Automated qualitative batch measurement of lipid droplets in the liver of bird using ImageJ

Summary

Oil red O staining is effective in detection and quantification of neutral lipid droplets in tissues such as the liver. However, converting images into testable data using ImageJ can be time-consuming and is prone to inaccuracy or bias. We describe a protocol for automated qualitative measurement of lipid droplets in the bird liver using a batch processing macro script. We explain steps for extracting tissue, cryosectioning, staining, and imaging, followed by script generation for quantification of lipid in the images.

Graphical abstract

Highlights

• •Oil red O stain is suitable for animal models with metabolic changes
• •Automated lipid droplet quantification of small/large datasets
• •Rapid detection of lipid droplet area by introducing a macro script to Image J

Publisher’s note: Undertaking any experimental protocol requires adherence to local institutional guidelines for laboratory safety and ethics.

Oil red O staining is effective in detection and quantification of neutral lipid droplets in tissues such as the liver. However, converting images into testable data using ImageJ can be time-consuming and is prone to inaccuracy or bias. We describe a protocol for automated qualitative measurement of lipid droplets in the bird liver using a batch processing macro script. We explain steps for extracting tissue, cryosectioning, staining, and imaging, followed by script generation for quantification of lipid in the images.

Before you begin

Metabolic illnesses like type 2 diabetes, non-alcoholic fatty liver disease, and the metabolic syndrome are characterized by pathological lipid accumulation.^1,2,3 Organ-accumulating lipids can obstruct the insulin signaling pathway,^4,5,6 reducing the amount of circulating glucose that is taken up. Insulin resistance is a condition that affects more than 90% of people with type 2 diabetes.¹ Liver, muscle, and endothelium are a few examples of cells and tissues with significant roles in glucose management that exhibit insulin resistance first.^7,8,9 This vicious cycle is reinforced by the development of insulin resistance in white adipose tissue and the large-scale release of free fatty acids, which build up in the peripheral tissues.^9,10 Therefore, it is critical to establish the degree of lipid build-up in organs like the liver and muscle in order to comprehend the illness progression and the general metabolic status of an organism. Although the causes of hepatic lipid accumulation are still up for debate,¹¹ it is widely acknowledged that insulin resistance and metabolic illness are closely related to liver and muscle steatosis.¹² The necessity for lipid-staining procedures that are simple to use and can be carried out in most laboratories is highlighted by the fact that further study is required to fully understand the mechanisms underlying lipid-induced metabolic dysfunction. The knowledge of general lipid biology also requires techniques that show the shape and intracellular location of lipid droplets within tissues.

A major drawback of imaging of neutral lipids by ORO staining is time consumption in large data set and requirement of large data set for correct quantification of LDs.¹³ In this protocol, we have tried to automate the processes through macro batch processing for large data set which massively reduces the time required to analyze data.

Institutional permissions

The animal study protocol (No. LU/ZOOL/IAEC/05/19/10) was approved by the Institutional Animal Ethical Committee (IAEC).

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Chemicals, peptides, and recombinant proteins
Polyvinyl pyrrolidone (PVP)	Himedia	RM854
Oil red O (ORO) stain	Sigma-Aldrich	O1391
Glycerol	Merck	61756005221730
Advanced PAP pen (Liquid blocker pen)	Merck	Z377821
Gelatin powder	Bombay Drug House	44045 4B
Chromium (III) potassium sulfate dodecahydrate	Bombay Drug House	10079
Molecular grade water	Avra	ALM1168
Experimental models: Organisms/strains
Male redheaded bunting (Emberiza bruniceps)	Ali and Ripley14	N/A
Software and algorithms
ImageJ	Schneider et al.15	https://doi.org/10.1038/nmeth.2089
Algorithm for ImageJ script	This paper	N/A
ImageJ macro script	This paper, GitHub	https://github.com/anurag-nishad/ImageJ_OROimage_Script
Prism 8.0.2	GraphPad	RRID: SCR_002798
Other
Extraction of liver and ORO staining	Mehlem et al.13	https://doi.org/10.1038/nprot.2013.055
Cryostat	Leica	CM 1850
Brightfield microscope	Leica	DM3000
Camera for microscope	Leica	DFC450C

Materials and equipment

Gelatin coated slides

Gelatin subbing solution

Reagent	Final concentration	Amount
Gelatin powder	0.5%	1.25 gm
Chromium (III) potassium sulfate dodecahydrate	0.05%	0.125 gm
Deionized ddH2O (at 40°C)	99.45%	250 ml
Total	N/A	250 ml

Slides were dipped 4–5 times for 5 s each in the gelatin subbing solution and were dried at room temperature (25°C) for 72 h in a clean environment.

