Hostname: page-component-cd9895bd7-gvvz8 Total loading time: 0 Render date: 2024-12-29T05:29:32.308Z Has data issue: false hasContentIssue false
  • English
  • Français

Quantification of Human Neural Stem Cell Fate Using Volocity® 3D Image Analysis Software

Published online by Cambridge University Press:  05 September 2012

Katja Piltti ,
Martin Daffertshofer  and
Brian Cummings
Katja Piltti
Affiliation:
Sue & Bill Gross Stem Cell Center, 845 Health Science Road, UC Irvine, Irvine, CA 92697-1705
Martin Daffertshofer
Affiliation:
PerkinElmer, Schnackenburgallee 114, Hamburg, Germany, 22525
Brian Cummings*
Affiliation:
Sue & Bill Gross Stem Cell Center, 845 Health Science Road, UC Irvine, Irvine, CA 92697-1705
*
cummings@uci.edu

Extract

The promise of stem cell therapy offers hope for the treatment of many human diseases and injuries. Spinal cord and traumatic brain injuries can cause the loss of neurons and glial cells. These lost cells may be partially replaced by transplanting multipotent human neural stem cells (hNSCs), which can differentiate into specialized neural cell types. The potential for these remarkable cells to promote recovery of neural function has made them a strong focus of researchers in this field.

Type
Instrumentation, Software, and Analysis
Information
Microscopy Today , Volume 20 , Issue 5 , September 2012 , pp. 34 - 36
Copyright
Copyright © Microscopy Society of America 2012

Introduction

The promise of stem cell therapy offers hope for the treatment of many human diseases and injuries [Reference Nelson, Martinez-Fernandez and Terzic1]. Spinal cord and traumatic brain injuries can cause the loss of neurons and glial cells. These lost cells may be partially replaced by transplanting multipotent human neural stem cells (hNSCs), which can differentiate into specialized neural cell types [Reference Anderson, Cummings, Hoosmand, Salazar, Ribak, Arámburo de la Hoz, Jones, Larriva Sahd and Swanson2, Reference Zhang, Wernig, Duncan, Brüstle and Thomson3]. The potential for these remarkable cells to promote recovery of neural function has made them a strong focus of researchers in this field.

Accurate and efficient quantification of human stem cell lineage and fate is essential for this research. In many laboratories, stem cells cultured and differentiated in vitro are routinely classified and quantified from 2D or 3D images. Recognition of hNSC progeny in 2D is challenging because of their heterogeneous cell morphology, long overlapping cell processes, and high cell density, which make cell fate quantification highly error prone [Reference Walton, Sutter, Chen, Chang, Roper, Scheffler and Steindler4]. In addition, manual methods are frequently used for hNSC fate analysis, which can be time-consuming, subjective, and inconsistent [Reference Huang and Murphy5]. The ability to visualize and analyze image data in 3D can overcome many of these classification errors and provide a deeper understanding of cellular samples [Reference Hamilton6]. The development of an automated method for cell fate analysis would also improve the efficiency of stem cell research.

Human NSCs have the ability to differentiate into (1) neurons, (2) astrocytes, which largely function to provide nutritional and structural support to neurons, and (3) oligodendrocytes, specialized cells that form the myelin sheath around axons and insulate neural signals [Reference Zhang, Wernig, Duncan, Brüstle and Thomson3]. These three specialized neural cell types can be identified using specific nuclear and cytoplasmic cell fate markers [Reference Rieske, Azizi, Augelli, Gaughan and Krynska7]. In a recent study published in Stem Cell Research, B. J. Cummings and colleagues from the Sue & Bill Gross Stem Cell Center at the University of California set out to create and optimize two image analysis protocols to quantify these cell fate markers using PerkinElmer's Volocity® 3D Image Analysis Software [Reference Piltti, Haus, Do, Perez, Anderson and Cummings8].

