Metabolomics

, 12:92

Metabolic fingerprints of human primary endothelial and fibroblast cells

  • Polona Žigon
  • Katjuša Mrak-Poljšak
  • Katja Lakota
  • Matic Terčelj
  • Saša Čučnik
  • Matija Tomsic
  • Snezna Sodin-Semrl
Open AccessOriginal Article

DOI: 10.1007/s11306-016-1024-7

Cite this article as:
Žigon, P., Mrak-Poljšak, K., Lakota, K. et al. Metabolomics (2016) 12: 92. doi:10.1007/s11306-016-1024-7

Abstract

Introduction

Human primary cells originating from different locations within the body could differ greatly in their metabolic phenotypes, influencing both how they act during physiological/pathological processes and how susceptible/resistant they are to a variety of disease risk factors. A novel way to monitor cellular metabolism is through cell energetics assays, so we explored this approach with human primary cell types, as models of sclerotic disorders.

Objectives

In order to better understand pathophysiological processes at the cellular level, our goals were to measure metabolic pathway activities of endothelial cells and fibroblasts, and determine their metabolic phenotype profiles.

Methods

Biolog Phenotype MicroArray™ technology was used for the first time to characterize metabolic phenotypes of diverse primary cells. These colorimetric assays enable detection of utilization of 367 specific biochemical substrates by human endothelial cells from the coronary artery (HCAEC), umbilical vein (HUVEC) and normal, healthy lung fibroblasts (NHLF).

Results

Adenosine, inosine, d-mannose and dextrin were strongly utilized by all three cell types, comparable to glucose. Substrates metabolized solely by HCAEC were mannan, pectin, gelatin and prevalently tricarballylic acid. HUVEC did not show any uniquely metabolized substrates whereas NHLF exhibited strong utilization of sugars and carboxylic acids along with amino acids and peptides.

Conclusion

Taken together, we show for the first time that this simple energetics assay platform enables metabolic characterization of primary cells and that each of the three human cell types examined gives a unique and distinguishable profile.

Keywords

Cellular metabolism Phenomics Phenotype MicroArrays OmniLog Human primary cells 

1 Introduction

Primary cells cultured from ex vivo isolated tissues originating from different spatial sources of the same organism may show specific phenotypes, including distinct fingerprints of cellular metabolism (Pearson 2007). One way to monitor cellular metabolism is through multiplexed cell energetics assays. A new approach explored in this study was to use Biolog Phenotype MicroArray™ technology to characterize cells for metabolically and bioenergetically-linked phenotypes (Bochner et al. 2011; Putluri et al. 2011). By simultaneously measuring utilization of hundreds of substrates, this assay system can reveal unique and insightful information on metabolic pathway activities and cellular responses to nutrients, hormones, cytokines, ions, and anti-cancer agents.

Systemic sclerosis (SSc) or scleroderma is a chronic, multiorgan autoimmune disease complicated by early vasculopathy and heart involvement (Boueiz et al. 2010; Hettema et al. 2008; Meune et al. 2010; Ngian et al. 2011) as well as progressive fibrosis of the skin and internal organs, including lungs (Kuwana et al. 2004; Reiseter et al. 2015). Thus, endothelial cells and fibroblasts are among the earliest and most affected cell types in SSc. The endothelium is heterogeneous depending on its vascular bed and tissue/embryological origin. Different endothelial cell phenotypes exhibit differential responses to changes in their environment and signaling (Rosenberg and Aird 1999). Thus, human endothelial cells derived from the umbilical vein (HUVEC) have a specific and intrinsic expression pattern of inflammatory molecules as compared to human coronary artery endothelial cells (HCAEC). Cardiovascular cells that contribute directly to atherosclerosis and cardiac dysfunction are known to exhibit metabolic flexibility, characterized by the ability to switch from generating ATP primarily through oxidative phosphorylation to using glycolysis as the predominant energy source, as well as to shift from one fuel source to another (Vallerie and Bornfeldt 2015). This flexibility occurs in endothelial cells, myeloid cells, and cardiomyocytes during normal development and physiology, and is thought to have evolved to protect cells with heightened energy demand from the increased oxidative stress that can be a result of elevated rates of oxidative phosphorylation (Galvan-Pena and O’Neill 2014). The cells shunt glucose to oxidative side branches of glycolysis (Eelen et al. 2015), to provide energy more rapidly (Galvan-Pena and O’Neill 2014), or to use the most abundant fuel available (Kolwicz et al. 2013). Studies support the concept that metabolic flexibility confers the advantage of ensuring ATP supplies for continual cardiac contraction under a variety of physiological conditions (Goodwin and Taegtmeyer 2000; Kaijser and Berglund 1992; Schonekess 1997; Wentz et al. 2010).

Cardiac metabolism also undergoes a reprogramming in response to pathological hypertrophy, characterized by increased reliance on glucose metabolism and decreased fatty acid oxidation (Kolwicz et al. 2013). The question then arises whether metabolic flexibility and dysfunction in vascular and cardiac cells themselves could contribute to cardiovascular pathologies. Since patients with SSc have an elevated risk for vasculopathy, our aim was to determine what kind of metabolism HCAEC exhibit in comparison to a standardly used, however less appropriate, HUVEC model.

In the past, we have shown that HCAEC have a higher responsiveness and susceptibility to cytokines, and can as such, represent an excellent gene and protein expression model for evaluating the effects of vascular stress (Lakota et al. 2013, 2007, 2009). Therefore, it is of interest whether HCAEC have a different metabolic phenotype at baseline levels than HUVEC.

SSc is also a complex chronic connective tissue disease characterized by progressive fibrosis (Laar and Varga 2015) which can lead to irreversible damage of external and internal organs, such as the skin and lungs, respectively. The major cells implicated in fibrosis are fibroblasts, with their main physiological function in wound healing and tissue remodeling. The fibrotic component is dominant in SSc, as compared to other autoimmune diseases and determines its prognosis and therapeutic refractoriness. Studies in cultured SSc skin fibroblasts have facilitated the identification of potential pathways involved in their profibrotic phenotype. Profibrotic fibroblasts characterized by abnormal growth and extracellular matrix synthesis may differentiate or expand from normal resident fibroblasts and multiple factors, including signaling pathways appear to be involved in the development and/or persistence of the SSc fibroblast phenotype (Usategui et al. 2011). We postulate that fibroblasts, when activated/stimulated, under stress microenvironment conditions, could function much like cancer cells, with unbridled proliferation, due to a redirection of their metabolism by the reverse Warburg effect (Pavlides et al. 2010). In cancer, fibroblasts/stromal cells can convert to myofibroblasts, which (using glycolysis) can produce large amounts of lactate and pyruvate, thus feeding the faster growing cancer cells (Pavlides et al. 2009). It is interesting that in SSc, where fibroblasts are also highly activated and responsible for abnormal extracellular matrix accumulation (Usategui et al. 2011), inhibition of autophagy and aerobic glycolysis was suggested as a primary strategy to reverse fibrosis (Castello-Cros et al. 2011). Thus, studying metabolic utilization of substrates is relevant in disease development, as well as for providing innovative strategies to combat disease activity.

No data to our knowledge is currently available on the metabolic activity of human primary cell types using Biolog Phenotype MicroArray™ technology, and specifically fibroblast or endothelial cells utilizing multiple substrates. So, we aimed to measure metabolic activity of primary human endothelial and fibroblast cells on multiple carbon and energy sources, as well as determine their specific metabolic fingerprints using the Biolog Phenotype MicroArray™ and OmniLog assay system.

2 Materials and methods

2.1 Cell culture

HCAEC, HUVEC and NHLF (normal human lung fibroblasts) are all primary cell types generated ex vivo from isolated tissue (Lonza, Inc. Basel, Switzerland). Cell characteristics as reported by the manufacturer are summarized in Supplementary Table 1. The cells were plated onto six well plates or 75  cm2 flasks (TPP, Trasadigen, Switzerland) at 37 °C in a humidified atmosphere with 5 % CO2. HCAEC were grown in EGM-2 M medium containing 5 % fetal bovine serum (FBS); for HUVEC we used EBM-2 medium containing 2 % fetal bovine serum (both from Lonza Inc., Basel, Switzerland). NHLF were cultivated in FGM-2 medium containing 2 % FBS. All media components were from Lonza Inc. (Basel, Switzerland). Prior to the experiments, cells were incubated in serum-free medium for 30 min.

2.2 Phenotype MicroArray and OmniLog optimization experiments

HCAEC and NHLF specifically, were used for optimization experiments. Optimal cell medium, reagent dye and seeding cell densities were tested. Two different cell media were also tested, specifically IF-M1 (with amino acids) and IF-M2 (without amino acids) with two different Redox Dye Mixes MA and MB using initially three different seeding densities. All Phenotype MicroArray assay components were from Biolog Inc. (Hayward, CA USA).