Note: Gelatin slides are charged slides and must be used within a week for best results. Coated slides should be stored at −20°C.

ORO stain

ORO stock solution

Reagent	Final concentration	Amount
Oil red O	0.625%	2.5 gm
Isopropyl alcohol	99.37%	400 ml
Total	N/A	400 ml

Mix ORO and isopropyl alcohol by magnetic stirring for 2 h at room temperature (25°C).

Note: The stock solution can be stored at room temperature (25°C) for 6 months or until precipitation occurs in a dark bottle away from direct sunlight.

ORO working solution

Reagent	Final concentration	Amount
Oil red O stock solution	60%	18 ml
ddH2O	40%	12 ml
Total	N/A	30 ml

After preparing ORO working solution, let the working solution stand for 5–10 min at 4°C before use, during this time the solution will thicken. Filter the working solution through a 45 μm filter to remove precipitate.

Note: Use the ORO working solution within 6 h. Store working solution at 4°C if needed.

CRITICAL: ORO stain can cause eye and skin irritation use protective glasses and gloves while working with the solution. It is also a highly flammable substance and should be kept away from ignition sources such as burners.

CRITICAL: ORO working stain must be freshly prepared for staining. Check for precipitation in stock solution and filter if required.

Polyvinylpyrrolidone (PVP) mold medium

PVP working medium

Reagent	Final concentration	Amount
Polyvinylpyrrolidone (PVP)	15%	6 gm
ddH2O	85%	34 ml
Total	N/A	40 ml

Note: PVP should not be stored for more than 48 h at 4°C.

CRITICAL: PVP prepared more than 48 h ago may cause tissue adherence issue.

Temporary slide mounting medium

Glycerol mounting medium

Reagent	Final concentration	Amount
Glycerol	50%	50 ml
ddH2O	50%	50 ml
Total	N/A	100 ml

Step-by-step method details

Tissue extraction

Timing: 1 day

Obtaining undamaged tissues yields good results after analysis, hence adequate measures should be taken while extracting tissue. This section accomplishes the aim of tissue extraction and storage according to good laboratory practice.

• 1.Collect the organ (liver) of interest from animal (here it is, Emberiza bruniceps) according to good laboratory practices.

CRITICAL: Informed consent should be sought from all donors of human biological resources in case of human tissue, and this process should be carried out in conformity with pertinent institutional and governmental ethics rules and regulations.

• 2.Immediately after collection, flash-freeze the organs with liquid N₂.

Alternatives:Instead of liquid nitrogen, dry ice and isopropanol bath can also be used.

CRITICAL: Liquid N₂ and dry ice can cause skin burns; wear protective garments.

CRITICAL: It is important to flash-freeze the organs in liquid N₂ and to keep them frozen in liquid N₂ or on dry ice throughout the entire collection procedure.

Note: Dissected tissue should not be thawed before or during the embedding of the tissue for cryosectioning as this will cause loss of tissue integrity and, consequently, inaccurate ORO staining.

Pause point: The tissues can be stored at −80°C for up to 3 years.

Cryosectioning

Timing: 2–3 days

This step describes the acquisition of cross-section of the liver tissue (Emberiza bruniceps) to get an appropriate overview of the lipid deposition.

• 3.Set the Cryostat at −25°C prior of sectioning.

CRITICAL: Correct temperature of the cryostat is important and deviation from the required temperature may cause damaging of tissue.

Note: For fatty tissue i.e., containing less amount of water, temperature around −25°C to −30°C must be used whereas for more water containing tissue, −15°C to −20°C must be used.

• 4.Prepare the half mount by freezing and flattening of PVP (see materials and equipment for details).
• 5.Place the tissue on the mount and pour the required amount of PVP to fully cover the tissue and freeze inside the cryostat.
• 6.Section the tissue at 15 μm thickness and collect five cross-sections from different depths of the organ onto the same gelatin coated glass slide in order to provide a good overview of the tissue as shown in the Figure 1A.

Note: Cryosections stick to the slides due to temperature difference between them.

CRITICAL: To avoid detachment of the sections during the ORO staining procedure, let the sections dry for 10 min at room temperature (25°C) before freezing them.

Optional: More cross-sections from different depths can be obtained for a more accurate analysis of the tissue.

