Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Feb 1.
Published in final edited form as: Brain Behav Immun. 2021 Nov 17;100:1–7. doi: 10.1016/j.bbi.2021.11.003

Prospective Associations Between Neighborhood Violence and Monocyte Pro-Inflammatory Transcriptional Activity in Children

Gregory E Miller 1, Edith Chen 1, Eric Finegood 1, Daichi Shimbo 2, Steve W Cole 3
PMCID: PMC8766930  NIHMSID: NIHMS1758661  PMID: 34800620

Abstract

Individuals exposed to persistent neighborhood violence are at increased risk for developing mental and physical health problems across the lifespan. The biological mechanisms underlying this phenomenon are not well understood. Thus, we examined the relationship between children’s exposure to neighborhood violence and inflammatory activity, a process involved in the pathogenesis of multiple health problems. 236 children from the Chicago area participated in a two-year longitudinal study (mean age at baseline, 13.9 years; 67% female; 39% White, 34% Black, 33% Hispanic). Neighborhood violence was measured as the homicide frequency in a child’s Census block group in the five years before study entry. Fasting blood was drawn at study entry and two years later (in eighth and tenth grade). The blood was used to quantify protein biomarkers of systemic inflammatory activity and perform genome-wide expression profiling of isolated monocytes. Neighborhood violence was associated with higher systemic inflammatory activity at both assessments. It also was associated with a monocyte transcriptional profile indicative of increased signaling along the nuclear factor-kappa B (NF-κB) and activator protein 1 (AP-1) control pathways, which are key orchestrators of pro-inflammatory effector functions. Neighborhood violence also was associated with transcriptional indications of higher beta-adrenergic and lower glucocorticoid signaling, which could function as neuroendocrine conduits linking threatening experiences with inflammatory activity. Neighborhood violence was not associated with two-year changes in protein biomarkers, although it did presage a transcriptional profile indicative of increasing AP-1 and declining glucocorticoid signaling over follow-up. Collectively, these observations highlight cellular and molecular pathways that could underlie health risks associated with neighborhood violence.

Keywords: violence, disparities, children, inflammation, monocytes

INTRODUCTION

Although violent crime has declined in the United States over the past several decades, the rate of progress has been uneven. Decades of economic disinvestment, residential segregation, and other discriminatory practices have resulted in marked geographical disparities in the prevalence of violence (Sharkey 2018). Data from Chicago illustrate this pattern vividly: there is a 65-fold difference in the frequency of murder between the city’s most and least violent neighborhoods (Papachristos et al. 2018).

Children who live in neighborhoods where violence is heavily concentrated are at increased risk for multiple forms of psychopathology, including major depressive disorder, generalized anxiety, post-traumatic stress, and substance misuse (Slopen et al. 2012; Zimmerman & Kushner 2017; Fowler et al. 2009). Of course, not all children from communities with high violence go on to have mental health problems (Foster & Brooks-Gunn 2009). But a sizeable minority do and, in some cases, the risks persist for years after the focal exposures (Zimmerman & Kushner 2017). Accumulating evidence indicates that neighborhood violence may also increase children’s risk for physical health problems (Wright et al. 2017; Suglia et al. 2015), including the development and worsening of asthma (Ramratnam et al. 2015; Wright et al. 2004), and the expression of early signs of cardiometabolic disease (Wilson et al. 2002; Theall et al. 2012; Theall et al. 2017). These health risks are observed among youth who have not been personally affected by violence, suggesting there may be physiological consequences of indirect or “vicarious” exposure (Finegood et al. 2020; Miller et al. 2018).

How could indirect exposure to neighborhood violence plausibly influence pathogenesis of this broad array of mental and physical health problems? Animal models show that chronic experiences of threat can accentuate inflammation in multiple tissues, including the heart, arterial wall, spleen, lungs, and brain (Heidt et al. 2014; Avitsur et al. 2006; McKim et al. 2018; Wohleb et al. 2013; Niraula et al. 2019). Acting via sympathetic nerves that innervate bone marrow, chronic threats mobilize classical monocytes into circulation (Powell et al. 2013; McKim et al. 2018). These cells migrate into tissue spaces where a sterile and/or microbial threat is present and display an aggressive inflammatory phenotype, marked by pronounced cytokine responses to challenge and insensitivity to inhibitory signals from glucocorticoids (Weber et al. 2017; Nathan & Ding 2010; Reader et al. 2015). In model systems, this threat-evoked inflammatory activity has led to anxiety-like behavior, mood disruption, severe respiratory illness, and atherosclerotic progression (Heidt et al. 2014; Avitsur et al. 2006; McKim et al. 2018; Wohleb et al. 2013; Niraula et al. 2019).

Despite the evidence from animal models, there has been limited empirical attention to these processes in humans. One study of adults observed higher circulating levels of C-reactive protein (CRP) among men living in neighborhoods with recent crime spikes, but this association was not apparent for women (Browning et al. 2012). Another focused on children, and found those living in high-crime and high-poverty neighborhoods were more likely to have CRP values > 3.0 mg/L compared to peers in safer, wealthier areas (Browning et al. 2012). Another study of youth (Finegood et al., 2020, #76166) reported that neighborhood homicide rates were marginally associated with several protein biomarkers of inflammation, reflected in a composite of CRP, IL-6, IL-8, IL-10, and TNF-α. Interestingly, neighborhood homicide was associated with higher circulating numbers of classical monocytes, which as noted above are key cellular players in threat-evoked inflammatory activity. (The latter paper was based on cross-sectional analyses of a sub-group of participants in the current study.)

