Immunomodulatory effects of colchicine on PBMC subpopulations in human obesity: data from a randomized controlled trial

Abstract

Objective:

Colchicine is known to reduce inflammation, improve endothelial-cell function and atherosclerosis in obesity, but there is little knowledge of the specific circulating leukocyte populations that are modulated by colchicine.

Methods :

We performed a secondary analysis of a double-blind, randomized controlled trial to colchicine 0.6mg or placebo twice-daily for 3-months on circulating leukocyte populations and regulation of the immune secretome in 35 adults with obesity.

Results :

Colchicine altered multiple innate immune cell populations, including dendritic cells and lymphoid progenitor cells (LP), monocytes, and natural killer (NK) cells when compared to placebo. Amongst all subjects and within the colchicine group, changes in NK cells were significantly positively associated with reductions in biomarkers of inflammation, including COX-2, SPD, myeloperoxidase, proteinase-3, IL-16, and resistin. Changes in dendritic cells were positively correlated with changes in serum hFABP concentrations. Additionally, colchicine treatment reduced CD4+T effector cells and CD8+T cytotoxic cells. Conversely, colchicine increased CD4+ and CD8+ T central memory cells and activated CD38^HighCD8+T cells. Changes in CD4+T effector cells were associated with changes in serum hFABP.

Conclusions :

In adults with obesity, colchicine significantly affects circulating leukocyte populations involved in both innate and adaptive immune systems along with the associated inflammatory secretome.

INTRODUCTION

Obesity has become a serious public health threat throughout the world. An estimated 13% of adults worldwide suffer from obesity, and in the U.S. alone, the percentage of adults with obesity increased from 30.5% in 2000 to 42.4% in 2018 (1). Elevated body mass index (BMI) and fat mass are linked to the development of chronic inflammation, insulin resistance, and the metabolic syndrome (MetS) (2, 3). Chronic inflammation produced by activated innate immune cells is often characterized by production of a set of acute-phase effector molecules, including IL-1β, TNF-α, IL-6, IL-8, and C-reactive protein (CRP), which in turn can exacerbate cell stress in many body systems (4). The presence of low-grade inflammatory signaling induces local tissue damage and further modulates the composition and profile of leukocytes, especially peripheral blood mononuclear cells (PBMC). Increased accumulation of PBMC, also known as “leukocytosis”, positively correlates with body mass index (BMI, kg/m²) (5), and with increased adipose tissue adipokines and systemic cytokines that establish a basal proinflammatory state in PBMC (6). In addition to leukocytosis, obesity-induced inflammatory signaling modulates PBMC phenotype to support both pro- and anti-inflammatory states, which vary in degree of appropriate response (7).

The pharmaceutical drug colchicine, classically used as an anti-inflammatory agent for gout, has recently been studied as a repurposed treatment for other illnesses. Colchicine functions to inhibit microtubule formation, thereby inhibiting leukocyte locomotion and diapedesis, as well as reducing the ability of the NOD-, LRR- and pyrin domain-containing protein 3 (NLRP3) and apoptosis-associated speck-like protein containing a caspase-recruitment domain (ASC) proteins to come into close proximity and form the NLRP3 inflammasome, thus reducing pro-inflammatory cytokine production. Large randomized controlled trials (RCTs) have shown that colchicine-treated patients suffer fewer myocardial infarction, stroke, and composite cardiovascular disease outcomes than placebo-treated patients (8, 9, 10).

We previously conducted a pilot RCT investigating colchicine’s metabolic and inflammatory effects in individuals with obesity, elevated CRP, and MetS. Colchicine demonstrated significant improvements in a fasting measure of insulin sensitivity and reductions in multiple circulating inflammatory markers (11, 12). However, colchicine’s effects on circulating leukocyte populations in obesity remain poorly understood and are largely unexplored even in the chronic inflammatory conditions for which it is commonly prescribed. Herein, we performed a secondary analysis of the aforementioned pilot RCT, examining colchicine’s immunomodulatory effects on PBMC subpopulations. We also assessed whether changes in circulating immune cell populations were associated with changes in inflammatory molecules, thus providing further insight into the functional effects of colchicine on systemic inflammation in human obesity.

METHODS

Participants

This analysis uses data from a single-center randomized, double-blind, placebo-controlled trial of 40 adults who were randomly assigned in a 1:1 ratio to two groups who took oral colchicine (Spectrum Chemical MFG Corp, New Brunswick, NJ) 0.6 mg or placebo capsules, twice daily for 3 months (11). Participants were recruited between 2014 and 2018. Eligibility requirements included age ≥ 18 years; obesity (body mass index [BMI] ≥30 kg/m²; presence of MetS as defined in the AHA/NHLBI 2005 statement; evidence of chronic inflammation (hsCRP ≥ 19.0 nmol/L) and insulin resistance (Homeostatic model of insulin resistance ([HOMA-IR] ≥ 2.6). Subjects were excluded from the study if they had other metabolic disorders including diabetes mellitus, uncontrolled hypertension, congestive heart failure, or were users of any medication potentially interfering with colchicine metabolism (e.g., statins, CYP3A4 inhibitors). At baseline, all subjects underwent body composition assessment by whole-body dual-energy x-ray absorptiometry (GE Lunar iDXA, GE Healthcare, Madison WI; software GE enCore 15 with CoreScan algorithm), height measured using a calibrated stadiometer, and body weight measured using a calibrated digital scale. BMI was calculated as weight/height² (kg/m²). Blood samples were collected in the morning after ~10 hours of fasting at baseline and after 12 weeks of study drug for study of peripheral blood mononuclear cells (PBMC), for chemistry analyses by the NIH Department of Laboratory Medicine, and for serum proteomic analysis via SOMAscan 1.3k Assay (SomaLogic, Boulder, CO) as described previously (11, 13). This aptamer-based untargeted proteomics analysis detected 1305 protein analytes in human serum. In this secondary analysis we examined correlations of changes in leukocytes subsets only with the 34 molecules that remained significantly changed by colchicine after controlling for the FDR (13).

