There is a drought in the development of medications for the treatment of acute and chronic pain. The National Institutes of Health report a mere 2% of candidate pain medications advance from Phase I clinical trials to FDA-approval (Thomas and Wessel, 2017), as opposed to 10% of medications for other conditions. Clinician scientists are challenged in identifying novel, effective medications for pain patients due to the lack of quantifiable, objective methods to measure pain in the clinic; pain sensation is multifaceted and subjective for each individual (Younger et al., 2009). The International Association for the Study of Pain defines pain as “an unpleasant sensory and emotional experience associated with actual or potential tissue damage” (Raja et al., 2020). Due the unavoidable subjectivity of the pain experience as a whole, diagnostic measurements in the clinic should instead focus on quantifying proteins involved in nociceptive inflammation that contribute to tissue repair and restoration processes. The current gold standard of pain measurement at the point-of-care (POC) relies on self-report measures such as the Numerical Rating Scale or Visual Analogue Scale. Self-report results can be distorted by such factors as emotional state, cultural background, and cognitive abilities (Krebs et al., 2007). Clinical studies that investigated the predictive relationship between self-report of pain and physical performance tasks (i.e., “Up and Down” test in osteoarthritis patients) within the same individual found a modest relationship between the two measures, with correlation coefficients of 0.3 or lower (Simmonds et al., 1998; Novy et al., 2002). Skin conductance and body temperature may be used as physiological indicators of pain (Harrison et al., 2006; Storm, 2008), however, the validity of their use has been challenged by inconsistent findings (Hu et al., 2019). Lastly, neuroimaging tools (e.g., functional magnetic resonance imaging) are utilized to identify a neurologic signature of physical pain in cortical regions involved in regulating pain sensation such as the insular cortex (Apkarian et al., 2005; Wager et al., 2011; Lee et al., 2021). While these imaging techniques can provide important information about pain processing, these tests are expensive, time intensive, and do not provide a quantitative measurement. There is a critical need for economical, rapid POC quantification of accessible inflammation biomarkers to provide objective analyses beyond the subjective pain scales currently employed in clinical care settings (Chizh and Hobson, 2007).

The concept of screening body fluids (e.g., blood, saliva, urine) for molecular biomarkers of disease (i.e., liquid biopsy) is not new (Mandel and Metais, 1948), however, recent technological advancements enable these approaches to become powerful and innovative diagnostic tools in the clinic (Mader and Pantel, 2017; Poulet et al., 2019). At present, several liquid biopsy assays are FDA-approved for early detection of circulating tumor cells or circulating tumor DNA using next-generation sequencing. For example, a recent study sought to determine the clinical validity of liquid biopsy screening in a cohort of 514 metastatic castration-resistant prostate cancer patients and found genomic aberrations related to prostate cancer in blood samples of 94% of the diagnosed patients (Sonpavde et al., 2019). This is one of many examples that demonstrate high concordance between positive liquid biopsy screening and the clinical diagnosis (for review) (Vaidyanathan et al., 2018). Additionally, liquid biopsies possess several advantages including minimal invasiveness, decreased risk, and importantly, the ability to sample repeatedly to monitor disease progression and treatment response. Liquid biopsy diagnostic tools for cancer biomarkers show great promise for clinical utility and are spurring interest in expanding application to different pathologies, including osteoarthritis and joint inflammatory diseases (Michela, 2021). Current methodologies for quantifying pain in the clinic are inadequate; the liquid biopsy approach would greatly benefit physicians seeking to accurately diagnose the etiology and severity of tissue inflammation, as well as treatment efficacy across time in patients experiencing nociceptive inflammatory pain.

The present preclinical study was designed to identify blood-born biomarkers that directly correlate with the timecourse and intensity of acute inflammatory nociception. In turn, these biomarkers could provide prioritized candidates for diagnostic testing in pain patients at the POC. To achieve this, we induced paw edema through subplantar λ-carrageenan injection (Otterness and Moore, 1988) and determined the temporal relationship between localized inflammation and expression of inflammatory mediators in rodent plasma. The xenobiotic λ-carrageenan is regularly used to induce an acute inflammation and behavioral hyperalgesia (Boyce-Rustay et al., 2010; Da Silva et al., 2012; Armbruster et al., 2017). Concurrently, we monitored for potential impairment of voluntary, spontaneous movement in a familiar environment as a non-reflexive, non-evoked surrogate measurement of inflammatory nociception (Zhu et al., 2012; Cho et al., 2013; Sheahan et al., 2017). Reduced mobility is a hallmark symptom in pain patients and thus is more translational in comparison to more commonly used preclinical nociceptive models, such as the Hargreaves radiant heat assay of thermal sensitivity (Hargreaves et al., 1988) or von Frey filament test of mechanical sensitivity (Chaplan et al., 1994). Our approach is to assess and establish the potential feasibility for detection of inflammation biomarkers at the POC in pain patients.

