Abstract
Immunotoxicology has developed into an integral regulatory requirement of the toxicological assessment of xenobiotics. Histopathological assessment of lymphoid tissues can provide genuine insight into perturbations of lymphoid cell populations. To facilitate retrospective examination of lymphoid organs should concerns over immunotoxicity be raised, we have endeavored to develop a panel of immunohistochemical techniques to demonstrate T-cells and T-cell subsets in formalin-fixed, paraffin-embedded rat lymphoid tissues. We were successful in developing methods for CD3 and CD8 but failed to arrive at a satisfactory technique for the direct demonstration of CD4 in these tissues. Taking the assumption that the majority of mature T-cells are either CD4+ orCD8+, we have combined our methods for CD3 and CD8 in a novel dual-labeling IHC method to simultaneously demonstrate CD3, CD8, and, by implication, CD4 in rat spleen, thymus, lymph node, and Peyer’s patch.
Keywords
Introduction
Immunotoxicology has developed over the past quarter of a century into an integral part of the safety evaluation of xenobiotics. This development is in part owing to the enormous body of evidence that medicinal products can exert immunotoxic effects, either by the suppression or stimulation of immune responses, and because of increasing awareness of the need for this complex issue to be consistently addressed.
Immunotoxicity is an important cause of drug-related adverse events, contributing to the withdrawal of a number of drugs from the market in the past 15 years, including chlormezanone (Trancopal) (EMEA 1997; Roujeau et al. 1995), glafenine (Glifanan) (Davido et al. 1989; van der Klauw 1996), and temafloxin (Omniflox) (Woodcock 2003).
Over recent years, a number of proposals have been made in an attempt to rationalize and improve histopathological assessment of the immune system (STP IWG 2005; Germolec et al. 2004; Kuper et al. 1995; Kuper et al. 2000; Kuper et al. 2002; Mitsumori et al. 1996; Ward et al. 1993). Most recently, in 2005, the International Conference on Harmonisation (ICH) Working Group, involving regulatory authorities of the European Union, Japan, and the United States, produced the ICH Harmonised Tripartite Guideline on Immunotoxicity Studies for Human Pharmaceuticals (S8) (ICH 2005). This document, restricted to unintendedimmunosuppression or immunoenhancement of human pharmaceuticals, recommends that all new human pharmaceuticals should be tested for immunotoxic potential via a tiered approach.
The initial screen for potential immunotoxicity involves standard toxicity studies, including a recommendation for the semiquantitative description of changes in the separate compartments of lymphoid tissue in sections stained routinely for morphology. Based on a weight of evidence review, including results of histopathology, additional functional studies of immunocompetence may be required. Immunophenotyping (identification/enumeration) of leukocyte subsets, conducted using either flow cytometry or immunohistochemistry, is included in the additional battery of tests to be carried out when immunotoxicity is suspected from the standard studies.
Some immunotoxicants, such as cylophosphamide, can induce a low dose depletion of a subset of T-cells (Male et al. 2006a) prior to exhibiting a more generalized lymphocyte toxicity at higher doses. The detection and counting of shifts in T-cell population distribution in preclinical studies is usually performed by flow cytometry on peripheral blood or disaggregated tissues or, occasionally, by immunofluorescence on sectioned organs (Janeway et al. 2005b). These techniques, however, require specialized preparative procedures with prospective sample collection and cannot be carried out retrospectively.
The advantage of immunohistochemistry over flow cytometry is that tissues from standard toxicity studies can be analyzed retrospectively if concerns over immunotoxicity are raised. Since this material is routinely formalin fixed and paraffin embedded (FFPE), we felt it would be necessary to develop immunohistochemical methods that would work on routine tissue blocks.
The success of IHC depends on the interaction between the primary antibody and the tissue antigen (Haines and Chelak 1991), and anything disrupting this interaction may have a detrimental effect on the quality of staining. Historically the immunohistochemical demonstration of immune tissue antigens has been performed on frozen tissue sections (Barclay 1981; Bland and Whiting 1993; Liou et al. 1996; Nano et al. 2003; Paterson et al. 1987; Robinson et al. 1986; van Ewijk et al. 1981), but the preservation of antigenic epitopes comes at the expense of morphology, and frozen sections may suffer from artifactual streaming of target sites (Beckstead 1994; Fee et al. 1996; González et al. 2001).
