Using Laser Scanning Cytometry to Measure PPAR-Mediated Peroxisome Proliferation and β Oxidation

Abstract

Laser scanning cytometry (LSC) is a new technology that combines the properties and advantages of flow cytometry (FC) and immunohistochemistry (IHC), thus providing qualitative and quantitative information on protein expression with the additional perspective provided by cell and tissue localization. Formalin-fixed, paraffin embedded liver sections from rats exposed to a Peroxisome Proliferator Activated Receptor (PPAR) agonist were stained with antibodies against peroxisomal targeting signal-1 (PTS-1) (a highly conserved tripeptide contained within all peroxisomal enzymes), Acyl CoA oxidase (AOX) (the rate limiting enzyme of peroxisomal β oxidation), and catalase (an inducible peroxisomal antioxidant enzyme) to evaluate peroxisomal β oxidation, oxidative stress, and peroxisome proliferation. The LSC showed increased AOX, catalase, and PTS-1 expression in centrilobular hepatocytes that correlated favorably with the microscopic observation of centrilobular hepatocellular hypertrophy and with the palmitoyl CoA biochemical assay for peroxisomal β oxidation, and provided additional morphologic information about peroxisome proliferation and tissue patterns of activation. Therefore, the LSC provides qualitative and quantitative evaluation of peroxisome activity with similar sensitivity but higher throughput than the traditional biochemical methods. The additional benefits of the LSC include the direct correlation between histopathologic observations and peroxisomal alterations and the potential utilization of archived formalin-fixed tissues from a variety of organs and species.

Introduction

The LSC is a microscope-based FC that can automatically measure fluorescence of cells on a microscopic slide (up to 5 different fluorochromes as well as light scatter) and may rapidly generate a huge database of biochemical and morphologic information for thousands of cells in a typical sample (Kamentsky et al., 2001). Because the measurements are microscope slide-based, the data generated by laser scanning cytometry combines the advantages of both FC and IHC by quantifying the total fluorescence per cell, while maintaining the morphologic features of the tissue examined with respect to the fluorescent intensity and localization. The position of each cell on the slide is recorded by the LSC so that any subpopulation of cells with particular features can be selected (relocation step) and examined to confirm that data histograms are correctly interpreted and allow correlation of cellular phenotype (biochemical and/or morphological) with patterns of tissue injury. Furthermore, a feature referred to as maximum pixel intensity (MPI) analysis allows for the measurements of punctate, highly localized biochemical events in a subcellular location that would be missed by flow cytometry integral analysis.

We exploited these features of the LSC to develop techniques to sensitively and specifically evaluate peroxisome proliferation and peroxisomal β oxidation in hepatocytes from rats exposed to a potent peroxisome proliferator (Moody et al., 1992). Peroxisomes are cellular organelles containing catalase (an inducible enzyme that ameliorates oxidative damage by neutralizing H₂O₂) and flavin-containing oxidases (such as those that facilitate β oxidation of fatty acids). Traditionally, IHC and immunoelectron microscopy (3′-methyl-2-methylamino-azobenzene reaction for catalase) have been used to evaluate increases in peroxisomal number and volume while biochemical methods (palmitoyl CoA and AOX mRNA assays) have been used to measure changes in peroxisomal β oxidation (Lazarow et al., 1981; Amacher et al., 1997).

In this manuscript, we describe the evaluation of peroxisomal proliferation and peroxisomal enzyme induction by LSC, using 3 specific peroxisomal markers; Acyl coA oxidase, catalase, and peroxisome targeting signal-1. Catalase is a well-established surrogate marker of peroxisomal β oxidation and proliferation which in reality measures the degree of oxidative stress occasioned by production of peroxide radicals during β oxidation (Nemali et al., 1988). Acyl CoA oxidase (AOX) is an inducible and rate-limiting enzyme of the β oxidation of medium, long, and very long-chain fatty acids within peroxisomes and provides a specific marker for peroxisomal enzyme induction (Osumi et al., 1980). Finally, all peroxisomal enzymes are directed to the peroxisomal matrix by virtue of a highly conserved COOH-terminal tripeptide, serine-lysine-leucine, known as the peroxisomal targeting signal type 1 (PTS-1). In contrast to proteins transported into the endoplasmic reticulum and mitochondria, posttranslational modification of peroxisomal proteins does not occur, allowing the use of probes such as PTS-1 to specifically target all peroxisome enzymes using immunohistochemistry (Gould et al., 1990; Usuda et al., 1999; Otera et al., 2002).

