Cellular heterogeneity poses challenges to understanding the function of complex tissues at a transcriptome level. Using cell-type-specific RNAs avoids potential pitfalls caused by the heterogeneity of tissues and unleashes the powerful transcriptome analysis. The protocol described here demonstrates how to use the Translating Ribosome Affinity Purification (TRAP) method to isolate ribosome-bound RNAs from a small amount of EGFP-expressing cells in a complex tissue without cell sorting. This protocol is suitable for isolating cell-type-specific RNAs using the recently available NuTRAP mouse model and could also be used to isolate RNAs from any EGFP-expressing cells.
High-throughput approaches, including RNA sequencing (RNA-seq) and microarray, have made it possible to interrogate gene expression profiles at the genome-wide level. For complex tissues such as the heart, brain, and testis, the cell-type-specific data will provide more details comparing the use of RNAs from the whole tissue1,2,3. To overcome the impact of cellular heterogeneity, the Translating Ribosome Affinity Purification (TRAP) method has been developed since early 2010s4. TRAP is able to isolate ribosome-bound RNAs from specific cell types without tissue dissociation. This method has been used for translatome (mRNAs that are being recruited to the ribosome for translation) analysis in different organisms, including targeting an extremely rare population of muscle cells in Drosophila embryos5, studying different root cells in the model plant Arabidopsis thaliana6, and performing transcriptome analysis of endothelial cells in mammals7.
TRAP requires a genetic modification to tag the ribosome of a model organism. Evan Rosen and colleagues recently developed a mouse model called Nuclear tagging and Translating Ribosome Affinity Purification (NuTRAP) mouse8, which has been available through the Jackson Laboratory since 2017. By crossing with a Cre mouse line, researchers can use this NuTRAP mouse model to isolate ribosome-bound RNAs and cell nuclei from Cre-expressing cells without cell sorting. In Cre-expressing cells that also carry the NuTRAP allele, the EGFP/L10a tagged ribosome allows the isolation of translating mRNAs using affinity pulldown assays. At the same cell, the biotin ligase recognition peptide (BLRP)-tagged nuclear membrane, which is also mCherry positive, allows the nuclear isolation by using affinity- or fluorescence-based purification. The same research team also generated a similar mouse line in which the nuclear membrane is labeled only with mCherry without biotin8. These two genetically modified mouse lines give access to characterize paired epigenomic and transcriptomic profiles of specific types of cells in interest.
The hedgehog (Hh) signaling pathway plays a critical role in tissue development9. GLI1, a member of the GLI family, acts as a transcriptional activator and mediates the Hh signaling. Gli1+ cells can be found in many hormone-secreting organs, including the adrenal gland and the testis. To isolate cell-type-specific DNAs and RNAs from Gli1+ cells using the NuTRAP mouse model, Gli1-CreERT2 mice were crossed with the NuTRAP mice. Shh-CreERT2 mice were also crossed with the NuTRAP mice aim to isolate sonic hedgehog (Shh) expressing cells. The following protocol shows how to use Gli1-CreERT2;NuTRAP mice to isolate ribosome-bound RNAs from Gli1+ cells in adult mouse testes.
All performed animal experiments followed the protocols approved by the Institutional Animal Care and Use Committees (IACUC) at Auburn University.
NOTE: The following protocol uses one testis (about 100 mg) at P28 from Gli1-CreERT2; NuTRAP mice (Mus musculus). Volumes of reagents may need to be adjusted based on the types of samples and the number of tissues.
1. Tissue collection
- Euthanize the mice using a CO2 chamber, sanitize the abdomen surface with 70% ethanol.
- Open the lower abdomen with scissors and remove the testes. Use liquid nitrogen (LN2) to snap-freeze the testes immediately.
- Store samples in the vapor phase of LN2 until use.
2. Reagents and beads preparation
- Prepare the homogenization stock solution: Add 50 mM Tris (pH 7.4), 12 mM MgCl2, 100 mM KCl, 1% NP-40, and 1 mg/mL heparin. Store the solution at 4 °C until use (up to 1 month).
- Prepare the homogenization working buffer from the stock solution (step 2.1) freshly before use: Add DTT (final concentration: 1 mM), cycloheximide (final concentration: 100 µg/mL), recombinant ribonuclease (final concentration: 200 units/mL), and protease inhibitor cocktail (final concentration: 1x) to the homogenization stock solution to make the required amount of the homogenization working buffer. Store the freshly prepared working buffer on ice until use.