Cryopreservation of PBMC

PBMC were isolated from heparinized samples within 1 hour of collection at the NIH Center for Human Immunology, Autoimmunity and Inflammation (CHI) using a Ficoll-Paque Plus solution (Amersham Pharmacia Biotech# 17-1440-03) gradient centrifugation and cryopreserved with a freezing medium containing the 10% intracellular cryoprotectant dimethylsulfoxide (Fisher Scientific #D128500) plus 90% heat inactivated Fetal Bovine Serum (Gemini #100-106H). Cells suspended in the freezing medium were cryopreserved initially in a controlled rate freezer to a temperature of −120°C to minimize cell damage and were then transferred into the vapor phase of a liquid nitrogen (LN2) tank until thawed for use (https://chi.niaid.nih.gov/web/new/our-research/SOP-Isolation.pdf).

Immunophenotyping

We performed high-dimensional flow cytometry immunophenotyping of PBMC from study participants with data before and at 3 months of colchicine treatment (n=18). Multi-color flow cytometry using a 22-color panel was designed to allow deep immunophenotyping of the predominant cell populations found in human PBMC at NHLBI flow cytometry core protocol with some modification (14). We thawed two vials of cryopreserved PBMC containing total ~10 million cells from both the baseline (pre-study medication) and final (on-medication) study visits. Isolated PBMC cells were incubated with FluoroBrite DMEM Media at 37 °C for 1 hour (Thermo Fisher Scientific # A1896701) as recommended to allow the data to be more similar to fresh PBMCs (15) and centrifuged at @ 350 x g for 10 mins to remove DMSO and debris and resuspended in 650 μL Stain Buffer with 15 μL FC block. We split cells into an unstained control, live/dead control, and the staining tube to which we added equal volumes of a 21-antibody cocktail in FACS stain buffer (0.25mM EDTA+ 1% mouse serum in 0.1% BSA in PBS) for surface-stained cells and incubated for 40 min on ice, with the antibodies listed in Table 1 followed by LIVE/DEAD Fixable Yellow stain (Invitrogen # L34959). We also included single color bead controls for each of the 21 antibodies. The stained cells were transferred to FACS tubes. After washing cells twice with 5 times volume FACS buffer and centrifuging at 350 x g at room temperature for 5 minutes, we fixed the cells using 1% paraformaldehyde in FACS buffer.

Table 1:

List of Antibodies used for Immunophenotyping

MARKER (ANTIBODY)	CLONE	FLUOROCHROME	COMPANY	CAT. NO.	DESCRIPTION
CD206	19.2	FITC	BD	551135	Macrophage mannose receptor (MMR) or C-type lectin domain family 13 member D (CLEC13D)
CD194	1G1	PerCP-Cy5.5	BD	560726	CC Chemokine Receptor type 4 (CCR4).
IGD	IA6-2	BUV395	BD	563813	Heavy chain of human Immunoglobulin D (IgD)
CD4	SK3 (also known as Leu3a)	BUV496	BD	564651	Transmembrane glycoprotein that belongs to the immunoglobulin superfamily
CD19	SJ25C1	BUV563	BD	565697	B lymphocyte-lineage differentiation antigen. CD19
CD38	HIT2	BUV661	BD	565069	T10, ADP-ribosyl cyclase 1, and cyclic ADP ribose hydrolase 1
CD64	10.1	BUV737	BD	564425	Human IgG (FcγRI), especially the IgG1 and IgG3 subclasses.
CD8	SK1	BUV805	BD	564912	CD8α is a type I transmembrane glycoprotein
CD16	B73.1	APC	BD	561304	IgG Fc receptor III (FcγRIII)
CD56	B159	AF700	BD	557919	CD56 is a heavily glycosylated adhesion protein
CD20	2H7	APC-H7	BD	560734	MS4A1 (Membrane-spanning 4-domains, subfamily A, member 1) gene.
CD14	MΦP9	V450	BD	560349	Glycosylphosphatidylinositol (GPI)-anchored
CD3	HIT3a	BV510	BD	564713	CD3/T cell antigen receptor complex found on 70-80%
CD27	L128	BV605	BD	562655	CD27 is a 55-kDa disulfide-linked dimer that is a member of the nerve growth factor (NGF) super family.
CD25	2A3	BV650	BD	740634	Interleukin-2 receptor alpha chain subunit (IL-2Rα)
HLA-DR	G46-6	BV711	BD	563696	HLA-DR, a major histocompatibility complex (MHC) class II antigen
CD45RO	UCHL1	BV786	BD	564290	Leukocyte Common Antigen
CD127	HIL-7R-M21	PE	BD	557938	IL-7 receptor alpha (IL-7Rα) subunit. The IL-7 receptor complex is a heterodimer composed of CD127
CD294	BM16	PECF594	BD	563501	CRTH2 (chemoattractant receptor-homologous molecule expressed on Th2 cells), GPR44 (G protein-coupled receptor 44),
CD45	HI30	PEcy5	BD	555484	Protein tyrosine phosphatase receptor type C
CD11C	B-ly6	PE-Cy7	BD	561356	Adhesion glycoprotein CD11c (p150, integrin α chain).
LIVE/DEAD™ FIXABLE YELLOW DEAD CELL STAIN KIT	NA	V585/405/575	Invitrogen	L34959

Data Acquisition

Data were acquired on a FACS Symphony™ SORP (BD Biosciences) equipped with 350, 407, 455, 488, 562, 633 and 780 nm laser lines. We acquired slow events from ~2 million cells at 2000-6000 events/second using the BD FACSymphony™ system. Data were acquired and compensation was performed with BD FACSDiva™ v8.0.1 software.

