Iron and Noncontrast Magnetic Resonance T2* as a Marker of Intraplaque Iron in Human Atherosclerosis

Abstract

Objective

Iron has been implicated in atherogenesis and plaque destabilization, while less is known regarding iron-related proteins in this disease. We compared ex vivo quantities to in vivo vessel wall T2*, which is a noncontrast magnetic resonance relaxation time that quantitatively shortens with increased tissue iron content. We also tested the hypothesis that carotid atherosclerosis patients have abnormal T2* times vs. controls that would help support a role for iron in human atherosclerosis.

Methods and Results

46 patients undergoing carotid endarterectomy and 14 subjects without carotid disease were prospectively enrolled to undergo carotid MRI. Ex vivo measurements were performed on explanted plaque and 17 mammary artery samples. Plaques vs. normal arteries had higher levels of ferritin (median = 7.3 [IQR=4 – 13.8] vs. 1.0 [0.6 – 1.3] ng/mg, P < .001) and oxidized low-density lipoprotein (0.17 [0.12 – 0.30] vs. 0.01 [0.003 – 0.03] ng/mg, P < .001) as well as hepcidin (8.7 [4.6 – 12.4] vs. 2.6 [1.3 – 7.0] ng/mL, P = .03); serum hepcidin levels did not distinguish atherosclerosis patients from controls (40.6 [18.8 – 88.6] vs. 33.9 [17.6 – 55.2], P = 0.42). Shorter in vivo T2* paralleled larger plaque volume (ρ = −.44, p = 0.01), and diseased arteries had shorter T2* values compared to controls (17.7 ± 4.3 vs. 23.0 ± 2.4 ms, P < .001).

Conclusions

Diseased arteries have greater levels of iron-related proteins ex vivo and shorter T2* times in vivo. Further studies should help define T2*'s role as a biomarker of iron and atherosclerosis.

Background

Atherosclerosis and its sequelae affect over 20 million adults in the United States alone, including 6.4 million with a history of stroke^{1, 2}. Accruing evidence suggests a pathogenic role for iron in atheroma formation and destabilization^3–7. This theory is supported by ex vivo studies demonstrating higher iron levels in atherosclerotic plaque than normal vessel tissue^{8, 9}. Levels of ferritin mRNA have also been shown to be significantly higher in atherosclerotic plaque vs. normal arteries¹⁰. Mechanistically, iron holds appeal as a contributor to atherosclerosis given its catalytic role in the formation of oxidized low-density lipoprotein (oxLDL) through Fenton chemistry^{11, 12}. This altered LDL is taken up via the macrophage scavenger receptor in an unregulated fashion leading to foam cell formation which is a key step in the formation of atherosclerotic plaque^{13, 14}. Despite these intriguing findings, studies relating total body iron levels to the risk of atherosclerotic disease and even trials of therapeutic iron depletion have yielded mixed results^15–19. Three possible explanations warrant consideration. First, blood iron levels and tissue iron content may be poorly correlated²⁰. This limits studies seeking to correlate blood iron levels and arterial wall iron as well as intervention trials targeting total body iron without impacting intraplaque iron. A second is that once atherosclerosis is advanced, iron depletion may be less effective as shown in the FeAST trial; however, a subsequent substudy²¹ did show reduced mortality with statin use, suggesting that iron depletion alone is unlikely to reverse atherosclerosis. A third possibility is that key iron homeostasis mediators have been neglected in prior work.

Not well studied to date in human atherosclerosis is hepcidin, a 25 amino acid peptide that is a central regulator of iron homeostasis²². It is an inhibitor of ferroportin, which is the only known iron exporter in human cells. It promotes iron sequestration within macrophages and blocks the absorption of iron in the small intestine²³. Via the increase in intracellular iron, hepcidin expression increases the generation of reactive oxygen species and decreases cholesterol efflux. Hepcidin is upregulated in states of iron loading or inflammation^24–27 and decreased by anemia, hypoxia, and most forms of hereditary hemochromatosis^{24, 26–28}. Both iron and hepcidin induce the release of inflammatory cytokines from monocytes²⁹. Thus, high levels of hepcidin associated with inflammation, an important factor in the pathogenesis of atherosclerosis³⁰, would be expected to lead to relatively high levels of intra-plaque iron via macrophage sequestration. This iron would then be available for the oxidation of LDL and the promotion of atherogenesis. Furthermore, the low levels of hepcidin seen in hereditary hemochromatosis may explain the lack of increased atherosclerotic risk in this iron-overloaded population.

