Abstract
Fluorescence in situ hybridization (FISH) combined with immunohistochemistry of tissue-specific markers provides a reliable method for characterizing the fate of somatic stem cells in transplantation experiments. Furthermore, the association between FISH and fluorescent gene reporter detection can unravel cell fusion phenomena, which could account for apparent transdifferentiation events. However, despite the widespread use of these techniques, they still require labor-extensive protocol adjustments to achieve correct and satisfactory simultaneous signal detection. In the present paper, we describe an improvement of simultaneous FISH and immunofluorescence detection. We applied this protocol to the identification of transplanted human and mouse hematopoietic stem cells in murine brain and muscle. This technique provides unique opportunities for following the path taken by transplanted cells and their differentiation into mature cell types.
T
To test these possible events, it is necessary to investigate the detection of donor cells along with the characterization of their phenotype. Donor cell tracing has been achieved through genetic labeling using either a gene reporter such as β-galactosidase (LacZ) (Alvarez-Dolado et al. 2003), a fluorescent protein such as green fluorescent protein (GFP) (Weimann et al. 2003b) and its spectral variants (Feng et al. 2000), or detection of DNA-specific sequences by fluorescence in situ hybridization (FISH) (Eglitis and Mezey 1997; Mezey et al. 2000,2003).
In several in vivo transplantation assays, a Y-chromosome-specific probe has been used to detect male cells transplanted into female recipients in both mice and in humans (Mezey et al. 2000,2003; Weimann et al. 2003a). Moreover, FISH for a specific human chromosome can distinguish human cells in xenotransplantation experiments (Liechty et al. 2000). In recent years, preclinical, functional in vivo models for human hematopoietic stem cell transplantation have been developed, using immune-deficient mice or preimmune fetal sheep as recipients. Donor cell engraftment and hematopoietic chimerism have been evaluated using FISH analysis for human cells (Lapidot and Kollet 2002). These studies have also indicated that human hematopoietic stem cells could differentiate into a multitude of non-hematopoietic cell lineages. Human hematopoietic cells (cord blood and bone marrow) were able to differentiate into hepatocytes when infused into non-obese diabetic–severe combined immunodeficient mice (NOD-SCID). FISH for mouse and human DNA was performed to demonstrate that human cells have the ability to engraft into NOD-SCID liver and become mature hepatocytes (Ishikawa et al. 2003; Newsome et al. 2003). A human mesenchymal stem cell population that was transplanted into fetal sheep engrafted and persisted in multiple tissues for as long as 13 months after transplantation. Transplanted human cells underwent site-specific differentiation into chondrocytes, adipocytes, myocytes and cardiomyocytes, bone marrow stromal cells, and thymic stroma. Skeletal muscle differentiation was demonstrated by staining with an antibody against human dystrophin combined with ISH for human ALU sequences (Liechty et al. 2000).
The combination of FISH and immunofluorescence staining of tissue-specific markers provides a highly specific method for characterizing the phenotype of donor cells in tissues. In addition, simultaneous detection of FISH and a gene reporter may help to unravel the dynamics of cell fusion events.
Despite the widespread use of these techniques, they still require labor-intensive protocol adjustments for correct and satisfactory analysis of both fluorescent signal from a gene reporter or from tissue-specific markers and FISH signal. In fact, the GFP signal is unstable to agents used for cell permeabilization and to the high temperatures required for the FISH protocol. In a previously described protocol combining GFP staining and FISH, most of the GFP signal was removed after the proteinase K digestion that is necessary in this specific FISH staining method (Weimann et al. 2003b). On the other hand, the FISH signal detection depends on adequate fixation and permeabilization, both of which must be selected for each tissue.
Here we describe an improved technique combining FISH detection and immunofluorescence staining of gene reporter and tissue-specific antigens. We applied this protocol to evaluate the effect of transplantation of human and murine hematopoietic cells into murine brain and skeletal muscle.
