SEARCH

SEARCH BY CITATION

Keywords:

  • microglia;
  • variation;
  • cell differentiation;
  • immunology;
  • surface antigens;
  • flow cytometry

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

In accordance with a high degree of spatial organization in the central nervous system (CNS), most CNS diseases display a regional distribution. Although microglia have been established as key players in various CNS diseases, it is not yet clear whether microglia display region-specific properties. Therefore, this study aimed to evaluate the existence of distinct microglia phenotypes in various regions of the healthy, adult mouse CNS. Using ex vivo flow cytometric analysis surface expression of CD11b, CD40, CD45, CD80, CD86, F4/80, TREM-2b, MHCII, CXCR3, CCR9, and CCR7 were analyzed. Most of these immunoregulatory markers were found on microglia and showed significant region-specific differences in expression levels. These findings considerably corroborate the existence of immunological diversity among microglia in the healthy, unchallenged CNS of adult mice. © 2008 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Microglia are generally considered as the resident immunocompetent cells of the CNS. Sharing their myeloid origin, microglia show resemblance to tissue macrophages, and dendritic cells. In the healthy CNS, microglia continually extends and retracts their processes, thereby surveilling their environment. Upon injury, microglia rapidly protrude their processes to the site of injury and subsequently transform into amoeboid cells (Davalos et al.,2005; Nimmerjahn et al.,2005). According to this rapid injury response and their capacity to migrate, proliferate, phagocytose, present antigen and produce pro- and anti-inflammatory mediators (Aloisi,2001; Hanisch and Kettenmann,2007; Kim and de Vellis,2005; Nakajima and Kohsaka,2004; van Rossum and Hanisch,2004), microglia activity is conceivably at the base of the immune response to CNS injury. A key role for microglia in innate and adaptive immune responses to CNS injury is strongly corroborated by recent in vivo studies impeding in microglia activity (Boillée et al.,2006; Bye et al.,2007; Cardona et al.,2006; El Khoury et al.,2007; Lalancette-Hebert et al.,2007; Neumann and Takahashi,2007; Streit,2006; Turrin and Rivest,2006; van Rossum and Hanisch,2004).

Although general knowledge of microglia phenotype and function is rapidly increasing, it is today largely unknown whether microglia are a homogenous population or whether the CNS contains different microglia phenotypes with distinct functions (Schwartz et al.,2006). Thus far, few immunohistochemical-based studies occasional report regional differences in microglia density, morphology, proliferation, and expression of markers (Ehninger and Kempermann,2003; Goings et al.,2006; Lawson et al.,1990; Mittelbronn et al.,2001; Phillips et al.,1999; Sheffield and Berman,1998; Wu et al.,1997). However, a direct comparison of microglia from various brain regions was yet not published. Since most CNS diseases manifest in specific regions rather than in the whole CNS, profound knowledge on microglia diversity might be instrumental to understand the role of microglia in brain disease.

This study aimed to evaluate the existence of regional diversity of microglia phenotypes. We present here the first immunological comparison of microglia from various regions of the healthy, adult mouse CNS by ex vivo flow cytometric analysis. This study demonstrates region-specific differences in microglia expression levels of several immunoregulatory proteins, strongly corroborating the existence of immunological diversity among microglia in the CNS.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Animals

Eleven- to twelve-week-old male C57BL/6 mice (Harlan UK, Bicester, England) were housed and handled according to the guidelines of the laboratory animal facility of the University of Groningen, as approved by the local animal experimental committee.

