Whole Mount Enzyme Histochemistry as a Rapid Screen at Necropsy for Expression of β-Galactosidase ( LacZ )–Bearing Transgenes: Considerations for Separating Specific LacZ Activity from Nonspecific (Endogenous) Galactosidase Activity

Abstract

Whole mount enzyme histochemistry to localize lacZ-bearing transgenes (lacZ-WMH) also detects endogenous β-galactosidases. The experiments reported here evaluated lacZ-WMH as a potential tool for transgene expression analysis during high-throughput rodent necropsies. A lacZ-WMH survey of organs from adult, wild-type, male and female mice (C57BL/6, FVB/N) and female rats (Sprague-Dawley) performed at the optimal pH (≥ 7.0) for bacterial lacZ yielded intense endogenous staining in the gonads, kidney, male accessory sex organs, salivary glands, submucosal glands in the duodenum, and thyroid. Substantial staining occurred in the adrenal cortex, lymph nodes, and linings of the gastrointestinal tract, the urinary bladder and uterus, and (for rat only) in the adenohypophysis, bone marrow, thymus, and trigeminal ganglia. Endogenous galactosidases were distributed similarly in sections of flash-frozen organs used for slide-based lacZ histochemistry (lacZ-SBH) at pH ≤ 5.0 (optimal for eukaryotic enzymes). Cerebral neurons were labeled only by lacZ-SBH. At pH 7.4, endogenous but not specific lacZ activity was abolished for lacZ-SBH, while endogenous activity was not halted without reducing specific activity for lacZ-WMH. These data demonstrate that lacZ-WMH is feasible during rodent necropsies for many but not all organs if species-, strain-, and sex-specific divergence in endogenous galactosidase activity is considered and special fixation (3% paraformaldehyde for 3 hours at 4°C) is used.

Introduction

Genetically engineered rodents are used increasingly to investigate molecular mechanisms of genetic diseases, to evaluate the efficacy of innovative therapeutic strategies on disease progression, and to screen drug candidates for potential toxicity. In addition to conventional morphological endpoints, the pathology assessment of such animals often includes in situ biochemical assays to detect tissue-specific and/or xenobiotic-modified expression patterns for the engineered gene. Common slide-based procedures to achieve this purpose are in situ hybridization and immunohistochemistry. These methods require substantial reagent preparation and tissue processing that often delay the evaluation for weeks. In contrast, use of “whole mount” histochemical stains at necropsy to assess expression of a functional transgene-derived protein in intact organs can greatly speed the analysis of new genes and novel gene-regulating agents.

One proven whole mount histochemical (WMH) technique suitable for this purpose uses the β-galactosidase gene (lacZ) of Escherichia coli as the reporter element for transgenic and gene-targeting constructs. The lacZ protein can be expressed in eukaryotic cells both in vitro and in vivo for extended periods without adverse effects on cell function (Murti and Schimenti, 1991; Tsien et al., 1996). This trait is desirable because the lack of adverse lacZ-associated behavioral and structural abnormalities allows for accurate identification of xenobiotic-induced changes in gene expression during efficacy and toxicity studies. The action of the lacZ enzyme on a suitable substrate results in deposition of insoluble indigo crystals in proportion to and at the site of transgene expression (Goring et al., 1987; Lojda, 1970). The enzyme histochemical assay for lacZ is rapid (1 to 24 hours), simple (4 steps), inexpensive, and works well with whole embryos, intact organs, and cryosections over a range of processing parameters (Alam and Cook, 1990; Mercer, 1995). Taken together, these attributes have made the lacZ procedure a standard system for examining gene expression in modern biomedical research.

A potential drawback of using whole mount lacZ histochemistry (lacZ-WMH) as a screen during transgenic rodent necropsies is its sensitivity, which differs among tissues because of the variable presence of endogenous β-galactosidase activity. Spurious lacZ-like activity has been described in many eukaryotic cell lines (An et al., 1982; Young et al., 1993) as well as in several tissues of wild-type (WT) mice (Abeliovich et al., 1992; Mercer, 1995; Shaper et al., 1994) and rats (Rutenburg et al., 1958; Sanchez-Ramos et al., 2000). Therefore, depending on the site to which a lacZ-bearing transgene has been targeted, nonspecific labeling because of endogenous galactosidase activity could prevent the detection of specific staining resulting from lacZ activity. The distribution and intensity of this background reaction has not been reported in the full battery of tissues that are assessed routinely during rodent phenotyping and toxicity bioassays. Thus, the current experiments were initiated (1) to define a reliable protocol for using lacZ-WMH as a rapid screen at necropsy for transgene expression in mammalian tissues and (2) to establish the distribution and intensity of endogenous galactosidase activity in organs of adults from mouse and rat strains commonly used in genetic engineering programs.