Specifically, cells were seeded into 96 well plates (TPP, Trasadigen, Switzerland) at seeding densities of 20.000, 10.000, 5.000, 2.500, 1.250 and 625 cells/well in a compete medium MC-0 (50 µL/well). The MC-0 medium was prepared using either IF-M1 or IF-M2 medium with added Pen/Strep, l-glutamine (0.3 mM) and dialyzed FBS (at final concentration 5 %) (all from Lonza Inc., Basel, Switzerland). Cells were then incubated for 1 h at 37 °C under 5 % CO2, before adding Biolog Redox Dye Mix MA (10 μL) or Biolog Redox Dye Mix MB (10 µL) and measuring tetrazolium reduction, resulting in formation of a purple colour.

2.3 Tetrazolium colorimetric measurements

All cell types (HCAEC, HUVEC and NHLF) for experiments on Biolog Phenotype MicroArray plates PM-M1 to PM-M4 (Biolog Inc., Hayward, CA, USA) were used between passages 4 and 5. A total of 367 different carbon and energy substrates were tested, as shown in Table 1. Cells were initially grown in 75 cm2 culture flasks to confluency, before they were resuspended at a density of 200.000 cells/ml in Biolog IF-M1 medium with Pen/Strep, l-glutamine and dialyzed-FBS. Cells were then washed 2× with PBS and detached with trypsin at 37 °C for 2 min. Trypsinization neutralizing solution was added and cells were centrifuged at 350×g for 5 min at room temperature. MC-0 assay medium was added and cell count determined on Countess cells counter (Thermo Fisher Scientific, USA). Additional MC-0 was added to achieve a cell density of 10.000 cells/50 µL, which was seeded into the plates and incubated at 37 °C under 5 % CO2 for 18 h. Following the incubation, 10 µL Redox Dye Mix MB/well was added, the plate covered with sealing tape to prevent CO2 loss and incubated at 37 °C in the OmniLog (Biolog Inc., Hayward, CA, USA) for 24 h to kinetically measure tetrazolium reduction. OmniLog measures the intensity of purple colour formation using a CCD camera to record digital images every 15 min. All experiments on PM-M1 to PM-M4 plates were done in triplicate (48 plates total).
Table 1

PM-M1 to PM-M4 substrate templates

 

1

2

3

4

5

6

7

8

9

10

11

12

PM-M1 MicroPlate™—carbon and energy sources

 A

Negative control

Negative control

Negative control

α-Cyclodextrin

Dextrin

Glycogen

Maltitol

Maltotriose

d-Maltose

d-Trehalose

d-Cellobiose

β-Gentiobiose

 B

d-Glucose-6- phosphate

α-d-Glucose-1-Phosphate

l-Glucose

α-d-Glucose

α-d-Glucose

α-d-Glucose

3-O-Methyl-d-glucose

α-Methyl-d-glucoside

β-Methyl-d- glucoside

d-Salicin

d-Sorbitol

N-Acetyl-d- glucosamine

 C

d-Glucosaminic acid

d-Glucuronic acid

Chondroitin-6-sulfate

Mannan

d-Mannose

α-Methyl-d-mannoside

d-Mannitol

N-Acetyl-β-d- mannosamine

d-Melezitose

Sucrose

Palatinose

d-Turanose

 D

d-Tagatose

l-Sorbose

l-Rhamnose

l-Fucose

d-Fucose

d-Fructose-6-Phosphate

d-Fructose

Stachyose

d-Raffinose

d-Lactitol

Lactulose

α-d-Lactose

 E

Melibionic acid

d-Melibiose

d-Galactose

α-Methyl-d-galactoside

β-Methyl-d-galactoside

N-Acetyl-Neuraminic acid

Pectin

Sedoheptulosan

Thymidine

Uridine

Adenosine

Inosine

 F

Adonitol

l-Arabinose

d-Arabinose

β-Methyl-d-xylopyranoside

Xylitol

Myo-Inositol

Meso-Erythritol

Propylene glycol

Ethanolamine

d, l-α-Glycerol-phosphate

Glycerol

Citric acid

 G

Tricarballylic acid

d, l-Lactic acid

Methyl d-lactate

Methyl pyruvate

Pyruvic acid

α-Keto-Glutaric acid

Succinamic acid

Succinic acid

Mono-methyl succinate

l-Malic acid

d-Malic acid

Meso-tartaric acid

 H

Acetoacetic acid (a)

γ-Amino-N-butyric acid

α-Keto-buytric acid

α-Hydroxy-butyric acid

d, l-β-Hydroxy-butyric acid

γ-Hydroxy-butyric acid

Butyric acid

2,3-Butanediol

3-Hydroxy-2- Butanone

Propionic acid

Acetic acid

Hexanoic acid

PM-M2 MicroPlate™—carbon and energy sources/nitrogen sources

      