Complementary to these ex vivo analyses are in vivo studies with magnetic resonance imaging (MRI), specifically MRI using the relaxation parameter T2* that has been validated as a noninvasive measure of tissue iron in organs such as the liver and myocardium: shorter T2* relaxation times have been shown to indicate higher iron content in these tissues^{31, 32}. This is because paramagnetic iron causes local magnetic field inhomogeneities that shorten the intrinsic T2* relaxation times of the tissue containing iron. Our group has previously shown that in vivo T2* imaging can be performed in the carotid arteries. However, it remains unknown what form of iron is responsible for carotid artery T2* changes. Thus, in this work we sought to further investigate the role of iron in atherosclerosis by examining both in vivo MRI measurements and ex vivo measures of iron and hepcidin in atherosclerotic plaque and healthy arteries. We hypothesized that atherosclerotic carotid arteries have greater levels of iron-related proteins vs. healthy arterial wall. We further sought to define whether protein-bound or insoluble iron is responsible for the T2* changes detected noninvasively in patients with carotid artery atherosclerosis. T2* measurements were then used to characterize arterial tissue in vivo in patients with atherosclerotic disease and compared to measurements in healthy controls.

Methods

Patients

Patients with atherosclerotic carotid artery disease scheduled for carotid endarterectomy based on standard clinical indications were prospectively enrolled to undergo preoperative MRI and plaque collection at the time of surgery. Patients with ferromagnetic material, active implants such as pacemakers, defibrillators, or neurostimulators, aneurysm clips, those with known claustrophobia, and those unable to provide informed consent were excluded from enrollment. No patients had a history of iron overload or significant blood transfusion. All subjects gave written informed consent to participate in this IRB approved protocol. Plaques were collected at the time of explantation during endarterectomy and immediately frozen in liquid nitrogen. They were then stored at −80°C until analysis. Any blood found on the surface of the plaque was carefully removed and the surface washed with normal saline before freezing and further analysis.

Controls

Seventeen sections of the internal mammary artery were obtained from the Ohio State University Tissue Procurement Lab (Cooperative Human Tissue Network, National Cancer Institute). The internal mammary artery (IMA) was selected for comparison due to its known resistance to atherosclerosis^{33, 34}. Imaging and serologies were also obtained in a cohort of 14 healthy women with no or 1 risk factor for atherosclerosis undergoing the identical carotid T2* imaging as part of a longitudinal study of iron and perimenopausal atherosclerosis risk³⁵.

Ferritin and Oxidized LDL Quantification

Plaques and control mammary artery samples were freeze-fractured under liquid nitrogen using a stainless steel pulverizer (BioSpec Products, Bartlesville, OK). The powdered samples were transferred to 50 mL conical tubes, and resuspended in (3 mL/mg sample) PBS containing cOmplete Protease Inhibitor cocktail (Roche Applied Science, Mannhelm, Germany). The suspension was bubbled gently with argon, and then extracted overnight with rocking at 4° C. Insoluble material was removed by centrifugation (2,500 g) at 4° C, and the supernatant was transferred to a new tube. The oxLDL concentration was quantified using an ELISA (catalog no. 10-1143-01) from Mercodia Inc. (Uppsala, Sweden), and ferritin was quantified using an ELISA (catalog no. FR065T) from Calbiotech (San Diego, CA). Protein concentration was determined using the Pierce BCA Protein Assay kit (Thermo Fisher, Rockford, IL) following the manufacturer's protocol, using BSA as the standard.

Hepcidin Quantification

Samples of hepcidin-25 standards in serial dilutions brought to 50 μl were added to the antiserum pre-coated plate for competitive immunoassay (hepcidin-25 peptide EIA kit S-1337, Bachem Group, Torrance, CA). Serum diluted at 1:10 and plaque samples diluted at both 1:5 and 1:10 in EIA buffer were used in initial measurement. Samples outside the standard range were repeated using appropriate dilutions. Plaque hepcidin concentration was averaged from both dilutions and normalized to plaque protein concentrations. The coefficient of variation (CV) using a hepcidin-25 concentration of 1.56 ng/ml was 3.49% intra-assay and 3.43% inter-assay. The hepcidin-25 standard Liver-Expressed Antimicrobial Peptide 1 (LEAP1) from Peptides International Inc (Lousiville, KY) was used for validation of this EIA kit. The R-squared was 0.997 for LEAP1 from 0 – 50 ng/ml using a sigmoid regression curve fitting algorithm.