Materials and Methods

Study Design

The current work consisted of several sequential experiments (Table 1). In a pilot study (performed in duplicate), lacZ-WMH revealed intense staining in many tissues of age- and gender-matched lacZ-bearing (positive control) and WT, (negative control) mice. A different pattern of staining was observed in some tissues between the two replicate studies, suggesting that optimizing the assay conditions might ameliorate this background reaction. Therefore, a larger follow-up trial (mouse experiment 1) was completed to define a lacZ-WMH staining protocol that reduced or prevented the spurious staining. In the next study (mouse experiment 2), the pattern and intensity of the background reaction in whole mounts was analyzed in major organs of 2 mouse strains (C57BL/6N, FVB/N, n = 9/sex/strain, all at 3.5 months of age) that are routinely used in transgenic animal production. Finally, a small rat pilot test of lacZ-WMH was also performed with selected organs of three 4-week-old, female Sprague-Dawley rats (a strain used frequently for transgenic rat fabrication) so that the patterns of endogenous galactosidase activity in mice and rats could be compared.

Staining conditions selected for the present work were chosen to mimic the usual protocols described for lacZ-WMH. Four main variables were built into the current experiments: fixative composition; incubation times, ranging from 30 to 60 minutes (Gardner et al., 1996; Murti and Schimenti, 1991) up to 72 hours (Shaper, et al., 1994); incubation temperatures, between 22°C and 50°C (Jain and Magrath, 1991; Shaper, et al., 1994; Young, et al., 1993); and solution pH, ranging from 7.4 to 8.0 (Jain and Magrath, 1991; Weiss et al., 1999). These variables were chosen because they are most amenable to modification for a bench-side analytical procedure to be used during conventional high-throughput rodent necropsies.

Animals

These studies were conducted in accordance with federal animal care guidelines.

Young adult, WT mice and rats were obtained from Taconic (Germantown, New York). Animals were housed in filter-capped, polycarbonate, microisolator-type cages. Animals were housed 4 males or 5 females per cage for mice, or 2 per cage for rats. All cages contained autoclaved corncob bedding (W.F. Fisher & Sons, Boundbrook, New York). Animals received pelleted rodent chow (Purina Mills, Inc., St. Louis, Missouri) and filter-purified (to 0.2 μm), acidified (pH 3.0 ± 0.1) tap water ad libitum. Cages were kept in biologically clean rooms with HEPA-filtered air and a 12-hour light/dark cycle. Each cage was individually ventilated. Room temperature (RT) and relative humidity were maintained at 23 ± 2°C and 50%, respectively. Cages, bedding, and food were autoclaved before biweekly changes.

Two different positive controls were used in each experiment to confirm that the lacZ staining method was functioning correctly. The first was inclusion of intact organs from a young adult Rosa26 mouse (C57BL6/J genetic background; kindly supplied by Dr. Thomas Vogt, Princeton University). All tissues of Rosa26 mice bear a lacZ-containing transgene, which yields strong specific staining for lacZ at all stages of prenatal and postnatal development by both WMH and slide-based (SBH) lacZ histochemistry (Friedrich and Soriano, 1991). The second positive control was colon from each WT animal, as this organ contains large, bacteria-rich, intralumenal fecal pellets. Lung or skeletal muscle from WT mice served as negative controls based on their low levels of endogenous β-galactosidase activity (Gow et al., 1992; Shaper, et al., 1994).

Reagents

N, N-dimethylformamide (DMF) and paraformaldehyde (PFA) powder were bought from Aldrich Chemical Co. (Milwaukee, Wisconsin). All other fixatives were purchased from Electron Microscopy Sciences (Ft. Washington, Pennsylvania). The lacZ substrate 5-bromo-4-chloro-3-indolyl-β-D-galactoside (X-gal) and Dulbecco’s phosphate buffered saline (PBS) without Ca²⁺ or Mg²⁺ were ordered from Life Technologies, Inc. (Grand Island, New York). All other reagents were obtained from Sigma Chemical (St. Louis, Missouri).

Solutions and Incubations

Both lacZ-WMH and lacZ-SBH involved similar incubation steps using comparable solutions.

Buffers

Three buffer recipes (used at pH 7.4 unless noted) were tested as the basis for fixatives, washes, and staining solutions. PIPES (a common buffer used for lacZ-WMH [Mercer, 1995]) consisted of 0.1 M piperazine-N,N’-bis[2-ethanesulfonic acid] (PIPES), 2 mM MgCl₂, and 1.25 mM ethylene glycol-bis [β-aminoethyl ether]-N,N,N,’N’-tetraacetic acid (EGTA) in distilled water. PBS without Mg⁺⁺ was assessed because of its ready availability in many pathology laboratories as a component of immunohistochemical and molecular biological solutions; Mg⁺⁺ was added from a 1 M MgCl₂ stock for some experiments. HEPES buffer containing 0.1 M N-[2-hydroxyethyl]piperazine-N’-[2-ethanesulfonic acid] (HEPES), 1 mM MgSO₄, 5 mM dl-dithiothreitol (DTT), and 2% t-octylphenoxypolyethoxyethanol (Triton X-100) was tested at pH 7.4 or pH 8.0 because a biochemical assay performed with cell culture homogenates (Young, et al., 1993) suggested that this solution quenches endogenous galactosidases when used at a more basic pH.