Aldehyde-based fixatives, such as buffered formalin, are favored for the preparation of tissues for routine histology and for IHC. Fixation in buffered formalin results in excellent morphological preservation but may decrease the antigenicity of the tissue, possibly because of the protein cross-links created during aldehyde fixation (Cattoretti et al. 1993), or perhaps because of tight complexing of calcium ions or other divalent metal cations with proteins during formaldehyde tissue fixation, masking certain antigens (Morgan et al. 1994). Epitope unmasking techniques may be used to expose the hidden epitopes (Shi et al. 1997, 2001), although these techniques offer no guarantees and the optimal procedure must be determined for each individual antibody (Shi et al. 1996).
This laboratory has endeavored to assemble a toolkit of IHC stains to identify immune cell populations in routinely prepared formalin-fixed, paraffin-embedded (FFPE) tissues. The demonstration of mature T-lymphocytes and the division of the totality ,of T-cells into the T-cytotoxic (TC), CD8 + and T-helper (TH) CD4+ subsets has been a challenge. Here we describe a novel method whereby T-cells may be identified in formalin-fixed, paraffin-embedded lymphoid tissues, and the relative proportions of TCto THcells demonstrated by immunohistochemistry. The equivalent staining of lymphoid tissues of nude rats was included, where possible, as a control and comparative exercise.
Materials and Methods
Tissue Sampling
Lymphoid tissues including spleen, thymus, mesenteric lymph node, and Peyer’s patch were sampled from four male seven- to eight-week-old Hanover Wistar rats (Crl:WI(GIx/BRL/Han)BR, Charles River, UK). Additional lymphoid tissues including spleen, mesenteric lymph node, and Peyer’s patch were sampled from two male nude rats (Nude:HSD Han:RNU-mu, Charles River, UK). The rats were killed using an overdose of the inhalant anaesthetic agent Halothane (Nicholas Piramal, Mumbai, India). Whole organs were fixed in 10% phosphate buffered formalin, pH 7.0, for forty-eight hours. After fixation, tissues were trimmed to blocks no thicker than 3 mm, dehydrated in graded industrial methylated spirit, cleared in xylene, and processed to paraffin wax on a Shandon Excelsior Tissue Processor (Thermo Fisher Scientific, Runcorn, UK) using our standard rodent tissue schedule. All animal procedures were conducted in accordance with Home Office (UK) and local ethical review committee guidelines and complied with the Animals Scientific Procedures Act of 1986.
Pan T-lymphocyte Immunohistochemistry (CD3)
Sections 4 μm thick were cut and mounted on SuperFrost Plus (Thermo Fisher Scientific, Runcorn UK) electrostatically charged glass slides. Sections were allowed to dry at 37°C overnight in an incubator. Sections were dewaxed in xylene, passed through graded alcohols, and rehydrated with water. Heat-mediated antigen retrieval was performed in a Milestone RHS-2 microwave (Milestone, Sorisole, Italy) at 110°C for two minutes in 1 mM EDTA buffer, pH 8.0. Immunohistochemical staining was performed, at room temperature, on a Lab Vision Autostainer 720 (Lab Vision, Newmarket, UK). The reagent and wash buffer was 0.05 M Tris buffered saline plus 0.05% Tween 20 (TBST), pH 7.6.
Endogenous peroxidase activity was quenched by incubation in 3% hydrogen peroxide in TBST for ten minutes. Slides were then washed and incubated for twenty minutes with 5% normal goat serum (Dako UK Ltd, Ely, UK) in TBST. Excess blocking serum was blown off, and the slides were incubated with polyclonal rabbit anti-human CD3 (cat. #180102, Invitrogen, Paisley UK) diluted 1:50 in TBST for sixty minutes. Following washing, the slides were incubated for thirty minutes in rabbit-specific EnVision+ System–HRP (Dako UK Ltd, Ely, UK), and visualized by incubation in diaminobenzidine (DAB), from the EnVision+ kit, for ten minutes. The slides were counterstained for one minute using Carazzi’s hematoxylin (Clin-Tech, Guildford, UK). Staining was negatively controlled by substituting rabbit immunoglobulin (Ig) fraction, diluted to the same Ig concentration, for the primary antibody. Stained sections were dehydrated and mounted under glass coverslips with Histomount (RA Lamb, Eastbourne, UK).