Materials and Methods
Mice
B6.Cg-TgN(Thy1-YFP)16Jrs mice (Jackson Laboratory; Bar Harbor, ME) that express spectral variants of GFP [yellow fluorescent protein (YFP)] at high levels in motor and sensory neurons and in subsets of central nervous system (CNS) neurons were used as donors of bone marrow cells. The transgenic construct contains a YFP gene under the direction of regulatory elements derived from the Thy1 mouse. These elements are composed of a 6.5-kb fragment obtained from the 5′ portion of the Thy1 gene, extending from the promoter to the intron following exon 4. Exon 3 and its flanking introns are absent. The deleted sequences are required for expression in non-neural cells but not in neurons. The remainder of the sequence is required for neuronal expression (Feng et al. 2000).
As recipient mice for brain transplantation we used 1-day-old wild-type C57BL/6J mice (Jackson Laboratory). Four-week-old CD-1 nude mice (Crl:CD-1-nuBR; Charles River Laboratory, Calco, Italy) were used as recipient animals in muscle transplantation experiments.
All animal experiments were performed according to institutional guidelines in compliance with national (D.I. no. 116, G.U. suppl. 40, Feb. 18, 1992, Circolare No. 8, G.U., 14 Luglio 1994) and international law and policies (EEC Council Directive 86/609, OJ L358, 1 Dec. 12, 1987; Guide for the Care and Use of Laboratory Animals, U.S. National Research Council, 1996).
Whole Murine Bone Marrow Transplantation into the Brain
Whole bone marrow (BM) was harvested from the limb bones of adult YFP mice as described (Corti et al. 2002a, b). YFP-BM cells were transplanted directly into the brain of neonatal mice. Sex-mismatched transplantation was performed by transplanting male BM cells into female recipients (n = 6). As control for Y-chromosome FISH specificity, female BM cells were transplanted into female recipients (n = 3).
Neonatal C57BL/6J mice were cold-anesthetized, and 2 μl of cell suspension derived from BM (3 × 104 cells) was transplanted using a Hamilton syringe into the brain through a hole in the cranium. The injection was performed in the parietal area, 1 mm caudal and lateral to the skull bregma point and 0.5 mm into the brain parenchyma from the dura. A control group of mice received 2 μl of saline solution with the same protocol (Bonilla et al. 2002). Mice were sacrificed 1 month after transplantation, and brain tissues were analyzed.
Human Hematopoietic Cell Culture
The mononucleated fraction of human healthy peripheral blood was collected through a density gradient (Ficoll/Hypaque; Sigma-Aldrich, St Louis, MO). The cells obtained at the interface by centrifugation (45 min at 1200 rpm) were removed and washed once with PBS. These cells were then separated by negative selection for the CD45 antigen using a magnetic separation column (MACS Miltenyi Biotec; Auburn, CA). The target cells were obtained by removing the magnetic field and were seeded into culture with GIBCO Dulbecco's modified Eagle's medium (Invitrogen; Carlsbad, CA) plus 15% fetal bovine serum (FBS) for 7–10 days.
Human Hematopoietic Cell Transplantation into the Mouse Brain
Human CD45-negative peripheral blood cells (5 × 104 cells) were transplanted directly into the brain of neonatal mice following the protocol described above for murine BM transplantation. Six neonatal C57BL/6J mice were transplanted. A control group of mice received 2 μl of saline solution under the same protocol (Bonilla et al. 2002). Mice were sacrificed 1 month after transplantation.
Human Hematopoietic Transplantation into Immunodeficient Mouse Muscle
Six CD-1 nude mice were transplanted with CD45-negative human peripheral blood cells (3 × 106 cells) by direct injection into the right tibialis anterior (TA) muscle. Mice were sacrificed by cervical dislocation 2 months after transplantation.
Brain Tissue Analysis
The animals were sacrificed, perfused, and fixed with 4% paraformaldehyde (PF) in PBS (pH 7.4). The brain was isolated, immersed in PF solution for 30 min and then in sucrose 20% solution in PBS (pH 7.4) for 30 min, and frozen in Tissue Tek OCT compound with liquid nitrogen. The tissues were cryosectioned (10 μm) and mounted on gelatinized glass slides. Cerebral tissue was cut in a coronal plane from the frontal lobes. All CNS sections were blocked with 1% FBS in PBS and permeabilized with 0.25% Triton X-100.