Gating

Post-acquisition analysis was performed with FlowJo v.9.9.6 and v10.7 (FlowJo LLC, Ashland, OR: Becton, Dickinson and Company; 2019). The total leukocytes were first gated for singlets and excluded debris and doublets by using forward scatter height versus forward scatter area, then further analyzed for their uptake of LIVE/DEAD Yellow stain to differentiate live versus dead cells in the total cell population. Gating strategies with some modification from the previously reported protocol for the optimized multicolor immunofluorescence panel OMIP-042 (16) and cell surface markers for all listed immune cell subsets are provided in full detail in Table S1. The CD45+ SSC profile was used to identify lymphocytes (CD45^high SSC^low) and granulocytes (Total CD45+). For pre-processing data, the acquired data were cleaned and normalized using FlowJo, FlowAl and the cytonorm plugin in the FlowJo software. Singlet events from several data recordings were down-sampled to 5000 gated live events using the Flow Jo down-sample plugin and then digitally concatenated; a subsampled file of 2 million flow cytometry events was created and used for analyses of the 22 parameter datasets. Our gating strategy diagram displayed all major total lymphocyte and non-lymphocyte populations for multiple innate immune cell populations, including monocytes, dendritic cells, LP cells, NK cells, and lymphocyte cells like CD4+ T helper cells and CD8+ T cytotoxic cells, as well as B cell subclasses (Figure 1).

Figure 1:

Gating strategies for flow cytometry characterization of PBMC subpopulations. Gating scheme of 22-color staining panel for high parameter flow cytometry analysis was used to identify cell subpopulations of T cells, B cells and monocytes. PB: plasma blast cells; SM: switched memory cells; SM-DN: switched memory double negative cells; UM: unswitched memory cells; Naïve B cells; GR: granulocytes, NK cells: Natural Killer (NK) cells; CD56Dim CD16+ NK: cytotoxic natural killer cells; CD56Bright CD16− NK: proliferative NK cells; DCs: dendritic cells, CD11c+CD206- cDC1s : conventional type 1 dendritic cells; CD11c+ CD206+ cDC2s : conventional type 2 dendritic cells; Non-cDCs : non-conventional dendritic cells; LP : lymphoid progenitor cells; Total Monocytes; Classical Monocytes; Intermediate Monocytes; Non-Classical Monocytes; CD8+ TC: CD8+ T cytotoxic cells; CD38 HI CD8+: Activated CD8+ CD38^High T cells ; CD8+ T Naïve: CD8+ T Naïve cells; CD8+ TE: CD8+ T effector cells; CD38 HI CD4+: Activated CD4+ CD38^High T cells; CD4+ T Naïve: CD4+ T Naïve cells, CD4+ TE: CD4+ T effector cells; CD4+ TCM: CD4+ T central memory cells; CD4+ TEM: CD4+ T effector memory cells; Treg: T regulatory cells; CD4+ Th2: CD4+ T helper 2 cells.

ViSNE plot

The ViSNE tool, which was developed from the t-Distributed Stochastic Neighbor Embedding (t-SNE) algorithm was used to map high-dimensional cytometry data onto a 2D visualization platform while conserving high-dimensional structure. All flow cytometry data were compensated with acquisition-defined compensation matrices and analyzed using the t-SNE analysis platform to transform and reduce dimensionality. All population relevant antigens were included in the clustering analysis. Astrolabe analysis was carried out with uploaded normalized data. Single-cell data were clustered using the FlowSOM R package. Cell subset definitions follow cluster labeling, method implementation, and visualization per the Astrolabe Cytometry Platform recommendations.

Statistical analysis

No a priori power calculation was performed for the outcomes reported in this secondary analysis. Based on the limited data available from prior studies , we had estimated that 40 randomized subjects would likely have 80% power to detect a 60% difference in the primary outcome (change in insulin sensitivity) with p<0.05 (2-sided) (11), Cell populations (%) are reported as medians unless otherwise specified. Standard error of median was calculated in R using bootstrapping with 1000-fold resampling. We compared the changes between baseline and follow-up visits between treatment arms using one-way repeated measures ANOVA. P values are shown for time x treatment group interaction for all PBMC data sets unless otherwise specified. For correlation analyses, Pearson and Spearman correlation were used for normally distributed and nonnormally distributed data, respectively. Two-sided significance tests were performed for all analyses. SPSS v28.0 (IBM Corp, Armonk, NY) was used for all statistical analyses.

Study approval

The study design and experimental protocol were approved by the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) Institutional Review Board and the trial was registered at ClinicalTrials.gov (NCT02153983). The study was conducted in compliance with the principles of the Declaration of Helsinki. The study was overseen by a Data and Safety Monitoring Board convened by NICHD. We obtained written informed consent from each participant prior to conducting study procedures and data collection.

RESULTS

36 of the 40 randomized subjects completed the study (11). One subject was excluded from the flow cytometry analysis due to technical issues with the sample. As a result, PBMC samples from 17 subjects in the colchicine group (Mean ± SD: age 47.5±13.3 y; BMI 39.2 ± 6.6 kg/m2; sex: female 71%) and 18 subjects in the placebo group (age 44.7±10.3y; BMI 41.8 ± 8.2 kg/m2; sex: female 78 %) were available for flow cytometry analysis (Supplemental Figure S1). Clinical characteristics of the subjects who were analyzed in this study are shown in Table 2 (13). Mean body mass index (BMI) showed nonsignificant increases (after 3 months for subjects in the placebo group compared to the colchicine group; p=0.063). Colchicine significantly decreased inflammation compared with placebo. High-sensitivity C-reactive protein (hsCRP) decreased by 2.9 ± 2.5 mg/L in the colchicine arm and increased by 2.4 ± 9.8 mg/L in the placebo arm (P = 0.039; Table 2). There were significant decreases in total blood neutrophils count (p=0.017) and absolute monocyte count (p<0.001) in the colchicine versus placebo group (Table 2).

Table 2:

Subject demographics and changes in clinical parameters in colchicine and placebo groups.

	Colchicine (n = 17)			Placebo (n = 18)			P-Value
VARIABLE	Pre-Treatment	End-Treatment	Δ (End-Pre)	Pre-Treatment	End-Treatment	Δ (End-Pre)
Age (y)	47.5 ± 13.3			44.7 ± 10.3
Race (n, %)
Black	5, 29%			4, 22%
Caucasian	8, 47%			6, 33%
Unknown	4, 24%			8, 45%
Sex (female; n, %)	12, 71%			14, 78%
Height (cm)	168.8 ± 9.8			167.6 ± 8.2
Weight (kg)	112.2 ± 22.5	111.9 ± 23.0	−0.3 ± 2.3	118.8 ± 31.8	120.2 ± 31.2	+1.4 ± 2.9	0.063
Body mass index (kg/m2)	39.2 ± 6.6	39.1 ± 6.8	+0.2 ± 1.1	41.8 ± 8.2	42.3 ± 7.9	+0.5 ± 12.8	0.925
Body fat (%)	48.1 ± 4.0	48.3 ± 3.9	+0.3 ± 1.1	48.6 ± 5.8	48.8 ± 5.5	+0.2 ± 1.1	0.458
SI (×10−5 min−1 mU−1 mL)	9.6 ± 4.0	10.4 ± 3.7	+0.5 ± 3.4	9.6 ± 4.4	9.9 ± 5.1	+0.2 ± 3.0	0.800
Fasting glucose (mg/dL)	106.2 ± 10.2	104.2 ± 9.7	−2.0 ± 10.3	100.8 ± 7.3	103.4 ± 11.0	+2.7 ± 8.1	0.145
Fasting insulin (μIU/mL)	25.4 ± 10.3	24.2 ± 16.0	−1.2 ± 10.3	24.5 ± 11.0	29.2 ± 14.6	+4.7 ± 8.2	0.071
HOMA-IR	6.7 ± 3.0	6.4 ± 4.7	−0.3 ± 3.3	6.1 ± 2.8	7.4 ± 3.7	+1.4 ± 2.3	0.089
HbA1c (%)	5.6 ± 0.4	5.6 ± 0.5	−0.02 ± 0.3	5.5 ± 0.5	5.6 ± 0.4	+0.06 ± 0.4	0.488
hsCRP (mg/L)	5.9 ± 3.9	3.0 ± 1.9	−2.9 ± 2.5	6.5 ± 4.2	8.9 ± 8.3	+2.4 ± 9.8	0.039
ESR (mm/h)	16.4 ± 11.6	10.8 ± 7.6	−5.4 ± 10.6	21.1 ± 13.1	22.1 ± 12.1	+0.9 ± 6.3	0.042
WBC (x1000/μL)	7.5 ± 1.9	6.2 ± 1.2	−1.2 ± 1.3	6.7 ± 1.8	6.7 ± 1.7	+0.4 ± 2.3	0.019
Neutrophils (x1000/μL)	4.6 ± 1.5	3.3 ± 0.9	−1.2 ± 1.7	3.5 ± 1.2	3.5 ± 1.1	+0.2 ± 1.6	0.017
Monocytes (x1000/μL)	0.6 ± 0.2	0.5 ± 0.2	−0.1 ± 0.1	0.5 ± 0.2	0.6 ± 0.2	+0.1 ± 0.1	<0.001
Lymphocytes (x1000/μL)	2.1 ± 0.5	2.2 ± 0.4	+0.1 ± 0.4	2.1 ± 0.5	2.2 ± 0.6	+0.1 ± 0.3	0.931
Platelets (x1000/μL)	260.2 ± 59.4	244.7 ± 65.3	−15.5 ± 17.0	259.2 ± 67.9	264.3 ± 68.7	+5.1 ± 97.5	0.389

Note : Data are reported as mean ± S.D., Δ (End-Pre) is change between Week 12 and week 0 result, and calculated p-values are shown for the change in colchicine versus the change in placebo groups.

Multicolor PBMC flow cytometry analysis and qualitative visualization

To obtain a qualitative visualization of the overall effects of colchicine versus placebo on immune cell populations, the single-cell data were transformed, reduced in dimensionality, and qualitatively visualized using viSNE plots at the start and end of the trial. The changes in all immune cell subpopulations from baseline to 12-week colchicine (n=17) treatment compared to placebo (n=18) are shown in Figure 2. Statistically significant changes are described for each subpopulation in subsequent paragraphs.

