Immunohistochemical Analysis of Acetylation,Proliferation,Mitosis,and Apoptosis in Tumor Xenografts Following Administration of a Histone Deacetylase Inhibitor—A Pilot Study

Abstract

Histone acetyltransferases and histone deacetylases are protein-modifying enzymes involved in addition and removal of acetyl groups on histone proteins, respectively. These molecules play a pivotal role in cellular functions such as chromosome remodelling, gene transcription and cell proliferation. Histone deacetylase inhibitors (HDIs) have been shown to cause cell cycle arrest, cellular differentiation and inhibition of cell proliferation in tumor cells in vitro and in vivo. Their potential use for cancer therapy is currently under evaluation in clinical trials. A pilot study was performed to immunohistochemically evaluate the effects of a HDI, “Compound 1,” on acetylation, proliferation, mitosis, and apoptosis in tumor xenografts (Calu-6, SW 620, Colo 205, and LoVo) in nude mice, at 6, 24, and 48 hours, following a single oral dose. Qualitative immunohistochemistry and computer-assisted image analysis demonstrated an increase in acetylation in all xenografts. Immunohistochemical analysis of acetylation in skin showed increased acetylation at 6 hours after HDI administration. In addition, image analysis showed a decrease in mitosis and an increase in metaphase mitotic figures in the SW 620 xenograft. These two findings were consistent with a G1/S cell cycle phase arrest. Increased apoptosis of SW 620 and LoVo xenografts was also observed following treatment.

Introduction

Histones are important regulators of gene activity through integration into gene transcriptional machinery and dynamic posttranscriptional modification. It is now well known that levels of histone acetylation, at a given region of chromatin, correlate with transcriptional activity (Grunstein, 1997). Histone acetyltransferases (HATs) interact with histones or with specific transcription factors to enable transcriptional activation of key cell cycle-associated genes involved in tumor-related pathways; such as cyclin-dependent kinase inhibitor p21^WAF/CIP1, or p53 and GATA-1 (Yang and Seto, 2003; Vigushin and Combes, 2002). Histone deacetylase inhibitors (HDIs) mimic the action of HATS in that their activity leads to chromatin open structure and further gene transcription. Interestingly, micro-array profiling has shown that HDIs modulate only a small subset of genes (2–5%). However, most of these genes are involved in several key pathways closely linked to cell cycle, and furthermore, can function aberrantly in cancer (Marks et al., 2004). Thus, HDIs, which antagonise histone deacetylases (HDACs) and induce histone acetylation, appear to have a huge potential for cancer therapy through disruption of cell cycle regulatory and survival-related pathways. In certain cancers, such as leukemias, HDIs also appear to act specifically on tumor cells, yet are relatively non toxic to normal cells (Insinga et al., 2005). Structurally diverse HDIs have shown anti-tumor effects in vitro, in in vivo xenograft models and in clinical trials (Jaboin et al., 2002; Marks et al., 2004). Although their precise mode of action is not fully understood, it appears to involve blockage of one or several apoptotic and cell cycle modifying pathways dependent on dose administered, time of treatment, or tumor type (Marks et al., 2004).

The purpose of this pilot study was to use immunohistochemistry to evaluate the pharmacodynamic (PD) effects of “Compound 1,” a HDI, in different tumor xenografts (Calu-6, Colo 205, SW 620, and LoVo). This was to form a basis for subsequent suitable immunohistochemical analyses over time in larger numbers of xenografts. Based on the known actions of HDIs with regard to increasing acetylation and subsequent downstream effects on cell cycle regulatory mechanisms, we hypothesized that this inhibitor could potentially act in tumor xenografts to increase acetylation, decrease proliferation, arrest cell cycle and induce apoptosis. These processes were quantitatively measured at different time points following compound administration. The markers used to follow these cellular processes were antibodies against acetylated histone H3 (AcH3), Ki-67 (proliferation), phosphohistone H3 (PH3) (mitosis), and cleaved caspase-3 (apoptosis).

Materials and Methods

Xenograft Models

Human primary colorectal cell lines of metastatic origin (Colo 205, SW 620, LoVo) and an anaplastic carcinoma probably of lung origin (Calu-6) were injected subcutaneously into nude mice as previously established under U.K. Home Office regulations. The cell sources for Colo 205, SW 620, LoVo and Calu-6 were the European Collection of Cell Cultures (EACC, Health Protection Agency, Porton Down, UK) CCL222, EACC CCL227, EACC CCL229 and American Type Culture Collection (ATCC, Manassas, VA) HTB-56, respectively. There were 2 animals per time point for each xenograft apart from the LoVo xenograft, which was implanted in 3 animals per time point. A small number of animals was used in this study in line with the purpose of the work, which was to develop methodology applicable for screening, by immunohistochemistry, for a PD effect of a potential drug. Due to the labour-intensive nature of the development of methodologies suitable for screening, together with the evaluation of antibodies for their potential use, and ethical issues regarding animal usage, only 2 or 3 animals were used.

The xenografts used were standard established quality-controlled in-house models used in early evaluation of PD markers and form a basis for further robust model development with compounds in the later lead optimisation stages of drug development. Tumors grew between 10 and 20 days: Calu-6 (10 days), Calu-6 (21 days), LoVo (11 days), Colo 205 (14 days), and SW 620 (18 days). When the tumors were palpable (reaching an approximate volume of 1 cm³), animals were dosed orally with “Compound 1” (50 mg/kg single dose). This dose was chosen based on doses used in previous tumor efficacy work-up and from current literature. Control animals received vehicle only. At 6, 24, and 48 hours post-dosing, animals were sacrificed (in 5% CO₂ chamber) and xenografts removed. A skin sample was also excised from each animal at a site distant from the tumor.

Histological Assessment

Tumors were fixed in 10% formalin for 48 hours and embedded in paraffin-wax. Hematoxylin and eosin (H&E) assessment for general morphology including presence or absence of necrosis was carried out on a single 4 μm section per timepoint.

Immunohistochemistry

Tumor and skin sections for immunohistochemistry were electrostatically bonded on to SuperFrost Plus slides (Menzel-Glaser, Braunschweig, Germany) for evaluation of acetylation, proliferation, mitosis, and apoptosis.

Primary Antibodies

Primary antibodies (in conjunction with their appropriate secondary antibodies) had previously been titrated in order to optimise specific staining both manually and with our autostainers. An anti-AcH3 rabbit polyclonal IgG (Upstate Biotechnology, Lake Placid, NY, catalog no. 06-599) was used at a dilution of 1:300 to ascertain acetylation in xenografts and in mouse skin sections. To assess proliferation in xenografts, a mouse monoclonal anti-human Ki-67 antigen, clone Mib-1 antibody (Dako Cytomation, Glostrup, Denmark, catalog no. M7240) was used at 1:50 dilution. A monoclonal rat anti-mouse Ki-67 antigen, clone TEC-3 antibody (Dako Cytomation, catalog no. M7249) at 1:50 dilution was used for the analysis of proliferating cells in mouse skin sections. Mitosis was detected using an anti-PH3 rabbit polyclonal IgG (Upstate Biotechnology, catalog no. 06-570), at 1:1000 dilution. In order to ascertain apoptosis in xenografts, a rabbit polyclonal cleaved-caspase-3 (Asp175) antibody (Cell Signalling Technology, Beverley, MA, catalog no. 9661L) was used at 1:100 dilution.

Secondary Antibodies

Rabbit polyclonal antibodies were detected using sheep anti-rabbit IgG biotin conjugated antibody (Serotec, Kidlington, Oxford, UK, catalog no. 2AB02B 5204) at 1:200 dilution. The Ki-67 Mib-1 clone was detected using a rabbit anti-mouse biotin conjugated antibody (Dako Cytomation, catalog no. E0464) and the Ki-67 TEC-3 detected with a rabbit anti-rat biotin conjugated antibody (Dako Cytomation, catalog no. E0468), both at 1:200 dilution.

Staining Procedures

AcH3, Ki-67, PH3 and cleaved-caspase-3 antibody staining was performed using a Ventana Discovery autoimmunostainer (Ventana Medical Systems, Strasbourg, France). A diaminobenzadine (DAB) detection kit (DABMap, Ventana Medical Systems) was used to complete the section staining. Negative controls for each antibody staining procedure were carried out using an equivalent amount of antibody diluent instead of primary antibody. Positive controls used were xenograft tissue shown previously to be positive following HDI administration (AcH3), human breast carcinoma (Ki-67), PH3, and cleaved caspase-3 (in-house xenograft blocks).

Ki-67 Mib1 immunohistochemistry was performed using a Lab Vision autoimmunostainer (Model LV-1, Lab Vision UK Ltd., Newmarket). Sections were deparaffinized through 2 washes of xylene and rehydrated in industrial methylated spirit (I.M.S.) (75%), I.M.S (95%) and deionized water (all spirits purchased from Fisher Scientific, Loughborough, UK). Endogenous peroxidase activity was blocked by 3% hydrogen peroxide (H₂O₂) (Fisher Scientific) for 10 minutes after the pretreatment and slides were washed out in deionized water (dH₂O). Ki-67 Mib-1 clone and relevant IgG1 negative control (Dako Cytomation Mouse IgG1, catalog no. X0931, diluted to the same mouse IgG concentration as the primary antibody) were used on appropriate tissue sections. The sections were re-hydrated in dH₂O, 95% I.M.S., 75% I.M.S., and 2 washes of xylene, then coverslip mounted using a CV5000 coverslipper (Leica Microsystems UK Ltd., Milton Keynes).