These human studies are suggestive of a relationship between neighborhood violence and inflammatory activity. However, they have primarily used cross-sectional designs and generally been limited to protein biomarkers of inflammation. As a result, questions remain about the durability of any relationships over time, and the cellular and molecular processes involved. Here, we address these open questions in a two-year prospective study of children who were ages 12–14 at baseline. This period is a time of life when many chronic health problems begin to manifest, e.g., mood disorders, or to develop silently, e.g., atherosclerosis. We estimated children’s exposure to neighborhood violence with geospatial data, and quantified inflammatory activity using standard protein biomarkers and transcriptional profiling of monocytes.

Based on animal models (Heidt et al. 2014; Powell et al. 2013; Weber et al. 2017; McKim et al. 2018), we hypothesized that neighborhood violence would be associated with higher systemic inflammatory activity, as reflected in the protein biomarkers, and a monocyte transcriptional profile indicative of higher sympathetic outflow, lower glucocorticoid sensitivity, and greater effector activity. We also hypothesized these patterns would become more pronounced with time, as children were cumulatively exposed to more violence.

METHODS

Sample

Children from the Chicago area were recruited through ads in media, public transit, and schools. To be eligible, children had to be in eighth grade, English-speaking, and in good health, defined as (a) non-pregnant, (b) without a history of chronic medical or psychiatric illness, (c) free of prescription medications for the past month, (c) without acute infectious disease for two weeks, and (d) without MRI scanning contra-indications. Each child gave written assent to participate, and a parent or guardian gave written consent. Northwestern University’s IRB approved the protocol.

277 children enrolled in the two-wave study. The Time 1 visits occurred when children were in eighth grade (mean age, 13.9 years, SD = .52), and entailed surveys, interviews, a fasting antecubital blood draw, and anthropometric measurements. A parent or guardian attended the visit, and provided data about household demographics. The Time 2 visits occurred roughly two years later, when children were in tenth grade (mean age, 16.0 years, SD = .54), and also entailed surveys, interviews, fasting antecubital blood, and anthropometric measures. The mean duration between Time 1 and 2 was 24.0 months.

Neighborhood Violence

At Time 1, each child’s residential address was geocoded at the block-group level of resolution. Block groups consist of 600–3000 people and are the smallest geographic units for which the US Census Bureau publicly reports information. For each block group, Applied Geographic Solutions estimates a neighborhood murder index (CrimeRisk), based on data that local police provide to the Federal Bureau of Investigation. We used values from 2010–14, the five-year period before the study began.

Systemic Inflammatory Activity

At both timepoints, we quantified six biomarkers of systemic inflammatory activity implicated in mood disorders and cardiovascular disease (Ridker 2016; Hodges et al. 2015; Miller & Raison 2016). They were C-reactive protein, IL-6, IL-8, IL-10, TNF-α, and soluble urokinase-type plasminogen activator receptor (suPAR). Biomarkers were assayed in serum collected between 8:00–10:00am under overnight fasting conditions. CRP was measured in duplicate by high-sensitivity immunoturbidimetric assay on a Roche/Hitachi cobas c502 instrument. Cytokines were measured in triplicate by four-plex immunoassay (Aldo et al. 2016) on an microfluidic platform (Simple Plex; Protein Simple). suPAR was measured in duplicate by immunoassay (Human Quantikine ELISA; R&D Systems). Intra-assay coefficients of variation ranged from 1.6% – 5.0%.

Most biomarkers were skewed and/or kurtotic, so we normalized their distributions with log-10 transformations. The logged values were standardized (mean = 0; SD = 1), and averaged into a protein biomarker composite, where higher values reflected more inflammatory activity. It had good internal consistency, with Cronbach’s alpha = .65 and .70 at Time 1 and 2. The composite reduces the number of statistical tests performed and accordingly the rate of false discoveries.

Monocyte Transcriptional Profiling

At both timepoints, 10 mL of fasting antecubital blood was drawn into a Cell Preparation Tube (Becton-Dickinson) between 8:00–10:00 am. After peripheral blood mononuclear cells had been isolated by centrifugation, an automated cell sorter (autoMACS Pro; Miltenyi Biotec) was used to positively select monocytes (CD14+). The isolated cells were disrupted and homogenized in QIAshredder tubes containing RLT Plus Buffer (Qiagen) and frozen at −80° C until the study ended. Total RNA was then extracted using PCR-clean and RNAse-free techniques (Qiagen RNeasy).

Genome-wide transcriptional profiling was conducted on 300 ng of total RNA (Cole et al. 2020). Samples were tested for suitable mass (≥ 10 ng by NanoDrop One spectrophotometry) and integrity (RNA Integrity Number ≥ 3 by Agilent TapeStation electrophoresis), converted to cDNA libraries using the Lexogen QuantSeq 3′ FWD enzyme system, and sequenced in multiplex on an Illumina HiSeq 4000, targeting > 10 million single-strand 65-nt sequence reads per sample. Samples yielded an average of 14.0 million sequence reads (SD = 2.5 million), each mapped onto the GRCh38 human transcriptome sequence with STAR aligner (Dobin et al. 2013) (average of 94.6% reads mapped successfully). Counts were pre-standardized to transcripts per million mapped reads, and normalized to equate expression of 11 standard reference transcripts (Eisenberg & Levanon 2013). Endpoint validity metrics confirmed high inter-sample profile consistency (mean correlation among samples averaged r = .92, SD .02). Transcript abundance values were floored at 1 normalized transcript per million mapped reads to minimize spurious variability and log2-transformed for analysis.