Cytotoxic T-lymphocyte Immunohistochemistry (CD8)
Sections were prepared, subjected to the same antigen retrieval procedure, and stained in a similar manner to that described above. In this instance, the serum block was 5% normal rabbit serum (Dako UK Ltd, Ely, UK), and the primary antibody was monoclonal mouse anti-rat CD8 (clone OX-8, cat. #CBL1507, Chemicon, Chandlers Ford, UK) diluted 1:400. The detection system was designed to minimize cross-reactivity of the secondary anti-mouse reagent with endogenous rat immunoglobulins; after incubation in the primary antibody, the slides were washed and incubated in rabbit anti-mouse immunoglobulins that had been pre-adsorbed against rat immunoglobulins (cat. #Z0456, Dako UK Ltd, Ely, UK) diluted 1:400 in TBST. After washing, the slides were incubated for thirty minutes in rabbit-specific EnVision+ System–HRP, then visualized and counterstained as above.
Helper T-lymphocyte Immunohistochemistry (CD4)
Nine antibodies against either rat or human CD4 were evaluated against our control tissue (Table 1). Sections were dewaxed and stained immediately on the Lab Vision Autostainer 720, or after one of the following antigen retrieval pretreatments: Proteinase K (ready-to-use, cat. #S3020, Dako UK Ltd, Ely, UK) for two minutes at room temperature; heat-mediated antigen retrieval in a Milestone RHS-2 microwave at 110°C for two minutes in 10 mM citrate buffer, pH 6.0; or heat mediated antigen retrieval in a Milestone RHS-2 microwave at 110°C for two minutes in 1 mM EDTA buffer, pH 8.0. Following quenching of endogenous peroxidase activity, the slides were incubated with the antibodies at a range of dilutions, guided by the manufacturer’s data sheets, for one hour at room temperature. Detection and visualization for the mouse antibodies was as outlined for the CD8 antibody above, whereas for the Santa Cruz rabbit polyclonal antibody (Table 1), Dako’s rabbit-specific EnVision+ System–HRP system was used as previously described.
Dual-labeling Immunohistochemistry (CD3/CD8)
The lack of a successful method for the explicit demonstration of CD4+ THcells demanded an alternative approach. A dual stain illustrating the total CD3+ T-cell population and the CD8+ TCsubpopulation would implicitly identify the CD4+ T-lymphocytes by subtraction of the CD8+ cells from the total T-cell population. The methods described above were adapted to reduce potential stearic hindrance between antibodies and detection systems during the dual-labeling procedure.
Sections 4 μm thick were prepared and dewaxed as described. Heat-mediated antigen retrieval was performed in a Milestone RHS-2 microwave at 110°C for two minutes in 1 mM EDTA buffer, pH 8.0 and endogenous peroxidase quenched. After washing, the sections were incubated for twenty minutes in serum-free protein block (cat. #X0909, Dako UK Ltd, Ely, UK). Monoclonal mouse anti-rat CD8 (clone OX-8, cat. #CBL1507, Chemicon, Chandlers Ford, UK), diluted 1:200, was applied to the sections for one hour at room temperature and the signal detected by incubation in horseradish peroxidase-conjugated donkey anti-mouse IgG (H+ L) F(ab’)2 fragments (cat. #715-036-151, Jackson Immunoresearch Laboratories, West Grove, PA, USA) diluted 1:100 for thirty minutes. Visualization was achieved by application of DAB (cat. #K3468, Dako UK Ltd, Ely, UK) for ten minutes.