Qualitative Microscopic Examination

An initial qualitative light microscopic screening of sections was performed to determine if there was a difference between control and treated sections. For acetylation and apoptosis, a simple 1–4 grading system of (1) minimal increase in staining, (2) mild increase, (3), moderate increase, and (4) marked increase was recorded. Acetylation was nuclear and specific with no observable background staining at this qualitative level. Staining was completely absent in necrotic areas previously identified by H&E analysis. Cleaved caspase-3 staining was confined to apoptotic cells and apoptotic bodies in viable tumor areas. Nonspecific staining was present in necrotic areas. The latter areas were excluded from scoring. A similar qualitative evaluation of xenografts or skin stained with the Ki-67 antibodies for minimal, mild, moderate, or marked decreases in Ki-67 positive cells was carried out. This staining was nuclear and nucleolar and nonspecific background staining was not apparent. The latter evaluation, however, did not reveal differences between control and treated animals. Hence, no further analysis was carried out on Ki-67 stained sections.

Quantitative Image Analysis

AcH3

The automated cellular imaging system (ACIS II, Clarient Chromavision, Capistrano, CA) computational image analysis package was considered the most appropriate to use to measure (i) the percentage area of immunopositivity and (ii) staining intensity of immuno-reaction product for each xenograft and skin section per timepoint. Slides were microscope-scanned at ×10 objective lens magnification and thresholds set using HLS (hue, luminosity, and saturation) for brown (immuno-positive) and blue (immuno-negative) tissue coloration. Nonspecific background pixels defined by the thresholding step were then excluded using a binary operator to close and open the segmented image. Measurements of both area and intensity were made within 10 randomly selected regions that included viable tumor area with the exclusion of necrotic zones. The average values for each animal at each time point were plotted. For skin AcH3 staining evaluation, the epidermal region and top third of the dermis thickness, including sebaceous glands, were interactively selected and examined.

Cleaved Caspase-3

The percentage area of positive staining was measured as described for the AcH3 antibody.

PH3 Mitosis

A second image analysis system was used for quantitation of PH3 positive nuclei as methodology was currently already in use for this antibody in-house. Zeiss KS400 computer image analysis software (Imaging Associates Ltd., Bicester, Oxon, UK) was used to count the number of PH3 positive nuclei per mm². Images were captured from a DMRB (Leica) microscope with an ×20 objective lens magnification via a KY-F55B 3CCD (JVC, UK Ltd., London) camera using a Matrox meteor frame grabber. Up to 10 random regions of interest were selected with segmentation of brown nuclei for each field being achieved by setting HLS threshold levels. Objects of a defined circularity and within the threshold levels set, which were greater than 5 pixels and less then 100 pixels in size, were classed as PH3 positive nuclei. The framed regions established the area measures to which the numbers of nuclei were expressed.

PH3 Metaphase Cells

Based on the results for mitosis, a DMRB microscope (Leica) equipped with ×40 magnification and JVC colour video camera (3-CCD model) was used for analysis of SW 620 xenografts only. PH3 stained cells were randomly distributed across the section. As the Zeiss system was not developed to identify and discriminate specific phases of the cell cycle by DAB detection, metaphase cells were identified through operator interaction. Up to 1000 positive cells were counted 3 times on each section, starting from different ends of the section. At this time, cells in metaphase were also recorded and the percentage of metaphase cells calculated as a mean of the 3 measurements. This method was similar to a previously documented procedure by Scott et al. (2004).

Results

Histological Description of Xenografts

All xenografts were predominantly in a dermal location and, in general, formed a well-defined spherical to ovoid mass.

1. Calu-6 (21 day)

   Control and treated tumors comprised solid sheets of cells with big nuclei and multiple nucleoli. Sections were interspersed with small central necrotic areas along the edges of which were presumptive apoptotic cells. A large number of mitotic figures were present and cells had abundant eosinophilic cytoplasm.

2. Calu-6 (10 days)

   Control and treated sections were similar to the 21 day Calu-6 tumor although fewer necrotic areas were present.