Covariates

To minimize residual confounding by variables associated with violence, we included a panel of covariates selected a priori in all statistical models. They were age (in years), sex (male = 0, female =1), dummy variables reflecting self-reported identity as White, Black, and Hispanic (in each case; no = 0, yes =1), as well as body mass index (BMI; percentiled based on age and sex; (Kuczmarski et al. 2002) and pubertal status, measured using a validated self-report measure (Peterson et al. 1988).

Sensitivity Analyses

Neighborhood violence often co-occurs with other social disadvantages, e.g., poverty, which themselves are risk factors for health problems (Manduca & Sampson 2019; Havranek et al. 2015). To determine incremental risks associated with neighborhood violence, we did a sensitivity analysis with a general indicator of social disadvantage. It was calculated by assigning children one point for each of these risks (Miller et al. 2014): living in a single-parent household; having an unemployed parent; family income-to-needs ratio < 2.00; receipt of government financial assistance; and parents with high school education or less.

Living in a dangerous neighborhood increases the chances that one will be personally affected by violence. To determine what role these direct exposures play, we used a validated questionnaire (Thomson et al. 2002) to characterize each child’s history of experiencing and witnessing violence. Severe exposures were quite rare in this sample; only 7 children endorsed having been shot at with a gun or attacked with a knife. Thus, we adopted a broader perspective on victimization, and used a count variable to reflect the number of different types of violence to which each child had been exposed. Types of violence included in this variable were being shot at or attacked with a knife; being punched, kicked, or pushed in a fight; witnessing a gun or knife attack, and having family or friends who had been harmed by violence. In the analytic sample, values ranged from 0–7, with a mean of 1.10 (SD = 1.35).

Missing Data

236 of the 277 enrolled children were included in cross-sectional analyses. Missing cases were a result of unsuccessful venipuncture (n=2), geocoding failures (n=9), technical problems with monocyte isolation and/or invalid RNA sequencing (n=30). 225 children returned for a follow-up visit two years later (95.4%); reasons for attrition included death, loss of contact, and relocation. At Time 2 specimens from another 7 children were excluded due to invalid RNA sequencing data. Thus, for longitudinal analyses the analytic sample included 218 children.

The subgroup with missing values was comparable to the broader sample on demographic and biobehavioral variables, as well as neighborhood murder (all p’s > .18). The only exception was sex; boys were more likely than girls to have missing data (p = .008).

Statistical and Bioinformatic Approach

Our first hypothesis, that neighborhood violence would covary with systemic inflammatory activity, was tested in linear regressions in the software package SPSS Version 27.0. Separate models were estimated for inflammatory activity measured at Time 1 and Time 2. A change model was also estimated, where Time 2 values were the outcome, and Time 1 values were an additional covariate.

Our second hypothesis, that neighborhood violence would covary with transcription activity in monocytes, was tested with the Transcription Element Listening System (TELiS), a promoter sequence-based bioinformatics algorithm (Cole et al. 2005). We focused on five transcription control pathways selected a priori: NF-κB/Rel and AP-1, which orchestrate pro-inflammatory signaling; MAF, which promotes myeloid cell activation; CREB, which mediates beta-adrenergic signaling, and the glucocorticoid receptor, which mediates cortisol signaling.

Input transcripts were those showing > 2-fold difference in average expression over a 4-standard deviation (SD) range of variation in neighborhood violence, following covariate adjustment. Point estimates of differential expression serve as TELiS input because they provide more reliable results than screening based on p-/q-values (Cole et al. 2003; Norris & Kahn 2006; Shi et al. 2008). Analyses scanned promoter sequences using the TRANSFAC V$CREL_01, V$AP1_Q2, V$VMAF_01, $GRE_C and V$CREB_02 matrices position-specific weight matrices. TELiS used 9 different combinations of promoter sequence length and transcription factor-binding motif (TFBM) detection stringency (Cole et al. 2005). Log2-transformed TFBM ratios were averaged across the permutations and tested for statistical significance using standard errors from bootstrap resampling of linear model residual vectors (controlling for potential correlation across genes).

Similar to above, we estimated separate Time 1 and Time 2 models, as well as a change model, where Time 2 values were the outcome and Time 1 values were a covariate. Because variability in gene expression increased markedly over the two-year follow-up, the change models used a higher threshold to maintain strong signal-to-noise ratios (> 2.4-fold difference in average expression over a 4-SD range of variation in neighborhood violence).

RESULTS

Table 1 describes the analytic sample. At Time 1, all children were in eighth grade, aged 12–14 years, and 67% were female. The sample was diverse - 39%, 34%, and 33% of the children identified as White, Black, and Hispanic, respectively. There was marked variability in exposure to neighborhood violence. The index is calculated so a value of 100 represents the country’s average block group. The mean in the sample was 247, but values ranged from 3–994, and were skewed to the right. The median was 158.

Table 1:

Characteristics of the sample at Time 1 (n=236).