The “Law of Pain,” introduced by Sota Omoigui, states that “the origin of all pain is inflammation and the inflammatory response” (Omoigui, 2007b;a). Initiation of inflammation and the nociceptive state occurs when basophils, mast cells, and platelets release nitric oxide (Cury et al., 2011), histamine (Obara et al., 2020), and serotonin (Sommer, 2004; Loyd et al., 2013) in immediate response to localized tissue damage. The combined cellular response to these stimuli increases local blood flow and leads eventually to the secretion of prostaglandins, which facilitate the attraction of specialized immune cells (e.g., macrophages) to initiate the inflammatory response. Macrophages release pro-inflammatory interleukins (IL) (e.g., IL-1β, IL-4, IL-6, IL-8), anti-inflammatory cytokines (e.g., IL-4, IL-10, IL-11, IL-13) and chemokine (C-C and C-X-C motif) ligands [e.g., C-X-C motif chemokine ligand 1 (CXCL1), C-C motif chemokine ligand 2 (CCL2)] (Hughes and Nibbs, 2018) that attract and activate other immune cells such as granulocytes (neutrophils, eosinophils), lymphocytes (T-helper cells), as well as fibroblasts and endothelial cells to the inflamed tissue. Using Bio-Plex technology, we screened plasma samples against a broad panel of these cytokines to determine the relationship of peripheral expression relative to the timecourse of acute inflammatory nociception.

The cocktail of proteases [e.g., granzyme B, matrix metalloproteinases (MMPs), neutrophil elastase, and many others], bactericidal peptides and proteins, that is, generated by immune cells upon injury leads inevitably to tissue damage and pain sensations (Raoof et al., 2018). The MMPs include collagenases (e.g., MMP-1, -8, -13), stromelysins (e.g., MMP-3, -10, -11), and especially gelatinases (e.g., MMP-2 and -9) which play important functions in acute and chronic inflammation (Lakhan and Avramut, 2012) and are regulated by the competition of post-translational proteolytic cleavage and endogenous inhibitors (Quintero-Fabian et al., 2019). In general, MMP-2 is constitutively expressed by numerous cell types, whereas MMP-9 is induced by tumor necrosis factor-α (TNF-α) and IL-1β released from inflammatory cells (neutrophils, macrophages, and lymphocytes) (Lakhan and Avramut, 2012). MMP-9 plays a major role in the migration of cells through the walls of blood vessels to inflammatory sites and degradation of collagen in the extracellular matrix to aid in tissue remodeling processes (Ahmed et al., 2011; Lakhan and Avramut, 2012; Ramirez et al., 2018). MMP-9 is transiently overexpressed and then activated during an early response to tissue damage, whereas MMP-2 is a signature enzyme of the late response. In contrast to virtually all previous studies, the nanobiosensor technology employed here permits the quantitative determination of proteolytic enzyme activities. This is of translational importance because MMPs require proteolytic activation (Lopez-Otin and Bond, 2008). Based on this paradigm, the quotient of MMP-9 divided by MMP-2 activity can be used as an indicator for the temporal course of inflammatory nociception (Kawasaki et al., 2008).

