points (Fig 1, C and E). Thrombus and bone marrow-derived cells have decreased M1 markers under conditions of complete blood stasis. To better define M1 and M2 phenotypic difference between stasis and nonstasis thrombosis, we examined the thrombus for inflammatory (M1) and anti-inflammatory (M2) cytokines using Luminex/BioPlex (Fig 2). We found that the stasis model had significantly decreased levels of M1 marker IL-1b (P < .01) and moderately decreased IL-12 at day 2 (P ¼ .17) compared with nonstasis thrombi (Fig 2, A and B). We then examined gene expression from cells derived at the same acute time point (day 2) from the bone marrow and cultured ex vivo into macrophages. IL1-b and IL-12 transcript levels in bone marrow macrophages after lipopolysaccharide stimulation were significantly decreased in stasis models (but not in nonstasis models) compared with sham controls (Fig 2, C and D). Chronically, at day 8, stasis thrombi retained decreased M1-like macrophages (Fig 2, E and F) but significantly increased anti-
inflammatory (M2; Fig 2, G) cytokine profile compared with nonstasis thrombi. Divergent responses in vein wall at early and mid time points after thrombosis result in altered vein wall phenotype chronically. Changes in the vein wall are characteristic of the post-thrombotic fibrotic response26,27 and were examined in conjunction with macrophage phenotype (Fig 3). We found that macrophages expressing an M2-like marker were more prevalent than cells expressing the M1associated IL-12 protein in the stasis model compared with the nonstasis model (Fig 3, A and B). The macrophage response did not heavily favor M1 or M2 phenotype in the nonstasis model but was predominantly M2 when the thrombi were formed under conditions of complete blood stasis (Fig 3, A, B, E, and G). Consistent with this observation, the intimal thickness in the stasis model was found to be greater compared with the nonstasis model at mid and late time points (Fig 3, C). Furthermore, vein wall fibrosis, as measured by collagen content, was elevated in the stasis VT model (Fig 3, D, F, and H), characteristic of an overly robust M2 response. Circulating and tissue anti-inflammatory macrophages decrease over time in human VT. Human tissue from acute (<1 year) or chronic ($1 year) VT was analyzed to determine if such an approach could be similarly applied across species (Fig 4). Thrombi samples were stained for M1-like marker iNOS (Fig 4, A, C, and D) and M2-like marker Arg-1 (Fig 4, A, G, and H). Levels of both M1and M2 macrophages decreased over time (Fig 4, A), without considerable predominance of one
JOURNAL OF VASCULAR SURGERY: VENOUS AND LYMPHATIC DISORDERS Volume 4, Number 4
Gallagher et al 469
Fig 5. Schematic illustrating results.
phenotype. Circulating human leukocytes were isolated and examined for gene expression of inflammatory markers (CCR2, iNOS) and anti-inflammatory scavenger receptor (CD206), using samples from a prior published study.25 Levels were compared between controls and patients with acute VT or at 6 months after VT. CD206 (M2 marker) was significantly decreased at 6 months after VT, consistent with histologic observations (Fig 4, B). M1-like markers (iNOS, P ¼ .08; CCR2, P ¼ .04) were decreased (Fig 4, E and F). DISCUSSION Venous thrombogenesis and resolution are dependent on leukocytes and their interactions with other cellular and noncellular elements within the local environment.9 Macrophages are essential for thrombus resolution and significantly contribute to vein wall remodeling and fibrosis by mediating the inflammatory response.26 In this study, we show for the first time in experimental VT that (1) thrombus resolution is characterized by M2-like macrophage predominance, (2) M1/M2 balance in the thrombus correlates with changes in bone marrowderived cells, and (3) divergent macrophage phenotypes between thrombosis models correspond to differing fibrotic response of the vein wall. We also demonstrate feasibility of characterizing human macrophage phenotypes in thrombus and blood samples and present data suggesting that phenotypic changes in thrombi are similar to changes in circulating monocytes over time. A schematic of our findings is shown in Fig 5.
Thrombi. Macrophages are essential to thrombus resolution. Previous studies have shown that direct injection of intraperitoneal macrophages into experimental caval thrombi dramatically decreases thrombus size and increases recanalization.15 Among murine monocytes, presence of CCR2 is characteristic of a “proinflammatory” (M1) phenotype and absence is indicative of an “anti-inflammatory” (M2) phenotype.28,29 Experimentally, gene deletion of CCR2 results in decreased macrophage recruitment and impaired thrombus resolution.11,15 Exogenous interferong (Th1 cytokine) can reverse impaired VT resolution in a CCR2/ mouse, perhaps by influencing macrophage phenotypes because interferon-g is well known to skew macrophages toward an M1-like phenotype.11 Therefore, the consequence of an early M1-like phenotype would be expected to lead to increased clearance of thrombosis. However, in our study, the thrombus was characterized by a predominantly M2-like phenotype, particularly during chronic resolution, suggesting that the presence of residual venous thrombi may be related to macrophage polarization. This may indicate that the process of thrombus formation in the venous system induces changes in the bone marrow progenitor cells that influence their plasticity as mature monocyte-macrophages in the periphery. The intrathrombus M2 cytokine phenotype was more pronounced and M1 diminished in our stasis model, corresponding to larger residual thrombus burden. Interestingly, this phenotype corresponded to that of bone marrowderived cells, suggesting that targeting of bone marrowderived progenitor cells might represent a method to alter macrophage phenotypes involved in thrombus clearance.