Insoluble Iron Quantification

The insoluble portions of the samples (0.5 g to 3 g wet weight) were split into ~0.5 g (wet weight) aliquots and weighed in pre-dried, pre-weighed PFA Teflon digestion vials (Savillex, Eden Prairie, MN). Samples were dried completely at 95°C in an oven. Dry sample weight was measured by difference (most were ~ 0.1 g dry). To each sample, 3 mL ultra-pure (Double Distilled “Veritas”, GFS Chemical, Powell, OH) nitric acid was added and the vessel closed. The samples were digested on a hotplate for 4 hours (or until clear for some). Digested samples were transferred to 125 mL LDPE bottles and diluted to 100 mL total volume. Cobalt was added during dilution to a final concentration of 10 ppb as an internal standard. Samples were analyzed using an Element 2 inductively coupled plasma-sector field mass spectroscope (ICP-MS, Thermo Finnigan, Bremen, Germany) used in medium resolution (R = 4000). The samples were introduced into the ICP-MS by a PFA-ST concentric nebulizer (Elemental Scientific, Omaha, NE) and a PFA spray chamber (Elemental Scientific). The sample was pumped at an uptake rate of 0.5 mL/min to the nebulizer. The amounts of Fe in each aliquot of the insoluble portions of each sample were summed to obtain the total amount of Fe in the insoluble portion of each sample.

MRI Protocol and Analysis

Patients underwent carotid magnetic resonance imaging approximately one week prior to endarterectomy. All MRI examinations were performed on the identical 3.0 Tesla scanner (MAGNETOM Verio, Siemens Healthcare; Erlangen, Germany) with a custom built 8-element (4 left and 4 right) coil for carotid imaging (Massachussetts General Hospital). After localization using steady state free precession images, dark blood images of the carotid arteries were obtained utilizing a T1-weighted three-dimensional turbo spin echo sequence known as SPACE (Sampling Perfection with Application optimized Contrast using different flip angle Evolution) that our group has optimized for carotid vessel wall imaging³⁶. The parameters for the SPACE sequence were an echo time (TE) of 22 ms, repetition time (TR) 700 ms, variable flip angle, 2 averages, and a resolution of 0.7 mm isotropic. This dataset was transferred to a three-dimensional viewing station and used to select the point of maximum carotid stenosis and to generate a set of slices through each plaque for plaque volume measurement. The maximum stenosis location was used for T2* measurement with a multiple-echo, gradient-echo acquisition with echo times (TE) of 2.7, 7.5, 12.3, 17.1, and 22.4 ms (Figure 1). Up to three slices were acquired through each plaque with in-plane spatial resolution of 0.5 mm × 0.5 mm and slice thickness of 4.0 mm. Using images from all TEs, in-house developed software³⁷ was used to estimate T2* at each pixel by a weighted least square estimation of the encompassing 3×3 neighborhood with outlier detection and deletion. The T2* values were then averaged across all slices obtained. Liver T2* measurement was also performed as a surrogate measure of total body iron stores³⁸.

Figure 1.

Magnetic resonance gradient echo images acquired at multiple echo times (TEs) are acquired to measure carotid artery T2* relaxation time. From left, TEs are 2.7 ms, 7.5 ms, 12.3 ms, 17.1 ms, and 22.4 ms. T2* is computed from the signal intensity decay over successive TEs within a region of interest encompassing the plaque.

To measure vessel wall volume, a stack of 20 contiguous slices was constructed from the three dimensional SPACE dataset on an offline workstation. The slices were 0.7 mm thick, oriented perpendicular to the long axis of the vessel and centered on the area of greatest carotid stenosis. In each slice the inner and outer lumen were manually contoured by an experienced observer. The vessel wall volume, a surrogate for plaque burden, was then calculated using the following formulae:

• $$\text{Vessel wall area}=\text{Outer vessel area - Lumen area}$$
• $${\text{Vessel wall volume}}_{\text{slice}}=\text{Vessel well area * Slice thickness}$$
• $${\text{Vessel wall volume}}_{\text{total}}=\Sigma {\text{Vessel well volume}}_{\text{slice}}$$