Fixatives

Fixation protocols for lacZ-WMH vary among laboratories (Gardner, et al., 1996; Kapur et al., 1991; Login et al., 1987; Mercer, 1995; Sanes et al., 1986; Shaper, et al., 1994). The typical procedure in the research setting (brief [30-minute to 8-hour] immersion in freshly made 3% PFA in PIPES buffer at 4°C) is quite divergent from the conditions available in a high-throughput pathology laboratory (necropsy-grade neutral buffered 10% formalin [NBF] overnight [at least 16 hours] or longer at RT). In the present study, tissues were fixed in NBF (consisting of 3.7% formaldehyde in PBS, stabilized with less than 1% [research-grade] or approximately 15% [necropsy-grade] methanol) or 3% to 4% PFA (freshly made in buffer from powder or from a commercially available 16% concentrate) based on their ready availability. All fixatives were used at pH 7.4 except for some PFA aliquots at pH 8.0. For pilot studies, fixation temperature and length were varied between 4°C or RT (~22°C) and from 1 hour to overnight (18 hours), respectively, to determine the optimal conditions for bulk processing of whole mouse tissues. All tissues from a single animal were fixed together in 1 container (a screw-cap plastic specimen cup with 100 ml capacity).

Washes

Each buffer was spiked with 2 mM MgCl₂ alone (Wash #1) or with 2 mM MgCl₂, 0.01% (w/v) sodium deoxycholate, and 0.02% (v/v) Nonidet P-40 (Wash #2). For WMH procedures, fixed organs were washed twice for 30 minutes at 4°C or at RT. Washes for SBH sections were performed at RT for 5 minutes.

Stains

The lacZ staining solutions incorporate (1) the oxidation catalysts potassium ferricyanide (K₃Fe[CN]₆) and potassium ferrocyanide (K₄Fe[CN]₆); (2) the lacZ cofactor Mg⁺⁺; (3) the lacZ activator Na⁺; (4) 1 or more membrane-permeabilizing detergents such as sodium deoxycholate, Nonidet P-40 (NP40), or Triton X-100; and (5) a chromogenic lacZ substrate such as X-gal. In the present work, 5, 20, or 35 mM each of K₃Fe(CN)₆ and K₄Fe(CN)₆, 2 mM MgCl₂, 0.01% (w/v) deoxycholate, and 0.02% (v/v) Nonidet P-40 were dissolved in buffer. Selected tissues were stained for eukaryotic galactosidases by incubation in McIlvaine’s solution (Goring et al., 1987) containing 120 mM NaCl mixed with 3 mM each of K₃Fe(CN)₆ and K₄Fe(CN)₆ in distilled water. All staining solutions were made fresh daily by adding 1 mg X-gal/ml to each buffer. For both WMH and SBH procedures, washed tissues were incubated overnight at 37°C in the dark. Slides were incubated in a vertical orientation while completely submersed in staining solution.

Removal of Reaction Product

Organs with intense deposition of spurious reaction product were washed in graded ethanol solutions (75%, 90%, 100%) followed by either DMF (2 washes) or xylene (2 washes) to ascertain whether or not staining by endogenous enzymes could be removed without affecting lacZ-derived deposits. Washes were performed for 5 minutes at RT using 10 volumes of solution for each volume of tissue.

Tissue Acquisition

Animals were killed using carbon dioxide, and major organs (Table 2) were removed within 15 minutes. Instruments were rinsed in soapy water and wiped dry between necropsies to prevent contamination with fecal bacteria. Organs were blotted on absorbent paper to remove tissue fluid. Larger organs with dense parenchyma (e.g., brain, liver) were incised once to provide for more adequate penetration of fixative, wash, and staining solutions (Kapur et al., 1991; Mercer, 1995). Specimens for WMH analysis were immersed in 10 or more volumes of fixative and processed at various times (Table 3) within the next 24 hours. Samples for SBH assays were snap-frozen in dry ice-cooled isopentane, wrapped in aluminum foil, and then stored in airtight containers at −20°C until sectioning (typically 5 to 7 days).

Galactosidase Assays

LacZ-WMH was performed as outlined (Table 3). Staining intensity was assessed on a graduated scale as either absent, minimal, mild, moderate, or marked (i.e., so intense that tissue features were obscured by opaque deposits of dark blue reaction product). Selected tissues were postfixed overnight in NBF, dehydrated in graded ethanols and xylene, embedded in paraffin, sectioned at 6 μm, and then counterstained with hematoxylin and eosin (HE). Where necessary, bone was decalcified after staining and postfixation using 10% formic acid, as this procedure is reported to have no effect on lacZ localization (Kapur et al., 1991).