Microglia Isolation

Mice were anaesthetized with isoflurane, nitrous oxide, and oxygen. To clear the intravascular compartment of blood cells, mice underwent transcardiac perfusion with ice-cold 0.9% saline until colorless fluid appeared from the inferior caval vein. Spinal cords were extracted by flushing the spinal canal with phosphate-buffered saline (PBS, Gibco Breda, The Netherlands 14190-094). Extracted spinal cords and brains were stored in ice-cold Hanks' balanced salt solution (HBSS, Gibco 14170-088), containing 15 mM N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid (Gibco 15630-056) and 0.5% glucose (Sigma Zwijndrecht, The Netherlands G8769). Brains were dissected into striatum, hippocampus, cerebellum, and cerebral cortex in ice-cold HBSS. Regions from five mice were pooled per experiment to increase microglia number. CNS tissue was mechanically grinded in a tissue homogenizer (glass Potter, Braun Melsungen, Germany) in 1–4 mL ice-cold HBSS, followed by trituration using fire-polished Pasteur pipettes of decreasing diameter (VWR international Roden, The Netherlands 612-1701). The cell suspension was filtered through a 70-μm cell strainer (BD Biosciences Alphen aan de Rijn, The Netherlands 352350). Finally, cells were pelleted at 1,000g for 10 min at 4°C and resuspended in ice-cold 75% Percoll (GE Healthcare Uppsala, Sweden 17-0891; 9 volumes of Percoll were mixed with 1 volume of 10× PBS to yield a stock isotonic Percoll solution and Percoll densities were obtained by dilution with 1× PBS). Units of 3.3 mL were divided among 12 mL polystyrene tubes (Greiner Bio-One Alphen aan de Rijn, The Netherlands 163160) and gently overlayered with 5 mL ice-cold 25% Percoll and 3 mL ice-cold PBS. The density gradient was centrifuged in a swinging bucket rotor (Falcon 6/300) at 800g (slow acceleration and no brake) for 25 min at 4°C. After centrifugation, a thick myelin-containing layer at the 0/25 interface was removed with a Pasteur pipette and the cells at the 25/75 interface collected with a fresh Pasteur pipette. To pellet these cells, the cell-Percoll suspension was diluted at least threefold with ice-cold PBS and centrifuged at 1,000g for 10 min at 4°C.

Flow Cytometry

To minimize aspecific binding, the cells were preincubated for 15 min at 4°C in 100 μL PBS containing 10% normal rat serum (Invitrogen 10710C) and anti-CD16/32 monoclonal antibodies (mAbs, eBioscience San Diego 14-0161, 1:100). To determine cell number, 5 μL of the cell suspension was stained 1:1 with 0.2% trypan blue (Sigma T8154) and counted under the microscope in a Bürker-Türk 0.1-mm counting chamber (Optik Labor, Germany). Approximately 2 × 104 cells were incubated for 15 min at 4°C in 100 μL PBS with mAbs (all from eBioscience except CXCR3 and TREM-2b mAbs) against CD11b (12-0112, 0.13 μg), CD40 (11-0402, 1.00 μg), CD45 (11-0451, 0.25 μg), CD80 (11-0801, 0.25 μg), CD86 (11-0862, 0.13 μg), F4/80 (11-4801, 0.50 μg), MHCII (11-5321, 0.13 μg), CCR7 (12-1971, 0.50 μg), CCR9 (11-1991, 0.25 μg), CXCR3 (R&D Minneapolis FAB1685P, 0.25 μg), TREM-2b (R&D MAB17291, 0.13 μg), or concentration-matched isotype controls (11-4301, 11-4321, 11-4331, 11-4724, 11-4888, 12-4321, and 12-4331). For TREM-2b staining, cells were subsequently incubated for 15 min with fluorescein isothiocyanate-labeled goat anti-rat Abs (Southern Biotech Birmingham, Alabama 3010-02, 0.50 μg). Subsequently, cells were rinsed in 3 mL PBS, pelleted at 1,000g for 10 min at 4°C, and resuspended in 300 μL PBS. To discriminate between living and dead cells, cells were incubated with DRAQ5 (Biostatus Limited Leicestershire, UK 1:1,500) 10 min prior to flow cytometry. Total time from sacrificing the animals to flow cytometry was 2–2.5 h. Cell size, granularity, and fluorescence intensity were analyzed using a 488-nm laser FACS Calibur (Becton Dickinson Franklin Lakes) with 530/30, 585/42, and 670 nm/LP filters. A minimum of 5 × 103 cells were analyzed using Cell Quest software (Becton Dickinson) and WinMDI 2.8 for Windows. Microglia purity was defined as the percentage of all living cells that showed CD11b+/CD45low expression.

