Sage Journals: Discover world-class research

Abstract

Background:

Long-term management of inflammatory bowel diseases (IBD) is challenging and the identification of reliable predictors for treatment outcomes is an unmet need. Neutrophil-related biomarkers have been mainly studied in the feces, but blood analyses have inherent advantages.

Objective:

To review the recent learnings on the ability of blood-based neutrophil-expressed biomarkers to predict treatment outcomes in IBD.

Design:

Systematic scoping review.

Data sources and methods:

We performed a literature search in Pubmed, EMBASE, SCOPUS, Web of Science, ScienceDirect, and Cochrane Central Register of Controlled Trials from inception until May 2022 according to Preferred Reporting Items for Systematic reviews and Meta-Analyses guidelines. All human studies associating blood-based neutrophil-related compounds with the prediction of disease progression, complication onset, or treatment outcomes were included.

Results:

From 1032 retrieved entries, 34 studies were selected, 32 published in 2013 or later. In all, 17 biomarkers from granules, cytoplasm, plasmatic membrane, and plasma were explored. In total, 1850 Crohn’s disease (CD) and 1122 ulcerative colitis non-duplicated patients were included. The most mentioned biomarkers were nCD64, serum calprotectin (SC), oncostatin M (OSM), neutrophil elastase-generated calprotectin fragment (CPa9-HNE), and triggering receptor expressed on myeloid cells 1 (TREM1). Six biomarkers showed promising results: OSM, SC, eNAMPT, nCD64, TREM1, and CPa9-HNE. Variable positive signals were found for human neutrophil peptide 1-3, LL-37, S100A12, and neutrophil gelatinase–associated lipocalin. No predictive ability was found for the remaining markers. Sharing a neutrophil compartment did not indicate similar behavior.

Conclusion:

Advances in the last decade began to unveil the untapped potential of the readily accessible blood neutrophil-expressed biomarkers, especially nCD64, TREM1, and CPa9-HNE. Current evidence suggests that future research should focus on well-defined subpopulations instead of a one-size-fits-all biomarker.

Registration:

https://osf.io/kes9a.

Keywords

calprotectin CPa9-HNE eNAMPT nCD64 OSM TREM1

Introduction

Crohn’s disease (CD) and ulcerative colitis (UC), grouped under the umbrella term of inflammatory bowel disease (IBD), are complex chronic diseases increasing worldwide.¹ Despite the dramatic improvement of care brought by biologics targeting anti-tumor necrosis factor alpha (TNFα) and, later, integrin α4β7, interleukin (IL)-12/IL-23, and JAK, the therapeutic management of IBD remains challenging.¹ In fact, one-third of patients will be either primary or secondary non-responders to anti-TNFα² and the withdrawal of anti-TNFα is associated with a risk of relapse ranging from 20% to 80%³. Identification of reliable predictors of treatment outcomes would foster the effective management of IBD.^3–5

Neutrophils are key players in the pathophysiology of IBD, and neutrophil infiltration in the gut correlates with disease activity in both UC and CD.⁶ Part of the innate immune response, they produce reactive oxygen species, proteases, and proinflammatory cytokines that combat the microbial pathogens, but also damage the epithelial barrier and recruit more inflammatory cells, perpetuating the gut inflammation.^1,6 Interestingly, while a neutrophil response is strongly prominent in UC, in CD there are observations of impaired migration, reduced production of cytokines and phagocytosis, and delayed/defective antimicrobial clearance.^1,6,7 In fact, fecal calprotectin (FC) is a neutrophil-related biomarker widely used to monitor disease⁸ and is recommended as a target of treatment.⁹ A multitude of other biomarkers, especially fecal,¹⁰ have been extensively studied, but with primary focus on differential diagnosis or disease activity monitoring.^11,12 In contrast, predictive biomarkers provide information on the anticipated response to a specific treatment intervention.^5,13 From all monitoring tools, venipuncture was considered the most acceptable and used by IBD patients. Stool collection scored lower than venipuncture and some imaging options, mainly due to the embarrassment related to collecting feces¹⁴; indeed, in some Asian countries, the collection of feces is not an option at all.¹⁵ Thus, focusing on readily accessible and acceptable blood-based biomarkers is logical and reasonable.

Recent reviews explored, with variable depths, biologic treatment response predictors covering patient and disease characteristics, changes in microbiome, pharmacogenomics and blood, intestinal, and fecal markers.^4,5,13,16,17 However, blood biomarkers coverage is mainly confined to blood count parameters, albumin, C-reactive protein, perinuclear anti-neutrophil cytoplasmic antibodies, and anti-Saccharomyces cerevisiae antibodies. This strategy leaves a host of upcoming biomarkers out of the spotlight. To our knowledge, the potential of blood-based neutrophil-expressed biomarkers to predict treatment outcomes in IBD was not yet fully revised and summarized. Therefore, we conducted a systematic scoping review to comprehensively map all available evidence on their ability to predict the evolution of IBD, both in terms of natural disease progression and treatment outcomes.

Methods

Protocol and registration

The scoping review protocol was drafted according to the Preferred Reporting Items for Systematic reviews and Meta-Analyses Protocol (PRISMA-P)¹⁸ and registered in the Open Science Framework (available here: https://osf.io/kes9a). Scoping reviews are not eligible for PROSPERO database registration. Search and reporting were conducted according to the PRISMA extension for Scoping reviews (PRISMA-ScR; see Supplemental Table 1 for checklist).¹⁹

Eligibility criteria

Eligible studies covered neutrophil-related biomarkers (defined as compounds expressed and/or secreted by neutrophils) in IBD in at least one of the following contexts: (1) prediction of disease progression and complication onset or (2) prediction of treatment outcomes.

Studies were excluded if met any of the following criteria: (1) did not refer to human studies; (2) did not refer to IBD; (3) were reviews, book chapters, guidelines, editorials, case reports, or comments; (4) had full text in languages other than English, Spanish, or Portuguese; (5) did not refer to neutrophil-related biomarkers (this criterion excludes compounds that act on neutrophils and/or influence their activity, but which are not expressed or secreted by them, such as antineutrophil cytoplasmic antibodies, and ratios like the neutrophil-to-lymphocyte ratio); (6) refer to tissues other than serum, plasma, or whole blood; and (7) refer to blood-based neutrophil-related biomarkers not studied in relation to disease progression or treatment outcomes [e.g. only (differential) diagnosis] or unable to assess predictive ability due to lack of temporal relationships (e.g. cross-sectional or single-observation studies). Journal articles, short communications, letters to the editor, and proceedings of scientific meetings were eligible if included enough information for data chart.

Search and selection process

Pubmed, EMBASE, SCOPUS, Web of Science, ScienceDirect, and the Cochrane Central Register of Controlled Trials were queried from inception up to 7 May 2022. Full queries are described in Supplemental Table 2. Reviewers also conducted a hand search for additional reports in the reference lists of the selected studies and recent reviews (2018–2022).

One reviewer screened all retrieved studies and applied eligibility criteria to title, abstract, and keywords. A second reviewer confirmed or questioned all variables extracted by the first reviewer. If eligibility was uncertain, full text was assessed. Doubts during the full-text phase were resolved through discussion between the reviewers, with possible escalation to a third reviewer. Data were systematically charted in a tailor-made form (available on request). For each entry, variables were assessed in hierarchical order, in agreement with exclusion criteria: if a study failed to fulfil one criterion, no further data were sought. Articles were assessed for species, disease, type of publication, and full-text language. Then, abstracts were scanned for neutrophil-related biomarkers by checking each compound for neutrophil expression against a reference compendium,²⁰ UniProt (https://www.uniprot.org/), the abstract content itself, and additional expression-focused literature references. If uncertain or conflicting literature persisted, the decision to include was taken by discussion between the reviewers. Next, biomarkers’ location was retrieved (e.g. blood, mucosa, feces, urine). Finally, study design and outcomes/endpoints were recorded.

As a scoping review, no a priori boundaries were established for the suitable outcomes, as long as they related with prediction of disease progression, complication onset, or treatment outcomes. Thus, clinical, endoscopic, histological, or time-based endpoints were all considered, and no restrictions were imposed on treatment type (e.g. biologics, immunomodulators, corticosteroids).

Summary of results

Results were grouped according to primary location of the biomarker in the neutrophil: granules (Table 1), cytoplasm (Table 2), plasmatic membrane (Table 3), or compounds only formed in the plasma (Table 4). The tables summarize study design, population, key endpoints, and main findings. Only the relevant results for this review were included, even if the article presented additional discoveries. When a single study reported on several biomarkers, each compound was presented in its own row, across different tables, if suitable.

Table 1.

Neutrophil biomarkers from granules.