Statistical Analysis

Nonparametric methods were used in most of our analyses because each measurement of interest (other than intraplaque T2*) had a skewed distribution. Wilcoxon rank sum tests were used to compare the medians of ex vivo carotid plaque measurements of diseased patients versus ex vivo mammary artery controls, and to compare in vivo diseased carotid artery measurements versus in vivo measurements in non-diseased carotid arteries of perimenopausal women. The average intraplaque T2* of diseased patients was compared to the average of non-diseased perimenopausal women using a two sample t-test. Relationships between measurements were quantified using Spearman correlations (ρ) and depicted using scatter plots and locally weighted scatterplot smoothing (Lowess, bandwidth=0.8). Bland-Altman analysis was used to evaluate agreement in T2* measurements obtained from two different readers³⁹. Non-normally distributed variables were expressed in terms of median [Interquartile range (IQR)] and normally distributed variables were expressed in terms of mean ± standard deviation. SAS Version 9.2 (SAS Inc., Cary, NC) and Intercooled Stata Version 11.2 (Stata Corp., College Station, TX) were used in the analyses.

Results

Clinical Characteristics

A total of 46 patients with atherosclerosis (11 symptomatic, 35 asymptomatic) were enrolled in the study (Table 1). The range of serum hepcidin values was greater in atherosclerosis patients [18.8–88.6] vs. control subjects [17.6–55.2], though the difference in median levels between groups did not reach statistical significance (40.6 [18.8–88.6] vs. 33.9 [17.6–55.2], P = .42). Seventeen control mammary arteries were obtained for analysis. Age and gender information was available for artery donors; the age of these subjects was similar to the diseased patients, though the proportion of female controls was considerably smaller. The majority (79%) of carotid disease patients were on a statin medication compared to 14% of control women without evident carotid disease (P < .001).

Table 1.

Baseline Patient Characteristics. Normally distributed variables expressed as mean ± SD, non-normally distributed variables expressed as median [IQR], categorical variables expressed as frequency (%).

	Atherosclerosis Patients (N = 46)	Mammary Artery Controls (N=17)	Healthy Women Controls (N=14)
Age, years	64.0 ± 10.3	63.1 ± 13.1	49.1 ± 3.7***
Body mass index, kg/m2	28.6 ± 6.5a	n/a	29.5 ± 8.3
Diabetes	15 (33%)	n/a	n/a
Female	17 (37%)	2 (12%)*	14 (100%)***
Smoker	20 (43%)	n/a	1 (7%)**
Statin use	36 (78%)	n/a	2 (14%)***
Serum ferritin, ng/mL	94 [56–159]c	n/a	n/a
Serum hepcidin, ng/mL	40.6 [18.8–88.6]	n/a	33.9 [17.6–55.2]
Hematocrit, %	38.5 ± 5.7	n/a	40.8 ± 2.4
Low-density lipoprotein, mg/mL	97.3 ± 39.2d	n/a	153.6 ± 42.5***
High-density lipoprotein, mg/mL	37.0 [31.5–45.0]e	n/a	62.5 [47.0–71.0]***
Triglycerides, mg/dL	52.0 [103.0–222.5]e	n/a	75.0 [51.0–106.0]***

^aN=45;

^cN=43;

^cN=47 (includes duplicate measure from patient with bilateral disease);

^dN=37, 3 missing due to triglycerides greater than 400;

^eN=40

^*0.05<p<0.1;

^**0.01< p < 0.05;

^***p < 0.01 compared to carotid artery patients

Ex Vivo Analyses

Ex vivo analysis of arterial samples included measurement of ferritin, hepcidin, oxidized LDL, and insoluble iron both in carotid artery plaques obtained by endarterectomy as well as in control mammary artery samples. Ferritin (median = 7.3 [IQR = 4–13.8] vs. 1.0 [0.6–1.3] ng/mg protein for control, P < .001), oxidized LDL (0.17 [0.12–0.30] vs. 0.01 [0.003–0.03] ng/mg protein, P < .001) and hepcidin (8.7 [4.6–12.4] vs. 2.6 [1.3–7.0] ng/mL, P = .03) levels were significantly higher in atherosclerotic plaques vs. control arteries (Figures 2–4). Due to the greater percentage of females among carotid artery patients than controls, we repeated our comparisons among males and found similar results with the only exception being that the difference in hepcidin levels was no longer statistically significant (P = .11). The sample size was not large enough to perform the corresponding analysis among females.

Figure 2.

Ferritin levels were significantly higher in carotid artery atherosclerotic plaques compared to measurements obtained in control mammary arteries (P < .001). N=46 atherosclerotic samples and N=17 control samples.

Figure 4.

Hepcidin levels were significantly higher in carotid plaques samples compared to measurements obtained in control mammary artery samples (P = .03). N=46 atherosclerotic samples and N=10 control samples.

