Semiquantitative immunohistochemical analysis for hypoxia in human tumors

Abstract

Objective: The goal of this study was to develop a semiquantitative scoring system for measuring hypoxia in human tumors by an immunohistochemical marker approach.

Methods and Materials: Eighteen patients diagnosed with squamous cell carcinoma of the uterine cervix or head and neck were infused intravenously with a solution of pimonidazole hydrochloride at a dose of 0.5 gm/m². Twenty-four hours later, four biopsies on average from each tumor were fixed in formalin, processed into paraffin blocks, and sectioned. Tissue sections were immunostained for the presence of pimonidazole adducts. Microscopic images (×200) of immunostaining were captured and quantitated by standard image analysis. Images with known amounts of hypoxia spanning ranges of > 0% to 5%, > 5% to 15%, > 15% to 30%, and >30% were assigned scores of +1, +2, +3, and +4, respectively. Three observers then used this calibrated scoring system to analyze hypoxia in tumor sections in a blinded fashion.

Results: Excellent interobserver reproducibility was obtained with the calibrated, semiquantitative, immunohistochemical assay for hypoxia in squamous cell carcinomas.

Conclusion: The calibrated, semiquantitative assay shows promise as an approach to simplifying the quantitation of human tumor hypoxia by immunohistochemical techniques.

Introduction

In 1976, Varghese et al. (1) reported that ¹⁴C-labeled misonidazole formed adducts in hypoxic cells in vitro and in vivo. It was subsequently found that adducts form with thiol groups in proteins, peptides and amino acids in a way that all atoms of the ring and side-chain of the 2-nitroimidazole are retained 2, 3, 4, 5. Binding is oxygen dependent but is independent of specialized P450 redox enzymes or changes in NADH and NADPH levels 6, 7. Chapman et al. (8) showed that the oxygen dependence of binding was fortuitously close to that for radiation resistance and suggested that misonidazole binding might be used as a hypoxia marker in solid tumors. The clinical feasibility of the hypoxia marker idea was demonstrated by means of autoradiographic analyses of ³H-misonidazole binding in a variety of human tumors (9). Although the ³H-misonidazole approach had limited clinical utility, it spurred the development of a variety of noninvasive assays for tissue hypoxia based on 2-nitroimidazoles. These included single photon electron capture tomography, positron emission tomography, nuclear medicine analysis, and magnetic resonance spectroscopy of suitably labeled 2-nitroimidazole analogues [for review see Raleigh et al. (10)].

During ¹⁹F MRS investigations of tumor hypoxia with the hexafluorinated 2-nitroimidazole, CCI-103F, it became clear that an histologic assessment of hypoxia would be a useful complement to noninvasive assays 11, 12, 13. This led to the invention of the immunochemical hypoxia marker technique based on monoclonal and polyclonal antibodies raised against protein adducts of reductively activated 2-nitroimidazoles 14, 15. Preclinical testing of immunochemical reagents in spontaneous canine tumors showed that immunochemical hypoxia markers would be useful in their own right 16, 17, 18, 19, 20, 21. In addition to providing a quantitative measure of hypoxia, immunohistochemical markers provide insights into microregional relationships between hypoxia and factors such as necrosis, proliferation, differentiation, apoptosis, and oxygen regulated protein expression. A variety of immunochemical hypoxia markers has now been used in clinical 22, 23, 24, 25, 26 and preclinical studies 27, 28, 29, 30, 31, 32 of such relationships. An example of the unique value of the immunohistochemical marker approach is the observation that neither metallothionein nor vascular endothelial growth factor are expressed in the majority of hypoxic cells in human squamous cell carcinomas 22, 24 even though in vitro studies would have predicted otherwise 33, 34.

With respect to quantifying hypoxia by the immunochemical technique, image analysis 22, 24 or flow cytometric analysis 35, 36 appear most promising. Preclinical studies of sampling error showed that stratification of patients is feasible with the immunohistochemical approach if 4 biopsies are obtained from geographically separate regions of each tumor. Precision can be increased by increasing the number of sections analyzed per biopsy from 1 to 3, but analysis of multiple biopsies is the most important factor. Interestingly, the accuracy of the immunochemical analysis increases as the amount of hypoxia decreases 19, 20. Quantitative image analysis of pimonidazole binding has been used clinically to measure tumor hypoxia 22, 24, 25, 26, 37, but it was anticipated that routine clinical use might require that the analysis be simplified by the elimination of sophisticated and time consuming image analysis techniques. The object of the present study was to explore the feasibility of developing a semiquantitative immunohistochemical assay based on the hypoxia marker, pimonidazole hydrochloride.

The rationale for developing pimonidazole hydrochloride (Hypoxyprobe-1, NPI, Belmont, MA, USA) as a hypoxia marker for translation to clinical use was based on its chemical stability, water solubility, wide tissue distribution and the fact that human toxicity data were available from earlier radiosensitizer trials. The availability of toxicity data facilitated the early clinical application of pimonidazole hydrochloride 10, 26 and the marker has now been used in over 200 patients worldwide. