quantitative comparison of human intestinal mononuclear leukocyte isolation techniques for flow cytometric analyses

In this study we evaluated cell yield, viability, and surface-molecule expression on mononuclear leukocytes, comparing three techniques to obtain a single immune cell suspension from hum

Quantitative comparison of human intestinal mononuclear

leukocyte isolation techniques for flow cytometric analyses

R.R.C.E Schreurs, A Drewniak, R Bakx, W.E Corpeleijn,

T.H.B Geijtenbeek, J.B van Goudoever, M.J Bunders

Please cite this article as: R.R.C.E Schreurs, A Drewniak, R Bakx, W.E Corpeleijn, T.H.B Geijtenbeek, J.B van Goudoever, M.J Bunders , Quantitative comparison of human intestinal mononuclear leukocyte isolation techniques for flow cytometric analyses The address for the corresponding author was captured as affiliation for all authors Please check if appropriate Jim(2017), doi: 10.1016/j.jim.2017.03.006

This is a PDF file of an unedited manuscript that has been accepted for publication As

a service to our customers we are providing this early version of the manuscript The manuscript will undergo copyediting, typesetting, and review of the resulting proof before

it is published in its final form Please note that during the production process errors may

be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Quantitative comparison of human intestinal mononuclear leukocyte isolation techniques for flow cytometric analyses

R.R.C.E Schreursa,b, A Drewniakb, R Bakxc, W.E Corpeleijna, T.H.B Geijtenbeekb, J.B van Goudoevera, M.J Bundersa,b*

a

Department of Pediatrics, Emma Children’s Hospital, Academic Medical Center (AMC), University of Amsterdam (UvA), Amsterdam, The Netherlands

b

Department of Experimental Immunology, AMC, UvA

c

Department of Pediatric Surgery, Pediatric Surgical Center, Emma Children’s Hospital, AMC & Free University Medical Center Amsterdam, Amsterdam, The Netherlands

* Correspondence should be addressed to M.J.B (m.j.bunders@amc.nl), department of Experimental Immunology, AMC, UvA, Meibergdreef 9, 1105 AZ, Amsterdam, The Netherlands

Keywords:

Human

Lymphocytes

Monocytes

Isolation method

Intestine

ABSTRACT

Studies on immune cells derived from the human intestine are needed to understand the pathogenesis of gastrointestinal diseases and to develop novel treatment strategies Isolation techniques to extract these immune cells from intestinal tissue are largely based on murine studies and comparative data on isolation from human intestine is scarce In this study we evaluated cell yield, viability, and surface-molecule expression on mononuclear leukocytes, comparing three techniques to obtain a single immune cell suspension from human intestine; low concentrations of either the enzymes Collagenase D or Liberase TL, and enzyme-free mechanical dissociation with the Medimachine Both enzymatic isolation techniques provided a higher cell yield than mechanical dissociation Expression of surface molecules remained intact after Collagenase D treatment, while Liberase TL digestion resulted in a strong decrease in the expression of the CD4 receptor on T cells Monocyte-derived cell surface molecules were not differentially affected by either enzyme Taken together, Collagenase D digestion provides the highest yield of T cells while keeping surface molecule

ACCEPTED MANUSCRIPT

expression intact Both Collagenase D and Liberase TL result in a good yield without differentially altering the expression of cell surface molecules when investigating monocyte-derived cells

1 Introduction

The human intestinal mucosa is a dynamic interface between the body and the environment The intestinal lumen provides an extensive surface area for the digestion and absorption of essential metabolites (Mowat & Agace, 2014) At the same time, adequate barrier function is required to prevent microbial translocation (Brown, Sadarangani, & Finlay, 2013; Cebra, 1999) To maintain homeostasis under continuous microbial exposure, tissue-resident intestinal immune cells such as lymphocytes and monocyte-derived cells play a crucial role However, protective immune responses need to be tightly regulated in order to prevent excess inflammation and collateral tissue damage (Siddiqui & Powrie, 2008; Veenbergen & Samsom, 2012)

Dysregulation of intestinal immune cells leads to increased susceptibility to severe infections, while deficient tolerant responses of lymphocytes and monocyte-derived cells provide the underlying conditions for intestinal inflammatory diseases such as ulcerative colitis (Troncone, Marafini, Pallone,

& Monteleone, 2013) In order to understand the pathogenesis of intestinal diseases, in-depth analyses of immune responses at the tissue level are essential as populations of immune cells in the blood and the intestinal compartments differ substantially (Kunkel & Butcher, 2002)