 A

Negative control

Negative control

Negative control

Tween 20

Tween 40

Tween 80

Gelatin

l-Alaninamide

l-Alanine

d-Alanine

l-Arginine

l-Asparagine

 B

l-Aspartic acid

d-Aspartic acid

l-Glutamic acid

d-Glutamic acid

l-Glutamine

Glycine

l-Histidine

l-Homoserine

Hydroxy-L-Proline

l-Isoleucine

l-Leucine

l-Lysine

 C

l-Methionine

l-Ornithine

l-Phenylalanine

l-Proline

l-Serine

d-Serine

l-Threonine

d-Threonine

l-Tryptophan

l-Tyrosine

l-Valine

Ala-Ala

 D

Ala-Arg

Ala-Asn

Ala-Asp

Ala-Glu

Ala-Gln

Ala-Gly

Ala-His

Ala-Ile

Ala-Leu

Ala-Lys

Ala-Met

Ala-Phe

 E

Ala-Pro

Ala-Ser

Ala-Thr

Ala-Trp

Ala-Tyr

Ala-Val

Arg-Ala (b)

Arg-Arg (b)

Arg-Asp

Arg-Gln

Arg-Glu

Arg-Ile (b)

 F

Arg-Leu (b)

Arg-Lys (b)

Arg-Met (b)

Arg-Phe (b)

Arg-Ser (b)

Arg-Trp

Arg-Tyr (b)

Arg-Val (b)

Asn-Glu

Asn-Val

Asp-Ala

Asp-Asp

 G

Asp-Glu

Asp-Gln

Asp-Gly

Asp-Leu

Asp-Lys

Asp-Phe

Asp-Trp

Asp-Val

Glu-Ala

Glu-Asp

Glu-Glu

Glu-Gly

 H

Glu-Ser

Glu-Trp

Glu-Tyr

Glu-Val

Gln-Glu

Gln-Gln

Gln-Gly

Gly-Ala

Gly-Arg

Gly-Asn

Gly-Asp

α-D-Glucose

PM-M3 MicroPlate™—carbon and energy sources/nitrogen sources

      

 A

Negative control

Negative control

Negative control

Gly-Gly

Gly-His

Gly-Ile

Gly-Leu

Gly-Lys

Gly-Met

Gly-Phe

Gly-Pro

Gly-Ser

 B

Gly-Thr

Gly-Trp

Gly-Tyr

Gly-Val

His-Ala

His-Asp

His-Glu

His-Gly

His-His (c)

His-Leu

His-Lys (d)

His-Met

 C

His-Pro

His-Ser

His-Trp

His-Tyr

His-Val

Ile-Ala

Ile-Arg (b)

Ile-Asn

Ile-Gln

Ile-Gly

Ile-His

Ile-Ile

 D

Ile-Leu

Ile-Met

Ile-Phe

Ile-Pro

Ile-Ser

Ile-Trp

Ile-Tyr

Ile-Val

Leu-Ala

Leu-Arg (b)

Leu-Asn

Leu-Asp

 E

Leu-Glu

Leu-Gly

Leu-His

Leu-Ile

Leu-Leu

Leu-Met

Leu-Phe

Leu-Pro

Leu-Ser

Leu-Trp

Leu-Tyr

Leu-Val

 F

Lys-Ala (d)

Lys-Arg (b)

Lys-Asp

Lys-Glu

Lys-Gly

Lys-Ile (b)

Lys-Leu (b)

Lys-Lys

Lys-Met (e)

Lys-Phe

Lys-Pro

Lys-Ser

 G

Lys-Thr

Lys-Trp (b)

Lys-Tyr (b)

Lys-Val (d)

Met-Arg (b)

Met-Asp

Met-Gln

Met-Glu

Met-Gly

Met-His

Met-Ile

Met-Leu

 H

Met-Lys (e)

Met-Met

Met-Phe

Met-Pro

Met-Thr

Met-Trp

Met-Tyr

Met-Val

Phe-Ala

Phe-Asp

Phe-Glu

α-d-Glucose

PM-M4 MicroPlate™—carbon and energy sources/nitrogen sources

      

 A

Negative control

Negative control

Negative control

Phe-Gly

Phe-Ile

Phe-Met

Phe-Phe

Phe-Pro

Phe-Ser

Phe-Trp

Phe-Tyr

Phe-Val

 B

Pro-Ala

Pro-Arg (b)