Biomarker	Design	Population	Endpoints	Conclusions	Ref
Primary/azurophil granules
HNP 1-3	HNP 1–3 (total concentration of HNP 1, 2, and 3) measured in plasma by ELISA before and after glucocorticoids induction therapy	15 active UC patients	Disease activity (active disease in UC: UCDAI score ⩾ 3; active disease in CD: IOIBD score ⩾ 2)	Significant lower levels at induction in future non-responders (n = 7) to glucocorticoids (p = 0.002)A cutoff of 100 ng/mL had a sensitivity of 89%, a specificity of 80% and an AUC of 0.89 to discriminate responders from non-responders	Kanmura et al.²¹
Neutrophil elastase MPO	Prospective study, median follow-up 54 weeksSerum levels assessed by ELISA at enrolment and end of treatment	32 patients (CD: 21; UC: 11) treated with anti-TNFα	Primary responder (CD): HBI score ⩽ 4 or 3-point reduction at week 8Patient remission: HBI score ⩽ 4 (CD) or pMayo score < 2 (UC) after 14 weeksSustained response: HBI score ⩽ 4 (CD) or pMayo score < 2 (UC) at end of follow-up	Change from baseline to week 14 was not an independent predictor of short- or long-term remission of IBD. Baseline levels of future remitters were not compared with non-remitters. Change from baseline to week 14 was not an independent predictor of short- or long-term remission of IBD. Baseline levels of future remitters were not compared with non-remitters.	Carbone et al.²²
Secondary/specific granules
NGAL	Prospective studySerum levels measured by ELISA at baseline, post-induction (8 weeks) and 6 months	17 patients (CD: 14; UC: 3): 7 IFX and 10 ADA)	Clinical response based on HBI (CD) and Mayo Index (UC)	Baseline levels similar in future clinical responders (n = 10; 20.48 ± 9.33 ng/ml) and non-responders (n = 7; 21.07 ± 13.36 ng/ml). Decrease in NGAL levels at 6 months in clinical responders (p = ns) but increase in non-responders (p = 0.043).	Rivero et al.²³
NGAL	Gene expression datasets of blood samples containing human tissue of UCAC (GEO94648)	8 UC patients of short duration (<10 years) and 14 UC patients of long duration (⩾10 years)	UCAC	NGAL serum expression was significantly higher in short-duration UC than long-duration UC and UCAC and negatively correlated with the risk of UCAC dependent on disease duration	Kou et al.²⁴
LL-37 CHI3L1	Retrospective cohort Serum levels measured by ELISA at baseline and after a median of 9.5 weeks	145 IFX and 31 ADA UC patients with active disease	Mucosal healing (MES ⩽ 1)	Pre-IFX and pre-ADA levels not significantly different between future healers and non-healersPre-IFX and pre-ADA levels not significantly different between future healers and non-healers	de Bruyn et al.²⁵
LL-37	Retrospective cohort Serum levels measured by ELISA	Cohort 1: 23 UC, 28 CD; Cohort 2: 57 UC, 67 CD Anti-TNFα, 5-ASA, AZA, corticosteroids	Clinical activity: pMayo score and HBIMucosal activity: MES and SES-CDStrictures: observed by imaging or colonoscopy	In UC and CD patients with moderate or severe disease activity, high LL-37 levels predicted greater improvement in clinical activity after 6–18 months, regardless of the current medication. In CD, low LL-37 levels predicted intestinal strictures.	Hoang-Ngoc Tran et al.²⁶
S100A12	Prospective study Serum levels measured by ELISA before and after 14 weeks of induction	32 CD biologic-naïve patients: 20 initiating ADA and 12 initiating IFX	Clinical remission (HBI ⩽4 points at weeks 14 and 52)Loss of response (relapse of clinical symptoms and/or occurrence of evident radiologic and/or endoscopic signs of disease relapse)	Baseline values in future responders (n = 21) (mean ± SEM = 469 ± 85 µg/mL) not significantly higher than non-responders (n = 11) (293 ± 98) (p = 0.19) Low levels at week 14 predicted clinical remission within 1 year after induction with moderate ability [AUC: 0.70 (95% CI, 0.43–0.96)]	Boschetti et al.²⁷
S100A12	Prospective multicenter study, median follow-up of 45 weeks Serum levels measured by ELISA	147 adults and 34 children with CD (n = 61) or UC (n = 120) Anti-TNFα, 5-ASA, AZA, glucocorticoids	Clinical remission: CDAI < 150, PCDAI < 11, UCAI < 5, and PUCAI < 10Relapse: CDAI > 250 over 2W or CDAI > 150 with ⩾70-point increase within 2W; PCDAI > 20 over 2 weeks or PCDAI > 10 with ⩾5-point increase within 2W; UCAI > 6 over 2 weeks or UCAI > 4 with ⩾3-point increase within 2W; PUCAI > 40 over 2 weeks or PUCAI > 10 with ⩾5-point increase within 2 weeks	No differences at inclusion between future relapsers (75 ng/mL; 1–1370 ng/mL) and non-relapsers (72 ng/mL; 3–9500 ng/mL; p = 0.229). No ability to predict relapse. Levels remained unchanged during a 9-month period before relapse, even though they increased significantly at the time of relapse (n = 59; p < 0.05)	Däbritz et al.²⁸
S100A12	ProspectiveMedian follow-up of 4.4 yearsSerum levels measured by ELISA	49 CD in clinical remission (CDAI < 150) Anti-TNFα, 5-ASA, AZA	Clinical relapse: CDAI CD-related hospitalization Treatment escalation (start AZA or anti-TNFα or rescue corticosteroids) CD-related intestinal resection	Serum levels did not predict clinical relapse, hospitalizations, medication step-up, or surgery	Lee et al.²⁹
Tertiary/gelatinase granules
MMP-9	3 studies (2 of which placebo-controlled)Variable follow-up at 2, 4, 6, 8, 10, 14, and/or 18 weeksSerum levels measured by ELISA	IFX-treated fistulizing (n = 10) or active disease (n = 7) CD patients from in-house study Fistulizing (n = 42) and active CD patients (n = 24) from 2 placebo-controlled studies	Response (reduction of half or more in the number of draining fistulas from baseline)	Pre-IFX levels were not significantly different between responders and non-responders in either fistulizing or active patients across the three studies	Gao et al.³⁰
MMP-9	Prospective study, median follow-up 54 weeks Serum levels assessed by ELISA at enrolment and end of treatment	32 patients (21 CD and 11 UC) treated with anti-TNFα	Primary responder (CD): HBI score ⩽ 4 or 3-point reduction at week 8 Patient remission: HBI score ⩽ 4 (CD) or pMayo score < 2 (UC) after 14 weeks Sustained response: HBI score ⩽ 4 (CD) or pMayo score < 2 (UC) at end of follow-up	Changes from baseline to week 14 did not demonstrate to be independent predictors of short- or long-term remission of IBD. Baseline levels of future remitters were not compared with non-remitters.	Carbone et al.²²
OSM	Observational, prospective, 1-year-long, single-center study OSM determined by ELISA in the plasma	40 anti-TNFα-naive CD patients treated with IFX	Biochemical response (primary outcome) (>50% reduction in FC at week 12) Biochemical remission (FC < 250 μg/g at week 12 or 52) Clinical remission (SPCDAI < 15)	Pre-IFX low plasma OSM (<143.5 pg/mL) significantly associated with higher rate of early (week 12) biochemical response (p = 0.03) and higher rates of late (week 52) clinical (p = 0.026) and biochemical (p = 0.019) remissionsNo association between OSM and time to surgery and/or IFX discontinuation	Minar et al.³¹
OSM	Prospective, multiple cohort study OSM mRNA expression measured in whole blood using qPCR	54 active patients (24 CD, 30 UC) initiating IFX or ADA	Endoscopic response/remission (SES-CD ⩽ 2 at week 24 for CD and Mayo endoscopic subscore ⩽ 1 at week 24 for UC)	Baseline whole-blood OSM expression was numerically but not statistically lower in future responders versus non-responders (p = 0.09)	Verstockt et al.³²
OSM	Serum protein expression (proximity extension technology)	(1) Treatment-naïve cohort (CD: 41, UC: 17) (2) anti-TNFα cohort (CD: 105, UC: 81) (3) vedolizumab cohort (CD: 79, UC: 95) (4) postoperative CD patients 6 months after surgery (n = 82)	Poor prognosis: need for biologic therapy within 2 years of diagnosis Endoscopic remission: CD: absence of ulcerations at 6 months; UC: MES ⩽ 1 at week 8/14 Postoperative recurrence: Rutgeerts score ⩾ i2b at month 6	Baseline serum levels were not able to predict early need for biologic therapy or to identify future anti-TNFα or vedolizumab endoscopic non-responders (p = 0.6; p = 0.54). Serum levels 6 months after surgery predicted early recurrent CD ileitis with an accuracy of 79.6% (95% CI, 67.6–91.5%)	Verstockt et al.³³
OSM	Prospective 1-year-long study. Serum levels measured at baseline and week 14	45 CD treatment-naïve patients starting IFX	Mucosal healing at week 54	Baseline serum OSM was significantly lower in patients with mucosal healing at week 54 (p < 0.001) and predicted mucosal healing with AUC, 0.91 (95% CI, 0.81–1.00)	Bertani et al.³⁴

5-ASA, mesalamine; ADA, adalimumab; AUC, area under the curve; AZA, azathioprine; CDAI, Crohn’s disease activity index; CDEIS, Crohn’s disease endoscopic inflammatory score; CHI3L1, chitinase 3-like 1; CI, confidence interval; ELISA, enzyme-linked immunosorbent assay; FC, fecal calprotectin; HBI, Harvey-Bradshaw index; HNP, human neutrophil peptide; IBDQ, Inflammatory Bowel Diseases Questionnaire; IFX, infliximab; IOIBD, International Organization for the Study of Inflammatory Bowel Disease; LL-37, cathelicidin; MES, Mayo endoscopic subscore; MMP-9, metalloproteinase 9; MPO, myeloperoxidase; NE, neutrophil elastase; NGAL, neutrophil gelatinase–associated lipocalin; OSM, oncostatin M; PCDAI, pediatric CDAI; qPCR, quantitative polymerase chain reaction; PUCAI, pediatric UCAI; SEM, standard error of the mean; SES-CD, simple endoscopic score for CD; SPCDAI, short PCDAI; TNF, tumor necrosis factor; UCAC, UC-associated carcinogenesis; UCAI, ulcerative colitis activity index; wPCDAI, weighted PCDAI.

Table 2.

Neutrophil biomarkers released from the cytoplasm.

Biomarker	Design	Population	Endpoints	Conclusions	Ref
Calprotectin	Prospective studySerum levels measured by ELISA before and after 14 weeks of induction	32 CD biologic-naïve patients starting ADA (n = 20) and IFX (n = 12)	Clinical remission: HBI ⩽ 4 points at week 14 and 52 Loss of response: clinical symptoms and/or occurrence of evident radiologic and/or endoscopic signs)	Baseline values in future responders (n = 21) (mean ± SEM=12.7 ± 6.5 µg/mL) not significantly higher than non-responders (n = 11) (4.4 ± 0.8) (p = 0.22). Low levels at week 14 predicted clinical remission within 1 year after induction with moderate ability [AUC, 0.68 (95% CI, 0.42–0.93)]	Boschetti et al.²⁷
Calprotectin	Prospective trial, median follow-up of 28 monthsSerum levels measured by ELISA at baseline, in clinical remission (pre-IFX withdrawal), at relapse and 1 month after retreatment	103 CD patients in remission (CDAI < 150) receiving IFX for ⩾1 year	Clinical relapse: CDAI > 250 or 70 points increase from baseline over 2 consecutive weeks	Levels at baseline were not significantly different between relapsers and non-relapsers. No association with time-to-relapse. However, in a multivariate analysis and at a threshold of 5675 ng/ml, it was complementary to CRP and FC to predict relapse after IFX withdrawal.	Meuwis et al.³⁵
Calprotectin	Serum levels measured by ELISA within 24 h of admission	45 acute severe UC (ACUS)	Colectomy	Higher levels at admission predicted colectomy in ACUS with an AUC of 0.69 (95% CI, 0.53–0.81) compared with FC (AUC, 0.58; 95% CI, 0.35–0.81) and CRP (AUC, 0.71; 95% CI, 0.56–0.86)	Hare et al.³⁶
Calprotectin	Prospective, 1-year-long, single-center studySerum levels measured by ELISA within 90 days from diagnosis	83 newly diagnosed patients: 35 CD on AZA, 45 UC on 5-ASA, 3 unclassified IBD	Treatment escalation (need for escalation and establishment of ⩾2 immunomodulatory therapies and/or surgery for flare after induction; in UC, included emergency colectomy during index admission)	High levels predicted treatment escalation and/or surgery (for ⩾1046 ng/ml, HR, 2.7, 95% CI, 1.1–4.9; p = 0.007), in particular CD (for > 991 ng/ml, HR, 4.2, 95% CI, 1.2–15.3; p = 0.03), at follow-up (median 342 days; interquartile range: 88–563) in a multivariate Cox analysis	Kalla et al.³⁷
Calprotectin	Prospective study, 12 months follow-up periodLevels measured in serum by ELISA at baseline, months 3 and 6	119 patients in deep remission (CD: 75; UC: 44) on biologic stable treatment (76% on IFX, 14% on vedolizumab)	Clinical remission: CDAI < 150 or total Mayo score < 3 Mucosal healing: CD: SES-CD < 3 or simplified MARIA < 1; UC: MES < 2 Relapse: CDAI > 220 & simplified MARIA score ⩾ 1 OR SES-CD ⩾ 4; Mayo total score > 5 & MES > 1	In patients with clinical symptoms, higher levels moderately predicted relapse (AUC 0.764, 95% CI, 0.68–0.88). Concentration in non-relapsers at inclusion was significantly lower than in relapsers (p = 0.02). However, median levels remained stable in future clinical relapsers (p = 0.063).	Veyrard et al.³⁸
eNAMPT	Three independent cohorts with follow-up of 22, 14, and 14 weeksSerum levels measured by ELISA	Cohort 1: 16 CD and 5 UC IFX-treated children Cohort 2: 29 CD and 2 UC treated with ADA; Cohort 3: 25 CD and 14 UC treated with IFX	Clinical response (reduction in >2 points in HBI (CD) and in pMayo (UC) from baseline)	High pre-treatment levels predicted poor response to IFX and ADA. Non-responders maintained high eNAMPT levels at week 14. All 52 patients with levels below 4.5 ng/ml were responsive to IFX or ADA in the three cohorts	Colombo et al.³⁹

ACUS, acute severe UC; ADA, adalimumab; AUC, area under the curve; CDAI, Crohn’s disease activity index; CI, confidence interval; CRP, C-reactive protein; ELISA, enzyme-linked immunosorbent assay; eNAMPT, extracellular nicotinamide phosphoribosyltransferase; FC, fecal calprotectin; HBI, Harvey–Bradshaw index; HR, hazard ratio; IFX, infliximab; SEM, standard error of the mean.

Table 3.

Neutrophil biomarkers from the plasmatic membrane.

Biomarker	Design	Population	Endpoints	Conclusions*	Ref
nCD64 ratio	Multicenter, observational study, follow-up of up to 2 years	78 CD anti-TNFα-naïve children and young adults (<22 years)	IFX concentrations	nCD64 ratio > 6 (OR, 4.5, 95%, CI 1.4–17.8) was a pre-treatment predictor of week 14 (end of induction) IFX subtherapeutic concentrations	Xiong et al.⁴⁰
nCD64 ratio	Prospective, observational, single-center study, follow-up of 1 year Blood collected before IFX infusion (baseline) and up to post-induction (IFX-4)	56 CD children treated with IFX	Clinical response (wPCDAI prior to IFX-4 <12.5 and/or a change > 17.5 points) Sustained clinical remission (wPCDAI < 12.5 from post-induction up to 6 months) Biomarker remission (FC < 250 µg/g at week 14)	Baseline levels comparison between achievers and non-achievers not reported nCD64 ratio at post-induction did not predict IFX clinical remission outcomes at 6 months, even though median decrease from baseline to post-induction was greater in patients achieving IFX target concentration (5–10 µg/mL) versus those who do not [48% (IQR, 27%–58%) versus 20% (IQR, –6% to 36%); p = 0.031] Greater decrease in nCD64 in clinical responders at infusion 4 versus non-responders [37% (IQR, 19%–49%) versus 9% (IQR, −21% to 23%; p = 0.07)]	Colman et al.⁴¹
nCD64 ratio	Prospective, multicenter cohort study	21 children (16 CD, 5 UC) starting vedolizumab	Vedolizumab’s clearance multivariate PK modelBiomarker remission (FC < 250 μg/g)	Higher nCD64 ratio levels (one of the model covariates) were associated with worse probability of remission with vedolizumab, because nCD64 ratio positively correlated with higher clearance (R = 0.68, p = 0.001) and low clearance predicted biomarker remission with an AUC of 0.7 (sensitivity 0.8, specificity 0.75)	Colman et al.⁴²
nCD64 ratio	Prospective, observational, single-center study, follow-up of 1 year	54 anti-TNFα-naïve CD children starting IFX	Clinical response (wPCDAI prior to IFX-4 <12.5 and/or a change > 17.5 points)IFX concentrations at the end of induction	The best early predictor for CR was a nCD64 decrease between baseline and infusions 2, 3, and 4 (significantly greater reduction in responders versus non-responders at all timepoints, p < 0.05). >40% improvement in nCD64 before IFX third infusion was associated with 91% of CR versus 55% for those with <40% improvement (p = 0.016) and higher median IFX levels of 2.6 μg/ml (0.7–9.3) versus 1.5 μg/ml (0.4–21, p = 0.04) at the end of induction. Higher mean baseline nCD64 levels in future clinical responders (1.6; 0.6–1.7) versus non-responders (0.96; 0.5–4.3; p = 0.0018). 73% with high nCD64 at baseline achieved CR compared to 30% of those without (p = 0.024).	Minar et al.⁴³
CD64 index	Prospective cohort study, follow-up of 12 months	37 asymptomatic (sPCDAI < 15) CD patients (⩽17 years) under mixed maintenance therapy	Disease severity (sPCDAI): inactive (remission) if <15, mild if 15–30, and moderate to severe if >30	CD64 index < 1.0 was associated with a sustained remission rate of 95% over 1 year compared with 56% if CD64 index > 1.0 (p < 0.01)	Minar et al.⁴⁴
CD64 index	1-year-long, observational, prospective study	37 CD children, asymptomatic, steroid-free, in IFX maintenance treatment	Clinical remission (sPCDAI < 15)Time-to-IFX failure (either a SPCDAI > 15 on 2 consecutive visits, discontinuation of IFX, hospitalization or surgery)	Asymptomatic subjects with high CD64 index (silent IBD) were significantly more likely to relapse compared to those with low CD64 (deep remission) (log-rank < 0.001) over 1 year. Median CD64 index at study entry for future IFX failures was 1.04 (0.5–4.9) versus 0.64 (0.38–1.16) for those maintaining remission (p < 0.01) after 52 weeks. CD64 index elevations at baseline correlated with fewer days until IFX failure (R = −0.66, p < 0.01).	Minar et al.⁴⁵
CD64 index	Observational, prospective, 1-year-long, single-center study	40 anti-TNFα-naive CD patients treated with IFX	Clinical remission (SPCDAI < 15)Biochemical response (primary outcome) (>50% reduction of FC at week 12)Biochemical remission (FC < 250 μg/g at week 12 or 52)	Pre-IFX high nCD64 significantly associated with biochemical non-remission (p = 0.03), reinforced by a multivariate regression (p = 0.011) No association between nCD64 and time to surgery and/or IFX discontinuation	Minar et al.³¹
TREM1	Prospective, multiple cohort study TREM1 mRNA expression measured in whole blood using qPCR	54 active patients (24 CD, 30 UC) initiating IFX or ADA 22 CD patients starting ustekinumab 51 patients (25 CD, 26 UC) initiating vedolizumab	Endoscopic response/remission (SES-CD ⩽ 2 at week 24 for CD and Mayo endoscopic subscore ⩽ 1 at week 24 for UC).	Lower baseline whole-blood TREM1 expression was a significant predictor of future anti-TNFα endoscopic remission (AUC, 0.78, p = 0.001) in UC (p = 0.001), CD (p = 0.007), anti-TNFα-naïve and anti-TNFα previously exposed. Prediction was anti-TNFα specific, with no associations seen in patients treated with ustekinumab (p = 0.82) or vedolizumab (p = 0.53).	Verstockt et al.³²
TREM1	Prospective cohortTREM1 mRNA expression measured in whole blood using qPCR	20 CD patients with active disease starting IFX (13) or ADA (7)	Mucosal healing: complete absence of ulcerations and erosions after 6 months	Significantly lower baseline whole-blood TREM1 levels in responders compared with non-responders (p = 0.02), resulting in a prediction AUC of 0.80 (0.60–1.00)	Verstockt et al.⁴⁶
TREM1	Retrospective cohort TREM1 mRNA expression measured in whole blood using qPCR	22 CD patients treated with IFX	Clinical remission (cessation of diarrhea and abdominal cramping or cessation of fistula drainage and complete closure of all draining fistulas at week 14, plus decision to continue IFX at the current regimen)	TREM1 mRNA expression significantly lower in non-responders at baseline (p = 0.007). TREM1 predicted IFX response with an AUC of 94%.	Gaujoux et al.⁴⁷
TREM1	Prospective study, follow-up up to 56 days after remissionTREM1 mRNA expression measured in whole blood using qPCR	20 CD patients under anti-TNFα and/or prednisone remission induction	Clinical response (improvement in: CDEIS at 8 weeks, global physician’s assessment, CDAI, IBDQ, and next to normalization of inflammation markers)	Significantly higher TREM1 expression in future responders (n = 10) versus non-responders (n = 10) within the first 2 weeks of remission induction therapy (p < 0.01)	Peters et al.⁴⁸