LacZ-SBH was adapted for use on cryosections from the WMH method to confirm the distribution of endogenous galactosidase activity in selected whole organs. Frozen sections were used because the reaction product was partially to completely removed during paraffin embedding (data not shown). Briefly, frozen tissues were surrounded by viscous mounting medium (Tissue Tek O.C.T., Sakura Finetek, Torrance, California), sectioned at 10 μm, applied to silanized slides (Polysciences Inc., Ft. Washington, Pennsylvania), and fixed in 4% PFA in PBS for 5 minutes at RT. After washing and staining (see above), sections were rinsed briefly in distilled water at RT and then counterstained for 30 seconds with 0.5% alcoholic eosin. Stained sections were rinsed twice for 5 minutes in distilled water, coverslipped with aqueous mounting medium (Fluoromount; Electron Microscopy Sciences), and examined by bright-field microscopy. Staining intensity was assessed using the tiered semi-quantitative scale used for WMH.

Results

Assessment of Processing Conditions for LacZ Histochemistry

Many combinations of conditions (Table 3) supported the detection of bacterial lacZ in control specimens (colon from WT mouse strains, and all tissues from Rosa26 mice). The most reproducible staining intensity (moderate to marked) was achieved using the following conditions: fixation in 3% PFA (freshly reconstituted from powder in PBS) at 4°C for 3 hours, buffer wash in PBS (with or without Mg⁺⁺) or PIPES at 22°C (i.e., RT) for 30 minutes, detergent wash in PBS or PIPES at 22°C for 30 minutes, and staining (with catalyst concentrations held at 35 mM) at 37°C for 6 hours. The best labeling was detected for solutions made in PIPES buffer. Substantial staining was also garnered if fixation was performed at 22°C for no more than 6 hours. The intensity of staining was not improved by using wash solutions kept at 4°C or 37°C. The stain intensity also was unaffected if samples were stained overnight at 37°C. Alternative means of fixation (either NBF or 4% PFA reconstituted from a commercial 16% stock solution) yielded mild to occasionally moderate staining of positive control organs. Staining was especially light if fixation was performed using necropsy-grade NBF.

The same processing conditions (Table 3) also effectively revealed the widespread expression of endogenous galactosidase activity in organs from WT mice (Table 4). Efforts to quench this unwanted activity (e.g., use of HEPES buffer at pH 8.0, with staining performed at 42°C or 50°C [Young et al., 1993]) were partially successful in that use of HEPES-based solutions reduced the intensity of spurious labeling by 1 to 2 grades (e.g., from marked to mild), even in such galactosidase-rich organs as the epididymis, kidney, and salivary gland. A lacZ-WMH staining protocol under which the superfluous activity of the eukaryotic enzymes could be reliably extinguished without affecting the detection of specific bacterial lacZ was not identified.

Reaction product could be effectively removed from tissues stained by lacZ-WMH using either DMF or xylene, but both spurious (stemming from endogenous enzymes) and specific (lacZ-derived) deposits were eliminated. The extent of color loss was greater for DMF. Ethanol alone did not ablate the staining, a finding consistent with the conventional practice of holding stained specimens in 70% ethanol for long-term storage (Mercer, 1995). If decolorized tissues were restained, background labeling from endogenous enzyme activity was much stronger than lacZ-derived staining.

When organs stained by WMH were embedded in paraffin and sectioned, reaction product was confined to cytoplasmic granules in many epithelial cells and some tissue macrophages (particularly osteoclasts). The cellular localization of the WMH reaction product was partially to completely disrupted when tissues were processed into paraffin (Figure 1); a qualitatively similar compartmentalization for endogenous galactosidase activity was observed in cells of sections from flash-frozen organs processed for lacZ-SBH using acidic solutions (pH 4.0 to 5.5, the preferred range for the eukaryotic enzyme). In addition, sections of flash-frozen mouse brain stained by SBH revealed cytoplasmic labeling in scattered cerebral neurons. The background reaction was abolished in all tissues when SBH was performed at physiological pH (7.4, the optimal state for bacterial lacZ).

Assessment of Endogenous Galactosidase Activity in Tissues of Wild-Type Rodents

The lacZ-WMH protocol that facilitated the optimal detection of bacterial lacZ also revealed pervasive activity of endogenous galactosidases. Spurious staining resulting from the activity of eukaryotic enzymes appeared as moderate to marked labeling of epithelial elements in many organs from both C57BL/6N and FVB/N WT mice (Table 4). Portions of the intra-abdominal digestive tract (Figure 1) and associated salivary glands (Figure 2) as well as the urogenital system (Figure 3) had the most intense reactivity. In both sexes of both strains, diffuse and moderate to marked staining occurred in the renal cortex (particularly along the corticomedullary junction; Figure 3), salivary glands (mandibular [Figure 2] as well as buccal elements near the base of the tongue), submucosal (Brunner’s) glands and to a lesser extent crypt epithelium of the duodenum (Figure 1), colonic mucosa, and thyroid gland (Figure 2D). The labeling pattern differed if organs with dense parenchyma (e.g., kidney, mandibular salivary gland) were bisected before immersion in the staining solution; for intact organs, staining was evident for only a few millimeters of the outermost tissue (Figures 2C and 3A), even if the capsule had been removed, while in cut organs, labeling extended into the interior (e.g., the entire renal cortex or all major ducts of the salivary gland). Additional gender-specific sites of diffuse, marked labeling that were observed in mice of both strains included the coagulating (Figure 3) and preputial glands, epididymis (Figure 3), and ventral prostate of males and the ovaries in females. Interestingly, the clitoral glands of female mice were not labeled.