Characteristic N (%) or Mean (SD)
Age, years 13.9 (0.52)
Sex, female 158 (67.0%)
Self-identified race, White (non-Latinx) 81 (34.3%)
Self-identified race, Black (non-Latinx) 80 (33.9%)
Self-identified race, Other (non-Latinx) 15 (6.4%)
Self-identified ethnicity, Latinx (any race) 78 (33.1%)
Socioeconomic disadvantages (count, 0–5) 1.5 (1.5)
Pre, Early, or Mid Puberty 83 (35.2%)
Late or Post Puberty 153 (64.8%)
Body mass index (percentile based on age and sex) 70.7 (26.4)
Neighborhood murder index for 2010–14 246.8 (261.8)
C-reactive protein (mg/L) 1.3 (3.4)
Interleukin-6 (pg/mL) 1.8 (2.7)
Interleukin-8 (pg/mL) 8.2 (2.7)
Interleukin-10 (pg/mL) 2.2 (1.6)
Tumor necrosis factor-α (pg/mL) 6.0 (1.6)
Soluble urokinase-type plasminogen activator receptor (ug/ml) 2.4 (4.8)
Open in a new tab

Note. Children can endorse multiple racial and ethnic identities, so values in these categories exceed 100 percent.

Systemic Inflammatory Activity

Table S1 displays regression models for systemic inflammatory activity. Greater BMI was associated with higher scores on the protein biomarker composite, in blood collected at both Time 1 and Time 2 (p’s < .01). None of the other covariates was consistently related to composite scores. More neighborhood violence was also associated with higher scores on the protein biomarker composite, in both Time 1 (B = .05; 95% CI = .01, .09; p = .009) and Time 2 (B = .05; 95% CI = .004, .09; p = .03) samples. Net of the covariates, neighborhood violence explained 4–5% of the variance in composite scores. Figure 1 illustrates this relationship by stratifying children into three categories of neighborhood violence, which reflect exposure levels below, near, and above the national average. Across categories, there was a stepwise increase in systemic inflammatory activity (Time 1: p < .001; Time 2: p = .04). However, neighborhood violence was not associated with changes over follow-up on the biomarker composite (p = .23).

Figure 1 -. Neighborhood violence and systemic biomarkers of inflammatory activity.

Figure 1 -

Open in a new tab

Six biomarkers of systemic inflammatory activity were quantified using immunoassay: C-reactive protein, interleukin 6, 8, and 10, tumor necrosis factor-α, and soluble urokinase-type plasminogen activator receptor. Values were standardized and averaged to form a composite. The figure shows the association between neighborhood violence and the biomarker composite in blood collected at (A) Time 1 (eighth grade) and (B) Time 2 (tenth grade). Values are adjusted for demographic and biobehavioral covariates.

These observations were unchanged in sensitivity analyses. Specifically, when the general indicator of social disadvantage was included as a covariate, neighborhood violence continued to be associated with higher composite scores. This was the case for specimens collected at both Time 1 (B = .04; 95% CI = .003, .08; p = .05) and Time 2 (B = .04; 95% CI = .001, .09; p = .05). Also, neighborhood violence was not simply acting as a proxy for personal victimization. Even when a covariate reflecting this exposure was included, the associations reported above remained statistically significant in both Time 1 (B = .04; 95% CI = .004, .08; p = .03) and Time 2 specimens (B = .05; 95% CI = .002, .10; p = .02).

Monocyte Transcriptional Activity

Cross-Sectional Analyses:

Covariate-adjusted models identified 258 transcripts associated with neighborhood violence (Table S2) in monocytes isolated at Time 1. 128 of those transcripts were relatively upregulated, including multiple chemokines, cytokines, and receptors involved in the mobilization, trafficking, and activation of myeloid cells (e.g., CCL3, CCL4, CCL4L2, IL3RA, CXCL8, CCL3L1). Among the 130 relatively downregulated transcripts were genes involved in cell adhesion (LAMA3, LAMB1, NRP1, CDH26, CAMPSAP3).

This pool of 258 transcripts was submitted to TELiS bioinformatics analysis (Cole et al. 2005), which indicated that neighborhood violence was associated with higher activity of the control pathways orchestrated by the MAF, AP-1, and NF-κB/Rel families (Figure 2). MAF is involved in the early stages of myeloid cell differentiation and mobilization, whereas the AP-1 and NF-κB families are key drivers of monocyte pro-inflammatory effector functions. TELiS also indicated that neighborhood violence was associated with higher activity of the CREB/ATF control pathway, which is involved in conveying adrenergic signals from the sympathetic nervous system to the monocyte genome. Contrary to hypotheses, neighborhood violence was unrelated to TELIS estimates of glucocorticoid-mediated transcriptional activity.

Figure 2 -. Cross-sectional association between neighborhood violence and monocyte gene expression:

Figure 2 -

Open in a new tab

Monocyte gene expression at Time 1 was modeled as a function of neighborhood violence plus demographic and biobehavioral covariates. Genes with 2-fold differential expression served as input to bioinformatic analyses quantifying the prevalence of transcription factor binding motifs. Values > 0 indicate up-regulation of pathway with neighborhood violence; < 0 indicate converse. NF-κB = nuclear factor kappa-B; AP-1 = activator protein-1; CREB = Cyclic AMP response element binding protein; GR = glucocorticoid receptor.

Durability Analyses:

Covariate-adjusted linear models identified 1559 violence-associated transcripts in monocytes isolated at Time 2 (Table S3). The vast majority (1402) were up-regulated, including multiple chemokine, cytokine, and receptor genes identified as more active in cross-sectional analyses (CCL3, CCL4, CCL4L2, IL3RA, CCL3L1). Newly identified up-regulated transcripts included molecules involved in mobilization and activation of myeloid-lineage cells (CCL5, IL1R1, IL1R2, TNFRSF9), prostaglandin metabolism (PTGR1, PTGR2), and hematopoiesis (IL3RA, IL11RA, MMP9, HOXA9).