Pro-Asn

Pro-Asp

Pro-Glu

Pro-Gln

Pro-Gly

Pro-Hyp

Pro-Ile

Pro-Leu

Pro-Lys (b)

Pro-Phe

 C

Pro-Pro

Pro-Ser

Pro-Trp

Pro-Tyr

Pro-Val

Ser-Ala

Ser-Asn

Ser-Asp

Ser-Glu

Ser-Gln

Ser-Gly

Ser-His(b)

 D

Ser-Leu

Ser-Met

Ser-Phe

Ser-Pro

Ser-Ser

Ser-Tyr

Ser-Val

Thr-Ala

Thr-Arg (f)

Thr-Asp

Thr-Glu

Thr-Gln

 E

Thr-Gly

Thr-Leu

Thr-Met

Thr-Phe

Thr-Pro

Thr-Ser

Trp-Ala

Trp-Arg

Trp-Asp

Trp-Glu

Trp-Gly

Trp-Leu

 F

Trp-Lys (e)

Trp-Phe

Trp-Ser

Trp-Trp

Trp-Tyr

Trp-Val

Tyr-Ala

Tyr-Gln

Tyr-Glu

Tyr-Gly

Tyr-His

Tyr-Ile

 G

Tyr-Leu

Tyr-Lys

Tyr-Phe

Tyr-Trp

Tyr-Tyr

Tyr-Val

Val-Ala

Val-Arg

Val-Asn

Val-Asp

Val-Glu

Val-Gln

 H

Val-Gly

Val-His

Val-Ile

Val-Leu

Val-Lys

Val-Met

Val-Phe

Val-Pro

Val-Ser

Val-Tyr

Val-Val

α-d-Glucose

367 biochemical substrates on PM-M1, PM-M2, PM-M3 and PM-M4 microplates are indicated. PM-M1 contains primarily carbohydrate and carboxylic substrates, while PM-M2, PM-M3 and PM-M4 contain individual L-amino acids and most dipeptide combinations (Biolog Inc.). The highlighted areas indicate substrates that were utilized by all three, or any one or two cell types HCAEC, HUVEC and NHLF

Acidic substances are sodium salts and basic substances are the chloride salts, except as noted (a): lithium salt, (b): acetate salt, (c): trifluoroacetate salt, (d): bromide salt, (e): formate salt, (f): sulfate salt

2.4 Statistical analysis

The Omnilog PMM Data Analyses Software was used to analyze metabolic phenotypes. The data were adjusted by subtracting average values of three negative control wells from all other samples at the end time point. Metabolic responses were normalized due to the differences in doubling times of the different cell types. Omnilog data exportable parameters were exported to Microsoft Excel and Bar graphs were made. Venn diagrams and digital images were processed using CorelDraw.

3 Results and discussion

3.1 Primary human cell types and growth conditions

In order to investigate cell morphology and growth of human primary endothelial (HCAEC, HUVEC), as well as fibroblast (NHLF) cells, inverted microscopy was used. Cells appeared phenotypically healthy, with endothelial cells exhibiting “cobblestone” morphology, while fibroblasts appeared typically elongated (Fig. 1). For measurement of the cellular metabolism Biolog Phenotype MicroArray was used. The kinetics of purple formazan accumulation was measured with OmniLog PMM System, operating as an incubator/reader that holds up to 50 microplates, and reads/quantifies colour density in each well, every 15 min over the user-specified time period. Prior to all measurements of cell metabolic activity, it is useful to first optimize which of the two dye mixes will work better for the cell type under study and how many cells will be needed per well to produce a given amount of reduced formazan product within an appropriate time frame. Dye preference and optimal number of cells for determining growth in complete medium has previously been determined for 15 different human, mouse and rat cell lines (http://biolog.com/pdf/pmmlit/00P%20133rC%20Redox%20Dye%20Mix%20Brochure%20JUL07.pdf), however no data is currently available on optimal conditions for primary human endothelial and fibroblast cells. We therefore performed experiments which determined the specific media (M1 or M2), cell densities and Biolog redox dye mixes (MA or MB) to be used (Fig. 2). No reaction was detected using Biolog Redox Dye Mix MA. M1 medium enabled slightly stronger reactions compared to M2 medium. The cell density of 10.000 cells/well was the first dilution to give maximal readings while 5.000 cells/well showed considerably decreased responses. The optimal conditions for growth of HCAEC and NHLF were therefore set on 10.000 cells/well in Biolog IF-M1 medium with 1 × Pen/Strep, 0.3 mM l-glutamine and 5 % dialysed FBS with Biolog Redox Dye Mix MB (10 µl/well) (Fig. 2). These conditions were used in all subsequent experiments determining cell growth on different substrates.
Fig. 1