For several organs, the distribution of endogenous galactosidase activity by lacZ-WMH differed between the 2 sexes of a given mouse strain and/or for animals of the same sex between the 2 mouse strains (Table 4). For example, the vas deferens was markedly and diffusely stained in all FVB/N males but was unlabeled in C57BL/6N males. Both the coagulating glands (diffuse labeling) and seminal vesicles (multifocal labeling, rather than the diffuse pattern seen in all other organs) were stained in males of both strains; however, the intensity was moderate to marked in almost all FVB/N mice, but minimal to mild when seen in just a few of the C57BL/6N animals. More than half of the female mice from both strains had mild labeling of mesenteric lymph nodes and urinary bladder mucosa, while few to no males exhibited labeling at these same sites. In general, organs in C57BL/6N males were stained less frequently than were the corresponding tissues of C57BL/6N females or FVB/N animals of either sex. An exception to this trend was noted for the Harderian gland, in which minimal staining was evident in almost all C57BL/6N animals of both sexes, while mild staining occurred in only 22% of FVB/N mice of both sexes.

In mice, the results of SBH and WMH agreed for most tissues (Figures 1, 2, and 3). The major exception was the brain. No endogenous labeling was evident in any brain structure by WMH, even those located on the surface of the coronal cut made to bisect the cerebrum before staining. In contrast, several cerebral regions at this same coronal level (as defined in Paxinos and Watson, 1997) exhibited labeling by SBH. Moderate to marked staining was observed in the cytoplasm of most neurons in the hippocampus (CA1, CA2, CA3, and dentate gyrus), the superficial layers of the piriform cortex, the ventromedial hypothalamic nucleus, and the choroid plexus and ependymal lining of the third ventricle. Minimal to mild labeling was also evident in several thalamic nuclei: centromedial, ventrolateral, ventroposterolateral, and ventroposteromedial. Neither the white matter nor the meninges contained labeled cells.

Intriguingly, the pilot study for lacZ-WMH in female rat tissues revealed substantial endogenous galactosidase activity in many organs (Table 5). Furthermore, the pattern of staining was more widespread than that seen in organs of female mice by lacZ-WMH (Table 4). Major examples of this divergence included more intense staining of certain rat organs relative to mouse organs (e.g., the adrenal cortex, lymph node, urinary bladder, uterus) and the presence of pronounced staining in some rat organs that was not evident in the corresponding mouse sites (e.g., bone marrow in metaphyses of long bones [resulting from osteoclasts’ lining bony trabeculae], trigeminal ganglia [neurons], liver [hepatocytes], adenohypophysis of the pituitary gland, trachea [mucosa], thymus). However, in some instances, staining was more widespread in female mice than in rats (e.g., Harderian gland, mandibular salivary gland).

Discussion

Several exogenous reporter proteins have been incorporated into fabricated genes to allow screening for transgene expression in mammalian tissues, including E. coli-derived β-galactosidase (lacZ), alkaline phosphatase (ALP), β-glucuronidase, choline acetyltransferase (CAT), firefly luciferase, green fluorescent protein (GFP), and lactamase (Cui et al., 1994; Spergel et al., 2001; Wood, 1995). LacZ was chosen for the present study because it has a long history of use for this purpose based on the fast, easy, and inexpensive nature of the detection system and its demonstrated lack of toxicity even after long-term in vivo overexpression (Murti and Schimenti, 1991; Tsien, et al., 1996). However, despite the divergence between the mammalian and bacterial enzymes (human and murine β-galactosidases are 80% homologous at the amino acid level but retain only 20% homology with the E. coli variant; Nanba and Suzuki, 1990), a common problem encountered in using lacZ-WMH to evaluate transgene expression in mammalian specimens is its inconsistent sensitivity among tissues because of the variable presence of endogenous galactosidase activity. The current experiments were launched (1) to identify a dependable protocol for using lacZ-WMH as a rapid screen at necropsy for transgene expression in tissues of engineered rodents and (2) to characterize the pattern of endogenous galactosidase activity in organs of adult rodents from several strains commonly subjected to phenotypic analysis in drug discovery and development programs.

Method Development

The standard lacZ-WMH method is well characterized (Alam and Cook, 1990; Goring, et al., 1987; Lojda, 1970; Mercer, 1995). However, the recommended assay conditions in published research reports do not seem to be readily adaptable to the rapid tempo required in high-throughput drug discovery and development programs. Thus, the initial thrust of the present study was to adjust the fixation procedure, ideally to incorporate immersion in NBF at RT (22°C). This endeavor was deemed feasible because typical WMH protocols to detect bacterial lacZ require samples to be fixed by immersion in buffered aldehyde solutions (usually PFA) for 5 minutes to a few hours (Gardner et al., 1996; Sanes et al., 1986). However, such aldehyde treatment is usually performed at 4°C because fixation at RT can lead to reduced WMH staining (Mercer, 1995).