Figure 2:

Multicolor dimensionality reduction (t-SNE) analysis of PBMC subpopulations comparing colchicine and placebo for pre- and post-treatment timepoints. ViSNE plots represent distributions of subpopulations at baseline (week 0) and after colchicine or placebo treatment (week 12) demonstrate that colchicine induced changes in immune cellular subpopulations. Colchicine-treated subjects: n=17. Placebo-treated subjects: n=18. Non-Classical Monocytes; Intermediate Monocytes; Lymphoid Progenitor cells; Monocytes; Classical Monocytes; Cytotoxic Natural Killer Cells (CD56^Dim CD16+); Proliferative Natural Killer Cells (CD56^Bright CD16−); Dendritic Cells; CD8+ CD38^High T: Activated CD8+ CD38^High T cells; CD8+ TCM: CD8+ T central memory cells; CD8+ TC: CD8+ T cytotoxic cells; CD4+ TEM: CD4+ T effector memory cells; CD4+ TE: CD4+ T effector cells; CD4+ TCM: CD4+ T central memory cells; CD4+ Th: CD4+ T helper cells.

Changes in circulatory monocytes in colchicine treated patients

After 12 weeks of treatment, the colchicine group demonstrated a reduction in total monocytes compared to the placebo group (Figure 3A, P=0.022), these data confirm our previously reported reduced monocyte cell numbers by automated cell counting (11). Of the major classes of monocytes, the classical subtype was significantly reduced in the colchicine-treated group with a significantly greater reduction compared to placebo at the end of the treatment period (Figure 3B, P=0.007). Neither intermediate monocytes (Figure 3C, P=0.495), nor non-classical monocyte populations (Figure 3D, P=0.606) were significantly affected in the colchicine-treated group compared to placebo at end of treatment.

Figure 3:

Monocyte subpopulations pre- and post-colchicine treatment. A) monocytes; B) classical monocytes; C) intermediate monocytes; D) non-classical monocytes; E) natural killer (NK) cells F) proliferative NK cells (CD56^Bright CD16−) G) cytotoxic NK cells (CD56^Dim CD16+) H) dendritic cells I) lymphoid progenitor cells, in colchicine (means are closed circles and bold lines, n=17 dotted individually in grey closed circles) vs placebo group (means are open circles and dotted lines, n=18, dotted individually in grey open circles).

Changes in circulatory natural killer cells in colchicine treated patients

Pre- and post-colchicine treatment compared with placebo showed that colchicine decreased a subset of innate leukocytes populations, natural killer (NK) cells (Figure 3E, P= 0.002). However, 12-weeks of colchicine did not significantly change proliferative CD56^Bright CD16− NK cells (Figure 3F, P=0.065) or cytotoxic CD56^Dim CD16+ NK cells (Figure 3G, P=0.565) in comparison to placebo.

Changes in circulatory dendritic cells in colchicine treated patients

Dendritic cells (DCs) are key immune sentinels linking the innate and adaptive immune systems. DCs recognize foreign or self-signals and initiate T-cell activation/tolerance, memory, and polarization. Following 12 weeks of treatment, colchicine increased total DC populations as compared to placebo (Figure 3H, P=0.048). Further analysis of specific DC subsets showed that non-conventional and conventional dendritic cells (CD206+ and CD206−) were not significantly affected by colchicine treatment compared to control (Table 3). Lymphoid progenitor cells (LP) significantly increased in colchicine-treated participants, as compared to placebo (Figure 3I, P=0.047).

Table 3:

All analyzed subsets of PBMC comparing between 0 week and 12 weeks of colchicine versus placebo treatment.