Different phenotypes of human primary endothelial and fibroblast cells. Micrographs show normal, healthy endothelial cells isolated from coronary arteries, umbilical vein and fibroblasts from the lungs. HCAEC human coronary artery endothelial cells, HUVEC human umbilical vein endothelial cells and NHLF normal human lung fibroblasts

Fig. 2

Optimization of Biolog media, redox dye mixes and cell density, for HCAEC human coronary artery endothelial cells and NHLF normal human lung fibroblasts. The OmniLog reading results, as area under curve, of the separate wells, are presented in green

3.2 Utilization of substrates on PM-M1, PM-M2, PM-M3 and PM-M4 microplates

The extent to which human primary cell types from various tissues use different carbon substrates for energy has, to our knowledge, not been systematically investigated. It is known that, in addition to glucose, animal cells can metabolize and grow on other substrates (Bochner et al. 2011). In 1976, a survey of nutrient metabolism (Burns et al. 1976) examined 93 carbohydrates of which (a) 15 supported mammalian cell proliferation and (b) 42 were toxic or growth inhibitory and concluded that carbohydrate preferences of cells can be utilized to biochemically distinguish between different mammalian cell lines. In the following years, culture media were developed using glucose, pyruvate, and glutamine as energy sources, which were shown to support growth of most cell types and the interest to investigate the diversity of possible nutrients metabolized by different cell types waned. Recently, there is renewed interest to understand how the metabolism of different cell types could contribute to pathological pathways and help develop new approaches in nutritional therapy to support and improve treatment of a wide range of systemic disorders.

Using the optimized assay conditions we measured metabolic utilization of different substrates by HCAEC, HUVEC and NHLF in four microplates PM-M1 to PM-M4 containing 367 substrate. Representative plates out of three independent experiments are shown (Fig. 3). Glucose is widely accepted as the primary nutrient for the maintenance and promotion of cell function and all cell types tested produced a strong response in wells containing glucose (Fig. 3, black boxes). Background, the average absorbance measured in negative controls, was subtracted from the data used for further analysis (Fig. 3, blue boxes). Exclusive substrates are indicated in HCAEC (Fig. 3, green boxes) and NHLF (yellow boxes), while HUVEC showed no exclusively utilized substrates.
Fig. 3

HCAEC, HUVEC and NHLF assayed in PM-M1 to PM-M4. The columns show the endpoint of a 24 h incubation of a representative replicate of three independent experiments performed. Positive wells with glucose utilization are indicated with black boxes, while negative, background wells are shown with blue boxes. Exclusive results obtained for HCAEC are shown in green boxes and for NHLF in yellow boxes. HCAEC human coronary artery endothelial cells, HUVEC human umbilical vein endothelial cells and NHLF normal human lung fibroblasts

3.3 Data analysis of substrate utilization

All three cell types highly utilized adenosine, inosine, d-mannose and dextrin as substrates (to tetrazolium reduction levels above 100 mOD), comparable to glucose. Using plates PM-M1 through PM-M4, patterns of substrate utilization were observed for HCAEC, HUVEC and NHLF, as indicated in green, red and yellow bars, respectively (Fig. 4). HCAEC (green bars), prevalently utilized substrates on PM-M1 and to a lesser extent on PM-M3, with only a few substrates utilized on PM-M2 and PM-M4. NHLF, on the other hand, showed high and exclusive utilization of certain glutamine-associated peptides on PM-M2, PM-M3 and PM-M4, not observed in either HCAEC or HUVEC (Figs. 3, 4).
Fig. 4

Comparison of substrate utilization in all three cell types are shown on plates PM-M1 and PM-M2, PM-M3 and PM-M4. Cells were assayed according to the standard protocol and data collected after 24 h using the OmniLog and PM software, with subtraction of the background. Average height (mOD) of tetrazolium reduction was measured in triplicate. HCAEC human coronary artery endothelial cells, HUVEC human umbilical vein endothelial cells and NHLF normal human lung fibroblasts

HCAEC were able to use a larger number of substrate nutrients, with higher tetrazolium reduction, as measured in mOD for energy production than HUVEC (such as dextrin, glycogen, maltotriose, d-maltose, α-d-glucose-6-phosphate, mannan, d-mannose, d-turanose, d-fructose-6-phosphate, d-galactose, pectin, uridine, adenosine, inosine, d, l-α-glycerol-phosphate, tricarballylic acid, d, l-lactic acid, pyruvic acid, α-keto-glutaric acid, succinamic acid, among others) (Fig. 4). The specific substrates metabolized only by HCAEC, and not by either HUVEC or NHLF were the polymeric substrates mannan, pectin, gelatin, Val-Val and prevalently tricarballylic acid, as shown in Figs. 3 and 4. HUVEC did not show any exclusive metabolism of any substrates, but it was differentiated from HCAEC by notably stronger metabolism of D-glucuronic acid.