470 Gallagher et al
JOURNAL OF VASCULAR SURGERY: VENOUS AND LYMPHATIC DISORDERS October 2016
Others have shown feasibility of such an approach in a rat VT model; stimulation of bone marrow with recombinant human granulocyte-stimulating factor increases circulating CCR2 and macrophage recruitment to venous thrombi and improves thrombus clearance.30 The discovery of ways to modulate thrombus resolution is beneficial because observations suggest that rapid VT resolution may be associated with a lower risk of PTS31 and may not confer the associated bleeding risks of anticoagulation or thrombolysis. Vein wall. Changes in macrophage phenotypes were also evident in vein wall over time, with a heavy predilection toward M2 phenotype in the stasis model at acute and chronic time points, correlating with increased vein wall thickness and fibrosis. This is consistent with observations in other organ systems, in which M2 polarization has been associated with pulmonary32 and renal fibrosis.33 Previous data have shown that the mechanism of VT, stasis vs nonstasis, contributes to altered vein wall phenotype.10 We propose that a contributing factor might be differences in macrophage polarization, with a predominantly M2 phenotype predisposing the vein wall to a more intense fibrotic response. Whether the M2 macrophages are set at the bone marrow level, as suggested here indirectly from the systemic response to venous thrombosis, or become M2 from an initial M1 influx is not clear from our studies. For example, recent work in a sterile inflammatory model of liver disease suggests an active change due to the local environment in macrophage function rather than separate recruitments.34 From a translational aspect, early interventions designed to alter macrophage phenotype toward an inflammatory state would be most important in altering the fibrotic response. Human. Our human data demonstrate for the first time in man that venous thrombi show variation in M1 and M2 markers. Although our study was not powered to detect differences among such a heterogeneous group of patients and our results should be verified with a prospectively identified, larger population, we did identify that tissue macrophages expressing either phenotype seemed to decrease over time, correlating with circulating macrophage phenotypes. For the first time, we are able to identify that variations in macrophage polarization exist in human VT. As human trials targeting macrophage response are already ongoing for inflammatory diseases such as nephritis, atherosclerosis, psoriasis, and cancer,35 these data represent a potentially new avenue of investigation for the treatment of PTS. Limitations. Sterile inflammation is a complex process, and although our data suggest that macrophages play a key role, we cannot exclude the effects of other cells and signaling pathways shown to play a role in inflammation. In addition to Toll-like receptor stimulation, nuclear factor kB or TRIF signaling pathways may play critical roles in macrophage activation during VT.14,36 Further, because iron has recently been shown to play a role in macrophage polarization in venous ulcers, and iron has been shown to accumulate during 8 days in VT, it is unclear what role this may play in influencing macrophage polarization in VT resolution.37
Although previous research supports that neutrophils and macrophages are among the most critical cells early in VT,11,38 we have not accounted for the role that platelets play in this process. It has been shown recently that platelets may promote necrotic inflammation.39-41 CONCLUSIONS In humans, the longer time that the thrombus is in contact with the vein wall, the greater the damage. Therefore, treatment strategies designed to minimize thrombus burden at early time points are imperative. Direct modulation of inflammation has not been explored in VT, and we speculate that altering the M1/M2 balance may accelerate thrombus resolution and allow the development of translatable novel therapies to treat VT. These new therapies may allow the thrombus to be more lysable, which may represent an alternative treatment avenue for reduction of PTS and avoidance of complications related to standard therapies. AUTHOR CONTRIBUTIONS Conception and design: KG, PH Analysis and interpretation: KG, AO, DC, PH Data collection: ME, EH, CL, SH Writing the article: KG, AO, PH Critical revision of the article: KG, AO, ME, CL, PH Final approval of the article: KG, AO, ME, EH, CL, SH, DC, PH Statistical analysis: KG, AO, CL, SH, PH Obtained funding: KG, PH Overall responsibility: PH REFERENCES 1. Coon WW, Willis PW 3rd, Keller JB. Venous thromboembolism and other venous disease in the Tecumseh community health study. Circulation 1973;48:839-46. 