Thin brain sections (10 μm) from YFP-transplanted mice were incubated for 1 hr with anti-GFP antibody rabbit serum conjugated with Alexa 488 (1:400 dilution; Molecular Probes, Eugene, OR), followed by incubation with a tyramide signal amplification (TSA) kit horseradish peroxidase (HRP) goat anti-rabbit IgG and then Alexa Fluor 488 Tyramide (Molecular Probes) (Wu et al. 2000).
Brain sections of neonatal mice transplanted with CD45-negative cells were incubated for 3 hr with 1:200 anti-neuronal nuclei (NeuN, mouse monoclonal, biotin-conjugated; Chemicon, Temecula, CA), 1:200 anti-class III β-tubulin (TuJ1, mouse monoclonal; Chemicon), and 1:200 anti-neurofilament (NF, mouse monoclonal; Chemicon). Cy3-conjugated streptavidin (1:600; Sigma-Aldrich) was used for 30 min at room temperature for biotin-conjugated antibody, and an R-phycoerythrin (R-PE)-conjugated goat anti-mouse (1:100; DAKO, Carpentiria, CA) was used as secondary antibody in the other cases. Incubation with TSA kit HRP goat anti-mouse IgG and Alexa Fluor 488 Tyramide (Molecular Probes) was performed to amplify the immunoreaction.
As control, we performed immunostaining with amplification but without primary and/or secondary antibodies.
Immunohistochemistry on Muscle Tissue
The animals were sacrificed, and the right TA muscle was isolated and frozen in Tissue Tek OCT compound with liquid nitrogen. TA muscle was transversely cryosectioned at 10 μm with serial sections. Tissue slides were fixed with 4% PF for 5 min. Immunohistochemistry (IHC) was performed with a primary monoclonal dystrophin antibody raised against the human NH2 terminus (1:40 NCL-Dys3; Novocastra, Newcastle upon Tyne, UK). This antibody is human specific and does not recognize murine dystrophin (Cooper et al. 2001). Incubation with a TSA kit HRP goat antimouse IgG and then Alexa Fluor 488 Tyramide (Molecular Probes) was performed to identify the immune complex.
Immunoreaction with amplification but without primary and/or secondary antibodies was performed as control.
FISH in Murine BM Transplantation into the Brain
After immunoreaction, sections were rinsed with 2 × SSC (standard saline citrate stock solution) for 15 min and denatured in 70% formamide in 2 × SSC at 60C for 10 min. The Cy3 Y-chromosome probe (15 ng/μl; Cambio, Cambridge, UK) was directly placed on the sections, denatured at 37C for 10 min, and incubated at 60C for 10 min. The sealed slides were placed horizontally in a humid chamber and hybridized at 37C for 18 hr. Then the probe was washed off in 50% formamide/2 × SSC and in 2 × SSC for 5 min at 37C. The nuclei were then counterstained with DAPI.
Sections from female mice transplanted with female BM and sections from untransplanted female mice were used as negative controls for Y-chromosome staining.
FISH in Human HSC Transplantation into the Brain
After the immunoreaction, the slides were processed as previously described for BM transplantation. In this case, 8 ng/μl of the Cy3 human pancentromeric chromosome probe (Cambio), was directly placed on the sections (22 × 22 mm), denatured at 37C for 10 min, and incubated at 85C for 10 min. The sealed slides were placed horizontally in a humid chamber and hybridized at 37C for 18 hr. Then the probe was washed off in 50% formamide/2 × SSC and in 2 × SSC for 5 min at 37C. The nuclei were then counterstained with DAPI.
Brain sections from untransplanted mice were used as negative controls for human FISH staining.