Taking advantage of the progress in the field of applied nanotechnology, Bossmann and colleagues developed nanoparticle-based biosensors for the ultra-sensitive detection of proteases in liquid biopsies (Wang et al., 2014; Udukala et al., 2016; Malalasekera et al., 2017; Kalubowilage et al., 2018). The proteolytic activity of numerous MMPs can be detected at sub-femtomolar levels and across a wide detection range (10⁻⁵–10⁻¹⁶ M) (Wang et al., 2014; Udukala et al., 2016). The core of the detection nanoparticle is a dopamine-coated iron/iron oxide core/shell nanoparticle with a diameter of 15–17 nm, including a 1–2 mm Fe₃O₄ shell (Wang et al., 2014) which is linked to the two Foerster Resonance Energy Transfer (FRET) nanobiosensors including tetrakis-carboxyphenyl-porphyrin (TCPP) and cyanine 5.5 (Cy 5.5). Whereas Cy 5.5 is permanently tethered, TCPP is linked via a protease-cleavable consensus sequence. Once this oligopeptide is cleaved, TCPP escapes plasmonic quenching by the Fe core and FRET by Cy 5.5, leading to a measurable fluorescence increase which is detected by a conventional plate reader. Here, we report the application of Fe/Fe₃O₄-based nanobiosensor technology to detect circulating biomarkers related to the inflammatory nociceptive state (Wang et al., 2014). We selected MMP-2 and MMP-9 activities to assess in the plasma collected from our rat model for inflammatory nociception due to the important role these enzymes play in tissue repair and restoration processes (Leira et al., 2007; Bruschi et al., 2016; Medeiros et al., 2017; Hannocks et al., 2019; Escolano-Lozano et al., 2021).

We postulated that inflammation and locomotor impairment would peak 5–8 h following λ-carrageenan treatment and return to baseline levels at 24 h (Otterness and Moore, 1988; Lin et al., 2015; Annamalai and Thangam, 2017). We also hypothesized that variation in expression or activity of inflammatory regulators will be target-specific, as each plays a unique role in the inflammatory response (Zhang and An, 2007). To our knowledge, this study is the first to capture the full timecourse of protein expression and activity of selected inflammatory regulators in alignment with inflammatory nociception within subjects. A better understanding of the timecourse of these fluctuations will aid in predicting onset, duration, and termination of the pain sensation in patients and ultimately be valuable to enhance management of coinciding symptoms. These efforts align with a key aspect of the Helping to End Addiction Long-Term (HEAL™) initiative, launched by the National Institutes of Health in 2019: acceleration of the discovery and validation of biomarkers, endpoints, and signatures for pain conditions.

Figure 1A illustrates the experimental timeline. Paw inflammation, non-evoked inflammatory nociceptive behavior, and biomarker expression were assessed prior to treatment with saline or λ-carrageenan. The anticipated timeline of λ-carrageenan on these measures at 1–24 h following treatment is noted.

The primary goal of the current study was to explore the feasibility of utilizing a robust sensing platform with liquid biopsy techniques to quantify biomarkers of inflammatory nociception. Specifically, we monitored the kinetic activity profile of detectable cytokines and MMP9/MMP2 ratio in the plasma of rats experiencing acute inflammation and identified correlations between these fluctuations and paw edema, as well as non-evoked inflammatory nociceptive behavior. To the best of our knowledge, no previous study has surveyed the timecourse of cytokine protein concentrations or MMP activity in plasma in a rodent model of acute inflammation. Paw thickness was significantly greater in λ-carrageenan-injected rats versus saline-injected rats at all time points, with peak inflammation observed the 8 h timepoint after treatment. While we did not detect a main effect of treatment on the timecourse of locomotor activity, we did detect a strong negative correlation between paw thickness and locomotor activity at 8 h after intraplantar injection. These data suggest the existence of a threshold of inflammation that must be surpassed to adversely impact mobility. We then narrowed subsequent correlational analyses between cytokine expression/MMP activity and paw inflammation/locomotor activity to the 8 h time point.

Cytokines CXCL1 and CCL2 are small peptide chemoattractants responsible for recruiting several classes of immune cells (i.e., monocytes, neutrophils, etc.) to the site of injury. Average plasma CXCL1 protein expression dramatically increased in λ-carrageenan-injected rats, relative to saline, at the 5 and 8 h timepoints after treatment. Further, greater CXCL1 expression at the 8 h timepoint was associated with increased paw thickness and reduced locomotor activity. These data agree with a previous study demonstrating increased Cxcl1 mRNA expression in inflamed tissue following intraplantar λ-carrageenan treatment (Goto et al., 2021). Increases in Cxcl1 mRNA were greatest 1–4 h post-treatment, whereas differences in plasma CXCL1 protein expression were most pronounced at 5–8 h in the current study. This discrepancy may be explained by the sequential orchestration of transcription and translation (Crick, 1958). Moreover, our ability to detect inflammatory biomarkers and the timecourse of their expression is undoubtedly dependent on the source of the sample (localized inflamed tissue versus circulating blood). Modest, yet significant, increases in CCL2 protein expression were seen at the eight and 24 h time points in the current study. Similar to CXCL1, we found increased CCL2 expression associates with heightened paw inflammation and reduced locomotor activity at the 8 h timepoint. Not surprisingly, Ccl2 mRNA expression was also increased in the λ-carrageenan-injected paw sooner than we detected the increases in plasma protein expression (Goto et al., 2021). We saw a trending increase in CCL2 plasma expression 5 h post treatment, in line with previous research reporting a significant increase in CCL2 serum expression 4 h later after intraplantar λ-carrageenan administration (Ou et al., 2019). In contrast to the present findings, Ou et al. (2019) reported significant changes in serum protein expression of IL-1β at 4 h post-treatment, whereas there was no main effect of carrageenan-injection on any of the other three cytokines detected (MIP-1α, MIP-3α, and RANTES; data not shown). This divergence might be explained by differences in the timing of blood draws for liquid biopsy as our timecourse analyses display distinct, time-dependent variation in cytokine expression.