The slides were then rinsed thoroughly in deionized water and TBST before the re-application of serum-free protein block for twenty minutes. Polyclonal rabbit anti-CD3 (cat. #180102, Invitrogen, Paisley, UK) diluted 1:10 in TBST was applied for sixty minutes followed, after washing, by incubation in goat anti-rabbit IgG (H+ L), alkaline phosphatase conjugate (cat. #G21079, Invitrogen, Paisley, UK), diluted 1:100 for thirty minutes. Visualization of primary antibody binding was achieved by incubation in Permanent Red (cat. #K0640, Dako UK Ltd, Ely, UK) for ten minutes. The slides were counter-stained for one minute using Carazzi’s hematoxylin, and stained sections were dehydrated and mounted under glass coverslips with Histomount.
Staining was negatively controlled by substituting either rabbit Ig fraction or mouse IgG1 isotype control for either one or both of the primary antibodies. The controls were diluted to the same immunoglobulin concentration as the diluted primary antibodies.
Validation of Dual-labeling Immunohistochemistry
To verify the validity of the CD8/CD3 dual-staining technique, 3-mm-thick spleen samples, from a male seven- to eight-week-old Hanover Wistar rat (Crl:WI(GIx/BRL/Han)BR, Charles River, UK), were sampled into Z7 zinc salt fixative (Lykidis et al. 2007). This fixative has been shown to permit the immunohistochemical demonstration of CD4 in paraffin-embedded rat tissue. Following forty-eight hours of fixation, the samples were processed to wax block as previously described.
Sections 4 μm thick were stained for either CD3 or CD8 using the method detailed above for dual labeling, omitting the CD3 or CD8 primary antibodies as necessary. A dual stain of CD8 and CD3 was also performed for comparison with that executed on the formalin-fixed material. High-temperature antigen retrieval was unnecessary for tissues fixed in Z7.
CD4 was demonstrated using a mouse monoclonal antibody (clone OX-38, cat. #CBL CBL1506, Chemicon, Chandlers Ford, UK), diluted 1:200. The detection method was that detailed for the single staining of CD8 given above.
Image analysis and quantification were performed with the Zeiss KS400 system (Imaging Associates, Bicester, Oxfordshire, UK), linked to a Leica DMRB microscope, using a bespoke macro program. Six fields of view from each of the sections were captured using a 10X objective lens, and the area of immunohistochemical staining, brown for CD4 and CD8 or red for CD3, was measured.
Results
Pan T-lymphocyte Immunohistochemistry (CD3)
Spleen
Intense membranous staining of lymphocytes was seen in the periarteriolar lymphoid sheath (PALS) in the white pulp of the Han Wistar rats stained for CD3 (Figure 1A). The follicles of the white pulp and the marginal zone, between the white and red pulp, contained approximately 10% to 20% of CD3+ cells. Almost 100% of the cells within the inner PALS were positive for CD3. The outer rim of the PALS was a combination of CD3+ and CD3− cells. There was a scattering of CD3+ lymphocytes throughout the red pulp.
In the nude rat, although the PALS area was almost completely absent, occasional CD3+ cells were observed running along the length of the splenic central arterioles (Figure1B). Within the follicles, marginal zones and red pulp CD3+ cells were identified in fewer numbers, less than 5% of the total cell population, than in the immunocompetent Han Wistar rats.
Thymus
Over 90% of the cells in both the thymic cortex and medulla were positive for CD3, with those cells in the medulla being more strongly stained than the cortical thymocytes. Overall, the intensity of staining of the lymphocytes was much less than that seen in the PALS of the spleen.
No results were obtained from the nude rats, as these animals are athymic.
Mesenteric Lymph Node
In the immunocompetent Han Wistar rats, more than 95% of cells within the deep cortical units (DCUs) of the paracortex were positive for CD3. Occasional CD3+ cells were scattered throughout the follicles, interfollicular regions of the cortex, and the medulla. Some clusters of CD3+ lymphocytes were present in the medullary sinuses.
The DCUs were absent from the lymph nodes of the nude rat, but the follicles, interfollicular cortex, and medulla contained occasional CD3+ cells.
Peyer’s Patch
In the Han Wistar rat, CD3+ cells were concentrated in the interfollicular regions (IFR), with fewer positive cells within the follicles, and scattered throughout the overlying epithelium.