In samples in which the insoluble fraction was available for analysis (38 carotid artery plaques and 7 mammary arteries), an unusually high insoluble iron value was obtained in one mammary artery that was an order of magnitude higher than that seen in any of the other control samples. Even after excluding this value (notably from a 76 year-old individual vs. a mean age of 55 years for the others controls), the somewhat higher insoluble iron levels in atherosclerotic artery samples [0.023–0.081] vs. mammary artery controls [0.028–0.048] did not reach statistical significance (0.040 vs. 0.036 mg/g dry weight, P = .55, Figure 5).

Figure 5.

The range of insoluble iron levels was greater in carotid plaques samples compared to measurements obtained in control mammary artery samples; however, even after excluding an unusually high outlier in a control sample obtained from a 76 year-old individual, median values were similar (P = .55).

In atherosclerotic carotid artery plaques, a significant correlation was found between ferritin and insoluble iron (ρ = .63, P < .001). Intraplaque hepcidin showed a modest correlation with intraplaque oxidized LDL (r = .31, P = .02).

In Vivo Carotid Plaque T2* Imaging and Ex Vivo Intraplaque Iron Measures

Of the 47 carotid arteries interrogated with in vivo MRI, 5 had motion artifact limiting T2* image quality and in one instance T2* scans were prescribed through the contralateral plaque vs. the side scheduled for endarterectomy leaving 41 in vivo plaque T2* measurements. There was excellent reproducibility between 2 independent observers for the measurement of plaque T2* relaxation times by Bland-Altman analysis (r = .84, P < .001) and no significant bias (r = .22, P = .14, Figure 6). Liver T2* values in atherosclerosis patients were similar to those in healthy controls (15.6 [12.7–19.7] vs. 15.9 [14.0–18.8] ms, P = .71), and fell within the range of previously reported values in healthy volunteers⁴⁰. Plaque T2* did not correlate significantly with either liver T2* (ρ = .16, P = .33) or serum ferritin (ρ = .15, P = .39).

Figure 6.

Bland-Altman analysis demonstrated excellent inter-observer agreement in carotid artery T2* measurements (r = .84, P < .001) and no significant bias (r = .22, P = .14), with an absolute mean difference of 0.45 ms and standard deviation of 4.26.

Comparing in vivo plaque MRI-T2* values to ex vivo measures suggested an inverse trend with ex vivo intraplaque insoluble iron, but did not reach statistical significance (ρ = −.19, P = .27). Plaque T2* was not significantly associated with plaque ferritin level (ρ = −.03, P = .87, Figure 7).

Figure 7.

Ferritin levels did not significantly correlate with intraplaque T2* in patients with carotid artery atherosclerosis.

Comparison of Diseased and Normal Carotid Arteries using In Vivo MRI

In patients with carotid artery atherosclerosis, plaque volume tended to be higher in patients with lower values for intraplaque T2* (ρ = .32, P = .06). One outlier with an unusually small plaque volume for the shortest plaque T2* value was noted; the plaque volume may have been underestimated due to the geometry of the lesion. Excluding this outlier yielded improved correlation and statistical significance (ρ = −.44, P = .01). When comparing in vivo T2* in diseased vs. plaque-free carotid arteries, T2* was significantly shorter in arteries of patients with atherosclerosis vs. those of relatively healthy but at-risk controls (17.7±4.3 vs. 23.0±2.4 ms, P < .001, Figure 8). This trend held when the comparison was restricted to females and to patients not on statin therapy.

Figure 8.

In vivo T2* measured with noncontrast MRI in the atherosclerotic carotid artery wall was shorter than in the healthy artery wall of control subjects (P < .001). Error bars represent standard deviations.

Discussion

In this work, we have shown that i) intraplaque iron and related proteins are present in greater quantities in atherosclerotic plaque vs. plaque-free arterial samples and ii) the parameter T2* that can be measured in vivo using noncontrast MRI is shorter in diseased arteries compared to arteries of at-risk individuals without apparent disease. Prior validation studies of noninvasive MRI have indicated iron as the substrate for shortened T2* in heart and liver^{20, 31, 41, 42}, though less understanding exists as to the form of iron responsible for the change in this measurable relaxation parameter. We found a possible linear relationship between insoluble iron and T2* among patients with larger levels of insoluble iron in plaque (>0.01 mg/g), and no appreciable relationship between intraplaque ferritin and T2*. This suggests that the `shielding' of iron within the large ferritin protein making intra-ferritin iron less influential on MRI T2*⁴³. We speculate that hemosiderin, a small and insoluble form of iron in biologic tissues not directly quantifiable by specific ELISA or related techniques due to its fragmented nature, is most likely responsible for T2* shortening in plaque. Alternatively, lack of correlation between ferritin and T2* may indicate relatively modest iron loading of ferritin or formation of antiferromagnetically-coupled clusters with minimal effect on magnetic resonance relaxation times⁴⁴. It is also possible that ferritin may have a protective effect by sequestering iron and preventing it from participating in LDL oxidation. Further studies are needed to address this area of uncertainty.