FISH in Human HSC Transplantation into Muscle
FISH analysis was performed immediately after the immunoreaction on muscle sections treated with Histochoice Tissue Fixative for 30 min (Sigma-Aldrich). A Cy3 human pancentromeric-labeled probe (Cambio) was used to identify human donor cells in the TA skeletal muscles under examination. FISH analysis was processed as previously described for human HSC brain transplantation. The nuclei were counterstained with DAPI. Human muscle sections were used as positive controls and muscle sections from non-transplanted mice were used as negative ones. The muscle biopsies used in this study derived from the Telethon Bank of DNA, nerve and muscle tissues (GTF01001; Department of Neurological Sciences, University of Milan, Milan, Italy). All slides were observed with a conventional fluorescence microscope (Zeiss Axioskop; Oberkochen, Germany).
Results
Combined FISH and IHC after BM Transplantation into the Mouse Brain
The sensitivity and specificity of the combined FISH-IHC method were first verified using as control brain sections of male Thy1-YFP mouse expressing a spectral variant of GFP (YFP) in neurons. The co-expression of Thy1-YFP (green signal) in neurons and the Cy3-Y chromosome (red dot) in nuclei counterstained with DAPI occurred in over 90% of cells (Figure 1A). We transplanted BM cells derived from male Thy1-YFP mice into the parietal cortex of a female neonatal brain. One month after transplantation we observed rare YFP+ neurons in the recipient cerebral cortex (Figures 1B–1D). In transplanted brains, almost all the identified YFP+ cells contained a Y-chromosome signal, whereas none of the negative controls (female mice transplanted with female YFP-BM and untransplanted female mice) showed any FISH signal. Therefore, the method of detection had high sensitivity and specificity.
(
Activation of the Thy-1 neuron-specific transgene in the cortex of female neonatal mice transplanted with male YFP BM cells provides direct evidence for the acquisition of a neuronal phenotype by BM transplanted cells.
To obtain these results, before tissue freezing, the brain tissue was fixed with PF for 30 min to allow adequate fixation of YFP without compromising FISH analysis. Prolonged exposure to PF fixative abrogates detection of the Y-chromosome.
A TSA kit is commonly used to considerably increase the signal intensity in various immunocytochemical and FISH applications (Kerstens et al. 1995; Speel et al. 1999).
To maintain and protect the YFP signal from degradation during the FISH protocol, we used a 488 fluorochrome-labeled tyramide amplification signal. In addition, the lack of proteinase K digestion preserves the YFP signal, allowing the processing of both techniques—immunostaining and FISH—on the same slide.
FISH in Human HSC Transplantation into the Brain
We injected peripheral blood CD45-negative cells into neonatal mouse brain and 1 month later donor tissues were observed to evaluate the presence of human cells expressing neuronal markers. During FISH protocol optimization, we observed that human chromosomes are less sensitive to the duration of PF fixation. As described for murine BM transplantation, we performed the analysis of neuronal markers with TSA amplification followed by FISH.
A few human-derived cells, identified by the FISH signal in the nuclei and expressing two neuronal antigens (NF and MAP2), were found in recipient brains (Figures 1E and 1F).
FISH in Human Hematopoietic Cells Transplanted into the Muscle
To investigate whether human hematopoietic cells can participate in muscle generation in vivo, we transplanted CD45-negative human cells from peripheral blood into the TA of immunodeficient mice (CD1 nude mice). The mice were sacrificed 2 months after cell injection and the right skeletal TA was analyzed for human dystrophin expression by IHC with the TSA kit. After IHC the sections were incubated in Histochoice tissue fixative for 30 min. To determine if dystrophin-positive cells derived from the human injected cells, FISH analysis for human chromosomes was used. The percentage of Dys3-positive fibers with human nuclei positive for FISH ranged from 0.05 to 2%.
Furthermore, we observed the presence of human cells that are fused with endogenous fibers, as demonstrated by detection of donor human pancentromeric FISH-positive nuclei and FISH- negative murine nuclei in the same muscle fiber (Figure 1H). The analysis of human dystrophin by a specific Dys3 antibody showed that fused human nuclei were transcriptionally active and reprogrammed towards a muscle phenotype.