Most TELiS patterns in cross-sectional analyses were still evident two years later (Figure 3). Neighborhood violence continued to be associated with higher activity of transcription control pathways coordinated by AP-1, NF-κB/REL, and CREB/ATF. In Time 2 mRNA, TELIS indicated that neighborhood violence had also become associated with glucocorticoid insensitivity, as reflected in relatively lower expression of transcripts with response elements for the glucocorticoid receptor.

Figure 3 -. Prospective association between neighborhood violence and monocyte gene expression:

Figure 3 -

Open in a new tab

Monocyte gene expression at Time 2 was modeled as a function of neighborhood violence and demographic and biobehavioral covariates, using the same approach as above.

Change Over Time:

The final analyses considered whether neighborhood violence presaged change in monocyte transcription over the subsequent two years. Covariate-adjusted models identified such 1131 transcripts; 931 were relatively up-regulated and 220 were relatively down-regulated (Table S4). The up-regulated transcripts included multiple chemokines (CCL2, CCL4, CCL5, CXCL1, CXCL11) identified in analyses above. Also up-regulated were transcripts encoding receptors (IL1R1, IL15RA, IL20RB, IL31RA) involved in the recruitment, differentiation, and activation of myeloid cells.

TELIS indicated that neighborhood violence was associated with increasing activity of the AP-1 pathway over the two-year follow-up period (p = .036), but not with the changes in expression of genes associated with NF-κB/REL or MAF signaling (p’s = .35 and .31). With regard to the neuro-hormonal pathways, neighborhood violence was associated with declining sensitivity to glucocorticoids over time, as reflected in fewer transcripts with glucocorticoid response elements (p = .0005). However, it was not related to changes in CREB/ATF activity (p = .66).

Sensitivity Analyses:

When social disadvantage and personal victimization were included in sensitivity analyses, TELiS results were similar to those reported above. None of the findings for neighborhood violence changed substantively, except for AP-1, which was no longer significant at either wave (p’s > .32).

DISCUSSION

Children exposed to persistent neighborhood violence are vulnerable to an array of health problems, which can include psychiatric (Slopen et al. 2012; Zimmerman & Kushner 2017; Fowler et al. 2009), respiratory (Wright et al. 2017; Ramratnam et al. 2015; Wright et al. 2004), and cardiometabolic conditions (Suglia et al. 2015; Wilson et al. 2002; Theall et al. 2012; Theall et al. 2017). Here, we leveraged discoveries from animal models of threat (Heidt et al. 2014; Powell et al. 2013; Weber et al. 2017; McKim et al. 2018) to formulate mechanistic hypotheses specifying how these associations could emerge. We found that neighborhood violence exposure was associated with higher levels of systemic inflammatory activity in late childhood. This association was present in cross-sectional analyses and at a follow-up assessment two years later.

To identify molecular signaling pathways underlying this pattern, we conducted transcriptional profiling of monocytes. Results indicated that neighborhood violence was associated with higher activity of the NF-κB and AP-1 transcriptional control pathways. As with the protein biomarkers, these associations were evident at study entry and durable at re-assessment two years later. Moreover, across that interval, children in high-violence neighborhoods displayed an increase in AP-1-driven transcriptional activity, suggesting the correlates of threat may compound with time.

NF-κB and AP-1 are master switches that control inflammatory effector functions in monocytes and macrophages (Natoli et al. 2011). Integrating our findings with animal models (Heidt et al. 2014; Avitsur et al. 2006; McKim et al. 2018; Wohleb et al. 2013; Niraula et al. 2019), one plausible hypothesis is that neighborhood violence contributes to health problems by sensitizing the NF-κB and AP-1 signaling pathways, so that after monocytes migrate into tissue, they mount exaggerated inflammatory responses to local stimuli, e.g., necrotic cells, damaged tissue, oxidized lipids, pathogens. The consequence of these exaggerated responses would likely differ across tissues. For example, in the brain, ongoing monocyte activity could modulate synthesis and metabolism of glutamate and dopamine (Miller & Raison 2016), with implications for activity of neural circuitries that underlie threat and reward processing (Haroon et al. 2012; Weber et al. 2017; Eisenberger et al. 2017). These circuits have trans-diagnostic relevance for mood, anxiety, and substance-misuse disorders (Nusslock & Miller 2015; Volkow et al. 2017). An ongoing mobilization of sensitized monocytes might also accentuate cardiometabolic risk. Indeed, monocyte-driven activity can induce lipolysis in adipose tissue, reduce insulin sensitivity in skeletal muscle, and promote growth of atherosclerotic plaque in the arterial wall (Nahrendorf 2018; Lackey & Olefsky 2016). Of course, at this juncture, these are just scenarios, intended to stimulate research that explicitly connects violence, monocytes, and disease. Follow-up studies with clinical outcomes are needed to evaluate their plausibility.

The transcriptional profiling results also suggest hypotheses about neuro-hormonal pathways that initiate and maintain pro-inflammatory activity in monocytes. When children entered the study, neighborhood violence was associated with higher activity in the CREB/ATF transcription control pathway, and this relationship was still apparent two years later. This pathway orchestrates beta-adrenergic signaling to the genome - among other things - and is thought to reflect cellular exposure to catecholamines (Cole & Sood 2012; Cole et al. 2020). If so, the findings here may reflect a scenario wherein neighborhood violence triggers sustained catecholamine release (Wilson et al. 2002), which selectively mobilizes classical monocytes into circulation, and further polarizes them towards an aggressive inflammatory phenotype (Heidt et al. 2014; Powell et al. 2013; Weber et al. 2017; McKim et al. 2018). Our results also indicate that over time, the monocytes of children in high-violence areas show progressively less glucocorticoid-mediated transcription. In humans, many chronic experiences of threat, including violence, diminish systemic release of cortisol (Miller et al. 2007; Lupien et al. 2009) and simultaneously interfere with its downstream transduction (Raison & Miller 2003; Miller et al. 2011). Both of these processes attenuate key regulatory signals that function to constrain inflammation (Weber et al. 2017). Collectively, these findings suggest that neighborhood violence initiates monocyte inflammatory activity via increased adrenergic signaling, and maintains it through a combination of this pathway and attenuated glucocorticoid regulation.