This work does not fill all gaps in knowledge regarding the location, forms and mediators of iron in plaque. Additional experiments controlling for the total insoluble fraction amount available per plaque may offer better insights into insoluble vs. other forms of iron in atherosclerosis. Also, we did not attempt to distinguish arterial wall layers in our analyses; further studies dissecting the intima and inner half of the media of control arteries compared to atheroma samples may be considered. Ideally, one would compare to healthy carotid arteries vs. IMAs, as different distributions among arterial beds may be a confounder.

Serum ferritin levels and liver iron have been used as surrogate markers for whole body iron status, however previous studies found no correlation between myocardial T2* and liver iron and none or only a weak correlation with serum ferritin^{45, 46}. Similarly, we found no correlation between plaque T2* and either liver T2* or serum ferritin levels. As such, these commonly used markers of body iron status should not be considered as markers of iron involved in atherosclerosis.

Prior studies have suggested that senescent red cells infiltrating damaged endothelium bring iron via heme to oxidize lipids⁶. Our results demonstrating shorter T2* in diseased arteries in comparison to healthy arteries supports the hypothesis that iron accumulation is pathologic.

The sample size did not afford consideration of gender as a covariate in analyses. Larger studies are needed to explore this potential influence, noting that iron balance in post-menopausal women is more similar to that of men. While the present study was not designed to assess age vs. arterial iron content (rather, we compared groups of individuals of similar age), further studies are warranted to address the possibility that aging itself leads to iron accrual in the arterial wall, fueling early atherosclerosis development. Similarly, we did not design the present work to predict self-reported symptom status by plaque characteristics; rather, our primary goal was to establish the presence of iron and related proteins in diseased vs. healthy arteries, and to show that a noninvasive MRI biomarker of iron is measurably different in diseased vs. healthy arteries. We hope that this work helps motivate the implementation of larger-scale trials in patients with carotid atherosclerosis of varying stenosis severity that couples MRI-based plaque biomarkers including T2* with careful adjudication of symptoms and events due to plaque instability. While ultrasonography performs well in identifying presence of carotid artery disease and stenosis severity, noncontrast MRI T2* characterization of carotid plaque could potentially improve outcomes for patients with intermediate stenosis carotid disease who still realize a 5% or greater annualized stroke risk with current management guidelines. In this subgroup, more refined imaging might afford not only timely relief of mechanical stenosis but also biologic interventions that target mechanistic factors in the plaque microenvironment proven to fuel events.

Our focus on total plaque iron and in vivo T2* measurements did not include separation of plaques into sections for histopathology, although our group and others have previously shown that intraplaque iron occurs in the intima in aggregate form or within macrophages^{6, 32}. Further histopathological studies on intraplaque hemorrhage may further inform our ex vivo findings. Saeed and colleagues demonstrated hepcidin-induced increased iron content within macrophages that had been previously exposed to hemoglobin-haptoglobin complexes, with decreased cholesterol efflux from these macrophages⁴⁷. This exciting work coupled with our studies suggesting utility of noninvasive MRI T2* as a marker of intraplaque iron behoove further investigations of novel anti-atherosclerosis therapies targeting intraplaque iron homeostasis.

Clinical Relevance.

Clinical evaluation of patients with carotid atherosclerosis routinely assesses luminal stenosis but not plaque characteristics that may predict instability. This work studied iron, which forms oxidized lipid central to atheroma growth, in carotid plaque. More ferritin was found ex vivo in diseased vs. healthy arteries. Diseased arteries also had more oxidized low-density lipoprotein and hepcidin, a protein that promotes iron retention in macrophages. Furthermore, an in vivo, noncontrast MRI measure of endogenous plaque iron content paralleled plaque volume. Further studies targeting intraplaque iron are warranted to refine management of patients with carotid artery disease.

Figure 3.

Oxidized LDL levels were significantly higher in carotid plaques compared to measurements obtained in control mammary arteries (P < .001). N=46 atherosclerotic samples and N=17 control samples.