The number of studies investigating tissue-resident immune cells is rapidly increasing (Annunziato et al., 2007; Bunders et al., 2012; Sathaliyawala et al., 2013; Thome et al., 2016) Throughout those studies, various techniques are applied to isolate the mononuclear leukocyte fraction from the human intestine, first pioneered by Wahl and Smith (1991) However, these techniques differ in terms of cell yield and expression of the molecules of interest on the cell surface (Chen et al., 2014; Shen et al., 2015)

Mucosal lymphoid and monocyte-derived cells are present in several intestinal compartments; the epithelium, the underlying lamina propria, the Peyer’s patches (PPs) embedded

in the small intestine, and the isolated lymphoid follicles (ILFs) embedded in the colorectum These compartments contain cell types with a phenotype and functionality unique to their anatomical location (Mowat & Agace, 2014) In this study we focus on the isolation of mononuclear leukocytes from the epithelium and from the lamina propria, excluding PPs and ILFs on macroscopic visual inspection of the tissues

The isolation of the intraepithelial mononuclear leukocytes (IELs) is homogenous among protocols; dithiothreitol (DTT) and ethylenediaminetetraacetic acid (EDTA) are used to remove

ACCEPTED MANUSCRIPT

mucus and to detach the epithelial cells from the tissue Subsequently, enrichment of the cell suspension for IELs is commonly achieved with gradient centrifugation, which separates cells based

on their buoyant density (Fuss, Kanof, Smith, & Zola, 2009) However, layer gradients vary between protocols from 60%-67.5% Percoll (Braunstein, Qiao, Autschbach, Schürmann, & Meuer, 1997; Ebert

& Roberts, 1995; Lundqvist, Hammarström, Athlin, & Hammarström, 1992) to using standard Ficoll-Hypaque (Comer, Ramey, Kotler, & Holt, 1986)

After detachment of the epithelial layer, the mononuclear leukocytes from the lamina propria (LPLs) can be obtained via mechanic or enzymatic disaggregation of the tissue The described isolation techniques to obtain this cell population vary greatly with regards to cell yield (Shacklett et al., 2003) Furthermore, the expression of molecules of interest on the cell surface (Chen et al., 2014; Shen et al., 2015) can be differentially altered by the techniques thereby impacting the feasibility of the study as well as the interpretation of the obtained results An isolation technique that harvests the maximal number of cells while limiting alterations to cell surface marker expression is particularly relevant to pediatric studies as collection of sufficient tissue is often more challenging than in adults

Taken together, comparative data of cell numbers and surface marker consistency of the most frequently used techniques to isolate mononuclear leukocytes from the epithelium and lamina propria is needed to design the appropriate methodology for performing intestinal immune cell analyses In this study, we compared three frequently used methods of lamina propria disaggregation

to obtain a single mononuclear leukocyte suspension: Medimachine (enzyme-free mechanical disaggregation), and low concentrations of the enzymes Collagenase D and Liberase TL Furthermore,

we investigated the efficiency of using different density gradients to enrich for IELs and LPLs The different isolation and enrichment techniques of leukocytes (T lymphocytes and monocyte-derived cells) from intestinal mucosal tissue were compared regarding cell yield, viability, and surface marker expression

2 Materials and methods

2.1 Tissue samples

Human tissues were collected after the donors or their guardians provided informed consent All experiments were performed on fresh large bowel tissue (colon) from ten individuals

obtained during surgery (Supplementary Table 1); four were infants (median age 6 months,

interquartile range (IQR) 4-11 months), of which three were surgically treated for Hirschsprung’s disease and one for anorectal malformation; six were adults (median known age 62 years, IQR 44-63 years), of which four were surgically treated for ulcerative colitis and two for colorectal cancer All

ACCEPTED MANUSCRIPT

tissue was derived from the surgical margin and based on visual assessment all tissues, except from one adult patient with ulcerative colitis, were presumed to be not severely affected Every tissue was processed within six hours after resection Because of the variation in donor age and diagnosis, direct comparisons were only made within individual donors, not between them The study was approved

by the medical ethical committee of our institute, the Academic Medical Center (University of Amsterdam) and in accordance with the Declaration of Helsinki

2.2 Tissue preparation

The tissue was rinsed by manually shaking 20 seconds in 20 ml sterile PBS in a 50 ml canonical centrifuge tube (Greiner, Sigma-Aldrich), followed by removal of the muscular layer using

scissors (Fig 1) To determine cell yield per cm2, the size of the mucosal tissue was measured after removal of the muscular layer The tissue was divided into pieces with equal surface area (0.5 cm2)