This study has several potential limitations to consider. Despite the longitudinal design and covariate adjustment, the associations could be affected by residual confounding. However, there is evidence from quasi experimental studies in humans (Baldwin et al. 2018) and truly experimental studies in animals (Hodes et al. 2014) that violence exposure up-regulates inflammatory activity. Thus, a causal effect is biologically plausible. Another potential limitation is the relatively brief follow-up. As subjects matured from late childhood into early adolescence, the variance in transcriptional activity increased substantially, as reflected in the larger number of violence-associated genes at Time 2. Studies with longer follow-up are needed to determine whether this association persists as youth mature and are exposed to further violence. A final potential limitation is the absence of clinical outcomes reflecting mental and/or physical health status. By including these measures, future studies can evaluate whether the patterns observed here translate into differences in disease, as our hypothesis would suggest.

Despite these limitations, the study had several methodological strengths, including a diverse sample of children, a longitudinal design, and the integration of neighborhood, protein biomarker, and transcriptional data. The results provide mechanistic clues suggesting how neighborhood violence - a common exposure for many American children - could instigate and maintain inflammatory activity in monocytes, and by doing so potentially increase long-term health risks. In light of preliminary evidence that stress-management interventions attenuate neurohormonal and inflammatory activity among at-risk youth (Miller et al. 2014; Slopen et al. 2014), an important next step in this literature would be a randomized clinical trial to evaluate whether such treatments can mitigate the presumptive impact of violence.

Supplementary Material

1

HIGHLIGHTS.

  • Individuals exposed to neighborhood violence have a higher risk of developing mental and physical health problems across the lifespan. Preclinical signs of this risk are evident starting in childhood, but the underlying cellular and molecular are unknown.

  • This paper suggests a mechanistic role for systemic inflammatory activity that is orchestrated by activated monocytes. It also highlights neurohormonal pathways involving glucocorticoid and adrenergic signaling that could initiate and maintain the inflammatory activity.

FUNDING

Supported by NIH grants R01 HL122328, R01 HL137818, F32 HL14600, and P30 AG017265.

Abbreviations:

AP-1

Activator protein 1

BMI

body mass index

CRP

C-reactive protein

IL

interleukin

NF-κB

nuclear factor-kappa B

TFBM

transcription factor-binding motif

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

DISCLOSURES

None of the authors reports a conflict of interest, financial or otherwise.