The frequency of CD3+ cells was much reduced in the nude rat and showed no concentration of positive cells within the interfollicular region.
Cytotoxic T-lymphocyte Immunohistochemistry (CD8)
Spleen
Intense membranous staining for CD8 was seen in the lymphocytes of the white pulp of the Han Wistar rats (Figure 1C). The follicles of the white pulp and the marginal zone contained up to 10% of CD8+ cells. Approximately 50% of the cells within the PALS were positive for CD8, the greater proportion being found closer to the arterioles and becoming less frequent toward the outer rim of the PALS. Occasional CD8+ lymphocytes were spread throughout the red pulp.
In the nude rat, occasional CD8+ cells were observed in the region of the splenic arterioles (Figure 1D). Within the follicles, marginal zones, and red pulp, CD8+ cells were observed, but at a reduced frequency compared with the immunocompetent Han Wistar rats.
Thymus
Almost 100% of the cells in the thymic cortex stained strongly for CD8, with the obvious exclusion of the endothelial cells. Roughly 30% to 40% of the cells in the medulla stained positive for CD8.
Mesenteric Lymph Node
Immunocompetent Han Wistar rats exhibited 50% CD8 positivity of cells within the deep cortical units (DCUs) of the paracortex; the majority of the strongly stained cells were located in the periphery of the DCU. Occasional CD8+ cells were scattered throughout the follicles, interfollicular cortex, and medulla, and, as with the CD3 results, some clusters of CD8+ lymphocytes were present in the medullary sinuses. In the nude rat, very small numbers of CD8+ cells were seen distributed throughout the follicles, interfollicular cortex, and medulla.
Peyer’s Patch
In the Han Wistar rat, CD8+ cells were concentrated in the IFRs. Approximately 25% of lymphocytes within this area were CD8 positive. Very few CD8+ cells were observed within the follicles, but the majority of the intraepithelial lymphocytes appeared positive. The frequency of CD8+ cells was much reduced in the nude rat and showed no concentration of positive cells within the interfollicular region.
Helper T-lymphocyte Immunohistochemistry (CD4)
No specific staining of T-helper cells was achieved with any of the antibodies listed in Table 1.
Dual CD8/CD3 Immunohistochemistry
For all of the tissues examined the dual staining technique appeared to accurately replicate the staining patterns observed following the individual CD8 and CD3 immunohistochemical procedures described previously (Figures 2and 3). Far fewer positively stained cells were apparent in the spleen of the nude rat in comparison with the Han Wistar rat (Figures 1Eand 1F). The general T-lymphocyte population (CD3+) could be recognized as the sum of the red- and brown-labeled cells. The CD8+ T-cells could be identified by the brown DAB reaction product, which overpowers and disguises the permanent red of the CD3 staining.
A summary of the immunohistochemical findings is given in Table 2.
Validation of Dual-labeling Immunohistochemistry
The dual-staining distribution for CD8 and CD3 on the Z7 zinc-fixed tissue was comparable with that achieved with formalin-fixed spleen (Figure 4A), although the morphological preservation of the tissue was inferior.
The areas stained and quantitated for CD3, CD8, and CD4 individually (Figure 4B–4D) were:
CD3: 214584.30 μm2
CD8: 123950.60 μm2
CD4: 86293.60 μm2
Taking the CD3 measurement as the total T-cell population, one could calculate the populations of CD4 and CD8 in this material to be 58% and 40%, respectively.
The distribution and staining patterns achieved with CD3 and CD8 in the different lymphoid tissues were consistent with our expectations. In Han Wistar rats, the pan T-cell marker CD3 stained areas of the spleen, thymus, lymph nodes, and Peyer’s patches known to be T-cell areas. CD3 is known to be weakly expressed in developing T-cells of the human thymic cortex (Male et al. 2006b), which is reflected in our results in the rat.
The frequency of CD3+ cells in lymphoid tissues from the nude rat was extremely reduced. Because these animals are athymic, one might have expected a complete absence of staining for T-cells in the nude rat. There is, however, evidence for the extrathymic development and maturation of T-cells in the epithelium of the small intestine (LeFrançois and Puddington 1995; Ramanathan et al. 2002; Rocha, Vassalli et al. 1992; Rocha et al. 1995; Rocha 2007), which may provide an explanation for this observation. The distribution of the CD8+ cells mirrors that of the CD3, but at a lower frequency, labeling only the cytotoxic subset of T-cells (Patel et al. 1988).