The discriminating power of lacZ histochemistry can be enhanced by modifying conventional lacZ-WMH processing conditions to better differentiate between activities of the specific (bacterial lacZ) and nonspecific (eukaryotic) β-galactosidases. The prokaryotic and eukaryotic enzymes differ in their need for divalent and monovalent cations, thermostability, and preferred pH range (Shimohama et al., 1989; Young, et al., 1993). Bacterial lacZ is stabilized by Mg⁺⁺ and is active at higher temperatures and more basic pH, so the signal-to-noise ratio for detecting lacZ in eukaryotic tissues laden with endogenous galactosidases can be improved by adding Mg⁺⁺ to most assay solutions, heating the staining solution to 50°C (Young et al., 1993), and carrying out the WMH reaction at pH 8.0 (Jain and Magrath, 1991; Young et al., 1993). The addition of protease inhibitors (Shaper et al., 1994) or D(+)-galactose (Hendrikx et al., 1994) to deactivate the eukaryotic enzyme also has been shown to improve the discriminatory power of lacZ-WMH in vitro. The second portion of the present study tested some of these modifications for their ability to improve the outcome of lacZ-WMH in rodent tissues.

Organs in which the gross distribution of lacZ has been localized using WMH may be subjected to additional microscopic analysis to better define the presence of a lacZ-bearing gene in specific cell types and cellular compartments. One option is to routinely process the prestained WMH specimens into paraffin (Gardner et al., 1996), although care must be taken to avoid long incubations in solvents (e.g., toluene, xylene), which can effectively remove the lacZ staining product from tissue (Lazik et al., 1996; Mercer, 1995). Prestained bony samples can be safely decalcified in 10% formic acid before processing without affecting the intensity of lacZ-WMH (Kapur et al., 1991). Another refinement is to use an ethanol/polyethylene glycol fixative followed by processing of unstained tissue into a special low-temperature paraffin to prevent deactivation of the enzyme, after which staining can be conducted on cut sections (Avé et al., 1997). These manipulations were not examined in the present experiments as their inclusion would not affect the ultimate performance of a tableside lacZ-WMH assay during high-throughput necropsies of transgenic rodents.

Specific lacZ Activity

The current data indicate that optimal detection of specific lacZ activity by lacZ-WMH in positive control specimens (colon from WT animals and all tissues from Rosa26 [lacZ-bearing] mice) with minimal eukaryotic-derived artifact requires fixation using 3% PFA for a limited time (3 hours or less) at 4°C and pH 7.4. LacZ staining intensity was also adequate if PFA fixation was performed at 22°C (RT). Maximal intensity of tissue staining was achieved in 6 hours. Overnight incubation did not enhance staining, although the staining solution itself progressively developed a blue hue (because of the migration of functional endogenous galactosidases from the tissue into the fluid; Rutenburg et al., 1958). Regardless of fixation time or temperature, use of NBF resulted in reduced and inconsistent specific lacZ activity, findings that mirrored those described in prior reports (Mercer, 1995; Rutenburg et al., 1958); the decrease in specific staining intensity was most pronounced if necropsy-grade NBF (i.e., containing 15% methanol as a stabilizer) was used. As demonstrated previously (Sanchez-Ramos et al., 2000), the best staining was obtained using solutions made in PIPES buffer, a result thought to arise because EGTA inhibits the activity of endogenous galactosidases. Neither PFA nor NBF fixation quenched the activity of the endogenous galactosidases.

Nonspecific (Endogenous) Galactosidase Activity

Unfortunately, the combinatorial experiments performed in the present study identified assay conditions that could reliably quench endogenous galactosidase activity during lacZ-WMH in some but not all tissues. Activity of endogenous β-galactosidase is doused at temperatures greater than 45°C and solution pH greater than 7.4, particularly in cells with high endogenous levels of enzyme activity (Young et al., 1993). In like manner, buffer composition has been suggested as a means of reducing the activity of endogenous galactosidases; solutions based on HEPES (pH 7.0) but not PBS or PIPES have been shown to significantly inhibit the eukaryotic enzymes (Young et al., 1993). Interestingly, the current experiments did not confirm this effect for higher incubation temperatures (42°C or 50°C), more basic pH (8.0), HEPES buffer, or for any or all of these factors in combination. This discrepancy likely reflects major differences in sample structure, as the present negative data were acquired in whole organs, while successful ablation of endogenous enzyme activity using high temperature or pH or special buffer was obtained in tissue extracts or isolated cells (Jain and Magrath, 1991; Young et al., 1993).

In general, apparent distributions of nonspecific (and also specific) galactosidase activities were similar for organs stained at necropsy by lacZ-WMH and tissue sections acquired from flash-frozen organs stained by lacZ-SBH (Figure 1). The sole exception was the brain, in which staining associated with endogenous activity was only seen using SBH. One feasible explanation for this discrepancy in brain localization of endogenous galactosidases between the SBH and WBH protocols is that the dense neuropil prevented penetration of the staining solution during WBH; another possible alternative is that the number and size of neurons that have cytoplasmic galactosidase activity are too small to deposit reaction product in quantities that can be seen by the naked eye. Brains in which a lacZ-bearing transgene is widely expressed (e.g., those of the positive control Rosa26 mice used in the current studies) do exhibit robust lacZ staining (data not shown), thereby indicating that the second explanation for this result has more credence than does the first.