Pause point: The sections can be stored at −80°C for up to 3 years.

Figure 1.

Liver region from where the sections were selected

(A) Dotted region (I, II, III, IV, V) represents the general area from where the sections were selected for the analysis of lipid droplets. Yellow line demonstrated the perimeter drawn with liquid blocker PAP pen.

(B) Oil-red O stained slide, red dot are lipid droplets. Scale bar = 50 μm.

ORO staining

Timing: 2 h

In this step we stain the neutral lipid droplets in the obtained liver cross-sections.

• 7.If the sections have been stored at −80°C, allow them to equilibrate for 10 min at room temperature (25°C).
• 8.Trace the perimeter of the tissue with the liquid blocker pen.
• 9.Add ∼1 mL of ORO working solution (freshly prepared) to cover the sections.

CRITICAL: The sections need to be completely covered by solution during the staining procedure; otherwise, the ORO stain will dry and become uneven; ORO working solution must be at 4°C and the slides at room temperature (25°C) before use.

Note: Staining of neutral lipid droplets through ORO is a physical phenomenon and hence requires temperature difference between the stain and the tissue. ORO is poorly soluble in isopropyl alcohol, and its solubility is being further decreased by adding water. Hence the hydrophobic dye moves from the solvent (isopropyl + water) to lipid in the target tissue.

• 10.Incubate the sections with ORO working solution at room temperature (25°C). Incubation time might vary with the tissue, model used and amount of lipid present in the tissue. For bird liver sections, we generally incubate them for 5 min with the ORO working solution.
• 11.Rinse the sections under running tap water for ∼30 min. Make sure the parts are not damaged while being rinsed.

Optional: Tie a muslin cloth on the tap to control the turbulence of water. Place the slides slightly away from tap water so as to prevent detaching of the tissue sections. Maintain a constant flow of fresh tap water onto the sections. This is the destaining stage.

• 12.Check the sections by using a brightfield microscope for a quick analysis to confirm adequate staining and washing, use Figure 1B as reference. If insufficient staining is observed, go back and repeat the staining procedure (step 9). For overstained tissue, increase the timing of destaining stage (step 11).

CRITICAL: Optimization of the stain must be done as described by Mehlem et al., 2013. Images of the slides must be captured within 24 h of preparation.

• 13.Mount the slides with a water-soluble mounting medium (glycerol) and place coverslips on them. Let the mounting medium set for 10 min at room temperature (25°C) and then seal the coverslip edges with nail polish, making them airtight.

Pause point: The mounted slides can be left at room temperature (25°C) for a maximum of 24 h.

Microscopy

Timing: 1 day

Image acquisition through microscope plays a vital role in quantification of the lipid droplets in the tissue, hence must be done carefully. This step describes about the magnification and image format used for export of the images from the microscope.

• 14.Place samples to be imaged in the same room as the microscope, so that they can equilibrate to the same temperature as the microscope (15 min–30 min).
• 15.Image using bright field and a 10× or 40× objective so that most of the lipid droplets are clearly visible.

Optional: Higher magnification images of droplets tends to give more accurate data and makes thresholding easier.

• 16.
  Capture the image with the highest possible illumination, such that the lipid droplets (red in color) are easily distinguishable from the background.

  • a.
    Acquire 1 image per liver section i.e., 5 images will be the representative of 1 bird liver.
  • b.
    Acquire data at 40× objective lens and 24-bit color depth with 2560 × 1920 pixels resolution.
  • c.
    Exported and imported images in ‘.tif’ file format from the microscope and ImageJ respectively.

CRITICAL: Standardization of illumination should be done and fixed illumination must be used to capture images so as to ease the process of processing afterward. Images should be slightly overexposed so that the border of the cells are minimally visible alongside without losing any detail of the droplet as shown in Figure 1B.

Image analysis in ImageJ

Timing: 3 h

This step describes how to quantify the lipid droplets in the obtained microscope images with an open-source software ImageJ. Large data set can be very time consuming and thus this step also states how to make a macro script which automatically calculate the lipid droplets in the image and save all the results in an excel file format. File management is important for the analysis of large image sets. A key component of this management is the naming of files and folders thus, this script also names the exported ‘.csv’ file with the corresponding image.

• 17.
  Open ImageJ.

  • a.
    Open a sample ORO image in ImageJ for creating macro code through File>Open…
• 18.
  Create a batch processing script for ImageJ.