Methods and Materials

The first image analysis protocol was designed to quantify hNSC progeny expressing the nuclear fate marker oligodendrocyte lineage transcription factor 2 (Olig2). The second protocol was designed to quantify progeny expressing the cytoplasmic fate markers β-Tubulin III (β-TubIII), which is neuron-specific, and glial fibrillary acidic protein (GFAP), which is expressed by mature astroglial cells and undifferentiated hNSCs. Fetal-derived hNSCs were differentiated for 14 days in vitro and then fixed, permeabilized, and stained with the nuclear counter stain Hoechst 33342 and antibodies against β-TubIII, GFAP, and/or Olig2.

Application of the Protocols

Two-dimensional and 3D (confocal) images were acquired and imported into Volocity® software to develop the protocols for quantifying the proportions of cells positive for each cell fate marker. In Volocity® software, protocols are built up through the addition of a series of measurement tasks that perform functions such as finding, relating, thresholding, filtering, etc. These allow the user to measure biological structures and the relationships between them.

To quantify the proportions of cells positive for each of the three cell fate markers, the protocols were started with a thresholding measurement task to identify objects of interest in the appropriate channels. Volocity® software includes a range of flexible thresholding options that do not rely on absolute intensity values, meaning that the same protocols may be applied to multiple data sets and time-resolved experiments. Further steps in the protocols filtered and excluded objects based on their color and size, removed background noise, and excluded cytoplasmic areas (objects that do not touch Hoechst). Finally, measurements were performed to identify cells as neuronal, astroglial, or oligodendroglial.

Results

Figure 1 shows the identification and quantification of neuronal, astroglial, and oligodendroglial cell types in Volocity® software after hNSC differentiation in vitro. The measurement feedback is shown as an overlay on the image data (Figure 1), and this was used for validation by manual inspection. In Volocity® software, objects can be viewed in different ways using a variety of image view modes. Figure 2 shows images taken from both the “XY Plane” view mode and the “3D Slice” view mode, the latter of which allows image data to be re-sliced in a different orientation, post-acquisition, to display a more informative visualization.

Figure 1: Classification of neuronal, astroglial, or oligodendroglial cell types after hNSC differentiation in vitro. Cells were stained with (A) the nuclear counter stain Hoechst 33342 and (B) an antibody against the oligodendroglial marker Olig2. (C) The nuclei of Volocity® software-recognized Olig2+/Hoechst+ cells are shown (pink). (D) Volocity® software 3D-rendered image of cells stained with the nuclear counter stain Hoechst 33342 (blue), and antibodies against the cytoplasmic markers β-TubIII (green) and GFAP (red). (E) Volocity® software image analysis showing objects classified as β-TubIII+, GFAP+, β-TubIII+/GFAP+, β-TubIII+/Hoechst+, GFAP+/Hoechst+, β-TubIII+/GFAP+/Hoechst+, or Hoechst+. (F) The quantification of cell type-specific markers required the exclusion of cytoplasmic areas (objects that do not touch Hoechst). The resulting image after Volocity® software quantification is shown. Analyzed nuclei are shown as overlapping purple, yellow, white, and blue.

Figure 2: Visualization of neuronal and astroglial cell lineages. (A, B) Volocity® software 3D renderings of neuronal (green) and astroglial (red) cell lineages, with nuclei shown in blue, before (A) and after (B) identification of β-TubIII+Hoechst+, GFAP+Hoechst+, β-TubIII+GFAP+Hoechst+, and Hoechst+ objects using Volocity software. (C) Identified objects are shown in the XY plane and enlarged area.

After optimizing the two protocols, researchers then evaluated the time-efficiency and validity of using this software-based image analysis approach compared to manual analysis. Study results demonstrated that software-based quantification of hNSC progeny was more accurate than manual methods and was at least two-fold faster, particularly in samples of high cell density [Reference Piltti, Haus, Do, Perez, Anderson and Cummings8].