	Colchicine (n = 17)		Placebo (n = 18)		P-Value
Subset/Cell type	Pre-Treatment	End-Treatment	Pre-Treatment	End-Treatment
Monocytes	7.78 ± 1.71	5.42 + 2.09	6.17 ± 2.56	5.85 ± 2.80	0.022
Classical Monocytes	6.16 ± 1.39	4.08 ± 1.42	4.83 ± 1.91	4.63 ± 2.28	0.007
Intermediate Monocytes	0.09 ± 0.08	0.07 ± 0.08	0.06 ± 0.07	0.06 ± 0.06	0.495
Non-Classical Monocytes	1.53 ± 0.89	1.27 ± 0.96	1.27 ± 0.99	1.16 ± 0.95	0.606
GR Cells	7.941 ± 3.407	8.808 ± 4.093	7.040 ± 2.853	7.685 ± 3.793	0.850
Natural Killer (NK) Cells	6.90 ± 2.61	4.67 ± 2.20	5.01 ± 1.56	6.25 ± 2.85	0.002
Proliferative NK Cells	0.33 ± 0.13	0.40 ± 0.22	0.14 ± 0.17	0.38 ± 0.21	0.065
Cytotoxic NK Cells	1.16 ± 0.47	1.13 ± 0.67	1.14 ± 0.87	1.31 ± 0.54	0.565
Dendritic Cells (DCs)	2.64 ± 0.85	3.64 ± 1.73	2.79 ± 1.31	2.54 ± 0.85	0.048
cDC1s Cells	0.483 ± 0.239	0.425 ± 0.173	0.406 ± 0.269	0.366 ± 0.144	0.837
cDC2s Cells	0.147 ± 0.080	0.136 ± 0.082	0.118 ± 0.085	0.096 ± 0.068	0.665
Non-conventional Dendritic Cells	1.496 ± 0.711	1.348 ± 0.754	1.443 ± 0.878	1.533 ± 0.968	0.315
Lymphoid Progenitor Cells (LP)	0.335 ± 0.236	0.520 ± 0.379	0.310 ± 0.281	0.303 ± 0.208	0.047
T Cells	36.20 ± 8.35	39.19 ± 11.72	45.22 ± 8.46	48.67 ± 11.00	0.897
CD4+ Th Cells	26.06 ± 5.22	30.68 ± 10.34	34.36 ± 7.93	33.27 ± 6.60	0.087
CD4+ Activated CD38 High Cells	0.0035 ± 0.0021	0.0029 ± 0.0032	0.0067 ± 0.0114	0.0026 ± 0.0033	0.265
CD4+ TCM Cells	17.23 ± 7.01	20.96 ± 7.99	25.94 ± 12.30	20.27 ± 7.17	0.036
CD4+ T effector Cells	27.75 ± 12.53	23.99 ± 12.36	17.71 ± 11.53	28.48 ± 12.55	0.031
CD4+ TEM Cells	21.17 ± 7.71	21.84 ± 10.67	22.56 ± 11.66	18.40 ± 5.75	0.236
CD4+ Th2 Cells	0.02 ± 0.04	0.14 ± 0.31	0.11 ± 0.23	0.24 ± 0.37	0.917
CD4+ Naïve Cells	28.65 ± 11.13	28.19 ± 14.35	30.07 ± 15.06	28.89 ± 12.64	0.913
CD4+ Treg Cells	6.57 ± 1.58	5.44 ± 2.08	5.51 ± 2.01	4.23 ± 1.22	0.861
CD8+ TC Cells	9.48 ± 3.07	8.17 ± 2.34	9.20 ± 2.72	12.60 ± 6.04	0.019
CD8+ Activated CD38 High Cells	0.08 ± 0.11	0.47 ± 0.44	0.14 ± 0.51	0.03 ± 0.02	0.004
CD8+ TCM Cells	10.22 ± 5.84	14.38 ± 5.16	18.28 ± 7.95	12.03 ± 4.66	0.004
CD8+ T Effector Cells	23.37 ± 13.31	15.19 ± 9.63	21.35 ± 10.65	21.39 ± 11.97	0.202
CD8+ TEM Cells	8.47 ± 3.90	8.14 ± 5.50	12.93 ± 6.10	7.04 ± 6.85	0.068
CD8+ T Naïve Cells	49.36 ± 20.76	53.00 ± 18.78	44.64 ± 16.89	55.42 ± 19.11	0.446
B Cells	4.96 ± 3.26	7.06 ± 3.58	5.35 ± 2.34	4.99 ± 2.85	0.105
Naïve B Cells	4.05 ± 3.04	6.19 ± 3.05	4.58 ± 2.32	4.30 ± 2.60	0.076
B SM Cells	0.10 ± 0.06	0.16 ± 0.13	0.14 ± 0.08	0.16 ± 0.10	0.417
B SM PB Cells	0.001 ± 0.001	0.001 ± 0.002	0.001 ± 0.002	0.001 ± 0.002	0.680
B SM DN Cells	0.74 ± 0.55	0.64 ± 0.43	0.54 ± 0.20	0.43 ± 0.27	0.989
B UM Cells	0.01 ± 0.01	0.03 ± 0.02	0.03 ± 0.03	0.04 ± 0.01	0.352

Note :Data are reported as mean ± S.D. Group*time interaction p-values from ANCVOAs are shown for colchicine versus the placebo treatment.

Changes in circulatory helper and cytotoxic T cells in colchicine treated patients

We calculated T cells percentage from CD4+ and CD8+ as parent population respectively. CD4+ T effector cell populations were reduced in colchicine treatment compared to significant increases in the placebo group along the 12 weeks of treatment, leading to a significant time x group interaction (Figure 4A, P=0.031); further interrogation identified no change in CD4+ T effector memory cell populations in the colchicine group (Figure 4B, P=0.236). CD4+ T central memory (TCM) cells population increased following treatment with colchicine compared to placebo (Figure 4C, P=0.036). The total population of CD4+ T-helper cells did not differ significantly between the groups (Figure 4D, P=0.087).

Figure 4:

Circulatory helper and cytotoxic T cells in colchicine treated patients. A) CD4+ T effector cells; B) CD4+ T effector memory cells; C) CD4+ T central memory cells; D) CD4+ T helper cells; E) CD8+ T cytotoxic cells; F) CD8+ T naïve cells; G) CD8+ T central memory cells; H) Immunosuppressive CD8+ Activated CD38^High T cells; in colchicine (means are closed circles and bold lines, n=17, dotted individually in grey closed circles) vs placebo group (means are open circles and dotted lines, n=18, dotted individually in grey open circles).

Among CD8+ T cytotoxic (TC) cell populations, total TC cells (Figure 4E, P=0.019) were significantly downregulated post-colchicine treatment, while CD8+ T central memory (TCM) (Figure 4G, P=0.004) and CD8+ CD38^High cells (Figure 4H, P=0.004) were increased. CD8+ T naïve (Figure 4F, P=0.446) were not changed. Changes in all other cell subpopulations examined comparing colchicine versus placebo treatment are shown in Table 3. Additionally, surface markers’ expression as mean fluorescence intensity per cell used to identify key cell populations were not altered upon colchicine treatment (Figure S2).

Association of changes in circulatory immune cells and in inflammatory biomolecules

We examined the relationships between changes in circulatory immune cells and changes in inflammatory biomolecules markers for the entire group, for the colchicine group alone, and for the placebo group alone to better understand how colchicine treatment affected inflammation circulatory markers of circulatory inflammation. Results are shown for correlations where the colchicine group showed similar trends to those for the total cohort and the slopes were significant.