- Prepare the low-salt and the high-salt wash buffers:
- To prepare low-salt wash buffer mix 50 mM Tris (pH 7.4), 12 mM MgCl2, 100 mM KCl, and 1% NP-40. Add DTT (final concentration: 1 mM) and cycloheximide (final concentration: 100 µg/mL) before use.
- To prepare high-salt wash buffer mix 50 mM Tris (pH 7.4), 12 mM MgCl2, 300 mM KCl, 1% NP-40. Add DTT (final concentration: 2 mM) and Cycloheximide (final concentration: 100 µg/mL) before use.
- Prepare protein G beads (Table of Materials):
- Each sample will need 50 µL of protein G beads. Place the required amount of beads in a 1.5 mL centrifuge tube and separate the beads from the solution using a magnetic rack by leaving the tube on the rack for 30-60 s.
- Remove the supernatant by pipetting. Wash the beads three times with 1 mL of ice-cold low-salt wash buffer each time.
3. Tissue lysis and homogenization
- Add 2 mL of ice-cold homogenization working buffer (freshly prepared from step 2.2) to a glass tissue grinder set. Quickly place the frozen sample into the grinder and homogenize the tissue with 30 strokes on ice using a loose pestle.
- Transfer the homogenate to a 2 mL round-bottom tube and centrifuge at 12,000 x g for 10 min at 4 °C.
- Transfer the supernatant to a new 2 mL tube. Save 100 µL to a 1.5 mL tube as the "input".
- Incubate the supernatant in the 2 mL tube with the anti-GFP antibody (5 µg/mL; 1:400) at 4 °C on an end-over-end rotator (24 rpm) overnight.
- Transfer the homogenate/antibody mixture to a new 2 mL round-bottom tube containing the washed protein G beads from step 2.4. Incubate at 4 °C on an end-over-end rotator (24 rpm) for 2 h.
- Separate the magnetic beads from the supernatant using a magnet rack. Save the supernatant as the "negative fraction". The negative fraction contains (1) RNAs in EGFP-negative cells and (2) RNAs in EGFP-positive cells that are not bound to ribosomes.
- Add 1 mL of high-salt wash buffer to the beads and briefly vortex the tube to wash the beads. Place the tube in a magnet rack.
- Remove the wash buffer. Repeat the washing step two more times. The beads now contain the beads-ribosome-RNA complex from EGFP-positive cells.
5. RNA extraction
NOTE: The following steps are adapted from the RNA isolation kit (Table of Materials). Treat each fraction (i.e., input, positive, and negative) as an independent sample and isolate RNAs independently.
- Incubate the beads from step 4.4 with 50 µL of Extraction Buffer (from the RNA isolation kit) in a thermomixer (42 °C, 500 rpm) for 30 min to release RNAs from beads.
- Separate the beads with a magnet rack, transfer the supernatant which contains the beads-ribosome-RNA complex to a 1.5 mL tube.
- Centrifuge the tube at 3000 x g for 2 min, then pipette the supernatant to a new 1.5 mL tube. This tube contains the "positive fraction" of the TRAP step.
NOTE: For the input and the negative fractions, extract RNA from 25 µL of samples using 1 mL of Extraction Buffer. Incubate in a thermomixer (42 °C, 500 rpm) for 30 min.
- Pre-condition the RNA purification column: Pipette 250 µL of Conditioning Buffer onto the purification column. Incubate for 5 min at room temperature (RT). Centrifuge the column at 16,000 x g for 1 min.
- Pipet equal volume of 70% EtOH into the supernatant from step 5.3 (around 50 µL of 70% EtOH for the positive fraction and 1 mL of 70% EtOH for the input and the negative fractions). Mix well by pipetting up and down.
- Pipette the mixture into the column from step 5.4.
- Centrifuge the column at 100 x g for 2 min to allow RNA binding to the membrane in the column, then continue centrifuge at 16,000 x g for 30 s immediately. Discard the flow-through.
NOTE: For the input and the negative fractions, add 250 µL of the mixture to the column each time. Repeat steps 5.6 and 5.7 until all mixtures are used.
- Pipette 100 µL of Wash Buffer 1 (W1) into the column and centrifuge at 8,000 x g for 1 min. Discard the flow-through.