  • a.
    Open Recorder window via Plugins>Macro>Record… and set Macro in Recorder option.

    Note: Every click will be now recorded as a predefined ImageJ command in the Recorder window.

  • b.
    Start analyzing the image.

    • i.
      Convert image to 8-bit, Image>Type > 8-bit.
    • ii.
      Manually threshold the image for highlighting the region of interest (lipid droplets), Image>Adjust>Threshold… by moving the sliders or using Auto Threshold function provided. Select the thresholding algorithm as default or whichever suits the situation of image (i.e., highlights all the lipid droplets) and untick all other arguments (i.e., Dark background, Stack histogram, Don’t reset range) as shown in the Figure 2B.

      CRITICAL: Black represents the region of interest and white determines the background of the image. Setting the correct thresholding value is important as this will define what area is to be calculated further in the procedure.

      Optional: Thresholding image can be cross-checked for correct selection of region of interest by creating a duplicate of the image before recording of steps. To duplicate the image use, Image>Duplicate…

    • iii.
      Applying the threshold will convert the image to binary image (i.e., black and white having 1 or 0 pixel value respectively).

      Note: All the measurements will be done on the black pixels of the image through ImageJ.

    • iv.
      Cut apart connected components into separate ones by watershed function, Process>Binary>Watershed…
    • v.
      Define the resolution ratio of the image in pixels/μm (here, 0.7 pixels/μm), Analyze>Set Scale…

      CRITICAL: Resolution ratio of image varies with respect to the equipment and image zoom used. In order define the pixel/μm ratio for quantification, capture an image of a known length (e.g., micrometre scale or technical ruler) at the identical magnification used for analysing lipid droplet. Click on one end of the known length, and drag a straight line to the other end of the known length. Next, select Analyze>Set Scale…, which will give the precise distance of the known length in pixels in the ‘Distance in pixels’ tab. Enter the real world known distance in the ‘Known distance’. Keep the Aspect ratio of the pixel 1. Enter μm in ‘Unit of length’ tab. Tick the ‘Global’ option to apply the resolution ratio to all the images.

    • vi.
      Select the parameters to be calculated by set measurement Analyze>Set measurements… Select the Area, Shape descriptors, Limit to threshold, Redirect to none and enter 3 in Decimal places tab.
    • vii.
      Open Analyze particle function, Analyze>Analyze Particles… and adjust circularity to 0.75–1.00 and tick Display results, Clear results and Exclude on edges and press ok.

      Note: If there is a hole or tear in the tissue custom area can be selected for measurement with the rectangle/freehand selection tool provided in the software.

  • c.
    Check the code generated in the recorder window and remove any unwanted sections in code by simply deleting it.
  • d.
    Give the name to the macro code and select create the code.
  • e.
    Save the macro code with File>Save As… with ‘.ijm’ extension.
  • f.
    Open the created macro code with Microsoft Notepad and add the following extra lines before and after the code generated as shown in the Figure 3 and save the file (Workflow of the algorithm is shown in Figure 4).

    Optional: Created script can be opened with any text file editing application such as Notepad++ etc.

• 19.Open created macro script with Plugin>Macro>Run… and select the source directory of images and destination directory of results when prompted.

Optional: ImageJ can be set as default opening application in Microsoft Windows operating system for opening ‘.ijm’ extension files and script can be opened without using Run… function by simple double clicking.

• 20.Obtain result from the excel sheets and compile them into one file.
• 21.Sort the data in excel sheet by circularity parameter in ascending order and extract the top 500 lipid droplets from each image data.

Note: Sorting the data according to the circularity helps because lipid droplets tends to have a circular shape. Default circularity formula in ImageJ is: 4π(area/perimeter²).

Optional: Selection of number of lipid droplets to be analysed is arbitrary and can be adapted to the image situation. It is minimum number of droplets which each image has, however it is important that this number must not change and should be constant in all images else the data generated will be biased.

• 22.Average the area size result of the top 500 lipid droplets extracted from each image analyzed. This will yield the average lipid droplet area size per image per bird liver. Similarly, we will receive a total of 2500 lipid droplet area per bird (500 lipid droplets/image/bird × 5 liver sections/bird).

Note: Greater number of images from each sample specimen will yield more accurate information about the lipid deposition content in the sample. Furthermore, increasing the sample size will also increase the accuracy of the result obtained.

Figure 2.