nCD64 ratio is the ratio between the CD64 mean fluorescent intensity (MFI) of granulocytes and the CD64 MFI of lymphocytes; nCD64 index is the ratio of the MFI of the granulocytes to the MFI of the calibration beads.

5-ASA, mesalamine; AUC, area under the curve; AZA, azathioprine; CI, confidence interval; FC, fecal calprotectin; IFX, infliximab; IQR, interquartile range; OR, odds ratio; PCDAI, pediatric Crohn’s disease activity index; qPCR, quantitative polymerase chain reaction; PK, pharmacokinetics; sPCDAI, short PCDAI; TNF, tumor necrosis factor; TREM1, triggering receptor expressed on myeloid cells 1; wPCDAI, weighted PCDAI.

Table 4.

Neutrophil biomarkers formed in the plasma.

Biomarker	Design	Population	Endpoints	Conclusions	Reference
Soluble CD64	Prospective, observational, single-center study, follow-up of 1 yearSoluble CD64 measured by ELISA in the plasma	54 anti-TNFα-naïve CD children starting IFX	Clinical response (wPCDAI prior to IFX-4 <12.5 and/or a change >17.5 points) IFX concentrations at the end of induction	Baseline sCD64 levels were similar between responders and non-responders. sCD64 suppression during induction was strongly associated with clinical response	Minar et al.⁴³
SolubleTREM1	Prospective, multiple cohort studySoluble TREM1 measured by ELISA in the serum	54 active patients (24 CD, 30 UC) initiating IFX or ADA	Endoscopic response/remission (SES-CD ⩽ 2 at week 24 for CD and MES ⩽ 1 at week 24 for UC)	Baseline serum TREM1 was numerically but not statistically lower in future responders versus non-responders (p = 0.09).	Verstockt et al.³²
SolubleTREM1	Prospective cohortSoluble TREM1 measured by ELISA in the serum	85 CD patients with active disease starting IFX (46) or ADA (39)	Mucosal healing: complete absence of ulcerations and erosions after 6 months	Significantly lower baseline sTREM1 levels in endoscopic responders compared with non-responders (p = 0.001), resulting in a prediction AUC of 0.68 (0.56–0.80) (p = 0.0009)	Verstockt et al.⁴⁶
CPa9-HNE	Prospective, 56-week-long studySerum CPa9-HNE measured by ELISA at baseline and weeks 1, 8, 26, and 56	30 CD patients, clinically active (HBI ⩾ 5), scheduled for ADA or IFX treatment	Clinical remission (HBI < 5 at week 8 and no flares at week 26 or week 56)	Higher levels in non-remitters at baseline [tertile 1,2 versus tertile 3; OR: 6.8 (CI, 1.22–36.5); p = 0.027] and week 8 [tertile 1, 2 versus tertile 3; OR: 24 (CI, 2.56–279), p = 0.0014]	Mortensen et al.⁴⁹
CPa9-HNE	Post-hoc analysis of two double-blind, randomized, placebo-controlled studiesSerum CPa9-HNE measured by protein fingerprint assays at baseline and day 78.	88 UC and 120 CD patients treated with eldelumab (anti-interferon-gamma-inducible protein-10 antibody).	Endoscopic remission in CD (SES-CD = 0) Clinical remission (UC: Mayo score 2 with no individual subscore > 1; CD: CDAI score < 150)	Lower levels in CD patients achieving endoscopic remission at both baseline and day 78 (values not available). No differences at baseline between patients with future clinical remission versus without remission (values not available).	Luo et al.⁵⁰
CPa9-HNE	Retrospective cohortSerum CPa9-HNE measured by protein fingerprint assays	32 CD patients starting vedolizumab	Long-term response (continuation of treatment beyond 1 year)	Higher levels at baseline in non-responders compared with responders (p = 0.003). CPa9-HNE predicted response at baseline [AUC, 0.81, 95% CI, (0.66–0.96)]. Lower levels associated with less discontinuations	Sorokina Alexdóttir et al.⁵¹
CPa9-HNE	Retrospective cohortSerum CPa9-HNE measured by ELISA at baseline and week 14	54 CD patients in IFX-induction therapy for 14 weeks	Clinical remission (HBI < 5 at week 14)	In responders (n = 50), CPa9-HNE levels were significantly lower at week 14 compared to baseline (p < 0.05). Baseline levels not reported.	Sorokina Alexdóttir et al.⁵²
NGAL-MMP-9	Retrospective cohortSerum levels measured by ELISA at baseline and after a median of 9.5 weeks	145 IFX and 31 ADA UC patients with active disease	Mucosal healing (MES ⩽ 1)	Pre-IFX and pre-ADA levels were not significantly different between future mucosal healers and non-healers.	de Bruyn et al.²⁵
NGAL-MMP-9	Serum sampling before and up to 5 years after first IFX infusion (median follow-up 13 weeks)	108 active CD patients treated with IFX	Complete endoscopic healing (absence of ulcers)Partial endoscopic healing (significant improvement but still with ulcerations)	No significant differences between pre-IFX levels in future completely [n = 38; median (IQR): 84.5 (36.7–138.4) ng/ml], partially [n = 34; 57.0 (30.0–136.5)], or non-healed [n = 36; 100.9 (43.4–152.6)] patients	de Bruyn et al.⁵³
NGAL-MMP-9	Two retrospective cohorts Serum levels measured by zymography and ELISA before and after first IFX administration	66 (cohort 1) and 132 (cohort 2) active UC IFX-naïve patients	Response: complete mucosal healing (MES = 0–1 after 8–14 weeks)	Median baseline levels were similar in responders (90.8 ng/mL) and non-responders (86.1 ng/mL) in the pooled data	de Bruyn et al.⁵⁴

ADA, adalimumab; CDAI, Crohn’s disease activity index; CI, confidence interval; CPa9-HNE, neutrophil elastase (HNE)-generated calprotectin (S100a9) fragment; ELISA, enzyme-linked immunosorbent assay; HBI, Harvey–Bradshaw index; IFX, infliximab; IQR, interquartile range; MES, Mayo endoscopic score; OR, odds ratio; sCD64, soluble CD64; SES-CD, simple endoscopic score for CD; sTREM1, soluble TREM1; TNF, tumor necrosis factor; TREM1, triggering receptor expressed on myeloid cells 1; wPCDAI, weighted pediatric CDAI.