Material and Methods

Animals

Sprague–Dawley rats (220–250 g) were exposed to a potent PPAR agonist by oral gavage at doses of 1, 3, and 10 mg/kg for 30 consecutive days. Rats were fasted overnight before necropsy. At necropsy, livers were weighed; sections of right and left liver lobes were collected, fixed in 10% neutral buffered formalin for 24 hours and embedded in paraffin. Fixed tissues were either processed for light microscopic examination or for immunofluorescence and LSC. The remaining liver was snap frozen in liquid nitrogen for biochemical analysis. The treatment and handling of the study animals was done in accordance with the Pfizer institutional animal care and use guidelines.

Palmitoyl-CoA β-Oxidation

Frozen sections of liver were gradually thawed in Dulbecco’s PBS. Following the addition of cold 0.25 M sucrose, samples were minced, homogenized, and centrifuged at 600g for 10 minutes in a refrigerated rotor. Supernatants were frozen in an Ultralow freezer at –80° C degrees centigrade. On the day of assays, supernatants were thawed and assayed separately for protein content and palmitoyl CoA activity. Total protein was quantitated by the bicinchionic acid method. Following the addition of Triton X-100, each sample was mixed and aliquots transferred to the Cobas sample cup with reaction mixture and 0.5 mM palmitoyl CoA separately placed in the appropriate reagent cups. A 60-second incubation at 37° C was followed by 30 absorbance readings at 340 nm taken every 10 seconds in order to determine the change in absorbance (Δ Abs) per minute. For each sample, the resulting Δ Abs/min was multiplied by a calculation factor F to yield μmol L⁻¹ min⁻¹. This value was then multiplied by the homogenate dilution factor and divided by 1,000 g/L to yield the final sample palmitoyl CoA activity as μmoles NAD reduced/g liver x min.

Antibodies

As previously described, PTS1 peptide NH2-CRYHLKPLQSKL-COOH was synthesized both as a free peptide and conjugated to keyhole limpet hemocyanine (KLH) (Gould et al., 1990; Usuda et al., 1999). Rabbit polyclonal antibodies were raised against the conjugated PTS1 peptide and the serum with the highest titers was used for all subsequent immunofluorescence experiments (antibody titers varied from 1:4,900 to 1:100,000). The specificity of PTS1 antibodies for the peroxisomal matrix has been illustrated through immunoblot against peroxisomal proteins, immnunohistochemistry, and immunoelectron microscopy in multiple species and organs (Usuda et al., 1999). The antibody for AOX was purchased from Dr. Nobuteru Usuda and was prepared by immunizing rabbits with AOX purified from rat liver tissues. The specificity of this antibody has been previously described (Osumi et al., 1980; Farioli-Vecchioli et al., 2001).

Other antibodies were purchased, including mouse anti-catalase (Sigma, St Louis, MO), goat anti-rabbit Alexa 488, and goat anti-mouse IgG fluorescein isothiocyanate (FITC) (Molecular probes Inc., Eugene, OR).

Histopathology and Immunofluorescence

Serial 5-μm liver sections were cut from the paraffin blocks, deparaffinized, and rehydrated. For light microscopic examination, sections were stained with hematoxylin and eosin (H&E). For AOX and PTS1 immunofluorescence only, proteolytic digestion was first performed by incubating the slides for 12 minutes at 37° C in Pepsin (Dako Corporation, Carpeinteria, CA). For all staining, slides were then washed in OptiMax wash buffer (OWB)(Biogenex, San Ramon, CA) 3 times for 5 minutes. Nonspecific binding sites were blocked by the addition of 5% normal donkey serum (AOX and PTS1) (Jackson Immuno Research Laboratories Inc., West Grove, PA) or 5% normal goat serum (catalase) (Vector Laboratories Inc., Burlingame, CA) in OWB for 30 minutes at room temperature. Sections were then incubated at room temperature with either of the following antibodies, dilutions, and times: rabbit anti-AOX (1:3,000, 1 hour), rabbit anti-PTS1 (1:500, 2 hours), and mouse anti-catalase (1:250, 3 hours). Sections were washed in OWB 3 times for 5 minutes at room temperature. Secondary goat anti-rabbit Alexa 488 IgG (AOX and PTS1, dilution 1:200) or goat anti-mouse IgG FITC (catalase, dilution 1:200) conjugated antibodies were added and slides were placed for 30 minutes at 37° (AOX and PTS1) or at room temperature (catalase). Slides were washed 3 times for 5 minutes in OWB, rinsed in distilled water, and coverslipped in Vectashield mounting medium with 4′ ,6-diamidino-2-phenylindole (DAPI) (Vector laboratories, Burlingame, CA). Negative control slides were obtained by omitting the incubation with the primary antibody.