Figure 1 Removal of the muscular layer (bottom) from the mucosa (top)

2.3 Epithelial layer detachment

To remove the remaining mucus and detach the epithelial layer containing the IELs, the intestinal fragments were incubated for 20 minutes in a shaking water bath (approximately 100 strokes/min) at 37°C in 10 ml Iscove’s Modified Dulbecco’s Medium (IMDM; Lonza, Verviers, Belgium) containing 2 mM DTT, 5 mM EDTA, 5% fetal calf serum (FCS; Biological Industries, Kibbutz Beit Haemek, Israel), 100 U/ml Penicillin, and 100 μg/ml Streptomycin (Gibco, Life Technologies) in a

50 ml canonical centrifuge tube The suspension including the tissue was then vortexed (IKA; MS3 basic Lab Shaker) at maximum speed (3000 rpm for this model) for 15 seconds after which the cell suspension without tissue was passed through a 70 µm single-cell strainer (Falcon, Corning, USA), rinsed with 4°C PBS and centrifuged at 400 G at 4°C for 10 minutes A second incubation with the above described DTT/EDTA solution followed for the remaining tissue using the same procedure and

ACCEPTED MANUSCRIPT

resulting in a second single cell suspension The two single cell suspensions were pooled after washing them with 4°C PBS and stored in PBS on ice until density gradient centrifugation

2.4 Lamina propria disaggregation

2.4.1 Mechanical disaggregation using the Medimachine

The BD Medimachine System (BD Biosciences) was used as described before (Bunders et al., 2012) Briefly, it is an automated mechanical disaggregation system of solid human tissues without enzymes The Medimachine spins the tissue in the Medicon filter (BD Biosciences, San Jose, USA) to disaggregate tissue fragments into single cells After detachment of the epithelial layer, one tissue fragment at a time (0.5 cm2) was placed in the Medicon filter with 1 ml PBS The single-cell suspension was obtained by placing a 5 ml syringe on the out-port of the Medimachine and extracting the cell suspension This procedure was repeated until each fragment was dissolved completely Cells were passed through a 70 µm single-cell strainer, rinsed with 4°C PBS, and centrifuged at 400 G (4°C) for 10 minutes The obtained single cell suspension was stored in PBS on ice until density gradient isolation Medimachine isolation was always compared with Collagenase D isolation; half of the tissue of one donor was mechanically disaggregated with Medimachine while the other half was enzymatically digested with Collagenase D as described in the next paragraph (and

in Fig 2)

2.4.2 Enzymatic disaggregation using Collagenase D

After detachment of the epithelial layer, tissue fragments were placed in a 50 ml tube with

10 ml disaggregation solution, which contains IMDM with 1 mg/ml (0.15 U/mg) Collagenase D (Roche, Mannheim, Germany), 1% FCS, and 1000 U/ml DNAse type I (Roche) The fragments were incubated in the medium in a shaking water bath at 37°C for 30 minutes (approximately 100 strokes/min) The supernatant containing the cell suspension was collected and passed through a 70

µm single-cell strainer, rinsed with 4°C PBS, and centrifuged at 400 G at 4°C for 10 minutes The remaining fragments were incubated in fresh disaggregation solution containing Collagenase D for another 30 minutes after which the remaining cell suspension was filtered through a 70 µm single-cell strainer The filtered single single-cell suspension was rinsed with 4°C PBS, centrifuged at 400 G at 4°C for 10 minutes and stored in PBS on ice till density gradient isolation

2.4.3 Enzymatic digestion using Liberase TL

Analogous to the above described isolation technique of lamina propria with collagenase D, the tissue fragments (after detachment of the epithelial cells) were placed in a 50 ml canonical

ACCEPTED MANUSCRIPT

centrifuge tube with 10 ml disaggregation solution containing IMDM with 125 μg/ml (0.65 U/ml) Liberase TL (Roche, Mannheim, Germany), 1% FCS, and 1000 U/ml DNAse type I The fragments were incubated in the medium in a shaking water bath at 37°C for 30 minutes (approximately 100 strokes/min) The cell suspension was removed and passed through a 70 µm single-cell strainer, rinsed with 4°C PBS and centrifuged at 400 G at 4°C for 10 minutes in a 50 ml canonical centrifuge tube The remaining fragments were incubated in fresh enzymatic solution containing Liberase TL for another 30 minutes after which the remaining cell suspension was filtered with a 70 µm single-cell strainer The filtered single cell suspension was rinsed with 4°C PBS, centrifuged at 400 G at 4°C for