3. Colo 205

   Control and treated tumors comprised solid sheets of epithelial cells. Some tumors extended multifocally into deep subcutaneous tissue and incorporated small numbers of intervening adipose or muscle cells at the invading margin. Nuclei were variable in size with multiple nucleoli. Large numbers of mitotic figures were present. A few scattered central necrotic areas were present.

4. SW 620

   The tumor comprised sheets of large epithelial cells with small nuclei. In areas, these formed separate cell islands surrounded by connective tissue. There were a high number of mitotic figures present and cells had abundant eosinophilic cytoplasm. Haemorrhage was present within some central necrotic areas.

5. LoVo

   Control and treated tumors comprised solid sheets of cells which occasionally extended deep into the sub-cutis and invaded underlying muscle and adipose tissue. The cells formed nests and circular patterns surrounded by connective tissue. Large nuclei with multiple nucleoli and a few scattered mitotic figures were present. Small foci of necrosis were present in the centre of the xenograft.

Acetylation

Xenografts

Administration of “Compound 1” into nude mice resulted in increased acetylation of histone H3 in tumor xenograft sections 6–24 hours postdosing, compared with controls (Figures 1 and 2). An increase in both percentage area of stained cells and intensity was observed 6–24 hours post dosing. Since the microscopic analysis of slides confirmed that the number of stained cells on each section was very similar between control and treated xenografts, although much weaker in controls, the intensity measurement was used to demonstrate drug effect over time. The increase in intensity over time in all xenograft types is shown in Figure 1.

Skin

Increased acetylation was observed in epidermal keratinocytes, hair follicles and sebaceous glands at 6 hours following compound administration (Figure 3).

Apoptosis (Cleaved Caspase-3)

“Compound 1” increased apoptosis in SW 620 xenografts and slightly in LoVo xenografts at 6–24 hours compared with controls (Figure 4). There was no difference in apoptosis between control and treated Colo 205 and Calu-6 xenografts. In addition, in the skin of animals from 1 study at the 6-hour time point, we observed an increased number of apoptotic cells in hair follicles compared to control animals (Figure 5).

PH3 (Mitosis)

PH3 is a marker for cells in mitosis. Following drug treatment, a decrease in PH3 expression occurred at 6–48 hours in the SW 620 xenograft only (Figure 6a). In other xenografts there was no difference between control and treated xenografts.

PH3 (Metaphase Assessment)

An increase in metaphase figures is an indication of G1 to S cell cycle phase arrest and may result from a backup of cells in metaphase unable to move forward (Figure 6b). A moderate increase in the number of mitotic metaphase figures was observed in the SW 620 xenograft only, between 24–48 hours following treatment (Figure 6c).

Discussion

The immunohistochemical evaluation of acetylation, apoptosis, mitosis and proliferation in xenografts following a single administration of a histone deacetylase inhibitor, “Compound 1,” highlights the benefits of such an approach for determining the mode of action of these compounds in the commonly used nude mouse xenograft model. This initial work was carried out as a pilot PD study using small numbers of animals to confirm acetylation in vivo following early indications that “Compound 1” acted chemically as an HDI. Based on current literature that HDIs may act to alter cell cycle progression, proliferation and apoptosis (Marks et al., 2001), we also examined these cellular functions using IHC.

“Compound 1” clearly increases acetylation of histone H3 within all xenografts examined and this effect is consistent with the predicted response. Using both tumor and skin for analysis, HDI effect is demonstrated by an increase in acetylation in xenografts at 6–24 hours post compound administration and also in skin at the 6-hour time point. The use of acetylation as a biomarker of pharmacodynamic activity of HDIs in the clinic has been applied to blood and bone marrow samples of patients with acute promyelocytic leukemia (APL) (Warrell et al., 1998). Our work indicates that tumor biopsy and skin samples might also be used for a similar purpose.

Immunohistochemical measurement of proliferation by measuring Ki-67 expression has been widely used as a method for assessment of tumor biological behaviour. Ki-67 antigens are present in proliferating cells (in G1, S, G2, and mitosis cell division phases) while quiescent or resting cells do not express Ki-67 protein. Determination of levels of Ki-67 expression can be of particular importance in cancers in which the clinical course is difficult to predict by histological criteria alone. Ki-67 expressing-cells were present in all control and treated xenograft sections evaluated in this study. A prominent redistribution of Ki-67 protein occurs during mitosis and we observed a nuclear and nucleolar staining pattern, also previously reported by Scholzen and Gerdes (2000). Qualitative assessment of control and treated xenografts did not reveal a clear difference in numbers of positive cells and further image analysis on Ki-67 stained sections was not carried out. The time points that were measured were acute relative to HDI administration, and it is likely that this antibody may be better suited to evaluate proliferation changes over longer duration efficacy studies. Ki-67 is also a difficult marker to interpret with regard to proliferation as it is not necessarily the case that a cell down-regulates Ki-67 just because DNA synthesis may be blocked. In addition, Ki-67 immunohistochemistry is particularly sensitive to the conditions used for fixation and staining and can sometimes prove difficult to interpret (Scholzen and Gerdes, 2000).