The distinctive attribute of a mature T-lymphocyte is the expression of one of the two types of T-cell antigen receptor (TCR) (Janeway et al. 2005a). The two types of TCR take the form of two similar heterodimers consisting of either α and β or γ and δpolypeptides. Over 95% of mature T-cells are αβ T-cells, and the remainder are γδ T-cells. αβ T-cells may be subdivided into distinct non-overlapping populations that are either CD4+ or CD8+. There is a very small proportion of αβ T-cells that are CD4−/CD8− and may serve a regulatory function (Chen et al. 2004; Thomson et al. 2006; Zhang et al. 2001) (Figure 5).
Accepting that the overwhelming majority of T-cells express either CD4 orCD8, we suggest that if one can label the general T-cell population and the CD8+ subpopulation simultaneously, then the labeled cells that are not CD8+, that is, the cells that have stained red not brown, must, on the whole, be CD4+ T-helper cells. A simple visual assessment of control spleen suggests that, out of the total T-cell population, slightly more CD4+ cells are present than CD8+ (Figures 2Aand 2B). This assessment is corroborated by the validation work performed on Z7 zinc-fixed tissue, which measured the populations as 58% CD4+ and 40% CD8+. This dual-labeling technique therefore offers a way of visualizing the general make-up of the T-cell population on a single slide of routinely processed formalin-fixed, paraffin-embedded tissue.
There are two caveats to be aware of:
There is, as previously mentioned, a small proportion of double-negative (DN) CD3+ CD4−/CD8− T-cells. There also exists, especially in the intraepithelial population, a minor subset of double-positive (DP) CD3+ CD4+/CD8+ T-cells (LeFrançois 1991; Poussier et al. 1992; Poussier and Julius 1994; Rocha, von Boehmer et al. 1992).
The logic behind the CD3 − CD8 =CD4 equation fails for these subsets, but for the majority of T-cells within lymphoid tissues the reasoning holds true.
In summary:
T-Lymphocytes (CD3 and CD8): red and brown cells Cytotoxic T-Lymphocytes (CD8): brown cells Helper T-Lymphocytes (CD4) : red cells
The investigation of the effects of xenobiotics on lymphocyte populations in FFPE tissue is an important part of the evaluation of their potential immunomodulatory effects. The comparison of staining in nude and immunocompetent Han Wistar rats (Figures 1Eand 1F) illustrates the power of this dual-labeling technique to demonstrate variations in T-lymphocyte populations between animals. Also, in addition to the semiquantitative assessment of a pathologist, we intend to investigate the feasibility of quantitative image analysis of the dual staining, either by traditional colorimetric means, or by employing multispectral imaging systems such as the Nuance (Channel Systems Inc, Pinawa, Canada). This dual-labeling IHC methodology has the advantage that it may be applied to archived wax blocks of formalin-fixed tissue many months, or even years, after the initial study, while preserving the T-cell populations within a morphological context.
Alternatives to aldehyde fixation have been proposed, such as periodate-lysine-paraformaldehyde (PLP) (Whiteland et al. 1995) or zinc-salt–based fixative (ZSF) (Beckstead 1994; Hicks et al. 2006; Lykidis et al. 2007), such as the Z7 fixative used in our validation experiment. Although these fixatives do appear to be better at preserving immune cell epitopes, we have found in this laboratory the morphological preservation of ZSF to be inferior to that of buffered formalin, and PLP is unsuitable for large toxicological studies because large volumes cannot be prepared in advance and stored, rather the fixative has to be prepared immediately before use. Also, one must sample tissues into these fixatives prospectively in anticipation of performing immune cell IHC.
We therefore consider this dual method of staining—which demonstrates T-cell numbers and relative populations of T-cell subsets all in one slide, from routinely prepared blocks of paraffin-embedded, formalin-fixed tissue—an informative and practical adjunct to investigations of immunotoxicity.