Distribution of Endogenous Galactosidase Activity in Mammals

Eukaryotic tissues contain several indigenous β-galactosidases (Abrahams and Robinson, 1969; Chytil, 1965; D’Agrosa et al., 1992; Furth and Robinson, 1965; Ho and O’Brien, 1969; Hubbes et al., 1992). Such endogenous β-galactosidases are widely distributed in rodent tissues by lacZ-WMH. Background labeling by lacZ-WMH occurs with particular intensity in the colon, epididymis (Figure 3), kidney (Figure 3), pancreas, salivary gland (Figure 2), spleen, testis (Figure 3), and thymus (Abeliovich et al., 1992; MacGregor et al., 1991; Mercer, 1995; Nowroozi et al., 1998; Palmiter and Brinster, 1986; Shaper et al., 1994). Prominent but less intense background staining has also been described for lacZ-WMH in the lacrimal gland, ovary, and mucosal layers of the stomach, small intestines (Figure 1), and proximal large intestines (Mercer, 1995). The current lacZ-WMH data (Table 4) concur with this endogenous activity pattern for many but not all of these organs; in particular, during the present experiments, labeling attributable to endogenous galactosidase activity was observed in the thymus of female rats but was absent in mice, while the pancreas and spleen were not labeled in either mice or rats. In addition, distinct endogenous galactosidase activity was also evident in some endocrine tissues (adrenal gland, thyroid), male accessory sex organs, Harderian gland, urinary bladder, and uterus. Faint background staining by lacZ-WMH has been reported in the choroid plexus, neonatal bone, and the endolymphatic diverticular appendage of the otic vesicle (E11.5 to 13.5 mouse embryo, strain unspecified; Mercer, 1995) and was observed in the current work (mouse experiment 2) in the choroid plexus of the fourth ventricle in 14.5-day-old FVB/N embryos (as conceptuses were serendipitously present in the uterus of one FVB/N female [data not shown]). Results obtained for a given tissue among various laboratories can be contradictory (e.g., ovary; compare Shaper et al. [1994] vs. Mercer [1995]). Intriguingly, some organs with quantifiably high endogenous levels of β-galactosidase activity (brain, liver; Gow et al., 1992) exhibit no labeling by lacZ-WMH (current data and Abeliovich et al., 1992). A sound explanation for this phenomenon has not been ascertained.

Interestingly, the pattern of endogenous galactosidase activity in tissues of WT mice varies by strain. One facet of this strain divergence involves the distribution of background labeling in different organs. For example, C57BL/6J × A/J mice are reported to express endogenous β-galactosidase by lacZ-WMH in the epididymis but not in the brain, heart, liver, lung, kidney, spleen, skeletal muscle, testis, duodenum, uterus, and ovary (Shaper et al., 1994); epididymal staining has also been reported in FVB/N mice (Langford et al., 1991). In contrast, the current lacZ-WMH work with C57BL6/N and FVB/N animals revealed clear endogenous galactosidase activity not only in the epididymis but also in the kidney, duodenum, uterus, and ovary (Table 4). Another aspect of this strain variance entails the intensity of background labeling in a given organ. For instance, C57BL/6J mice are reported to have 2-fold to 5-fold greater activity in their tissues than do DBA mice, with the extent varying by tissue (Hara et al., 1994). In the present WMH study, accessory sex glands and associated ducts as well as adrenal cortices in almost all FVB/N males were intensely labeled, while the same structures in C57BL6/N males exhibited limited labeling (in only a few animals) or no labeling at all (Table 4). In like manner, almost all C57BL/6N mice of both sexes exhibited minimal staining of Harderian glands, while only a few of their FVB/N counterparts developed mild labeling of the same tissue. To date, a reason that explains the differing localization of endogenous galactosidase activity in identical tissues from distinct mouse strains has not been defined. One possibility is that the galactosidase proteome in some mouse lineages has evolved rapidly as an adaptive response to a strain-specific environmental stress. However, this prospect seems unlikely given the relatively recent divergence of such strains from the standard Mus musculus genetic background. Another option is that the apparent variation in endogenous galactosidase activity reflects subtle changes in the assay conditions among laboratories. This latter alternative, if true, underscores both the need for great attention to detail in performing and interpreting lacZ-WMH assays as well as the need for further biochemical studies to better define the strain-specific quantities and distribution of endogenous galactosidases among major mouse strains.

In like manner, the pattern of endogenous galactosidase activity in tissues of WT mice assayed by lacZ-WMH varies by gender (Table 4). For example, preputial glands from both C57BL6/N and FVB/N males were prominently labeled, while clitoral glands (the female analog of the preputial gland) were not labeled in either strain. In contrast, more female C57BL6/N and FVB/N mice exhibited background labeling in the mesenteric lymph nodes and urinary bladder. The obvious rationale, a divergent hormonal milieu, may play a role in such discrepancies, although the exact nature of the sex-specific mechanisms remains to be determined.