Discussion
In this work we have described two different protocols enabling the detection of FISH and immunofluorescence in experiments of human and murine hematopoietic stem cell transplantation into brain and skeletal muscle. These techniques have proved highly sensitive and specific, allowing the detection of donor cells that are integrated into the tissue along with reliable opportunities to follow the fate of transplanted cells.
After YFP-BM direct transplantation into the brain, we observed rare YFP+ neurons in recipient animals. FISH analysis for the Y-chromosome confirmed that these cells derive from male donors.
This report covers the research needed for a method that can achieve adequate simultaneous detection of a fluorescent gene reporter and FISH by using TSA amplification to preserve the YFP signal.
The fluorescence of GFP and its spectral variants is very sensitive to fixatives, high temperatures, and other treatments. TSA has been shown to considerably increase the signal intensity in various immunocytochemical and FISH applications (Kerstens et al. 1995). It employs peroxidase-catalyzed deposition of fluorochromized or haptenized tyramides for localized signal enhancement, combining the sensitivity features of fluorescence with the peroxidase enzymatic activity (van Gijlswijk et al. 1997).
To detect the transplanted cells in the brain slides, we developed a new optimized protocol based on a previously described procedure (Mezey et al. 2000,2003). Their protocol was applied to brain murine frozen sections to detect sex chromosomes along with neuronal protein without further use of any gene reporter such as GFP. Moreover, the FISH technique was performed by using “home-made” probes.
Other FISH methods combined with immunofluorescence for tissue-specific proteins were described with commercially available probes, but only on paraffin-embedded tissues and not on frozen tissues (Poulsom et al. 2001).
In our protocol we avoided the proteinase K digestion step. Protein digestion pretreatments, usually required for tissue FISH, significantly limit the ability to detect cell type-specific markers by IHC. Actually, protein K digestion can be omitted when applied to frozen sections (Mezey et al. 2000). Despite this observation, in an attempt to simultaneously detect GFP and FISH signals in BM donor-derived Purkinje neurons, Weimann and colleagues (2003a, b) had to perform extensive proteinase K digestion on frozen sections. The results shown illustrated adequate FISH signal detection, yet with an almost complete loss of GFP signal.
In this FISH protocol adjustment, we observed that one of the most important factors influencing detection of the Y-chromosome signal is the reduction of the PF fixation time. Human chromosomes have been demonstrated to be less sensitive to this step.
In muscle transplantation experiments we combined the detection of human dystrophin, revealed by the specific antibody Dys3, with FISH for human-specific DNA pancentromeric sequences.
Dystrophin staining must be performed on frozen muscle sections; however, a short fixation (5 min) with 4% PF is necessary to preserve an adequate nuclear morphology. A tyramide amplification kit is used to detect the primary antibody. After IHC, the slides are incubated with Histochoice tissue fixative. The use of Histochoice has already been suggested in the FISH protocol by Gussoni et al. (1997) to follow the fate of myoblasts after transplantation into the muscles of Duchenne patients. In our experience, the entire three-step procedure described above is crucial to obtain adequate tissue permeabilization and nuclear preservation for FISH analysis.
In these experiments, we described an efficient and easy way to combine FISH and immunofluorescence staining to track trace and characterize the fate of somatic stem cells after transplantation into brain and muscle tissues. Our methodology could contribute to future advances in stem cell-mediated therapy.
Footnotes
Acknowledgements
The financial support of the following research grants to G.P.C. is gratefully acknowledged: Italian Ministry of Health, Ricerca Finalizzata 2001 “Isolamento, espansione e caratterizzazione di cellule staminali a scopo di trapianto e riparazione cellulare”; MIUR (Ministero Istruzione Università e Ricerca Scientifica) Italian Ministery FIRB 2002 “Animali geneticamente modificati per lo studio di patologie neurodegenerative; Progetto a Concorso Ospedale Maggiore Policlinico 2003.” We wish to especially thank the “Associazione Amici del Centro Dino Ferrari” for their support.