The assays indicate that cardiovascular cells, such as HCAEC exhibit extensive metabolic flexibility enabling them to produce energy more rapidly and use the most abundant fuel available, as well. These characteristics might also contribute directly to atherosclerosis and cardiac dysfunction, which are both known to be associated with changes in metabolism and obesity.

NHLF produced saturation of tetrazolium reduction in glucose wells during the 24 h kinetic measurements and utilized many substrates. Specifically, dextrin, glycogen, maltotriose, d-maltose, d-mannose, d-fructose-6-phosphate, inosine, d, l-lactic acid, d-glucose-6-phosphate, α-d-glucose-6-phosphate and adenosine produced strong reductive responses. Strikingly, NHLF also showed higher metabolism of the polyols d-sorbitol and xylitol along with l-glutamic acid, l-glutamine and all of the glutamine-containing dipeptides (Fig. 4, PM-M2, PM-M3 and PM-M4, yellow bars).

The most relevant glutamine-producing tissue is the muscle, accounting for about 90 % of all glutamine synthesized (Newsholme et al. 2003). Glutamine is also released, in small amounts, by the lung and the brain. In contrast, the biggest consumers of glutamine are the cells of intestines (Brosnan 2003), the kidney cells for the acid–base balance, activated immune cells (Newsholme 2001), and many cancer cells (Yuneva et al. 2007). We show for the first time, that normal healthy primary lung fibroblasts (in contrast to either HCAEC or HUVEC) have higher utilization of glutamine and glutamine-associated dipeptides. Glutamine has been shown in the past to be essential for growth of human embryonic diploid lung fibroblasts, when 10 % undialysed calf serum was used as a medium supplement (Litwin 1979). Interstitial fibroblasts within a biomatrix are exposed to varying levels of amino acids (Rishikof et al. 1998) and it was previously found that expression of α1(I) collagen mRNA was directly dependent on amino acid availability (Krupsky et al. 1997). Since type I collagen is a major structural protein in the lung known to participate in tissue fibrosis during systemic sclerosis, the effects of glutamine on its levels would be crucial to study. Rishikof et al. reported in 1998, on the regulation of type I collagen mRNA in human embryonic lung fibroblasts and found that the addition of the combination of glutamine and cysteine increased α1(I) collagen mRNA levels 6.3-fold (Rishikof et al. 1998). Importantly, glutamine, utilized for nucleotide synthesis (Engstrom and Zetterberg 1984) also increased α1(I) collagen mRNA in dermal fibroblasts by elevating gene transcription (Bellon et al. 1995).

The tetrazolium reduction measured in l-tryptophan during the 24 h measurements was detected in low levels only in NHLF. Neither HCAEC nor HUVEC metabolized l-tryptophan during this time frame. It is of interest that the metabolism of l-tryptophan is involved, via the kynurenine pathway, in patients with the eosinophilia-myalgia syndrome (Silver et al. 1992). Ingesting tryptophan may lead to a syndrome characterized by scleroderma-like skin abnormalities, fasciitis, in addition to eosinophilia. Plasma concentrations of l-kynurenine and quinolinic acid, both metabolites of tryptophan, were significantly higher in patients with active disease, as compared to patients studied after eosinophilia had resolved, or normal subjects (Silver et al. 1990). Since in late, diffused scleroderma, lungs can be prevalently affected, it could be hypothesized that this could be due to changed metabolism of tryptophan pathways in lung fibroblasts from patients. This would be relevant to test in the future.