Statistical Analysis

For all stainings, the geometric mean fluorescence intensity (MFI) was calculated with WinMDI. As indicator of protein expression levels, MFIs of isotype stainings were subtracted from the MFIs of protein stainings. Differences in the MFIs for the various regions were tested for significance using paired-samples t-tests in SPSS (SPSS 12.0.2 for Windows, SPSS Inc., Chicago). P values of <0.05 were considered significant.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Analysis of Microglia Surface Marker Expression

An accurate way to determine the expression of surface markers in microglia is by ex vivo flow cytometric analysis. For the isolation of microglia, we introduced an optimized protocol with respect to microglia purity and a minimal disturbance of in vivo cell properties (De Haas et al.,2007). Upon flow cytometry, microglia were identified based on cell size and granularity (Fig. 1A, R1) and viability by DRAQ5 staining (Fig. 1B, R2). The identification of microglia by these parameters was verified using their characteristic CD11b+/CD45 low-expression profile, revealing a microglia purity of 98.61% ± 0.92% (mean ± SD, Fig. 1C).

thumbnail image

Figure 1. Flow cytometric selection based on cell size, granularity, and viability identified microglia with a purity of 98.61%. (A) Microglia appeared homogeneous in cell size and granularity (R1). (B) DRAQ5 staining, a DNA-binding dye that is actively taken up by living cells, distinguished the living cells (R2) from dead cells and debris. (C) Combined selection based on cell size, granularity, and viability (R1 + R2) resulted in a microglia purity of 98.61% ± 0.92% (mean ± SD), as identified by their characteristic CD11b+/CD45 low expression.

Download figure to PowerPoint

Microglia Show Regional Diversity in Immunoregulatory Protein Expression

To test for regional diversity of protein expression, microglia were isolated from the striatum, hippocampus, spinal cord, cerebellum, and cerebral cortex of healthy, young adult mice, and ex vivo analyzed by flow cytometry. The proteins chosen for analysis have all been described to be expressed on in vivo microglia under control and/or inflammatory conditions. We determined a clear expression of the proteins CD11b, CD40, CD45, CD80, CD86, F4/80, TREM-2b, CXCR3, and CCR9 (see Fig. 2), whereas no expression was found for either MHCII or CCR7 (data not shown). Interestingly, CD11b, CD40, CD45, CD80, CD86, F4/80, TREM-2b, CXCR3, and CCR9 were generally observed on microglia from all brain regions. Thus, none of the expressed markers was present or absent in a brain region-dependent manner. Interestingly, the protein expression levels were markedly different on microglia derived from different brain regions (for each marker Fig. 2 shows two exemplified brain regions; please refer to Supplementary Fig. 1 for an overview of all data).

thumbnail image

Figure 2. For all regions investigated, a clear expression was found of the proteins CD11b, CD40, CD45, CD80, CD86, F4/80, TREM-2b, CXCR3, and CCR9. To illustrate the distribution of the protein expression levels, representative fluorescence intensities are shown for microglia from various regions. Protein expression levels are represented by black lines and background fluorescence with isotype controls by gray histograms. Events are normalized.