Rat tissues also harbor rich endogenous β-galactosidase activity, predominantly in epithelial cells (Rutenburg et al., 1958). Particular sites of high expression are kidney, liver (chiefly centrilobular domains), male genital tract (especially epididymis; Tulsiani et al., 1995), stomach, leukocytes, salivary gland, thyroid follicular cells, and submucosal (Brunner’s) glands in the duodenum. Endogenous galactosidase activity is also reported in neurons, glia, and endothelial cells in the brainstem, and to a lesser degree, in other brain regions (Rosenberg et al., 1992). Our present pilot data in young adult, female Sprague-Dawley rats (Table 5) confirm this description in some organs (e.g., kidney, liver, duodenum, buccal salivary glands) but not others (e.g., mandibular salivary glands). The possible explanations brought forth above to account for such variations in mice—differences in strain (which have generally not been specified in prior reports on lacZ-WMH in rat tissues) or the assay conditions among laboratories—would also apply to rats. Given the paucity of rat studies with lacZ-WMH, the most efficient way to quickly discern endogenous galactosidase activity patterns in rats would be to perform a full-fledged WMH survey of major organs from those rat strains that are commonly used in genetic engineering programs.

Interestingly, the pattern of endogenous galactosidase activity in rodent tissues also differs by species. For instance, a comparison of lacZ-WMH data generated for female mice (Table 4, footnote) and female rats (Table 5) in the current study demonstrates that the aorta, liver, pituitary gland, thymus, and trigeminal ganglia exhibit no background staining in the 2 mouse strains, while they are labeled in the WT Sprague-Dawley rats. In contrast, certain other organs (e.g., duodenum, kidney, ovary, stomach, urinary bladder, and uterus) were stained in females of both species; the staining intensity in the rats (Table 5) was sometimes more intense than that observed in mice (Table 4). One potential explanation for these observations is that the size of the organs placed in the staining solution affects the efficiency of the X-gal to the colored product. The rat organs would be expected to exhibit greater labeling for a given staining period because of their larger volume (i.e., greater endogenous galactosidase content) and higher surface area over which the enzyme can encounter the staining solution. Another possible explanation is that the larger size of rat organs allows the internal tissue pH to decline to the level (≤ 5.0) for optimal activity of the eukaryotic enzyme, while the smaller mouse organs fix before the tissue becomes acidified. However, this latter line of reasoning cannot explain the current finding that endogenous staining of some organs was more widespread in mice than rats (e.g., Harderian gland, mandibular salivary gland), an observation that suggests the existence of real differences in the distribution and/or the amount of endogenous galactosidases between homologous tissues in mice and rats. Prolonged incubation (36 to 48 hours) of mouse tissues in lacZ staining solution does not increase the labeling intensity, even if the tissues are periodically transferred to fresh solution (unpublished data), presumably because the enzymes become inactivated as energy reserves in the isolated organs become depleted.

The problem of endogenous galactosidase activity as a potential confounding factor in trying to detect lacZ-bearing trans-genes in tissues of engineered rodents is magnified by certain aspects of current molecular biological technology. For example, the level of specific lacZ expression depends on the transgene integration site and the associated tissue-specific enhancer elements (Cui, et al., 1994). In addition, expression of the lacZ reporter in adult tissues can be unpredictable even if lacZ production driven by the identical promoter (even a ubiquitous one such as actin) is good in embryonic tissues (Cui et al., 1994). Finally, detection of some ectopic lacZ expression is a relatively common occurrence (Cui et al., 1994). These features of the lacZ-WMH assay provide the background for the main take-home message of the current work: careful control of staining conditions and detailed evaluation of labeled regions (including, in some cases, microscopic examination of the reaction product) is needed to successfully discriminate specific lacZ expression from spurious staining caused by endogenous galactosidase activity (Sanchez-Ramos et al., 2000).

In summary, the current findings show that lacZ-WMH can be performed profitably for many organs under the environmental conditions that prevail in a standard high-throughput rodent necropsy suite. However, success in using lacZ-WMH as a rapid transgene expression assay at necropsy will require special processing conditions (3% PFA fixation at 4°C for no more than 3 hours), which will likely limit its utility to special satellite studies rather than as a standard endpoint in routine phenotyping experiments. Furthermore, the present data coupled with the demonstrated divergence in the ability of WMH and SBH lacZ staining assays to reveal endogenous galactosidase activity among many historical studies indicate that the use of lacZ as a reporter gene should be undertaken with caution in bioassays in which the transgene must be assessed in a full battery of rodent tissues. This caveat is particularly important when the transgene is targeted to tissues of either epithelial or monocyte/macrophage lineage, which will contain abundant stores of endogenous galactosidases. For organs with such rich supplies of endogenous galactosidase (e.g., kidney, male reproductive tract, thyroid), a different reporter gene likely should be incorporated into the engineered transgene to eliminate any confusion.