Results

Selection of sources of evidence

After removal of duplicates, a total of 1032 citations were identified from searches of electronic databases and manual review of reference lists. From these, 951 entries were excluded based on the title, abstract, and keywords, following a hierarchical application of selection criteria. The most common reason for exclusion was ‘did not refer to neutrophil-related biomarkers’ (n = 274). Of the 228 referring them, 119 were excluded as they were studied in non-blood tissues, mainly feces. In all, 81 full-text articles were assessed for eligibility. Of these, 47 were excluded for the following reasons: not-neutrophil-related biomarker (annexin 1, n = 1), lack of association with disease progression/treatment outcomes (n = 15), impossibility to assess predictive ability due to lack of temporal relationships (n = 26), and presence of duplicate datasets (n = 5). In total, 34 studies were included in the qualitative synthesis (Supplemental Figure 1).

Neutrophil biomarkers

The 34 selected records explored 17 different biomarkers (Figure 1), including 9 molecules from the granules^21–34; 2 found in the cytoplasm,^27,35–39 2 in the plasmatic membrane,^{31,32,40–48} and 4 only formed in the plasma.^{25,32,43,46,49–54} The most mentioned biomarker was nCD64 (index or ratio; 7 times),^31,40–45 followed by calprotectin (5 times)^27,35–38 and oncostatin M (OSM),^31–34 neutrophil elastase (HNE)-generated calprotectin (S100a9) fragment (CPa9-HNE),^49–52 and triggering receptor expressed on myeloid cells 1 (TREM1)^32,46–48 (4 times each).

Figure 1.

Location of biomarkers in the neutrophil.

In all, 18 studies included only CD patients, ^{27,29–31,34,35,40,41,43–49,51–53} 5 just UC patients,^{21,24,25,36,54} and 11 covered both UC and CD.^{22,23,26,28,32,33,37–39,42,50} In these, unless otherwise described, results were assumed to apply to both UC and CD. In total, 1850 non-duplicated CD patients and 1122 non-duplicated UC patients were included. Eight studies involved pediatric populations (n = 338).^28,39–45

Of the 34 selected studies, 19 had key endpoints based on disease activity,^{21–23,26–29,31,35,38,39,41,43–45,48–50,52} 8 on endoscopic response/mucosal healing,^{25,32,34,38,46,50,53,54} 8 on treatment-related endpoints,^{29,33,37,40,42,43,45,51} 7 on clinical manifestations and/or complications,^{24,29,30,33,36,47,26} and 3 on biochemical response (all FC-based definitions).^31,41,42

In total, 19 studies focused exclusively on patients taking infliximab (IFX) and/or adalimumab (ADA),^{22,23,25,27,30,31,34,35,39–41,43,45–47,49,52–54} and only five did not include anti-TNFα treatments.^{21,37,42,50,51}

Neutrophil granules

Biomarkers from the azurophil granules were assessed in one study each (Table 1). In a population of active UC patients, human neutrophil peptide (HNP) 1–3 levels were significantly higher at glucocorticoids induction in future responders (n = 8) than in non-responders (n = 7) (p = 0.002).²¹ The baseline levels and change over time of both HNE and myeloperoxidase (MPO) did not seem to predict remission in 21 CD and 11 UC patients treated with anti-TNFα.²²

In a small prospective study with anti-TNFα users, the evolution trend of serum neutrophil gelatinase–associated lipocalin (NGAL) or lipocalin-2 concentration after 6 months of treatment was different in clinical responders (non-significant decrease) and in non-responders (significant increase; p = 0.043). However, levels were similar at baseline and could not predict future clinical response.²³ In a gene expression study, NGAL serum expression was identified as a proxy of disease duration in UC, which, in turn, correlates with the risk of UC-associated carcinogenesis (UCAC), suggesting NGAL might be a useful early predictor and risk monitoring biomarker of UCAC.²⁴

LL-37 (cathelicidin) and chitinase 3-like 1 (also known as YKL-40) were evaluated in the same retrospective cohort of 176 UC patients, which did not find differences in pre-anti-TNFα levels between future mucosal healers and non-healers.²⁵ However, in a double-cohort study including 80 UC and 95 CD patients, elevated LL-37 serum levels predicted higher improvement in clinical activity after 6–18 months in UC and CD patients with moderate–severe disease activity, whereas low LL-37 levels predicted intestinal strictures in CD.²⁶

Three prospective studies reported data on S100A12 (calgranulin C). Boschetti et al.²⁷ observed that low serum levels, at week 14 post-induction, could predict clinical remission within 1 year with moderate ability [area under the curve (AUC), 0.70 (95% confidence interval (CI), 0.43–0.96)] in CD patients starting biologics. However, pre-treatment levels were not useful.²⁷ A 180-patient multicenter study did not detect differences between future relapsers and non-relapsers, both at baseline (p = 0.229) and during the period before relapse.²⁸ Serum levels also did not predict future clinical relapse, hospitalizations, medication step-up, or surgery in 49 CD subjects in remission followed for a median of 4 years.²⁹

Two articles reported on serum metalloproteinase 9 (MMP-9),^22,30 which is released by tertiary granules; one of them included three independent studies.³⁰ None of them could identify any ability of MMP-9 to predict short- or long-term remission, either defined by reduction in draining fistulas³⁰ or disease activity scores.²²

OSM is also found in tertiary granules. In two 1-year-long observational studies of anti-TNFα-naïve IFX-treated CD patients (n = 40³¹ and n = 54³⁴), low OSM pre-treatment levels were significantly associated with higher rates of early (week 12) FC-based biochemical response (p = 0.03),³¹ higher rates of late (week 52) clinical (p = 0.026) and biochemical (p = 0.019) remission,³¹ and mucosal healing (p < 0.001).³⁴ A similar trend was observed for OSM whole-blood expression in 54 patients starting anti-TNFα therapy, although the lower baseline concentrations in future endoscopic responders did not reach statistical significancy (p = 0.09).³² In contrast, in a larger study, the same authors reported that pre-treatment serum concentrations were not able to predict early need for biologic therapy or identify future anti-TNFα (n = 186) or vedolizumab (n = 174) endoscopic non-responders (p = 0.60; p = 0.54). Notwithstanding, serum OSM levels 6 months after surgery predicted early recurrent CD ileitis with an accuracy of almost 80%.³³

Cytoplasm

Serum calprotectin (SC) was evaluated in five studies (Table 2). In four of those, lower concentrations were associated with better future outcomes.^27,36–38 High levels predicted colectomy at admission in acute severe UC [AUC, 0.69 (95% CI, 0.53–0.81)]³⁶ and treatment escalation and/or surgery in newly diagnosed CD patients after a median of 342 days.³⁷ Low levels at week 14 post-induction could predict clinical remission with moderate ability within 1 year [AUC, 0.68 (95% CI, 0.42–0.93)].²⁷ Veynard et al. argued that SC could not predict clinical relapse since median levels did not change over time in clinical relapsers versus non-relapsers, but the same authors also reported that, at inclusion, the SC concentration in non-relapsers was significantly lower than in relapsers (p = 0.02) and that, in patients with clinical symptoms, it had a good predictive value for relapse [AUC, 0.764 (95% CI 0.68–0.88)].³⁸ On the other hand, SC baseline levels did not predict relapse or time-to-relapse in CD patients with long-term IFX remission.³⁵

In the only report on serum extracellular nicotinamide phosphoribosyltransferase (eNAMPT), which included three independent cohorts of anti-TNFα-treated patients, pre-treatment concentrations of <4.5 ng/ml consistently predicted response to IFX and ADA.³⁹

Plasmatic membrane

Neutrophil surface expression of CD64 (Fc-gamma receptor 1 – FcγRI) can be measured by peripheral blood quantitative flow cytometry and is represented by either the nCD64 ratio [ratio between the CD64 mean fluorescent intensity (MFI) of granulocytes and the CD64 MFI of lymphocytes^40,41,43] or the nCD64 index (ratio of the granulocytes MFI to the calibration beads MFI^31,44,45). Since the two ratios provide the same information, nCD64 refers to both in this review. Five out of seven studies found that lower pre-treatment nCD64 predicted better outcomes (Table 3). All five were prospective and four included children or young adults; 98% of the patients had CD. Higher nCD64 pre-treatment was associated with (i) increased odds of end-of-induction IFX subtherapeutic concentrations [odds ratio, 4.5 (95% CI, 1.4–17.8)]⁴⁰; (ii) decreased probability of vedolizumab remission (indirectly, by correlating with higher vedolizumab clearance)⁴²; (iii) significantly lower remission rates after 1 year of mixed maintenance treatment (56% if CD64 index > 1.0 versus 95% if <1.0, p < 0.01)⁴⁴; (iv) more frequent relapse over 1 year (log-rank < 0.001), more IFX failures (p < 0.01) and less days until failure (p < 0.01)⁴⁵; and (v) FC-based non-remission at week 12 of IFX (p = 0.03).³¹