The onset of MMP2 activity is delayed following tissue damage and thus was used as an internal calibration of MMP9 activity in our acute model of inflammatory nociception (Kawasaki et al., 2008; Nascimento et al., 2013). We observed the expected rapid and transient induction of MMP9/MMP2 activity at the 1 h timepoint in λ-carrageenan-injected rats, relative to saline-injected rats. This effect was reversed at the 8 h timepoint. Previous studies investigated the regulatory role of MMP inhibition on cytokine and chemokine expression through use of broad-spectrum MMP inhibitors [i.e., doxycycline, GM6001 (Galardin^®)] (Supasorn et al., 2016; Tang et al., 2017). Tang and colleagues found that pan-MMP inhibition resulted in a decrease in tissue expression of CXCL1 and CCL2 (Supasorn et al., 2016). However, at the height of peak inflammation in the current study, MMP9/MMP2 activity is significantly lower in λ-carrageenan-injected animals while CXCL1 and CCL2 expression is increased. Future studies will be designed to disentangle the role of each functional class of MMPs [e.g., collagenases (MMP1), gelatinases (MMP9, MMP2)] on cytokine/chemokine induction with more selective pharmacological compounds. It should also be mentioned that this is the first study in which the proteolytic activity of MMP2 and MMP9 was measured, not simply the protease concentrations.

The present studies were conducted in male rats and given sex-specific distinctions in production of cytokines and the overall immune response (Klein and Flanagan, 2016), future studies are necessary to determine if the temporal pattern of biomarker expression and activity is generalizable to females. The Bio-Plex assay detected a limited number of cytokines across all timepoints in both saline- and λ-carrageenan-injected rats (six of 23 assayed). It is probable that the expression of circulating cytokines in plasma is much lower in comparison to the localized site of inflammation. Nonetheless, our ability to capture differences in plasma CXCL1 and CCL2 expression suggests these cytokines will be useful as reliable markers of inflammatory nociception in liquid biopsies.

We provide foundational data for identifying quantitative markers of inflammatory nociception in liquid biopsies of inflammatory pain patients. We propose that CXCL1 shows the most promise as a readily detected peripheral biomarker of nociceptive inflammation as we observed robust changes in carrageenan-injected animals. Further, a recent clinical trial observed a significant reduction in serum CXCL1 protein expression following treatment with an anti-inflammatory agent in rheumatoid arthritis patients (Tao et al., 2021). This observation coincided with reduced clinical symptoms of pain in these patients. Relative changes observed in CCL2 expression and MMP9/MMP2 activity in the current study were modest. Previous clinical trials identified moderate changes in potential biomarkers of pain via the liquid biopsy approach (i.e., cystatin C in cerebrospinal fluid of women in active labor) (Mannes et al., 2003). However, cystatin C was later found to be an unreliable marker of pain in a follow-up study with a greater number of participants; cystatin C expression also did not correlate with pain levels of neuropathic conditions (Eisenach et al., 2004; Kalso, 2004). Future studies are required to expand on the current findings by identifying robust biosignatures of inflammatory regulators unique to various origins of pain (e.g., neuropathic) and how they might differ across species and sex. Our ultimate goal is to develop improved diagnostic POC devices to objectively quantify inflammatory nociception with increased reliability, relative to current the current standard-of-care practices that rely on subjective measures of self-report (Younger et al., 2009). The successful development of POC devices will enable health care providers in clinical settings to guide appropriate treatment strategies and optimize management of pain patients.