- Pipette 75 µL of DNase solution mix directly into the purification column membrane. Incubate at RT for 15 min.
- Pipette 40 µL of W1 into the column and centrifuge at 8,000 x g for 30 s. Discard the flow-through.
- Pipette 100 µL of Wash Buffer 2 (W2) into the column and centrifuge at 8,000 x g for 1 min. Discard the flow-through.
- Pipette 100 µL of W2 into the column and centrifuge at 16,000 x g for 2 min. Discard the flow-through. Re-centrifuge the same column at 16,000 x g for 1 min to remove all traces of wash buffer prior to the elution step.
- Transfer the column to a new 1.5 mL microcentrifuge tube.
- Pipette 12 µL of RNase-free water directly onto the membrane of the purification column. The pipet tip should not touch the membrane. Incubate at RT for 1 min and centrifuge at 1000 x g for 1 min. Then continue centrifugation at 16,000 x g for 1 min to elute the RNA.
6. RNA concentration and quality
- Use a bioanalyzer to assess the quality and quantity of the extracted RNA10.
7. Storage and further analysis
- Store the RNA at -80 °C (up to 1 year) until further analysis (e.g., microarray, quantitative PCR (qPCR), and RNA-seq, etc.).
NOTE: For details of the qPCR analysis, including cDNA synthesis, refer to Lyu et al.11. Primers for qPCR are listed in the Table of Materials.
Gli1-CreERT2 mouse (Jackson Lab Stock Number: 007913) were first crossed with the NuTRAP reporter mouse (Jackson Lab Stock Number: 029899) to generate double-mutant mice. Mice carrying both genetically engineered gene alleles (i.e., Gli1-CreERT2 and NuTRAP) were injected with tamoxifen once a day, every other day, for three injections. Tissue samples were collected on the 7th day after the 1st day of the injection. Immunofluorescence analysis showed that the EGFP was expressed in interstitial cells in testes (Figure 1). The adrenal gland capsule has been known to be another cell population positive of Gli112,13. EGFP was also found in adrenal capsular cells in Gli1-CreERT2;NuTRAP mice (Figure 1). Our lab also carries Shh-CreERT2;NuTRAP mice. Note that in Shh-CreERT2;NuTRAP mice, the EGFP+ cell population resides in the outer cortex of the adrenal gland underneath the capsule (Figure 1), the same expression site of EGFP+ cells in Shh+ area in the adrenal gland13 confirming the expression of EGFP in Cre-expressing cells.
After the extraction of the cell-type-specific RNAs, the quantity and quality of isolated RNAs in each fraction from two independent extractions (one testis used in each extraction) were assessed using a bioanalyzer (Figure 2). The bioanalyzer result indicated that this protocol is able to obtain high-quality RNAs from all three fractions. All fractions had a similar RNA Integrity Number (RIN).
To further test whether the extracted RNA is cell-type-specific, extracted RNA was sent for microarray analysis using a commercial microarray service (see Table of Materials). The microarray result showed that about 4,000 genes were enriched in the positive fraction comparing to the negative fraction, whereas about 3,000 genes were enriched in the negative fraction (Figure 3). Among these differentially expressed genes, Leydig-cell-associated genes Hsd11b1 and Hsd3b614,15 were enriched in the positive fraction. In the negative fraction, the Sertoli-cell-associated genes Dhh and Gstm616,17 were enriched. Only a few differentially expressed genes were identified when comparing the negative fraction with the input.
Real-time quantitative RT-PCR (qPCR) was also used to confirm the expression of key genes in the positive fraction and the negative fraction. Similar to what was found in the microarray assay, steroidogenic enzymes 3β-Hydroxysteroid dehydrogenase (encoded by Hsd3b) and cholesterol side-chain cleavage enzyme (encoded by Cyp11a1) were enriched in the positive fraction. Whereas the Sertoli cell marker Sox9 (SRY-box Transcription Factor 9), and the germ cell marker Sycp3 (Synaptonemal Complex Protein 3), were enriched in the negative fraction (Figure 4). Together these data demonstrated that the transcriptomes in Gli1+ cells were successfully pulled down and enriched by the above protocol.