Download figure to PowerPoint

To compare the protein expression between the various regions, we used the MFI as indicator of the average protein expression level per cell (Fig. 3A). Most of the investigated proteins displayed highly reproducible differences in MFI between regions. For example, we show in detail the results for CD11b (as a protein with high expression), CD40 (as a protein with moderate expression), and CXCR3 [as an example for a G-protein coupled receptor that is important for microglia function (Rappert et al.,2004)]. The MFI for CD11b on microglia from cerebral cortex was repeatedly lower than from spinal cord (n = 4) (Fig. 3B), the MFI for CD40 on microglia from cerebral cortex lower than from cerebellum (n = 4) (Fig. 3C) and the MFI for CXCR3 on microglia from hippocampus lower than from cerebral cortex (n = 7) (Fig. 3D). The points connected by lines in Figs. 3B–D are the data pairs derived from one experiment. Note that most of the lines have similar slopes revealing the high degree of reproducibility of the data.

thumbnail image

Figure 3. Differences in mean fluorescence intensities between regions were reproducible throughout experiments, as illustrated for CD11b, CD40, and CXCR3. (A) To compare the protein expression levels from the various regions, the mean fluorescence intensity (MFI) was used as an indicator of the average protein expression level per cell. Protein staining is represented by the black line and the background fluorescence with isotype control staining by the gray histogram. Differences in MFI between regions were highly reproducible, as illustrated for (B) CD11b (cerebral cortex vs. spinal cord, n = 4), (C) CD40 (cerebral cortex vs. cerebellum, n = 5), and (D) CXCR3 (hippocampus vs. cerebral cortex, n = 7). The points connected by lines in (B)–(D) are data pairs derived from one experiment. Note that most of the lines have similar slopes revealing the high degree of reproducibility of the data.

Download figure to PowerPoint

To quantify the differences in regional protein expression, for each protein the quotients of the regional MFIs were calculated. The maximal differences per protein, i.e., the quotient from the highest and lowest regional MFI, varied between 1.49 times for CD11b (spinal cord vs. cerebral cortex) and 3.20 times for CD40 (cerebellum vs. cerebral cortex) (Table 1). Paired-samples t-tests were subsequently used for significance testing of these regional differences. It showed that the maximal differences per protein were highly consistent according to the levels of significance (Table 1).

Table 1. Maximal Regional Differences in Fluorescence Levels per Protein
 High MFIaLow MFIQuotientb95% CIPcn
  • a

    Mean fluorescence intensity.

  • b

    Division of high MFI by low MFI.

  • c

    Paired-samples t-test P value.

CD11bSpinal cordCerebral cortex1.491.10–1.880.0284
CD40CerebellumCerebral cortex3.202.29–4.120.0035
CD45Spinal cordHippocampus1.621.40–1.840.0034
CD80Cerebral cortexHippocampus1.390.97–1.800.0584
CD86Spinal cordCerebellum1.861
F4/80HippocampusCerebellum1.841
TREM-2bCerebellumCerebral cortex1.521.00–2.040.0504
CXCR3CerebellumHippocampus2.260.71–3.810.0733
CCR9Spinal cordHippocampus2.261.40–3.120.0342

Since we observed considerably more significant differences and trends in regional protein expression than predicted on basis of pure chance it is unlikely that these results were merely the consequence of multiple testing. For example, six out of 30 comparisons made for CD11b, CD40, and CXCR3 showed significant (P < 0.05) differences in regional protein expression (dark gray cells Table 2) and another six showed a trend (0.05 < P < 0.10) (light gray cells Table 2).

Table 2. Multiple Regional Differences as Detailed Out for CD11b, CD40, and CXCR3a
CD11bStriatumHippocampusSpinal CordCerebellumCerebral cortex
  • a

    Cells contain the MFI-quotient (horizontal regions divided by vertical regions), paired-samples t-test P-value (P < 0.05 in dark gray, 0.05 < P < 0.10 in light gray), and n in brackets.