2. Anderson FA Jr, Wheeler HB, Goldberg RJ, Hosmer DW, Patwardhan NA, Jovanovic B, et al. A population-based perspective of the hospital incidence and case-fatality rates of deep vein thrombosis and pulmonary embolism. The Worcester DVT Study. Arch Intern Med 1991;151:933-8. 3. Weitz JI, Bates SM. New anticoagulants. J Thromb Haemost 2005;3: 1843-53. 4. Heit JA, Silverstein MD, Mohr DN, Petterson TM, Lohse CM, O’Fallon WM, et al. The epidemiology of venous thromboembolism in the community. Thromb Haemost 2001;86:452-63. 5. Henke PK, Comerota AJ. An update on etiology, prevention, and therapy of postthrombotic syndrome. J Vasc Surg 2011;53:500-9. 6. Deatrick KB, Luke CE, Elfline MA, Sood V, Baldwin J, Upchurch GR Jr, et al. The effect of matrix metalloproteinase 2 and matrix metalloproteinase 2/9 deletion in experimental post-thrombotic vein wall remodeling. J Vasc Surg 2013;58:1375-1384.e2. 7. Engelmann B, Massberg S. Thrombosis as an intravascular effector of innate immunity. Nat Rev Immunol 2013;13:34-45. 8. Saha P, Humphries J, Modarai B, Mattock K, Waltham M, Evans CE, et al. Leukocytes and the natural history of deep vein thrombosis: current concepts and future directions. Arterioscler Thromb Vasc Biol 2011;31:506-12. 9. Wakefield TW, Myers DD, Henke PK. Mechanisms of venous thrombosis and resolution. Arterioscler Thromb Vasc Biol 2008;28: 387-91.
JOURNAL OF VASCULAR SURGERY: VENOUS AND LYMPHATIC DISORDERS Volume 4, Number 4
10. Henke PK, Varma MR, Moaveni DK, Dewyer NA, Moore AJ, Lynch EM, et al. Fibrotic injury after experimental deep vein thrombosis is determined by the mechanism of thrombogenesis. Thromb Haemost 2007;98:1045-55. 11. Henke PK, Pearce CG, Moaveni DM, Moore AJ, Lynch EM, Longo C, et al. Targeted deletion of CCR2 impairs deep vein thombosis resolution in a mouse model. J Immunol 2006;177:3388-97. 12. Soehnlein O, Lindbom L. Phagocyte partnership during the onset and resolution of inflammation. Nat Rev Immunol 2010;10:427-39. 13. Nathan C, Ding A. Nonresolving inflammation. Cell 2010;140: 871-82. 14. Henke PK, Mitsuya M, Luke CE, Elfline MA, Baldwin JF, Deatrick KB, et al. Toll-like receptor 9 signaling is critical for early experimental deep vein thrombosis resolution. Arterioscler Thromb Vasc Biol 2011;31:43-9. 15. Ali T, Humphries J, Burnand K, Sawyer B, Bursill C, Channon K, et al. Monocyte recruitment in venous thrombus resolution. J Vasc Surg 2006;43:601-8. 16. Deatrick KB, Eliason JL, Lynch EM, Moore AJ, Dewyer NA, Varma MR, et al. Vein wall remodeling after deep vein thrombosis involves matrix metalloproteinases and late fibrosis in a mouse model. J Vasc Surg 2005;42:140-8. 17. Henke PK, Varga A, De S, Deatrick CB, Eliason J, Arenberg DA, et al. Deep vein thrombosis resolution is modulated by monocyte CXCR2mediated activity in a mouse model. Arterioscler Thromb Vasc Biol 2004;24:1130-7. 18. Diaz JA, Hawley AE, Alvarado CM, Berguer AM, Baker NK, Wrobleski SK, et al. Thrombogenesis with continuous blood flow in the inferior vena cava. A novel mouse model. Thromb Haemost 2010;104:366-75. 19. Junqueira LC, Bignolas G, Brentani RR. Picrosirius staining plus polarization microscopy, a specific method for collagen detection in tissue sections. Histochem J 1979;11:447-55. 20. Cuttle L, Nataatmadja M, Fraser JF, Kempf M, Kimble RM, Hayes MT. Collagen in the scarless fetal skin wound: detection with picrosirius-polarization. Wound Repair Regen 2005;13:198-204. 21. Baldwin JF, Sood V, Elfline MA, Luke CE, Dewyer NA, Diaz JA, et al. The role of urokinase plasminogen activator and plasmin activator inhibitor-1 on vein wall remodeling in experimental deep vein thrombosis. J Vasc Surg 2012;56:1089-97. 22. Wojcik BM, Wrobleski SK, Hawley AE, Wakefield TW, Myers DD Jr, Diaz JA. Interleukin-6: a potential target for post-thrombotic syndrome. Ann Vasc Surg 2011;25:229-39. 23. Laser A, Elfline M, Luke C, Slack D, Shah A, Sood V, et al. Deletion of cysteine-cysteine receptor 7 promotes fibrotic injury in experimental post-thrombotic vein wall remodeling. Arterioscler Thromb Vasc Biol 2014;34:377-85. 24. Comerota AJ, Oostra C, Fayad Z, Gunning W, Henke P, Luke C, et al. A histological and functional description of the tissue causing chronic postthrombotic venous obstruction. Thromb Res 2015;135: 882-7. 25. Deatrick KB, Elfline M, Baker N, Luke CE, Blackburn S, Stabler C, et al. Postthrombotic vein wall remodeling: preliminary observations. J Vasc Surg 2011;53:139-46. 26. Obi AT, Diaz JA, Ballard-Lipka NL, Roelofs KJ, Farris DM, Lawrence DA, et al. Plasminogen activator-1 overexpression decreases experimental postthrombotic vein wall fibrosis by a non-
27.