Colman et al.⁴¹ observed in a single center, 56-patient study that nCD64 ratio post-IFX-induction did not predict clinical remission outcomes at 6 months; pre-induction ratios were not reported. Both clinical responders at the end of infusion and patients achieving IFX target concentrations displayed a greater decrease in nCD64 levels.⁴¹ This is similar to the study reported by Minar et al.⁴³ in a 1-year long, single-center study on 54 anti-TNFα-naïve CD children starting IFX, where the best clinical response early predictor was the decrease in nCD64 between baseline and infusions 2, 3, and 4 (significantly higher reduction in responders than in non-responders at all timepoints, p < 0.05). Higher baseline nCD64 was observed in future clinical responders than in non-responders (p = 0.002); 73% of patients with high baseline nCD64 and 30% without high baseline nCD64 achieved clinical response (p = 0.024).⁴³

Two small studies in CD patients mostly treated with anti-TNFα – one retrospective⁴⁷ and one prospective,⁴⁸ both measuring whole-blood mRNA expression using quantitative polymerase chain reaction – reported significantly higher TREM1 expression in clinical responders than non-responders at baseline (p = 0.007)⁴⁷ and within the first 2 weeks of remission induction therapy (p < 0.01).⁴⁸ However, Verstockt and colleagues observed in two instances that, in anti-TNFα-treated patients, lower baseline whole-blood TREM1 expression was a significant predictor of future endoscopic remission (AUC, 0.78³² and 0.80,⁴⁶ in both UC³² and CD^32,46). No prediction power was found with ustekinumab (n = 22, p = 0.82) or vedolizumab (n = 51, p = 0.53).³²

Compounds formed in the plasma

Membrane-bound CD64 and TREM1 can have their extracellular segments cleaved, shedding soluble CD64 (sCD64) and soluble TREM1 (sTREM1) into the circulation. sCD64 was reported in one prospective study, and its pre-anti-TNFα concentrations could not predict future clinical response.⁴³ In one prospective study, lower pre-IFX sTREM1 was associated with future endoscopic response.⁴⁶ Another prospective study with smaller sample reported the same trend, but just missed the statistical threshold.³²

Recently, complex biomarkers have been identified. CPa9-HNE can be measured in the serum by enzyme-linked immunosorbent assay or protein fingerprint assays. Three of four entries on CPa9-HNE found that, in CD, lower concentrations were associated with better outcomes. In a 1-year study on 30 CD patients starting anti-TNFα treatment, higher CPa9-HNE levels pre-treatment (p = 0.027) and at week 8 (p = 0.001) were found in future relapsers.⁴⁹ In 32 CD patients on vedolizumab, higher baseline levels were observed in non-responders compared with responders (p = 0.003); CPa9-HNE predicted response at baseline with an AUC (95% CI) of 0.81 (0.66–0.96).⁵¹ Serum CPa9-HNE of 88 UC and 120 CD patients treated with eldelumab (anti-interferon-gamma-inducible protein-10 antibody) was analyzed in two double-blind, randomized, placebo-controlled trials. CD patients achieving endoscopic remission had lower baseline values, even though results did not translate to UC patients nor predicted clinical remission.⁵⁰ Finally, Sorokina et al. reported that responders presented with significant decreasing levels over time (p < 0.05); however, only four non-responders were included.⁵²

The NGAL-MMP-9 complex (formed by NGAL and MMP-9) is also only found in the serum. Results across the three identified entries (one of which included two independent retrospective cohorts⁵⁴) consistently showed a lack of endoscopic healing predictive ability of NAGL-MMP-9 pre-treatment levels in anti-TNFα-treated UC^25,54 and CD⁵³ patients.

Discussion

Effective prediction of disease progression and treatment outcomes is an unmet need in the setting of chronic debilitating IBDs.^3–5,16 Biomarkers are prime candidate tools to that end. However, concerning neutrophil-related biomarkers, the focus has been on fecal molecules,^10,55 which present well-known convenience limitations.¹⁴ Indeed, of the 119 neutrophil, but non-blood-related excluded entries, 89 studied fecal compounds. This systematic scoping review aimed to summarize the current evidence on the treatment outcome predictive ability of the less approached blood-based biomarkers. Six of the 17 biomarkers from the 34 included reports stood out with promising results – OSM, SC, eNAMPT, nCD64, TREM1, and CPa9-HNE (Figure 2), despite the heterogeneity of methods, populations and endpoints, and the general low number of subjects per study.

Figure 2.

Association between biomarker levels and future outcomes by disease, neutrophil location, and number of subjects studied. Each box represents the total number of patients included in all studies that reported that association for a certain molecule and disease. Molecules with different reported associations appear in one than one quadrant. Numbers between brackets below each biomarker represent the references by order of appearance in the main manuscript.

Degranulation is an hallmark of activated neutrophils in IBD,^21,56 but granule-related biomarkers were generally found to be non-predictors of treatment outcomes.

HNP 1–3 (α-defensins 1–3), HNE and MPO were evaluated in a single small study each, precluding any meaningful conclusions.^21,22

NGAL is an iron-binding bacteriostatic protein that has also tissue destructive effects by stabilizing MMP-9. Its serum levels distinguished active IBD from healthy controls and irritable bowel syndrome patients, and correlated with disease activity.^23,57 This review identified two small studies on NGAL serum levels, one of which suggested a potential usefulness in forecasting UC-associated carcinogenesis.²⁴ Although encouraging, data are too scarce to make any conclusions.

LL-37 is an antimicrobial peptide stored in the secondary granules whose serum levels inversely correlated with clinical activity in UC and CD and with endoscopic activity in UC.²⁶ One robust study observed a clear ability to predict treatment clinical outcomes (higher levels are better),²⁶ while other solid study failed to find LL-37 predictable of mucosal healing.²⁵ Considering that the studies had similar size, design, and population, the explanation for this disparity may lie in the time of follow-up, that was significantly longer in the first study (6–18 months²⁶ versus a median of 9.5 weeks²⁵). Additional studies are required to settle the question.

S100A12 belongs to the S100 family that also includes calprotectin. As S100A12 is released almost exclusively during neutrophil degranulation, it has been hypothesized as a more specific inflammatory marker than calprotectin,²⁹ but data on fecal S100A12 versus FC as a biomarker for treatment outcomes are conflicting.^27,29 Serum levels of S100A12 did not predict future outcomes (clinical relapse, hospitalization, treatment escalation, or surgery) in two studies (total n = 230),^28,29 and despite a small study (n = 32) showing a moderate ability to predict clinical remission 1-year post-induction²⁷, the current limited body of evidence does not confirm its utility.

Two reports on MMP-9, an enzyme whose serum levels are known to correlate with disease activity,⁵⁸ both pointed to an inability in predicting short- and long-term remission.^22,30 Although not definite, the consistent results suggest that MMP-9 research should at least be deprioritized to this end.

OSM is an IL-6 family cytokine quickly released during degranulation.^59,60 OSM fosters inflammatory behavior in gut-resident stromal cells.⁶¹ Two studies found out that lower serum concentrations of OSM were associated with better anti-TNFα treatment outcomes^31,34; the results of a third study lacked statistical significance.³² However, the largest study to date failed to show such a relationship for serum levels, although it confirmed that high pre-treatment OSM mucosal expression predicted endoscopic non-response to anti-TNFα and vedolizumab.³³ Therefore, OSM is a promising biomarker that deserves further exploration.

Calprotectin is a heterodimer of S100A8 and S100A9 and comprises more than 40% of the total protein content of the neutrophil cytoplasm.⁸ FC is an established biomarker in IBD, correlating with endoscopic and clinical activity,⁸ as well as predicting relapse.⁶² SC is increased in IBD and distinguishes active from inactive disease in CD, but perhaps not in UC.^35,37,63 Although we found mixed results regarding its predictive ability for disease complications and treatment outcomes, lower levels of SC were generally associated with better future outcomes.^27,36–38

Circulating eNAMPT (visfatin) is predominantly produced and secreted by granulocytes.⁶⁴ NAMPT is involved in multiple key metabolic pathways, such as DNA repair and control of NAD/NADH pools.⁶⁵ When released, it shows proinflammatory functions.⁶⁵ Serum eNAMPT levels are significantly higher in UC^65,66 and CD^65,66 and closely correlated with disease activity.⁶⁵ In UC, serum NAMPT also correlated with mucosal healing. These results highlight eNAMPT as an exciting biomarker to explore. However, so far only one study reported on the ability to predict responses to anti-TNFα.³⁹

CD64 is a cell surface glycoprotein that binds to the Fc portion of IgG antibodies. In the resting neutrophil, only a few copies are expressed. However, inflammatory cytokines such as interferon-γ quickly elicit an up to 10-fold increase, making it a sensible marker of inflammation.⁶⁷ Lower pre-treatment nCD64 ratios predicted better outcomes in five studies.^{31,40,42,44,45} In two others, a greater reduction of nCD64 throughout the induction phase was associated with improved IFX concentrations^41,43 and clinical outcomes.^41,43 This represents a trend consistent with the concept that ‘lower nCD64 is better’; however, since it means that higher baseline nCD64 levels were found in future clinical responders⁴³ (the other study did not compare baseline levels),⁴¹ it underscores the need to explore the correct moment of assessment before applying the biomarker in clinical practice, under penalty of contributing to undesirable outcomes.

The membrane-bound TREM1 is constitutively expressed on neutrophils and monocytes/macrophages and amplifies inflammation⁶⁸; sTREM1 is generated by cleavage of the extracellular domain by MMP-9⁶⁸ or produced via a specific splicing variant.³² The apparently opposing reported results (lower whole-blood TREM1 pre-treatment levels were associated with clinical non-response^47,48 but also with better endoscopic response,^32,46 supported by sTREM1 data^32,46) might be explained by the well-described lack of correlation between clinical, endoscopic, and histological scores,⁶⁹ and/or by the small sample sizes. While promising, further studies are highly required to shed light into the usefulness of TREM1.