We thus decided to use the PH3 antibody to evaluate mitosis in xenografts. This antibody only detects cells in mitosis and thus is not a broad indicator of effect on proliferation. However, our results have shown that the PH3 antibody produces specific, reproducible nuclear staining and is thus ideal for image analysis of nuclear immunopositivity. There was a significant decrease in mitotic cells in only one xenograft, SW 620. This was observed at 6–48 hours after “Compound 1” administration.

HDIs have been shown to block cell proliferation by up-regulating the expression of the cyclin dependent kinase inhibitor, p21, thus inducing a G1 phase arrest in a range of tumor types (Saito et al., 1999). We thus decided to evaluate the number of metaphase cells as a potential measure of G1/S arrest in the SW 620 xenograft in which a decrease in mitotic cells following HDI administration had been demonstrated (Figure 6a). An increase in mitotic figures was observed at 48 hours posttreatment in the SW 620 xenograft supporting the view that a G1 to S phase arrest is taking place and that cells cannot move forward from metaphase. Although there is evidence that HDIs can trigger a G2 to M phase arrest (Qiu et al., 2000), a cell block at this stage would result in a decrease in metaphase figures and an increase in polyploid cells with larger than normal nuclei. This was not observed in the present study. However, this data requires further support by particular stage-specific analysis of mitotic cells, for example using Feulgen’s technique (Sudbo et al., 2001).

Apoptosis is another potential mechanism through which HDIs may act although there are different and often conflicting conclusions as to which apoptotic pathways are evoked. Some studies indicate a death receptor pathway, such as the induction of Fas or Fas ligand, while other studies suggest that the mitochondrial apoptotic pathway is involved (Rosato et al., 2003). Caspase activity has also been indicated as an integral part of HDI-induced apoptosis, while other papers dispute this (Kwon et al., 2002; Insinga et al., 2005). In the current study, there was evidence of increased apoptosis in one xenograft line, SW 620. The fact that one xenograft, SW 620, shows both an increase in G1-arrested cells and apoptotic cells, may indicate that those cells that undergo cell cycle arrest (cytostasis) do not go through mitosis any further, but die in that stage.

Interestingly, apoptotic cells in hair follicles were observed in the skin of 1 study following HDI administration, suggesting the possible use of cleaved caspase-3 immunohistochemistry as a potential skin biomarker in the clinic. Hair follicles staining for cleaved caspase-3 have also been noticed with some other HDI compounds (J. McKay, personal communication, June 2004).

Conclusions

Our pilot study has given an insight into the mode of action of one histone deacetylase inhibitor on human xenografts implanted into nude mice. Immunohistochemistry was used to measure acetylation, proliferation, mitosis and apoptosis 6–48 hours following a single dose of one compound, to investigate possible mode of action and provide background information for further tumor PD and efficacy studies with similar compounds in a more advanced stage of development. These particular cellular processes were chosen based on current published information available on in vitro and clinical activity of HDIs in development for cancer therapy. In addition, image analysis of these processes provided a quantifiable basis of measuring such processes in the future with larger animal numbers. Time-point evaluation following a single dose also provided a model to dissect out direct effects on particular aspects of cell cycle and progression without the complications of longer-term compound administration that could mask these early responses.

Acetylation has been reproducibly increased in all xenografts examined. However, the variable effect on proliferation, mitosis, apoptosis, and cell cycle arrest in different xenografts highlights how the downstream actions of the HDIs are highly dependent on cell type and time of administration of HDI. Our work has highlighted that in the evaluation of these types of compounds in efficacy studies, consideration must be given to the fact that although their anti-tumor effects appear to involve common mechanisms, particularly disruption of cell cycle events and disruption of cytoprotective signalling pathways or mitochondrial injury and apoptosis, the actions are multiple and can be variable. Other histone modifications are becoming apparent such as phosphorylation, methylation and ubiquitination (Kristeleit et al., 2004) and these, in time, may also have to be evaluated in our mode of action studies. This study has also highlighted the use of immunohistochemistry to detect acetylation and apoptosis in skin following HDI administration, and is clearly of relevance in the development of other PD markers to monitor in the clinic.