REFERENCES

  1. Aldo P, Marusov G, Svancara D, David J, Mor G. Simple plex(™) : A novel multi-analyte, automated microfluidic immunoassay platform for the detection of human and mouse cytokines and chemokines. Am J Reprod Immunol 2016; 75:678–693. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Avitsur R, Hunzeker J, Sheridan JF. Role of early stress in the individual differences in host response to viral infection. Brain Behav. Immun 2006; 20:339–348. [DOI] [PubMed] [Google Scholar]
  3. Baldwin JR, Arseneault L, Caspi A, Fisher HL, Moffitt TE, Odgers CL, Pariante C, Ambler A, Dove R, Kepa A, Matthews T, Menard A et al. Childhood victimization and inflammation in young adulthood: A genetically sensitive cohort study. Brain Behav. Immun 2018; 67:211–217. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Browning CR, Cagney KA, Iveniuk J. Neighborhood stressors and cardiovascular health: Crime and C-reactive protein in Dallas, USA. Soc Sci Med 2012; 75:1271–1279. [DOI] [PubMed] [Google Scholar]
  5. Cole SW, Galic Z, Zack JA. Controlling false-negative errors in microarray differential expression analysis: A prim approach. Bioinformatics 2003; 19:1808–1816. [DOI] [PubMed] [Google Scholar]
  6. Cole SW, Shanahan MJ, Gaydosh L, Harris KM. Population-based RNA profiling in Add Health finds social disparities in inflammatory and antiviral gene regulation to emerge by young adulthood. Proc. Natl. Acad. Sci. U S A 2020; 117:4601–4608. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Cole SW, Sood AK. Molecular pathways: Beta-adrenergic signaling in cancer. Clin Cancer Res 2012; 18:1201–1206. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Cole SW, Yan W, Galic Z, Arevalo J, Zack JA. Expression-based monitoring of transcription factor activity: The TELiS database. Bioinformatics 2005; 21:803–810. [DOI] [PubMed] [Google Scholar]
  9. Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, Jha S, Batut P, Chaisson M, Gingeras TR. Star: Ultrafast universal RNA-seq aligner. Bioinformatics 2013; 29:15–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Eisenberg E, Levanon EY. Human housekeeping genes, revisited. Trends Genet. 2013; 29:569–574. [DOI] [PubMed] [Google Scholar]
  11. Eisenberger NI, Moieni M, Inagaki TK, Muscatell KA, Irwin MR. In sickness and in health: The co-regulation of inflammation and social behavior. Neuropsychopharmacology 2017; 42:242–253. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Finegood ED, Chen E, Kish J, Vause K, Leigh AKK, Hoffer L, Miller GE. Community violence and cellular and cytokine indicators of inflammation in adolescents. Psychoneuroendocrinology 2020; 115:104628. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Foster H, Brooks-Gunn J. Toward a stress process model of children’s exposure to physical family and community violence. Clin Child Fam Psychol Rev 2009; 12:71–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Fowler PJ, Tompsett CJ, Braciszewski JM, Jacques-Tiura AJ, Baltes BB. Community violence: A meta-analysis on the effect of exposure and mental health outcomes of children and adolescents. Dev Psychopathol 2009; 21:227–259. [DOI] [PubMed] [Google Scholar]
  15. Haroon E, Raison CL, Miller AH. Psychoneuroimmunology meets neuropsychopharmacology: Translational implications of the impact of inflammation on behavior. Neuropsychopharmacology 2012; 37:137–162. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Havranek EP, Mujahid MS, Barr DA, Blair IV, Cohen MS, Cruz-Flores S, Davey-Smith G, Dennison-Himmelfarb CR, Lauer MS, Lockwood DW, Rosal M, Yancy CW et al. Social determinants of risk and outcomes for cardiovascular disease: A scientific statement from the American Heart Association. Circulation 2015; 132:873–898. [DOI] [PubMed] [Google Scholar]
  17. Heidt T, Sager HB, Courties G, Dutta P, Iwamoto Y, Zaltsman A, von Zur Muhlen C, Bode C, Fricchione GL, Denninger J, Lin CP, Vinegoni C et al. Chronic variable stress activates hematopoietic stem cells. Nat. Med 2014; 20:754–758. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Hodes GE, Pfau ML, Leboeuf M, Golden SA, Christoffel DJ, Bregman D, Rebusi N, Heshmati M, Aleyasin H, Warren BL, Lebonté B, Horn S et al. Individual differences in the peripheral immune system promote resilience versus susceptibility to social stress. Proc. Natl. Acad. Sci. U S A 2014; 111:16136–16141. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Hodges GW, Bang CN, Wachtell K, Eugen-Olsen J, Jeppesen JL. suPAR: A new biomarker for cardiovascular disease. Can J Cardiol 2015; 31:1293–1302. [DOI] [PubMed] [Google Scholar]
  20. Kuczmarski RJ, Ogden CL, Guo SS, Grummer-Strawn LM, Flegal KM, Mei Z, Wei R, Curtin LR, Roche AF, Johnson CL. 2000 CDC growth charts for the United States: Methods and development. Vital Health Stat 11 2002; 1–190. [PubMed] [Google Scholar]
  21. Lackey DE, Olefsky JM. Regulation of metabolism by the innate immune system. Nat Rev Endocrinol 2016; 12:15–28. [DOI] [PubMed] [Google Scholar]
  22. Lupien SJ, McEwen BS, Gunnar MR, Heim C. Effects of stress throughout the lifespan on the brain, behaviour and cognition. Nat. Rev. Neurosci 2009; 10:434–445. [DOI] [PubMed] [Google Scholar]
  23. Manduca R, Sampson RJ. Punishing and toxic neighborhood environments independently predict the intergenerational social mobility of black and white children. Proc. Natl. Acad. Sci. U S A 2019; 116:7772–7777. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. McKim DB, Yin W, Wang Y, Cole SW, Godbout JP, Sheridan JF. Social stress mobilizes hematopoietic stem cells to establish persistent splenic myelopoiesis. Cell Rep 2018; 25:2552–2562.e3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Miller AH, Raison CL. The role of inflammation in depression: From evolutionary imperative to modern treatment target. Nat Rev Immunol 2016; 16:22–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Miller GE, Brody GH, Yu T, Chen E. A family-oriented psychosocial intervention reduces inflammation in low-SES African American youth. Proc. Natl. Acad. Sci. U S A 2014; 111:11287–11292. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Miller GE, Chen E, Armstrong CC, Carroll AL, Ozturk S, Rydland KJ, Brody GH, Parrish TB, Nusslock R. Functional connectivity in central executive network protects youth against cardiometabolic risks linked with neighborhood violence. Proc. Natl. Acad. Sci. U S A 2018; 115:12063–12068. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Miller GE, Chen E, Parker KJ. Psychological stress in childhood and susceptibility to the chronic diseases of aging: Moving toward a model of behavioral and biological mechanisms. Psychol. Bull 2011; 137:959–997. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Miller GE, Chen E, Zhou ES. If it goes up, must it come down? Chronic stress and the hypothalamic-pituitary-adrenocortical axis in humans. Psychol. Bull 2007; 133:25–45. [DOI] [PubMed] [Google Scholar]