Analysis of Cell Fluorescence by LSC

Slides were inspected under an epifluorescence microscope (Olympus BX50, Olympus America Inc., Melville, NY) using a mercury lamp illumination. Fluorescence of a minimum of 20,000 hepatocytes (in previous unpublished studies, we have determined this sample size to be the minimum to produce statistically significant and reproducible results) per section of liver was measured using the LSC (CompuCyte Corp., Cambridge, MA) at 20X magnification. Briefly, FITC and DNA-associated DAPI fluorescence were excited with an argon ion laser (488 nm) and a violet diode laser (400 nm), and emission detection measured using the standard band pass 515–545 nm and long pass 460–485 nm filters, respectively. Individual hepatocytes were identified with nuclear DAPI fluorescence, and hepatocellular perimeters were selected at 10 pixels outside the nuclear contour on the basis of the FITC fluorescence (contour analysis). For each hepatocytes, the values of green and blue fluorescence intensities were measured and automatically processed to generate integral analysis (sum of all the fluorescent dots within each cell), MPI (maximal fluorescence intensity or “brightness” for every one of the fluorescent dots in the data set), XY coordinates of individual hepatocytes. Negative control slides were used to determine the level of background fluorescence and define gating prior to analysis. FITC-labeled hepatocytes were mapped to visualize tissue architecture (X-Y plot) and examined visually within the section (relocated) to confirm that data scattergrams were correctly interpreted. Fluorescein isothiocyanate or Alexa 488 MPI versus FITC-integral scattergram was displayed for each scan.

Results

Animals and Histopathology

At necropsy, group mean absolute and relative (to body weight) liver weights were increased at ≥ 1 mg/kg PPAR agonist (1.93–2.51x control mean) in both sexes and were generally similar across dose groups. Microscopically, the increases in liver weight correlated with mild-to-moderate hypertrophy of centrilobular hepatocytes.

Palmitoyl-CoA β-Oxidation

As shown in Table 1, exposure of both males and females to ≥ 1 mg/kg PPAR agonist for 30 days resulted in significant elevation of palmitoyl-CoA β oxidation without change in liver protein content compared to control rats. Values were generally higher in males than in females. Results were similar or greater in magnitude to values obtained for concurrently processed samples from rats dosed with 300 mg/kg/day clofibrate in a previous, unrelated study.

Immunofluorescence

In control rats, occasional centrilobular or midzonal hepatocytes of normal size showed intense AOX, catalase, and PTS1 expression that was characterized by a cytoplasmic, 0.5–1 μm in diameter, granular fluorescence consistent with peroxisomes (Figure 1a). In PPAR-exposed rats, there was a significant increase in immunofluorescence in centrilobular hepatocytes that sometimes extended out to the periportal areas in high-dose animals (Figure 1b). The hepatocytes were enlarged due to increased number of prominent fluorescent peroxisomes, and this observation correlated with light microscopic observation of centrilobular hypertrophy. Negative control sections incubated with only the secondary antibody were devoid of staining, indicating that the reaction was specific for the primary antibody.