Figure 1: Immunofluorescence images of the EGFP expression in NuTRAP reporter mouse models. The Gli1-CreERT2;NuTRAP and the Shh-CreERT2;NuTRAP mice were treated with tamoxifen to activate the EGFP expression. In the testis, Gli1+ cells were in the interstitium, whereas in the adrenal gland, Gli1+ cells were in the adrenal capsule. In the adrenal gland, cells underneath the capsule were positive of SHH (EGFP+ cells in Shh-CreERT2;NuTRAP mice). SHH is the ligand of the SHH signaling pathway which elicits its function in Gli1+ capsular cells13. Scale bars: 50 µm. Please click here to view a larger version of this figure.
Figure 2: RNA quality and quantity from the TRAP extraction. RNAs of the positive fraction, the negative fraction, and the input were evaluated using a bioanalyzer. The positive fraction contains RNAs extracted from protein G beads after the incubation with the GFP antibody (step 4.1). The negative fraction contains RNAs that remain in the supernatant at step 4.2. The input contains RNAs from the homogenate (step 3.3). In the electropherogram, because the concentrations of the lower marker (displayed as the first peak at 20-25 s of the migration time of samples shown on the X-axis) and the ladder (not shown in these electropherograms) are known, the concentration of each sample can be calculated. The two major peaks at 40-50 s represent the 18S and 28S rRNA. The ribosomal ratio (based on the fluorescence intensity shown on the Y-axis) is used to determine the integrity of the RNA sample. The RNA Integrity Number (RIN) of each sample is shown on the top right corner of each plot. Each dot in the dot plot represents the amount of RNA that was extracted from one single testis. The amount of RNA of each sample in the negative fraction and the input was extracted from 25 µL of samples. Please click here to view a larger version of this figure.
Figure 3: Microarray analysis for TRAP samples. Results of two extractions (one testis each) are shown. The microarray analysis identified a similar number of differentially expressed genes from two independent extractions (Testis A and Testis B). Around 4,000 genes were enriched (red dots) in the positive fraction, whereas ~3,000 genes were enriched (green dots) in the negative fraction. Hsd11b1 and Hsd3b6 were enriched in positive fractions. Dhh and Gstm6 were enriched in negative fractions. Only a few genes were identified as differentially expressed genes between the negative fraction and the input, suggesting the testis only contains a very small number of Gli1+ cells. Please click here to view a larger version of this figure.
Figure 4: qPCR analysis for TRAP samples. qPCR analysis showed the relative expression of cell-type-specific genes in the positive fraction and the negative fraction. The expression of each gene was first normalized with Actb. The relative expressions of genes within the positive fraction were then calculated based on the expression of Sox9 (set as 10). For the negative fraction, Cyp11a1 (set as 10) was used to normalize the relative expression. Note that the relative expression of target genes can only be compared within each fraction. Genes that encode the steroidogenic enzymes (Hsd3b and Cyp11a1) were enriched in the positive fraction. Whereas the marker gene for germ cells (Scyp3) and Sertoli cells (Sox9), were enriched in the negative fraction. Three biological replicates (three independent extractions, one testis each) were shown. Please click here to view a larger version of this figure.
The usefulness of the whole-tissue transcriptome analysis could be dampened, especially when studying complex heterogeneous tissues. How to obtain cell-type-specific RNAs becomes an urgent need to unleash the powerful RNA-seq technique. The isolation of cell-type-specific RNAs usually relies on the collection of a specific type of cells using micromanipulation, fluorescent-activated cell sorting (FACS), or laser capture microdissection (LCM)18. Other modern high-throughput single-cell collection methods and instruments have also been developed19. These methods usually employ the microfluidics techniques to barcode single cells followed by single-cell RNA-seq. Cell dissociation is the required step to obtain the suspended cell solution, which then will go through either a cell sorter or a microfluidic device to barcode each cell. The cell dissociation step introduces challenges to these methods for cell-type-specific studies because the enzymatic treatment not only breaks down tissues but also affects cell viability and transcriptional profiles20. Moreover, the expense for single-cell RNA-seq is usually high and requires specialized equipment on site.
Recently, two studies successfully isolated RNA/DNA of specific cell types from whole tissues using the NuTRAP mice8,21. Without using specific equipment and tools, the NuTRAP mouse model allows obtaining RNAs and DNAs from specific types of cells. The NuTRAP allele could target Cre-expressing cells without the cell dissociation step, avoiding changing cell's viability and transcriptional profiles. Rol et al. used the NuTRAP mouse model to isolate nuclei and translate mRNA simultaneously from adipose tissue. The other study also demonstrated that the NuTRAP mouse model could work for glial cells in the central nervous system8.