Striatumx0.851.071.100.78
  0.131 (3)0.045 (3)0.365 (2)0.086 (3)
Hippocampus1.18x1.301.150.87
 0.131 (3) 0.026 (4)0.273 (3)0.103 (4)
Spinal cord0.940.77x0.890.67
 0.045 (3)0.026 (4) 0.517 (3)0.028 (4)
Cerebellum0.910.881.10x0.73
 0.365 (2)0.273 (3)0.517 (3) 0.122 (3)
Cerebral cortex1.301.151.491.38x
 0.086 (3)0.103 (4)0.028 (4)0.122 (3) 
CD40StriatumHippocampusSpinal cordCerebellumCerebral cortex
Striatumx0.820.972.100.68
  0.290 (5)0.895 (4)0.062 (4)0.051 (5)
Hippocampus1.21x1.272.650.68
 0.290 (5) 0.063 (4)0.082 (4)0.264 (6)
Spinal cord1.030.79x2.650.70
 0.895 (4)0.063 (4) 0.158 (3)0.445 (4)
Cerebellum0.470.380.38x0.31
 0.062 (4)0.082 (4)0.158 (3) 0.003 (5)
Cerebral cortex1.471.481.403.20x
 0.051 (5)0.264 (6)0.445 (4)0.003 (5)
CXCR3StriatumHippocampusSpinal cordCerebellumCerebral cortex
Striatumx0.811.181.841.99
  0.299 (3)0.824 (2)0.173 (3)0.114 (3)
Hippocampus1.23x1.512.262.13
 0.299 (3) 0.496 (2)0.073 (3)0.003 (7)
Spinal cord0.850.66x1.722.06
 0.824 (2)0.496 (2) 0.043 (2)0.274 (2)
Cerebellum0.540.440.58x1.08
 0.173 (3)0.073 (3)0.043 (2) 0.746 (3)
Cerebral cortex0.500.470.490.93x
 0.114 (3)0.003 (7)0.274 (2)0.746 (3) 

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Most, if not all CNS functions depend on cellular activity in distinct areas of the brain and are therefore region-specific. This is most likely related to the presence of various neuronal phenotypes with specific functions and also reflected by the fact that most CNS diseases display a regional distribution. Although microglia is considered as key players in various CNS diseases, surprisingly little is known about possible region-specificity of microglia. To address the question whether microglia from various CNS regions display different phenotypes, we analyzed microglia from various regions for the expression of 11 immunoregulatory proteins by ex vivo flow cytometry. The 11 markers used here have been chosen because all of them are known to be expressed on in vivo microglia under control and/or inflammatory conditions.

Throughout striatum, hippocampus, spinal cord, cerebellum, and cerebral cortex, we found a clear expression of CD11b, CD40, CD45, CD80, CD86, F4/80, TREM-2b, CXCR3 and CCR9, whereas no expression was found for either MHCII or CCR7. These findings are, to large extent, comparable with the results described in studies that used ex vivo flow cytometric analysis for microglia isolated from whole mouse brain (Aloisi et al.,2000; Carson et al.,1998; Fischer and Reichmann,2001; Havenith et al.,1998; Mack et al.,2003; Zhang et al.,2002). In addition to these studies, we quantified and compared the expression levels of these immunoregulatory proteins on microglia isolated from different CNS regions. Strikingly, most of the investigated proteins showed region-specific expression levels. The maximal differences for all investigated proteins varied between 1.49 times for CD11b (spinal cord vs. cerebral cortex) and 3.20 times for CD40 (cerebellum vs. cerebral cortex) and were highly consistent and significant. There is no post hoc analysis for multiple paired t-tests. Given a P-value of 0.05, one would expect to find one significant difference out of 20 tests based on pure chance. Since we observed considerably more significant differences and given the fast and cold isolation procedure and the resulting average microglia purity of 98.61% used in this study, it is very likely that the reported differences reliably reflect the expression profile of in vivo microglia in the CNS.