Analysis of Cell Fluorescence by LSC

The FITC or Alexa 488-MPI data versus FITC or Alexa 488 integral scattergrams indicated that there was a significant dose-related increase in the percentage of AOX, catalase, and PTS-1 positive hepatocytes in all PPAR treated rats compared to control rats (Table 2). In control rats, the mapping (X-Y plot), scattergrams and relocation galleries (Figures 2 and 3) indicated that the rare hepatocytes with very high MPI and integral values were randomly located in the hepatic lobule or adjacent to centrilobular veins and contained a low number of fluorescent peroxisomes. In PPAR-treated rats, the large numbers of hepatocytes with very high MPI and integral values were centrilobular to midzonal. These hepatocytes were also enlarged by numerous and prominent fluorescent peroxisomes. In control rats, the LSC data are consistent with occasional individual hepatocytes presenting high levels of peroxisomal enzymes activity but no peroxisomal proliferation, whereas in PPAR-treated rats, the LSC data are consistent with PPAR-mediated induction of peroxisomal enzymes (β oxidation and antioxidant pathways) and peroxisome proliferation.

The results from LSC analysis were also comparable to the palmitoyl CoA assay. For example, there was almost no increase in palmitoyl CoA activity (μmoles NAD reduced/g/min) or in the percentage of PTS1 positive hepatocytes in rats at 1 and 3 mg/kg. Also, females were approximately half as responsive to PPAR agonists as males for both the biochemical assay and the LSC analysis. PTS1 analysis, compared to catalase and AOX, provided the best correlation with the biochemical assay. This result was expected due to the fact that PTS1 antibodies bind to all peroxisomal proteins instead of only one enzyme (such as AOX and catalase), thereby generating a stronger fluorescent signal.

Discussion

In this manuscript, we described the validation of the LSC for the measurement of peroxisomal enzyme induction and peroxisome proliferation in rats. The LSC provided qualitative and quantitative information on the expression of 3 critical peroxisomal markers of the β oxidation and antioxidant pathways, AOX, catalase, and PTS1. Acyl CoA oxidase expression, compared to catalase (oxidative stress enzyme), had the advantage of being more specific for peroxisomal fatty acid β oxidation. We also have shown in a recent toxicology study with a known peroxisome inducer that AOX expression also occurred at earlier time points (unpublished data). In this study, AOX and PTS1 were already increased 24 hours after treatment, whereas catalase was only increased at 96 hours, likely in order to remove the toxic species generated by the perosisomal oxidases. The LSC analysis of PTS1 expression seems to be more sensitive than catalase and AOX, probably because PTS1 antibodies bind to all peroxisomal proteins instead of only one enzyme. This sensitivity is especially advantageous for the detection of peroxisomal induction in non-rodent species, in control animals, and in organs other than the liver where basal expression of peroxisomal enzymes can be so low that it would not be detected by the biochemical assay and light microscopic evaluation would be unlikely to be successful in detecting an effect (data unpublished). Because it is based on immunostaining, the LSC analysis is also more sensitive than biochemical assays requiring whole-tissue measurements in which many hepatocytes would not exhibit induction. Together with the published information on immunoelectron microscopy, the hepatocellular cytoplasmic granular staining associated with the three antibodies and the predilection of the staining for centrilobular areas supported the specificity of these antibodies for the peroxisomal matrix since little or no leakage of the antigen (visualized as a diffuse cytoplasmic staining) occurred during the immunohistochemistry procedure.

Quantification of peroxisomal β oxidation by LSC directly correlated with the palmitoyl-CoA biochemical assay. However, when compared to the biochemical assay, the LSC analysis provided additional morphologic information on peroxisome proliferation and tissue pattern of activation. In toxicologic studies, the biochemical assay requires the sampling, during necropsy, of frozen tissue sections followed by several days of organelle isolation and protein measurements. On the other hand, measuring peroxisomal induction/proliferation by LSC can be completed within 48 hours (1 day of immunohistochemistry, 1 day of LSC analysis) and does not require special processing during necropsy other than the routine formalin fixed, paraffin embedded sections. Peroxisome analysis can therefore be performed retrospectively from archival formalin-fixed tissues or prospectively when the liver is a known toxicologic target. Also, when the expression of a marker is very limited to occasional cells, LSC being immunohistochemistry-based, becomes a very sensitive tool compared to protein or gene expression profiling because the later measures whole-tissue levels in which many cells do not exhibit induction.

In conclusion, quantification of peroxisomal β oxidation by LSC was sensitive, specific and relatively high throughput compared to the palmitoyl CoA biochemical assay. In addition, the LSC analysis provided further information on peroxisome proliferation, tissue patterns of expression, correlation with hepatocellular hypertrophy observed using light microscopy, and can be performed on archival material from a variety of organs and species.