In our lab, we are interested in studying the stem cell populations in steroidogenic tissues such as Gli1+ interstitial cells in the testis22 and Gli1+ capsular cells in the adrenal gland13. The challenge of studying Gli1+ cells in these two organs is that the number of Gli1+ cells in the testis and the adrenal gland is small. For example, the proportion of Leydig cells, which are the major population of Gli1+ cells in the testis, only occupy about 3.8% of the total testis volume in adult mice23. Because the TRAP technique aims to specifically pull down translating ribosome-bound RNAs in complex tissue, the NuTRAP mouse model could be a powerful tool suitable for studying a rare cell population in a complex tissue. The previously published protocols using NuTRAP mice target adipocytes and glial cells that are more abundant in the brain and adipose tissue compared to Gli1+ cells in the testis and in the adrenal gland. To ensure obtaining required RNAs from a small number of cells in a complex tissue, we revised the existing protocols by (1) increasing the incubation time with the GFP antibody from 1 h to overnight; (2) using another type of RNA extraction kit which aims to isolate a small amount of RNA at a picogram level.
We demonstrated that this protocol could obtain high-quality cell-type-specific RNAs from a small number of cells in a complex tissue. The quality and quantity of extracted RNAs are capable for qPCR and a commercial microarray service. Results from microarray and qPCR confirmed that Leydig-cells-associated genes are enriched in the positive fraction coming from one testis. The protocol here provides a detailed approach to isolate cell-type-specific translating ribosome mRNAs using the NuTRAP mouse model. This protocol may also be used to isolate RNAs from any EGFP-expressing cells.
The authors declare no conflict of interest.
This work was partially supported by NIH R00HD082686. We thank the Endocrine Society Summer Research Fellowship to H.S.Z. We also thank Dr. Yuan Kang for breeding and maintaining the mouse colony.
|Actb||eurofins||qPCR primers||ATGGAGGGGAATACAGCCC / TTCTTTGCAGCTCCTTCGTT (forward primer/reverse primer)|
|Bioanalyzer||Agilent||2100 Bioanalyzer Instrument|
|cOmplete Mini EDTA-free Protease Inhibitor Cocktail||Millipore||11836170001|
|Cyp11a1||eurofins||qPCR primers||CTGCCTCCAGACTTCTTTCG / TTCTTGAAGGGCAGCTTGTT (forward primer/reverse primer)|
|dNTP||Thermo Fisher Scientific||R0191|
|DTT, Dithiothreitol||Thermo Fisher Scientific||P2325|
|DynaMag-2 magnet||Thermo Fisher Scientific||12321D|
|Falcon tubes 15 mL||VWR||89039-666|
|Glass grinder set||DWK Life Sciences||357542|
|Hsd3b||eurofins||qPCR primers||GACAGGAGCAGGAGGGTTTGTG / CACTGGGCATCCAGAATGTCTC (forward primer/reverse primer)|
|Microcentrifuge tubes 2 mL||Thermo Fisher Scientific||02-707-354|
|Mouse Clariom S Assay microarrays||Thermo Fisher Scientific||Microarray service|
|PicoPure RNA Isolation Kit||Thermo Fisher Scientific||KIT0204|
|Protein G Dynabead||Thermo Fisher Scientific||10003D|
|RNaseOUT Recombinant Ribonuclease Inhibitor||Thermo Fisher Scientific||10777019|
|Sox9||eurofins||qPCR primers||TGAAGAACGGACAAGCGGAG / CTGAGATTGCCCAGAGTGCT (forward primer/reverse primer|
|Superscript IV reverse transcriptase||Invitrogen||18090050|
|SYBR Green PCR Master Mix||Thermo Fisher Scientific||4309155|
|Sycp3||eurofins||qPCR primers||GAATGTGTTGCAGCAGTGGGA /GAACTGCTCGTGTATCTGTTTGA (forward primer/reverse primer)|
- Yang, K. C., et al. Deep RNA sequencing reveals dynamic regulation of myocardial noncoding RNAs in failing human heart and remodeling with mechanical circulatory support. Circulation. 129, (9), 1009-1021 (2014).
- Soumillon, M., et al. Cellular source and mechanisms of high transcriptome complexity in the mammalian testis. Cell Reports. 3, (6), 2179-2190 (2013).