Within the total cohort and group-wise analysis (Figure 5A–G), changes in NK cells were significantly associated with changes in biomarkers markers of inflammation, including cyclooxygenase-2 (COX-2; r=.560, P=0.001 for the entire group; P=0.002 for the colchicine group), surfactant protein D (SPD; r=.558, P=0.001 for the entire group; P=0.017 for the colchicine group), myeloperoxidase (r=.513, P=0.002 for the entire group; P=0.050 for the colchicine group), proteinase 3 (r=.508, P=0.002 for the entire group; P=0.042 for the colchicine group), interleukin 16 (IL-16; r=.489, P=0.003 for the entire group; P=0.064 for the colchicine group), resistin (r=.472, P=0.005 for the entire group; P=0.026 for the colchicine group), and phosphodiesterase 5A (PDE5A; r=.475, P=0.005 for the entire group; P=0.007 for the colchicine group). Changes in dendritic cells populations were positively associated with changes in serum heart-type fatty acid-binding protein (hFABP; Figure 5H; r=.518, P=0.002 for the entire group; P=0.043 for the colchicine group). In our study, changes in CD4+ T effector cells were negatively correlated with changes in hFABP (Figure 5I; r=−.568, P=0.0001 for the entire group; P=0.012 for the colchicine group).

Figure 5:

Association of changes (Δ) in non-lymphatic/lymphatic and changes (Δ) in inflammatory biomolecules among all study participants (dark red lines; p-values shown). Correlations between A) Δ NK cells and Δ COX2; B) Δ NK cells and Δ SPD; C) Δ NK cells and Δ myeloperoxidase; D) Δ NK cells and Δ proteinase 3; E) Δ NK cells and Δ IL-16; F) Δ NK cells and Δ resistin; G) Δ NK cells and Δ PDE5A; H) Δ Monocytes and Δ GDF15; H) Δ Dendritic cells and Δ hFABP; I) Δ CD4+ TE and Δ hFABP. Y axes are measured in arbitrary relative fluorescence units. Colchicine-treated subjects: n=16; closed circles, black lines. Placebo-treated subjects: n=18; open circles, dotted lines; all subjects: red lines.

DISCUSSION

Pharmacological interventions such as colchicine that target inflammatory pathways have been proposed to alleviate complications induced by obesity and progression of metabolic syndrome, insulin resistance, diabetes, and cardiovascular disease (17). We have previously reported that colchicine administration significantly affected serum concentrations of proteins involved in activation of the innate immune system, endothelial cell function, and atherosclerosis, including reductions in C-reactive protein, IL-6, GlycA, neutrophil count, and monocyte count in obesity-induced inflammation (11, 13, 18). We also previously reported improvements in fasting insulin resistance and rates of lipolysis that suggest potential metabolic improvements after colchicine versus placebo treatment (11, 19).

Colchicine is known to accumulate in high concentrations in granulocytes and mononuclear cells (20, 21). Other in vitro and mouse studies have suggested that colchicine specifically has direct effects on monocytes, T cells, and B cells (22). In the present study, we report that colchicine significantly modulated circulating leukocyte populations in multiple cell lineages in patients with obesity and inflammation; these include monocytes, natural killer (NK) cells, dendritic cells, and lymphocytes. We found reductions in circulating monocytes and classical monocytes but not intermediate, and non-classical monocyte percentages among colchicine-treated subjects when compared to the placebo group. This demonstrates that in this study, colchicine was effective in reducing abundance of multiple monocyte lineages that produce pro-inflammatory cytokines known to be upregulated in obesity-induced inflammation (23, 24), thereby, potentially improving the inflammatory state observed in severe obesity.

Previous studies have demonstrated that NK cells may serve as an important link between obesity and ATM M1-polarization and accumulation, and that reducing NK cell populations may improve adipose tissue inflammation and insulin resistance (25). Moreover, it has been shown that NK subsets vary considerably in their cytokine production profile: the predominant proliferative (CD56^Bright) NK cells produce high concentrations of cytokines, while the cytotoxic (CD56^Dim) cells produce low cytokine concentrations (26). Our study found that colchicine specifically induced reductions in NK cells, which were associated with multiple circulating inflammatory biomarkers, including the pro-inflammatory molecules COX-2, IL-16, and resistin. Further research is warranted to investigate whether any of these correlations are causative, and if so, in which direction.

In the obesity-mediated pro-inflammatory state, activated neutrophils infiltrate into adipose tissue, thus secreting high levels of circulating neutrophil-derived proteins, mainly myeloperoxidase (MPO), which further induces high CD66b expressing circulating neutrophils to secrete pathological levels of superoxide (30). MPO expression in individuals with obesity is positively associated with NK cells, suggesting further activation of neutrophils. MPO and CD162 levels are reportedly significantly increased in “intermediate” CD14+CD16+ as well as “classical” CD14+CD16− monocytes in severe peripheral artery occlusive disease and atherogenesis (31). PBMC monocytes and MPO are major targets for colchicine in suppression of inflammation and progression of obesity pathology. A possible mechanism for the activation of neutrophils could be the increased plasma concentrations of leptin and tumor necrosis factor-α observed in severe obesity (32). MPO is a contributing factor for pathogenesis of atherosclerosis in hypercholesterolemic transgenic mice that express MPO in their macrophages (33). Therefore, actions of colchicine, which we have reported reduced MPO (13), may be studied as a potential strategy for prevention and treatment of obesity and metabolic syndrome.

Circulating DCs are known to be involved in the chronic inflammation of obesity (34). Although our study found increased total DCs populations in the colchicine-treated groups when compared to placebo, nevertheless, the myeloid (or conventional) DCs subpopulations were not significantly altered, suggesting that the increased total DCs could be from a lymphoid-derived subset, such as plasmacytoid dendritic cells (pDCs); this hypothesis is supported by our reported increased lymphoid progenitor cells upon colchicine treatment, which may be giving rise to mature pDCs (35). Although pDCs are generally considered to be involved in viral responses,(36) these have also been reported to induce Tregs cell subsets (37), which could explain our observed increased CD8+ T reg population in the colchicine-treated group. Interestingly, some data suggest pDCs may regulate energy metabolism and promote the development of obesity via type I interferons. Alternatively, the observed increased total DCs with colchicine treatment could also be from a subset of tolerogenic conventional DCs that could elicit the anti-inflammatory effects shown in this study. Tolerogenic DCs can promote an anti-inflammatory phenotype in visceral adipose tissue while also delaying obesity-induced inflammation and insulin resistance (38).