3.4 Comparison of exclusive and overlapping substrates

Comparison of all three cell types together revealed that 49 of the 367 tested substrates were utilized by all three cell types, however both HCAEC and NHLF utilized several substrates uniquely, with a distinct pattern. On the other hand, HUVEC did not exhibit any exclusively utilized substrates. The specific substrates metabolized only by HCAEC, and not by HUVEC or NHLF were mannan, pectin, gelatin, Val-Val and prevalently tricarballylic acid, as shown in the Venn diagram (Fig. 5). NHLF were the only cells utilizing d-sorbitol, xylitol, l-glutamic acid, d-glutamic acid, l-tryptophan and the dipeptides Ala-Glu, Arg-Gln, Arg-Phe (b), Asp-Gln, Asp-Trp,Ile-Trp, Leu-Trp, Phe-Gly, Pro-Arg (b), Pro-Glu, Pro-Gln, Pro-Trp, Trp-Ala and Trp-Glu for their energy production (Fig. 5).
Fig. 5

Unique substrates metabolized exclusively by HCAEC, HUVEC and NHLF, as well as overlapping substrates, are represented by the Venn diagram. HCAEC human coronary artery endothelial cells, HUVEC human umbilical vein endothelial cells and NHLF normal human lung fibroblasts

Metabolism of substrates was measured within a 24 h timeframe. In the future, it could be advantageous to measure metabolic activity of different human primary cell lots, passages and preparations at different time periods with altered CO2 and oxygen levels in the incubator. This could be especially important for utilization of certain substrates, such as l-Glutamine, l-Tryptophan, Met- and Val- associated dipeptides, which indicated firm delineation between the fibroblast and endothelial cells.

Our results confirm that human primary cells can exhibit a unique and distinct metabolic fingerprint which could make them differentially susceptible to environmental changes that could subsequently lead to differential pathological processes.

4 Concluding remarks

Taken together, this is the first report to date exhibiting specific utilization of different carbon and energy sources of human primary endothelial and fibroblast cells on Biolog PM-M1 to PM-M4 plates. We showed that HCAEC have a higher overall metabolic rate than HUVEC, and that mannan, pectin, gelatin, Val-Val and prevalently tricarballylic acid were utilized exclusively by HCAEC. NHLF exhibited high metabolic activity by utilizing many substrate nutrients for energy production, especially glutamine associated dipeptides (Ala-Gln, Arg-Gln, Asp-Gln, Gln-Gly, Gln-Gln, Gln-Glu, Ile-Gln, Met-Gln, Pro-Gln, Ser-Gln, Thr-GlnTyr-Gln and Val-Gln). Also, the lack of utilization of mannan, pectin and gelatin in NHLF, coupled with their unique utilization of d-sorbitol, xylitol, l-tryptophan and tryptophan-associated dipeptides (Asp-Trp, Ile-Trp, Leu-Trp, Pro-Trp, Trp-Ala and Trp-Glu) points to a distinct lung fibroblast metabolic phenotype. Taken together, substrate utilization patterns on PM-M1 to PM-M4 plates provide a baseline metabolic characterization of different human primary cell types, which could be important in distinguishing early and cell-type specific metabolic changes that could initiate pathophysiological processes. In the future, it would be crucial to perform further experiments on activated cells or cells isolated from patient tissues, in order to elucidate early cellular changes relevant to cardiac complications, as well as the process of lung fibrosis in systemic sclerosis.

Acknowledgments

The authors would like to acknowledge Biolog Inc., Hayward, CA, USA and BIA, d.o.o., Ljubljana, Slovenia for kindly providing the Biolog OmniLog incubator/reader needed to perform the experiments.

Funding

This work was supported by the Slovenian Research Agency, with the National Research Program Grant #P3-0314.

Compliance with ethical standards

Conflict of interest

The authors state no conflict of interest.

Ethical approval

The work was performed within the National Research Program P3-0314, with approval from the Slovene Ethical committee #99/04/15.

Human and Animal Rights

This article does not contain any studies with human participants or animals performed by any of the authors.

Supplementary material

11306_2016_1024_MOESM1_ESM.docx (11 kb)
Supplementary Table 1: Donor demographics and cell characteristics (DOCX 11 kb)

Copyright information

© Springer Science+Business Media New York 2016

Authors and Affiliations

  • Polona Žigon
    • 1
  • Katjuša Mrak-Poljšak
    • 1
  • Katja Lakota
    • 1
  • Matic Terčelj
    • 1
  • Saša Čučnik
    • 1
    • 2
  • Matija Tomsic
    • 1
  • Snezna Sodin-Semrl
    • 1
    • 3
  1. 1.Department of RheumatologyUniversity Medical Centre LjubljanaLjubljanaSlovenia
  2. 2.Faculty of Pharmacy, Chair of Clinical BiochemistryUniversity of LjubljanaLjubljanaSlovenia
  3. 3.Faculty of Mathematics, Natural Sciences and Information TechnologyUniversity of PrimorskaKoperSlovenia