10 minutes in a 50 ml canonical centrifuge tube and stored in PBS on ice till density gradient isolation Isolation with Liberase TL was always compared with Collagenase D isolation; half of the tissue of one donor was digested with Liberase TL while the other half was digested with Collagenase

D as described in the previous paragraph (and in Fig 2)

2.5 Mononuclear leukocyte cell enrichment by density centrifugation

The single cell suspensions isolated from the epithelium and lamina propria were divided into

equal volumes (to be able to compare different gradients; Fig 2) and enriched for mononuclear

leukocytes by density gradient centrifugation using either standard Lymphoprep (Axis-Shield, Oslo, Norway) or a 60% Standard Isotonic Percoll solution (SIP; GE Healthcare, Uppsala, Sweden); the SIP solution was prepared by supplementing 100% Percoll with 10% 10X PBS after which it was diluted to 60% with 1X PBS The cell suspensions were pelleted and then resuspended in 10 ml IMDM after which they were layered on top of 4 ml Lymphoprep or 4 ml 60% SIP solution in a 15 ml canonical centrifuge tube The gradient tubes were centrifuged at 1000 G for 22 minutes with no break and at room temperature The interphases containing mononuclear leukocytes were collected by careful aspiration with sterile polystyrene pipettes (Falcon, Corning, USA) The single cell suspension was washed with 4°C PBS and centrifuged at 400 G for 10 minutes Cell yield and viability were determined by Trypan blue (Sigma-Aldrich) exclusion; samples (10 µl) were mixed with an equal amount of 4% Trypan blue, placed in a Bürker chamber (Blau, Wertheim, Germany) and only live cells were counted Concentrations of 10·106 cells/ml in 1X PBS were prepared for immune phenotyping and kept on ice

2.6 PBMC isolation from whole blood and treatment with enzymes

Whole blood (5 ml) from three healthy adult donors was collected in heparine vacuettes (Greiner) The blood was diluted 1:1 with IMDM and enriched for peripheral blood mononuclear cells (PBMC)

by density gradient centrifugation over 4 ml Lymphoprep (as described in section 2.5) After washing the PBMC twice (with 4°C PBS and centrifugation at 400 G for 10 minutes), each PBMC suspension

ACCEPTED MANUSCRIPT

was divided equally over three 15 ml canonical centrifuge tubes and treated with either Collagenase

D (section 2.4.2), Liberase TL (section 2.4.3), or left untreated (IMDM with 1% FCS, and 1000 U/ml DNAse type I) for 1 hour in a shaking water bath at 37°C The PBMC samples were washed twice more and concentrations of 10·106 cells/ml in 1X PBS were prepared for immune phenotyping and kept on ice

2.7 Cell staining and flow cytometry

Samples of 1·106 cells were placed in 96-well plates and stained immediately with a cocktail of monoclonal antibodies for 30 minutes at 4°C Cells were stained with previously optimized antibody-fluorochrome combinations to the following markers (all anti-human) to identify T cells, monocyte-derived, and dendritic cell sub populations: LIVE/DEAD Fixable Red (Invitrogen), CD45-FITC, CD3-FITC, CD3-AF700, CD27-APC-eFluor780, CD4-PerCP-Cy5.5, CD103-PerCP-eFluor710, CD28-PE, CD19-FITC, CD20-FITC, CD56-FITC, HLA-DR-APC-eFluor780, CD83-PE-Cy7, CD56-PE-Cy7 (all eBioscience), CD45-V500, CD3-CD45-V500, CD4-BUV737, CCR7-BUV395, CD69-BV421, CD14-BUV737, CD11c-BUV395, CD141-BV711, CD86-BV650 (all BD Horizon), CD4-BV570, CD8a-BV785, CD11b-BV785 (Biolegend), CCR7-PE, CD27-FITC, CD123-PE (all BD Pharmingen) After staining, the cells were washed with PBS and resuspended in 1X stabilizing fixative (BD Biosciences, San Jose, USA) Flow cytometric analyses were

performed on an LSR Fortessa (BD) within 24 hours after fixation

Figure 2 Workflow of protocol comparisons Tissues of

individual donors were divided in half and treated with different disaggregation protocols (Collagenase D, Liberase TL,

or enzyme-free Medimachine) Resulting cell suspension were divided in half and enriched for mononuclear leukocytes with density gradients (Lymphoprep of 60% Standard Isotonic Percoll (SIP)) Within donor comparisons of disaggregation method or density gradient were made (indicated in bold)