CPa9-HNE is thought to accurately reflect neutrophil activity⁷⁰ and strongly correlated with endoscopic disease activity in both CD and UC.⁷⁰ Three out of four retrieved records found that lower CPa9-HNE predicted favorable future outcomes in patients treated with anti-TNFα,⁴⁹ vedolizumab,⁵¹ or an anti-P10 antibody.⁵⁰ The fourth study, which included just 4 IFX non-responder patients,⁵² reported significant decreasing levels in clinical responders over time. CPa9-HNE seems to be a specific proxy of neutrophil activity and if further data confirm the positive signals, it may become a staple in clinical care.

Some important limitations are acknowledged. Summarizing information generally leads to an oversimplification of the underlying reports’ heterogeneity, especially in a scoping review where there is no quantitative look at data and the sources of evidence are not critically appraised,¹⁹ as the goal is to report all available literature. This is further compounded by potential biases on the studies. In fact, most reports were single-arm cohorts (several of which retrospective) or post-hoc analyses, often in a single center, and generally included a low number of subjects without a formal sample size calculation. While uncontrolled settings are typical of real-world studies, the inability to control for confounding factors calls for caution in taking any results as definitive. Indeed, the collected sum of evidence highlights many knowledge and quality gaps. Nevertheless, the comprehensive coverage of six international article databases using broad search queries plus the additional hand search gives confidence that no major findings were left out. Another caveat is that, even though some compounds may be produced by neutrophils, these might represent a minor source of their blood concentration, thus actually being a proxy for other cells/tissues. Literature review and authors expertise was used to rule on unclear cases, but sometimes such detail is not yet known. Finally, blood measurements might also reflect other systemic conditions.

Conclusion

In conclusion, OSM, SC, eNAMPT, nCD64, TREM1, and CPa9-HNE are promising biomarkers to predict treatment outcomes in IBD. Their appeal is further enhanced by being easily accessible. However, results are still too fragmented to draw definite conclusions. Notwithstanding, it seems clear that sharing the same compartment within neutrophils is not an indicator for similar behavior. Given the heterogeneity that characterizes CD and UC at the clinical, genomic, epigenomic, and molecular levels,⁵ a one-size-fits-all biomarker might be an elusive goal. Instead, further advances on this topic will likely depend on the identification of accurate biomarkers for well-defined subpopulations.

Supplemental Material

sj-docx-1-tag-10.1177_17562848231155987 – Supplemental material for Pursuing neutrophils: systematic scoping review on blood-based biomarkers as predictors of treatment outcomes in inflammatory bowel disease

Supplemental material, sj-docx-1-tag-10.1177_17562848231155987 for Pursuing neutrophils: systematic scoping review on blood-based biomarkers as predictors of treatment outcomes in inflammatory bowel disease by Diogo Magalhaes, Laurent Peyrin-Biroulet, Maria Manuela Estevinho, Silvio Danese and Fernando Magro in Therapeutic Advances in Gastroenterology

Footnotes

Acknowledgements

The authors thank Paula Pinto, PharmD, PhD (PMA – Pharmaceutical Medicine Academy) for providing medical writing and editorial assistance.

Declarations

ORCID iD

Fernando Magro

Supplemental material

Supplemental material for this article is available online.

References

Roda

Chien Ng

Kotze

, et al. Crohn’s disease. Nat Rev Dis Primers 2020; 6: 22.

Qiu

Chen

B-L

Mao

, et al. Systematic review with meta-analysis: loss of response and requirement of anti-TNFα dose intensification in Crohn’s disease. J Gastroenterol 2017; 52: 535–554.

Louis

Stopping anti-TNF in Crohn’s disease remitters: Pros and cons: The pros. Inflamm Intest Dis 2021; 7: 64–68.

Atreya

Neurath

Siegmund

Personalizing treatment in IBD: Hype or reality in 2020? Can we predict response to anti-TNF?

Front Med (Lausanne) 2020; 7: 517.

Giachero

Jenke

Zilbauer

Improving prediction of disease outcome for inflammatory bowel disease: progress through systems medicine. Expert Rev Clin Immunol 2021; 17: 871–881.

Drury

Hardisty

Gray

, et al. Neutrophil extracellular traps in inflammatory bowel disease: pathogenic mechanisms and clinical translation. Cell Mol Gastroenterol Hepatol 2021; 12: 321–333.

Segal

AW.

The role of neutrophils in the pathogenesis of Crohn’s disease. Eur J Clin Invest 2018; 48: e12983.

Jukic

Bakiri

Wagner

, et al. Calprotectin: from biomarker to biological function. Gut 2021; 70: 1978–1988.

Turner

Ricciuto

Lewis

, et al. STRIDE-II: An update on the selecting therapeutic targets in inflammatory bowel disease (STRIDE) initiative of the international organization for the study of IBD (IOIBD): determining therapeutic goals for treat-to-target strategies in IBD. Gastroenterology 2021; 160: 1570–1583.

10.

Vernia

Viscido

di Ruscio

, et al. Fecal lactoferrin and other putative fecal biomarkers in Crohn’s Disease: do they still have a potential clinical role? Digestion 2021; 102: 833–844.

11.

Marlicz

Skonieczna-Żydecka

Dabos

, et al. Emerging concepts in non-invasive monitoring of Crohn’s disease. Therap Adv Gastroenterol 2018; 11. Epub ahead of print April 2018. DOI: 10.1177/1756284818769076.

12.

Nardone

Shivaji

Ferruzza

, et al. Soluble blood markers of mucosal healing in inflammatory bowel disease: the future of noninvasive monitoring. Inflamm Bowel Dis 2020; 26: 961–969.

13.

Wagatsuma

Yokoyama

Nakase

Role of biomarkers in the diagnosis and treatment of inflammatory bowel disease. Life (Basel) 2021; 11: 1375.

14.

Buisson

Gonzalez

Poullenot

, et al. Comparative acceptability and perceived clinical utility of monitoring tools. Inflamm Bowel Dis 2017; 23: 1425–1433.

15.

Shiga

Abe

Onodera

, et al. Serum C-reactive protein and albumin are useful biomarkers for tight control management of Crohn’s disease in Japan. Sci Rep 2020; 10: 511.

16.

Gisbert

Chaparro

Predictors of primary response to biologic treatment [Anti-TNF, Vedolizumab, and Ustekinumab] in patients with inflammatory bowel disease: from basic science to clinical practice. J Crohns Colitis 2020; 14: 694–709.

17.

Chen

Zhou

Lin

, et al. Serum biomarkers for inflammatory bowel disease. Front Med 2020; 7: 123.

18.

Moher

Shamseer

Clarke

, et al. Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015 statement. Syst Rev 2015; 4: 1.

19.

Tricco

Lillie

Zarin

, et al. PRISMA extension for scoping reviews (PRISMA-ScR): checklist and explanation. Ann Intern Med 2018; 169: 467–473.

20.

Lominadze

Powell

Luerman

, et al. Proteomic analysis of human neutrophil granules. Molecular and Cellular Proteomics 2005; 4: 1503–1521.

21.

Kanmura

Uto

Numata

, et al. Human neutrophil peptides 1-3 are useful biomarkers in patients with active ulcerative colitis. Inflamm Bowel Dis 2009; 15: 909–917.

22.

Carbone

Bodini

Brunacci

, et al. Reduction in TIMP-2 serum levels predicts remission of inflammatory bowel diseases. Eur J Clin Invest 2018; 48: e13002.

23.

Rivero

Garcia

Garcia-Unzueta

, et al. Serum neutrophil gelatinase-associated lipocalin in inflammatory bowel disease: a new activity marker. United European Gastroenterol J 2016; 4: A625.

24.

Kou

Cheng

Shi

, et al. LCN2 as a potential diagnostic biomarker for ulcerative colitis-associated carcinogenesis related to disease duration. Front Oncol 2022; 11: 1–11.

25.

de Bruyn

Ringold

Martens

, et al. The ulcerative colitis response index for detection of mucosal healing in patients treated with anti-tumour necrosis factor. J Crohns Colitis 2020; 14: 176–184.

26.

Hoang-Ngoc Tran

Wang

, et al. Circulating cathelicidin levels correlate with mucosal disease activity in ulcerative colitis, risk of intestinal stricture in Crohn’s disease, and clinical prognosis in inflammatory bowel disease. BMC Gastroenterol 2017; 17: 63.

27.

Boschetti

Garnero

Moussata

, et al. Accuracies of serum and fecal S100 proteins (calprotectin and calgranulin C) to predict the response to TNF antagonists in patients with Crohn’s disease. Inflamm Bowel Dis 2015; 21: 331–336.

28.

Däbritz

Langhorst

Lügering

, et al. Improving relapse prediction in inflammatory bowel disease by neutrophil-derived S100A12. Inflamm Bowel Dis 2013; 19: 1130–1138.

29.

Lee

S-H

Hwang

Park

, et al. Fecal S100A12 is associated with future hospitalization and step-up of medical treatment in patients with Crohn’s disease in clinical remission: a pilot study. Intest Res 2022; 20: 203–212.