  30. Nahrendorf M Myeloid cell contributions to cardiovascular health and disease. Nat. Med 2018; 24:711–720. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Nathan C, Ding A. Nonresolving inflammation. Cell 2010; 140:871–882. [DOI] [PubMed] [Google Scholar]
  32. Natoli G, Ghisletti S, Barozzi I. The genomic landscapes of inflammation. Genes Dev. 2011; 25:101–106. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Niraula A, Witcher KG, Sheridan JF, Godbout JP. Interleukin-6 induced by social stress promotes a unique transcriptional signature in the monocytes that facilitate anxiety. Biol Psychiatry 2019; 85:679–689. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Norris AW, Kahn CR. Analysis of gene expression in pathophysiological states: Balancing false discovery and false negative rates. Proc. Natl. Acad. Sci. U S A 2006; 103:649–653. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Nusslock R, Miller GE. Early-life adversity and physical and emotional health across the lifespan: A neuroimmune network hypothesis. Biol Psychiatry 2015; 80:23–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Papachristos AV, Brazil N, Cheng TT. Understanding the crime gap: Violence and inequality in an American city. City Commun 2018; 17:1051–1074. [Google Scholar]
  37. Peterson A, Crockett L, Richards M, Boxer A. A self-report measure of pubertal status: Reliability, validity, and initial norms. J Youth Adoles 1988; 17:117–133. [DOI] [PubMed] [Google Scholar]
  38. Powell ND, Sloan EK, Bailey MT, Arevalo JM, Miller GE, Chen E, Kobor MS, Reader BF, Sheridan JF, Cole SW. Social stress up-regulates inflammatory gene expression in the leukocyte transcriptome via beta-adrenergic induction of myelopoiesis. Proc. Natl. Acad. Sci. U S A 2013; 110:16574–16579. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Raison CL, Miller AH. When not enough is too much: The role of insufficient glucocorticoid signaling in the pathophysiology of stress-related disorders. Am. J. Psychiatry 2003; 160:1554–1565. [DOI] [PubMed] [Google Scholar]
  40. Ramratnam SK, Han YY, Rosas-Salazar C, Forno E, Brehm JM, Rosser F, Marsland AL, Colón-Semidey A, Alvarez M, Miller GE, Acosta-Pérez E, Canino G et al. Exposure to gun violence and asthma among children in Puerto Rico. Respir. Med 2015; 109:975–981. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Reader BF, Jarrett BL, McKim DB, Wohleb ES, Godbout JP, Sheridan JF. Peripheral and central effects of repeated social defeat stress: Monocyte trafficking, microglial activation, and anxiety. Neuroscience 2015; 289:429–442. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Ridker PM. From C-reactive protein to interleukin-6 to interleukin-1: Moving upstream to identify novel targets for atheroprotection. Circ. Res 2016; 118:145–156. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Sharkey P An Uneasy Peace: The Great Crime Decline, the Renewal of City Life, and the Next War on Violence. New York: WW Norton; 2018 [Google Scholar]
  44. Shi L, Jones WD, Jensen RV, Harris SC, Perkins RG, Goodsaid FM, Guo L, Croner LJ, Boysen C, Fang H, Qian F, Amur S et al. The balance of reproducibility, sensitivity, and specificity of lists of differentially expressed genes in microarray studies. BMC Bioinformatics 2008; 9 Suppl 9:S10. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Slopen N, Fitzmaurice GM, Williams DR, Gilman SE. Common patterns of violence experiences and depression and anxiety among adolescents. Soc Psychiatry Psychiatr Epidemiol 2012; 47:1591–1605. [DOI] [PubMed] [Google Scholar]
  46. Slopen N, McLaughlin KA, Shonkoff JP. Interventions to improve cortisol regulation in children: A systematic review. Pediatrics 2014; 133:312–326. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Suglia SF, Sapra KJ, Koenen KC. Violence and cardiovascular health: A systematic review. Am J Prev Med 2015; 48:205–212. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Theall KP, Drury SS, Shirtcliff EA. Cumulative neighborhood risk of psychosocial stress and allostatic load in adolescents. Am. J. Epidemiol 2012; 176 Suppl 7:S164–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Theall KP, Shirtcliff EA, Dismukes AR, Wallace M, Drury SS. Association between neighborhood violence and biological stress in children. JAMA Pediatr 2017; 171:53–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Thomson CC, Roberts K, Curran A, Ryan L, Wright RJ. Caretaker-child concordance for child’s exposure to violence in a preadolescent inner-city population. Arch Pediatr Adolesc Med 2002; 156:818–823. [DOI] [PubMed] [Google Scholar]
  51. Volkow ND, Wise RA, Baler R. The dopamine motive system: Implications for drug and food addiction. Nat. Rev. Neurosci 2017; 18:741–752. [DOI] [PubMed] [Google Scholar]
  52. Weber MD, Godbout JP, Sheridan JF. Repeated social defeat, neuroinflammation, and behavior: Monocytes carry the signal. Neuropsychopharmacology 2017; 42:46–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Wilson DK, Kliewer W, Teasley N, Plybon L, Sica DA. Violence exposure, catecholamine excretion, and blood pressure non-dipping status in African American male versus female adolescents. Psychosom. Med 2002; 64:906–915. [DOI] [PubMed] [Google Scholar]
  54. Wohleb ES, Powell ND, Godbout JP, Sheridan JF. Stress-induced recruitment of bone marrow-derived monocytes to the brain promotes anxiety-like behavior. J. Neurosci 2013; 33:13820–13833. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Wright AW, Austin M, Booth C, Kliewer W. Systematic review: Exposure to community violence and physical health outcomes in youth. J Pediatr Psychol 2017; 42:364–378. [DOI] [PubMed] [Google Scholar]
  56. Wright RJ, Mitchell H, Visness C,M, Cohen S, Stout J, Evans R, Gold DR. Community violence and asthma morbidity: The Inner-City Asthma Study. Am J Pub Health 2004; 94:625–632. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Zimmerman GM, Kushner M. Examining the contemporaneous, short-term, and long-term effects of secondary exposure to violence on adolescent substance use. J Youth Adolesc 2017; 46:1933–1952. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1758661-supplement-1.pdf (55.8KB, pdf)

RESOURCES