FABPs interface between inflammation and metabolic signals. Classically, FABP4 (aP2) is mainly expressed in adipose tissue depots and immune cells (39), while heart-type fatty acid binding protein (hFABP) is released from myocytes during ischemic injury. hFABP can function similarly to FABP4 to facilitate fatty acid transport though the plasma membrane in immune cells (40). Colchicine significantly increased several molecules including hFABP, that are known to be altered in human obesity (13). Changes in total dendritic cells populations were associated with changes in hFABP concentrations. hFABP also plays roles in DC functions; it regulates cytokines production and activation of T naïve to T helper effectors cells (39). We observed changes in CD4+ T effectors cells were negatively correlated with changes in hFABP.

Low-grade chronic inflammation is also linked to the physiologic adaptive immune response known as “para-inflammation,” which can contribute to disease progression and affect recruitment of inflammatory cells in tissue. These changes lead to variation in total populations and activation states of T-cells in obesity (41), with Th1 polarization and recruitment of CD8+ cytotoxic T-cells (42) and thus, the pro-inflammatory T-cell profile of adipose tissue contributes to the chronic inflammatory state of obesity (43). We found increased peripheral CD4+ T helper (only marginally) and TCM cells, along with lower CD8+ cytotoxic T and CD8+ T effector cells, after colchicine versus placebo treatment. Importantly, our work also showed increased abundance of CD8+ CD38^high T cell subsets, which are known for their immunosuppressive properties (44), potentially contributing to improved inflammation observed in the colchicine treated group.

A strength of our study was the randomized, double-blind placebo-controlled design conducted in a racially diverse cohort of adults with elevated inflammation and obesity. To our knowledge, we have conducted the first placebo-controlled study examining colchicine’s immunomodulatory effects on peripheral blood mononuclear cell subpopulations and the associated changes in inflammatory molecules. Limitations of the study include the relatively small sample size, which may have limited the ability to detect differences between groups, that no a priori power calculation was performed for this analysis, and that the results presented are from a secondary analysis of the trial without adjustments in significance tests for multiple comparisons. In particular, the many correlational analyses conducted in this small cohort likely include some false positive results (Type 1 errors). Thus, these results require additional definitive studies and should be considered hypothesis generating. We reported reductions in total blood neutrophils as measured from automated blood counts previously (11, 19), however, the isolation and cryopreservation of PBMCs prevented flow cytometry analysis of neutrophils, therefore presenting another limitation to our study. As neutrophils are major targets for colchicine’s anti-inflammatory mechanisms in human obesity, flow cytometry analyses of neutrophils collected in colchicine vs. placebo-controlled, randomized clinical trials, are needed. As we have not compared immunophenotyping results in fresh and frozen colchicine-exposed samples, we do not know if colchicine may alter the resistance of some PBMC subpopulations to freezing/thawing, which theoretically might affect results. Further studies are needed to assess whether colchicine effected changes in particular inflammatory cell populations, or may impair the cell surface localization of certain proteins critical for accurate flow cytometric quantification, although it did not affect expression of major subclasses (Figure S2). Larger, definitive analyses from placebo-controlled trials are needed to confirm and extend these findings and allow more mechanistic evaluations such as pathway analysis. Finally, some of the reported effects appear caused as much by changes in the placebo group as from changes in the group randomized to colchicine. Although the differences in change of body mass index observed in this study were not significant, the placebo group did gain somewhat more than the colchicine group, suggesting progressive obesity may have led to some cell populations trending towards pro-inflammatory in the placebo group while colchicine restrained or stopped progression towards worsening of the obesity phenotype.

CONCLUSION

In conclusion, among adults with obesity, inflammation and insulin resistance, colchicine significantly affects circulating leukocyte populations involved in both innate and acquired immune systems. These changes are also associated with alterations in the inflammatory secretome. Further studies are required to determine which of these changes are most linked to improvements in metabolism induced by colchicine treatment.

Study Importance.

What is already known?

• Colchicine, an anti-inflammatory medication classically used in gout, blocks microtubule and NLRP3 inflammasome formation, thus reducing pro-inflammatory cytokine production. Colchicine has gained considerable interest for its potential cardioprotective benefits and potential to reduce acute myocardial infarction.

• Colchicine significantly affects serum concentrations of proteins relevant for the innate immune system, endothelial cell function, and atherosclerosis including C-reactive protein and IL-6.

What does this study add?

• Colchicine 0.6mg, taken twice daily for three months by 35 adults with obesity, changed circulating leukocyte populations, including total monocytes, classical monocytes, natural killer cells, T effector and central memory cells when compared to placebo.

• Key Inflammatory molecules, including COX2, SPD, Myeloperoxidase, proteinase 3, IL16, resistin and PDE5A correlated positively with NK cells, dendritic cells; T effector cells correlated negatively with hFABP.

How might these results change the direction of research?

This clinical research study advances our understanding of the systemic anti-inflammatory mechanisms of colchicine in humans with obesity and metabolic syndrome.

DATA AVAILABILITY STATEMENT:

The individual participant data that underlie the results reported in this article, after deidentification (text, tables), will be made available on request to the corresponding author immediately after publication, to researchers who provide a methodologically sound proposal for any purpose. To gain access, data requestors will need to sign a data access agreement.