30.

Gao

Meijer

MJW

Schlüter

, et al. Infliximab treatment influences the serological expression of matrix metalloproteinase (MMP)-2 and -9 in Crohn’s disease. Inflamm Bowel Dis 2007; 13: 693–702.

31.

Minar

Lehn

Tsai

, et al. Elevated pretreatment plasma oncostatin M is associated with poor biochemical response to infliximab. Crohns Colitis 360 2019; 1: otz026.

32.

Verstockt

Dehairs

, et al. Low TREM1 expression in whole blood predicts anti-TNF response in inflammatory bowel disease. EBioMedicine 2019; 40: 733–742.

33.

Verstockt

MacHiels

, et al. Oncostatin M is a biomarker of diagnosis, worse disease prognosis, and therapeutic nonresponse in inflammatory bowel disease. Inflamm Bowel Dis 2021; 27: 1564–1575.

34.

Bertani

Fornai

Fornili

, et al. Serum oncostatin M at baseline predicts mucosal healing in Crohn’s disease patients treated with infliximab. Aliment Pharmacol Ther 2020; 52: 284–291.

35.

Meuwis

Vernier-Massouille

Grimaud

, et al. Serum calprotectin as a biomarker for Crohn’s disease. J Crohns Colitis 2013; 7: e678–e683.

36.

Hare

Kennedy

Kalla

, et al. P349 Serum calprotectin: a novel biomarker to predict outcome in acute severe ulcerative colitis? J Crohns Colitis 2014; 8: S210.

37.

Kalla

Kennedy

Ventham

, et al. Serum calprotectin: A novel diagnostic and prognostic marker in inflammatory bowel diseases. Am J Gastroenterol 2016; 111: 1796–1805.

38.

Veyrard

Roblin

Pansart

, et al. Serum calprotectin is useful to confirm inflammatory bowel disease activity but not to predict relapse. Clin Immunol Commun 2022; 2: 33–38.

39.

Colombo

Clemente

Zito

, et al. Neutralization of extracellular NAMPT (nicotinamide phosphoribosyltransferase) ameliorates experimental murine colitis. J Mol Med (Berl) 2020; 98: 595–612.

40.

Xiong

Mizuno

Colman

, et al. Real-world infliximab pharmacokinetic study informs an electronic health record-embedded dashboard to guide precision dosing in children with Crohn’s Disease. Clin Pharmacol Ther 2021; 109: 1639–1647.

41.

Colman

Tsai

Jackson

, et al. Achieving target infliximab drug concentrations improves blood and fecal neutrophil biomarkers in Crohn’s Disease. Inflamm Bowel Dis 2021; 27: 1045–1051.

42.

Colman

Mizuno

Hyams

, et al. Real-world vedolizumab pharmacokinetic study in children identifies two novel biomarkers of drug clearance. Gastroenterology 2022; 162: S100.

43.

Minar

Clarkson

Jackson

, et al. CD64 suppression predicts early clinical response to infliximab induction. J Pediatr Gastroenterol Nutr 2017; 65: S218–S219.

44.

Minar

Haberman

Jurickova

, et al. Utility of neutrophil Fcγ receptor I (CD64) index as a biomarker for mucosal inflammation in pediatric Crohn’s disease. Inflamm Bowel Dis 2014; 20: 1037–1048.

45.

Minar

Jackson

Tsai

Y-T

, et al. O-020 PMN CD64: Therapeutic target for sustained remission during infliximab maintenance. Inflamm Bowel Dis 2016; 22: S7.

46.

Verstockt

Blevi

, et al. TREM-1, the ideal predictive biomarker for endoscopic healing in anti-TNF-treated Crohn’s disease patients? Gut 2019; 68: 1531–1533.

47.

Gaujoux

Starosvetsky

Maimon

, et al. Cell-centred meta-analysis reveals baseline predictors of anti-TNFα non-response in biopsy and blood of patients with IBD. Gut 2019; 68: 604–614.

48.

Peters

van Dooren

Duijvis

, et al. Peripheral blood biomarkers for therapy response in Crohn’s disease. United European Gastroenterol J 2013; 1: A362.

49.

Mortensen

Karsdal

Grønbæk

, et al. Human neutrophil elastase derived fragment of calprotectin (S100A9) is a serum biomarker (CPA9-HNE) for monitoring of anti-tnfa treatment response in Crohn’s disease. United European Gastroenterol J 2021; 9: 465.

50.

Luo

Saini

Mortensen

, et al. P154 Evaluation of serum protein FingerPrint biomarkers of collagen, citrullinated vimentin and calprotectin in patients with inflammatory bowel disease. J Crohns Colitis 2022; 16: i231.

51.

Sorokina Alexdóttir

Bourgonje

Karsdal

, et al. Biomarkers of neutrophil activity and extracellular matrix turnover predict long-term response to vedolizumab in patients with Crohn’s disease. J Crohns Colitis 2022; 16: i468–i469.

52.

Sorokina Alexdóttir

Bourgonje

Karsdal

, et al. Biomarkers reflecting extracellular matrix turnover and inflammation can be used to monitor disease activity and treatment response in patients with Crohn’s disease undergoing infliximab induction therapy. United European Gastroenterol J 2021; 9: 458.

53.

de Bruyn

Arijs

de Hertogh

, et al. Serum neutrophil gelatinase B-associated lipocalin and matrix metalloproteinase-9 complex as a surrogate marker for mucosal healing in patients with Crohn’s Disease. J Crohns Colitis 2015; 9: 1079–1087.

54.

de Bruyn

Arijs

Wollants

, et al. Neutrophil gelatinase B-associated lipocalin and matrix metalloproteinase-9 complex as a surrogate serum marker of mucosal healing in ulcerative colitis. Inflamm Bowel Dis 2014; 20: 1198–1207.

55.

di Ruscio

Vernia

Ciccone

, et al. Surrogate fecal biomarkers in inflammatory bowel disease: Rivals or complementary tools of fecal calprotectin? Inflamm Bowel Dis 2017; 24: 78–92.

56.

Mayadas

Cullere

Lowell

CA.

The multifaceted functions of neutrophils. Annu Rev Pathol 2014; 9: 181–218.

57.

Oikonomou

Kapsoritakis

Theodoridou

, et al. Neutrophil gelatinase-associated lipocalin (NGAL) in inflammatory bowel disease: association with pathophysiology of inflammation, established markers, and disease activity. J Gastroenterol 2012; 47: 519–530.

58.

Matusiewicz

Neubauer

Mierzchala-Pasierb

, et al. Matrix metalloproteinase-9: Its interplay with angiogenic factors in inflammatory bowel diseases. Dis Markers 2014; 2014: 643645.

59.

Bolignano

della Torre

Lacquaniti

, et al. Neutrophil gelatinase-associated lipocalin levels in patients with crohn disease undergoing treatment with infliximab. J Investig Med 2010; 58: 569–571.

60.

Grenier

Dehoux

Boutten

, et al. Oncostatin M production and regulation by human polymorphonuclear neutrophils. Blood 1999; 93: 1413–1421.

61.

West

Hegazy

Owens

BMJ

, et al. Oncostatin M drives intestinal inflammation and predicts response to tumor necrosis factor-neutralizing therapy in patients with inflammatory bowel disease. Nat Med 2017; 23: 579–589.

62.

de Vos

Louis

Jahnsen

, et al. Consecutive fecal calprotectin measurements to predict relapse in patients with ulcerative colitis receiving infliximab maintenance therapy. Inflamm Bowel Dis 2013; 19: 2111–2117.

63.

Mori

Mitsuyama

Sakemi

, et al. Evaluation of serum calprotectin levels in patients with inflammatory bowel disease. Kurume Med J 2021; 66: 209–215.

64.

Friebe

Neef

Kratzsch

, et al. Leucocytes are a major source of circulating nicotinamide phosphoribosyltransferase (NAMPT)/pre-B cell colony (PBEF)/visfatin linking obesity and inflammation in humans. Diabetologia 2011; 54: 1200–1211.

65.

Neubauer

Bednarz-Misa

Walecka-Zacharska

, et al. Oversecretion and overexpression of nicotinamide phosphoribosyltransferase/pre-B colony-enhancing factor/visfatin in inflammatory bowel disease reflects the disease activity, severity of inflammatory response and hypoxia. Int J Mol Sci 2019; 20: 166.

66.

Waluga

Hartleb

Boryczka

, et al. Serum adipokines in inflammatory bowel disease. World J Gastroenterol 2014; 20: 6912–6917.

67.

Wang

Jönsson

Expression, role, and regulation of neutrophil Fcγ receptors. Front Immunol 2019; 10: 1958.

68.

Roe

Gibot

Verma

Triggering receptor expressed on myeloid cells-1 (TREM-1): a new player in antiviral immunity?

Front Microbiol 2014; 5: 627.

69.

Tajra

Calegaro

de Paula

, et al. Correlation and concordance measures between clinical, endoscopic and histological scores activity in Crohn’s disease under treatment. Scand J Gastroenterol 2019; 54: 441–445.

70.

Mortensen

Sinkeviciute

Manon-Jensen

, et al. A specific calprotectin neo-epitope (CPa9-HNE) in serum from inflammatory bowel disease patients is associated with neutrophil activity and endoscopic severity. J Crohns Colitis 2022; 16: 1447–1460.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

1.12 MB