ACCEPTED MANUSCRIPT

2.8 Statistical analysis

The flow cytometry data were analyzed with FlowJo vX.0.7 software (TreeStar) Due to variability in the age and diagnosis of the donors, which can affect the frequency of cell populations

as well as cell surface molecule expression, comparisons between isolation and enrichment techniques were always made within individual donors Subsequently, in order to compare results between donors, differences in results per donor were expressed in individual fold changes; Lymphoprep was divided by 60% SIP; Collagenase D was divided by Medimachine, Liberase TL, or Untreated Statistical significance of differences was then assessed using one sample (compared to a value of ‘1’ indicating no change), paired T tests (Wilcoxon where appropriate) or ANOVAs The software package GraphPad Prism 7.02 (GraphPad Software, San Diego, California) was used for data

and statistical analyses Values of p<0.05 were considered significant Median frequencies ±

interquartile range (IQR) are given in figures

3 Results

3.1 Optimal gradient densities to enrich for intestinal mononuclear leukocytes

In order to enrich for mononuclear IELs and LPLs from the epithelium and lamina propria (respectively), half of the single cell suspensions obtained from individual donors were layered on Lymphoprep and the other half on SIP (60%) as described in the methods section Comparative analyses of cell yield, cell viability, and monocyte-to-lymphocyte ratio between these two gradient solutions were made

There were no differences detected regarding total yield of viable IELs as determined by

Trypan blue staining (Fig 3A) or mononuclear leukocytes as determined by flow cytometry

(LIVE/DEAD-CD45+; Fig 3B) (one-sample Wilcoxon T tests against ‘1’; p=0.75; p=0.25) per cm2

epithelium (non-inflammatory pediatric tissues were used; for data on individual donors see

Supplementary Fig 1), nor did the density gradient solution alter the 3:2

myeloid-to-lymphoid-derived cell ratio of this population (as determined by flow cytometric forward and sideward scatter

(Fig 3C), and confirmed with CD14+ and CD3+ stainings (not shown))

As for IEL, enrichment with either gradient solution resulted in a similar yield of viable LPLs as

determined by Trypan blue staining (Fig 4A), however a trend towards increased mononuclear

leukocytes as determined by flow cytometry (LIVE/DEAD-CD45+, Fig 4B) per cm2 lamina propria tissue in SIP versus Lymphoprep gradient isolation was observed (one-sample Wilcoxon T tests

against ‘1’; p=0.59; p=0.06) (non-inflammatory pediatric tissues were used; for data on individual

donors see Supplementary Fig 2) Myeloid-to-lymphoid-derived cell ratio, which was 2:3 did not

ACCEPTED MANUSCRIPT

differ between gradient isolation techniques (Fig 4C) The use of Lymphoprep further resulted in the

clumping of non-immune cells (based on visual assessment at 40X magnification of cell aggregate)

below the interphase within the gradient solution (Fig 4D), thereby increasing the chances of

contaminating the mononuclear cell fraction during aspiration

Figure 3 Cell yield after density gradient enrichment of the epithelium using Lymphoprep (LP) or 60% Standard Isotonic

Percoll (SIP) (fold change N=3; all pediatric donors) Both gradients yield similar numbers of (A) viable cells as determined

by Trypan blue staining, as well as (B) similar numbers of mononuclear leukocytes per cm2 tissue (one-sample Wilcoxon T

tests; p=0.75; p=0.25) and (C) a similar 3:2 myeloid-to-lymphoid derived cell ratio, both determined by flow cytometry Fold

change equals LP divided by SIP; 1 (black line) equals no change Median frequencies (± IQR) are given

Figure 4 Cell yield after density gradient enrichment of lamina propria (LPL) using Lymphoprep (LP) and 60% Standard

Isotonic Percoll (SIP) (fold change N=5; all pediatric donors) (A) Similar numbers of viable cells as determined by Trypan blue staining, as well as (B) similar numbers of mononuclear leukocytes (LIVE/DEAD-CD45+ as determined by flow

cytometry) (one-sample Wilcoxon T tests; p=0.59; p=0.06) Furthermore, (C) both LP and SIP resulted in a similar 2:3

myeloid-to-lymphoid derived cell ratio (D) LP usage resulted in the clumping of non-immune cells below the interphase,

increasing the risk of contaminating the LPL during aspiration Fold change equals LP divided by SIP; 1 (black line) equals no

change Median frequencies (± IQR) are given