28.
29.
30.
31.
32.
33.
34.
35. 36.
37.
38.
39.
40. 41.
Gallagher et al 471
vitronectin-dependent mechanism. J Thromb Haemost 2014;12: 1353-63. Obi AT, Diaz JA, Ballard-Lipka NL, Roelofs KJ, Farris DM, Lawrence DA, et al. Low-molecular-weight heparin modulates vein wall fibrotic response in a plasminogen activator inhibitor 1dependent manner. J Vasc Surg Venous Lymphat Disord 2014;2: 441-450.e1. Willenborg S, Lucas T, van Loo G, Knipper JA, Krieg T, Haase I, et al. CCR2 recruits an inflammatory macrophage subpopulation critical for angiogenesis in tissue repair. Blood 2012;120:613-25. Geissmann F, Jung S, Littman DR. Blood monocytes consist of two principal subsets with distinct migratory properties. Immunity 2003;19: 71-82. Chen YK, Jiang XM, Gong JP. Recombinant human granulocyte colony-stimulating factor enhanced the resolution of venous thrombi. J Vasc Surg 2008;47:1058-65. Enden T, Haig Y, Klow NE, Slagsvold CE, Sandvik L, Ghanima W, et al. Long-term outcome after additional catheter-directed thrombolysis versus standard treatment for acute iliofemoral deep vein thrombosis (the CaVenT study): a randomised controlled trial. Lancet 2012;379:31-8. Sun L, Louie MC, Vannella KM, Wilke CA, LeVine AM, Moore BB, et al. New concepts of IL-10-induced lung fibrosis: fibrocyte recruitment and M2 activation in a CCL2/CCR2 axis. Am J Physiol Lung Cell Mol Physiol 2011;300:L341-53. Anders HJ, Ryu M. Renal microenvironments and macrophage phenotypes determine progression or resolution of renal inflammation and fibrosis. Kidney Int 2011;80:915-25. Dal-Secco D, Wang J, Zeng Z, Kolaczkowska E, Wong CH, Petri B, et al. A dynamic spectrum of monocytes arising from the in situ reprogramming of CCR2þ monocytes at a site of sterile injury. J Exp Med 2015;212:447-56. Hamilton JA, Achuthan A. Colony stimulating factors and myeloid cell biology in health and disease. Trends Immunol 2013;34:81-9. Dewyer NA, El-Sayed OM, Luke CE, Elfline M, Kittan N, Allen R, et al. Divergent effects of Tlr9 deletion in experimental late venous thrombosis resolution and vein wall injury. Thromb Haemost 2015;114:1028-37. Sindrilaru A, Peters T, Wieschalka S, Baican C, Baican A, Peter H, et al. An unrestrained proinflammatory M1 macrophage population induced by iron impairs wound healing in humans and mice. J Clin Invest 2011;121:985-97. Henke PK, Varma MR, Deatrick KB, Drewyer NA, Lynch EM, Moore AJ, et al. Neutrophils modulate post-thrombotic vein wall remodeling but not thrombus neovascularization. Thromb Haemost 2006;95:272-81. Brill A, Fuchs TA, Chauhan AK, Yang JJ, De Meyer SF, Kollnberger M, et al. von Willebrand factor-mediated platelet adhesion is critical for deep vein thrombosis in mouse models. Blood 2011;117:1400-7. Jackson SP, Schoenwaelder SM. Procoagulant platelets: are they necrotic? Blood 2010;116:2011-8. Panigrahi S, Ma Y, Hong L, Gao D, West XZ, Salomon RG, et al. Engagement of platelet toll-like receptor 9 by novel endogenous ligands promotes platelet hyperreactivity and thrombosis. Circ Res 2013;112:103-12.
Submitted Nov 17, 2015; accepted Mar 12, 2016.