Interestingly, we could not observe a pattern in the expression profiles of all proteins. Thus, there was no region in which protein expression levels were generally high or low. For example, although microglia from hippocampus displayed the lowest expression for four of the proteins (CD45, CD80, CXCR3, and CCR9), they showed highest expression for F4/80. The only microglia that displayed no maximum or minimum expression level was derived from striatum. These data indicate that the immunological profile (phenotype) of microglia might be shaped by region-specific microenvironmental signals. Microglia is known to express numerous classes of receptors (Hanisch and Kettenmann,2007). Next to receptor families that are generally found on myeloid cells microglia express numerous receptors that recognize neuronal proteins and signaling molecules, making neurons important cells to control microglia function (Biber et al.,2007). Microglia are known to sense electric neuronal activity (Neumann,2001) most likely via neurotransmitter receptors of which microglia express numerous subtypes (Farber et al.,2005). The stimulation of these neurotransmitter receptors affects various microglial functions (Pocock and Kettenmann,2007). Since neurotransmitters are found in the brain in a spatially highly regulated manner, these neuronal molecules might contribute to the microenvironmental signals that lead to the development of microglia phentotypes. Moreover, microglia function is controlled by neuronal membrane-bound immunoglobulin-like molecules like CD200 or CD22 (Hoek et al.,2000; Mott et al.,2004) or the yet to be identified ligand for the microglia receptor TREM-2 (Biber et al.,2007; Neumann and Takahashi,2007). Whether or not CD200 or CD22 are differentially expressed in different brain regions is not well understood at the moment. However, the data described here and findings of others (Schmid et al.,2002) suggest that TREM-2 is expressed on microglia in a brain-region dependent manner. Another factor that might influence microglia might be differences in the leakiness of the blood–brain barrier as it is known, for example, for the periventricular organs (Pedersen et al.,1997; Schmid et al.,2002). Moreover, microglia phenotypes might be shaped by the white matter index since it was recently shown that only microglia in white matter regions express Tim-3 (Anderson et al.,2007). Thus various factors in the microenvironment might contribute to the development of different microglia phenotypes in different brain regions.

This work was aimed to address the question whether the healthy CNS would host different microglia populations. The data provided here strongly corroborate this assumption and subsequently lead to the question whether region-specificity of microglia in the healthy CNS might influence the inclination and the way microglia responds to injury. Interestingly, most of the proteins investigated here have an established or expected role in the initiation or perpetuation of the immune response to CNS injury. For example, deficiency of CD45 in a mouse model of Alzheimer's disease was associated with enhanced microglia activity (Tan et al.,2000). In mouse models of multiple sclerosis blockade of CD80 and deficiency of CD40 on microglia prevented the development of autoimmune encephalomyelitis (Becher et al.,2001; Miller et al.,1995), whereas overexpression of CD86 in mice caused autoimmune demyelinating disease (Zehntner et al.,2003). Different expression levels of CD80 or CD86 might influence the antigen-presenting capacities of microglia and subsequently affect cell–cell interaction between microglia and T cells (Greter et al.,2005; Heppner et al.,2005; Korn et al.,2007; McMahon et al.,2005, 2006).

Taken together, numerous recent findings have considerably changed our view of microglia. Today, it is clear that these cells have the capacity to specifically and heterogeneously respond to changes in the healthy and diseased CNS (Hanisch and Kettenmann,2007). The data presented here add up the concept of different microglia populations in the CNS. Although it is yet unknown whether distinct microglia are involved in CNS antigen presentation, T cell interaction, or region-specific CNS diseases, our results emphasize this possibility and may form an incentive for future studies concerning diversity in microglia phenotype and function, both in the healthy CNS and in CNS disease.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

The authors are grateful to Prof. U.K. Hanisch for critical reading of the manuscript and valuable comments.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

This article contains supplementary material available via the Internet at http://www.interscience.wiley.com/jpages/0894-1491/suppmat

FilenameFormatSizeDescription
glia20663-SupFig1.eps939K Supplementary Figure 1: For all regions investigated, a clear expression was found of the proteins CD11b, CD40, CD45, CD80, CD86, F4/80, TREM-2b, CXCR3 and CCR9.To illustrate the distribution of the protein expression levels, representative fluorescence intensities are shown for microglia from various regions. Protein expression levels are represented by colored lines and background fluorescence with isotype controls by grey histograms. Events are normalized.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.