Example of thresholding by ImageJ

(A–C) Thresholding the 8-bit image for lipid droplet area measurement during the migratory life history stage of the Emberiza bruniceps, (A) 8-bit image before thresholding; (B) During thresholding; (C) After completing thresholding of the 8-bit image; black and white represents foreground and background respectively. Scale bar = 50 μm.

Figure 3.

Final macro script

Automation script for quantification of lipid droplets in ImageJ for oil red O stained slide photomicrographs. Red color script shows the macro generated through the Recorder window and Black color script shows the edited part in the macro script for batch processing.

Figure 4.

Algorithm workflow

Flowchart showing the algorithm used during the lipid droplet area quantification with oil-red O stained images through automated ImageJ script.

Expected outcomes

Male migratory redheaded buntings¹⁴ were caught from the wild and were acclimatised in the outdoor aviary for about 1 week. After acclimatisation, the birds were transferred to 8L:16D photoperiod (Short-day, SD) in ∼100 ± 20 lux intensity. Birds were given ad libitum food and water supply. After 4 weeks 5 birds were sacrificed and the required tissue (liver) was extracted and the rest of the birds were transferred to 13L:11D photoperiod (Long-day, LD) having the same light intensity as in SD. Birds were sacrificed at mid-day Zt = 6 (zeitgeber) and required tissue was extracted after they showed 7 cycles of migratory restlessness i.e., Zugunruhe.

The current protocol was used to compare the fat deposition¹⁶ in the migratory redheaded bunting during two different life history stages i.e., non-migratory and migratory phases. The script worked as expected, fat deposition was significantly higher in the liver during the migratory phase compared to the non-migratory phase¹⁷ (t = 36.93, df = 8, p < 0.001, g = 23.356; Two-tailed, Unpaired t-test followed by Hedge’s g) Figure 5.

Figure 5.

Results as mean ± SEM of average lipid droplet area in redheaded bunting

Quantitative analysis of fat accumulation as mean ± SEM in liver during non-migratory (Short-Day, SD; 8L:16D) and migratory (Long-Day, LD + mig; 13L:11D) life-history phase of redheaded bunting (Emberiza bruniceps). An asterisk (∗) on the bar indicates the significance between the life history stages i.e., SD and LD + mig. Significance was considered at p < 0.05, Unpaired t-test. Scale bar = 50 μm.

Quantification and statistical analysis

Data from 5 images per bird per life-history stage were averaged for the estimation of average lipid droplet size. The statistics were done using GraphPad Prism software, version 8.0, San Diego, CA, USA. Data were first checked for gaussian distribution (Shapiro-Wilk’s test for normality) and test for comparing the two life history means was selected accordingly (Unpaired Students t-test for parametric data and Mann-Whitney for non-parametric data) followed by effect size estimation via Hedge’s g.¹⁸

Limitations

ORO stain only the most hydrophobic part which is the core containing non-polar lipid droplets (i.e., triglycerides, diacylglycerols and cholesterol esters). The outer surface which contains polar lipids (i.e., phospholipids, sphingolipids and ceramides) and protein are not stained. Paraffine-based and alcohol-based solvents can’t be used to fix tissue because they tend to make lipid droplets stick together. As white and brown adipose tissues require paraffinization and/or stiff fixing to retain tissue shape, ORO staining cannot be used on sections from these tissues. ORO can be challenging to stain in tissues with limited endogenous lipid deposition, such as the pancreas. Sometimes, non-specific staining may give absurd values hence, checking of value before the final calculation is necessary.

Troubleshooting

Problem 1

Precipitation of ORO during preparation of ORO stain.

Potential solution

Prepare a new stock and acquire images within 24 h of mounting.

Problem 2

Damaged tissue sections during Cryosectioning.

Potential solution

Freeze the tissue with N₂ or dry ice and alcohol bath; avoid freeze-and-thaw cycles; increase the section size and thickness; use a very sharp cryosectioning knife.

Problem 3

Non-specific staining during ORO Staining.

Potential solution

Reduce the staining time; increase washing time; optimize the sectioning.

Problem 4

Too high/low value of the area in the results during Image Analysis in ImageJ.

Potential solution

Check the thresholding values in the script; look for non-specific staining, damage in tissue and ORO precipitation.

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Shalie Malik (malik_shalie@lkouniv.ac.in).

Materials availability

This study did not generate new unique reagents.

Data and code availability

The protocol does not include all datasets generated but are available upon reasonable request to lead contact or technical contact. The ImageJ batch processing script generated in the protocol is available at https://github.com/anurag-nishad/ImageJ_OROimage_Script.