- Lake, B. B., et al. Neuronal subtypes and diversity revealed by single-nucleus RNA sequencing of the human brain. Science. 352, (6293), 1586-1590 (2016).
- Heiman, M., Kulicke, R., Fenster, R. J., Greengard, P., Heintz, N. Cell type-specific mRNA purification by translating ribosome affinity purification (TRAP). Nature Protocols. 9, (6), 1282-1291 (2014).
- Bertin, B., Renaud, Y., Aradhya, R., Jagla, K., Junion, G. J. J. TRAP-rc, translating ribosome affinity purification from rare cell populations of Drosophila embryos. Journal of Visualized Experiments: JoVE. (103), e52985 (2015).
- Thellmann, M., Andersen, T. G., Vermeer, J. E. Translating ribosome affinity purification (trap) to investigate Arabidopsis thaliana root development at a cell type-specific scale. Journal of Visualized Experiments: JoVE. (159), e60919 (2020).
- Moran, P., et al. Translating ribosome affinity purification (TRAP) for RNA isolation from endothelial cells in vivo. Journal of Visualized Experiments: JoVE. (147), e59624 (2019).
- Roh, H. C., et al. Simultaneous transcriptional and epigenomic profiling from specific cell types within heterogeneous tissues in vivo. Cell Reports. 18, (4), 1048-1061 (2017).
- Varjosalo, M., Taipale, J. Hedgehog: functions and mechanisms. Genes & Development. 22, (18), 2454-2472 (2008).
- Mueller, O., Lightfoot, S., Schroeder, A. RNA integrity number (RIN)-standardization of RNA quality control. Agilent Technologies. Application Note 1 1-8 (2004).
- Lyu, Q., et al. RNA-seq reveals sub-zones in mouse adrenal zona fasciculata and the sexually dimorphic responses to thyroid hormone. Endocrinology. 161, (9), (2020).
- King, P., Paul, A., Laufer, E. Shh signaling regulates adrenocortical development and identifies progenitors of steroidogenic lineages. Proceedings of the National Academy of Sciences of the United States of America. 106, (50), 21185-21190 (2009).
- Huang, C. C., Miyagawa, S., Matsumaru, D., Parker, K. L., Yao, H. H. Progenitor cell expansion and organ size of mouse adrenal is regulated by sonic hedgehog. Endocrinology. 151, (3), 1119-1128 (2010).
- Benton, L., Shan, L. -X., Hardy, M. P. Differentiation of adult Leydig cells. The Journal of Steroid Biochemistry and Molecular Biology. 53, (1-6), 61-68 (1995).
- Monder, C., Hardy, M., Blanchard, R., Blanchard, D. Comparative aspects of 11β-hydroxysteroid dehydrogenase. Testicular 11β-hydroxysteroid dehydrogenase: development of a model for the mediation of Leydig cell function by corticosteroids. Steroids. 59, (2), 69-73 (1994).
- Bitgood, M. J., Shen, L., McMahon, A. P. Sertoli cell signaling by Desert hedgehog regulates the male germline. Current Biology. 6, (3), 298-304 (1996).
- Beverdam, A., et al. Sox9-dependent expression of Gstm6 in Sertoli cells during testis development in mice. Reproduction. 137, (3), 481 (2009).
- Gross, A., et al. Technologies for single-cell isolation. International Journal of Molecular Sciences. 16, (8), 16897-16919 (2015).
- Ziegenhain, C., et al. Comparative analysis of single-cell RNA sequencing methods. Molecular Cell. 65, (4), 631-643 (2017).
- Nguyen, Q. H., Pervolarakis, N., Nee, K., Kessenbrock, K. Experimental considerations for single-cell rna sequencing approaches. Frontiers in Cell and Development Biology. 6, 108 (2018).
- Chucair-Elliott, A. J., et al. Inducible cell-specific mouse models for paired epigenetic and transcriptomic studies of microglia and astroglia. Communications Biology. 3, (1), 693 (2020).
- Barsoum, I., Yao, H. H. Redundant and differential roles of transcription factors Gli1 and Gli2 in the development of mouse fetal Leydig cells. Biology of Reproduction. 84, (5), 894-899 (2011).
- Mori, H., Shimizu, D., Fukunishi, R., Christensen, A. K. Morphometric analysis of testicular Leydig cells in normal adult mice. The Anatomical Record. 204, (4), 333-339 (1982).