These two image analysis protocols could also be applied to the quantification of other human stem cell lineages, with only minor modifications such as adjustments in object size, color intensities, and z-stack intervals. These protocols can be downloaded [http://www.sciencedirect.com/science/article/pii/S1873506111000663] from the supplementary material associated with Piltti et al. [Reference Piltti, Haus, Do, Perez, Anderson and Cummings8] and tested or adapted in Volocity® software. A complimentary demo version can be found at http://cellularimaging.perkinelmer.com/downloads/.

Conclusion

These results show that, compared to human-based manual 2D image analysis, Volocity® software-based 3D analysis allowed for two-fold faster analysis and more accurate analysis of cytoplasmic and nuclear cell fate markers after in vitro differentiation.

References

[1]Nelson, JT, Martinez-Fernandez, A, and Terzic, A, Nat Rev Cardiol 7(12) (2010) 700–710.CrossRefGoogle Scholar
[2]Anderson, AJ, Cummings, BJ, Hoosmand, MJ, and Salazar, DL, “Human neural stem cell mediated repair of the contused spinal cord: Timing the microenvironment” in From Development to Degeneration and Regeneration of the Nervous System, eds. Ribak, CE, Arámburo de la Hoz, C, Jones, EG, Larriva Sahd, JA, and Swanson, LW, Oxford University Press, New York, 2008, p. 368.Google Scholar
[3]Zhang, SC, Wernig, M, Duncan, ID, Brüstle, O, and Thomson, JA, Nat Biotechnol 19(12) (2001) 1129–33.CrossRefGoogle Scholar
[4]Walton, NM, Sutter, BM, Chen, H-X, Chang, L-J, Roper, SN, Scheffler, B, and Steindler, DA, Development 133(18) (2006) 3671–81.CrossRefGoogle Scholar
[5]Huang, K and Murphy, RF, J Biomed Opt 9(5) (2004) 893–912.CrossRefGoogle Scholar
[6]Hamilton, N, Traffic 10(8) (2009) 951–61.CrossRefGoogle Scholar
[7]Rieske, P, Azizi, SA, Augelli, B, Gaughan, J, and Krynska, B, Eur J Neurosci 25(1) (2007) 31–37.CrossRefGoogle Scholar
[8]Piltti, KM, Haus, DL, Do, E, Perez, H, Anderson, AJ, and Cummings, BJ, Stem Cell Research 7(3) (2011) 256–63.CrossRefGoogle Scholar
Figure 0

Figure 1: Classification of neuronal, astroglial, or oligodendroglial cell types after hNSC differentiation in vitro. Cells were stained with (A) the nuclear counter stain Hoechst 33342 and (B) an antibody against the oligodendroglial marker Olig2. (C) The nuclei of Volocity® software-recognized Olig2+/Hoechst+ cells are shown (pink). (D) Volocity® software 3D-rendered image of cells stained with the nuclear counter stain Hoechst 33342 (blue), and antibodies against the cytoplasmic markers β-TubIII (green) and GFAP (red). (E) Volocity® software image analysis showing objects classified as β-TubIII+, GFAP+, β-TubIII+/GFAP+, β-TubIII+/Hoechst+, GFAP+/Hoechst+, β-TubIII+/GFAP+/Hoechst+, or Hoechst+. (F) The quantification of cell type-specific markers required the exclusion of cytoplasmic areas (objects that do not touch Hoechst). The resulting image after Volocity® software quantification is shown. Analyzed nuclei are shown as overlapping purple, yellow, white, and blue.

Figure 1

Figure 2: Visualization of neuronal and astroglial cell lineages. (A, B) Volocity® software 3D renderings of neuronal (green) and astroglial (red) cell lineages, with nuclei shown in blue, before (A) and after (B) identification of β-TubIII+Hoechst+, GFAP+Hoechst+, β-TubIII+GFAP+Hoechst+, and Hoechst+ objects using Volocity software. (C) Identified objects are shown in the XY plane and enlarged area.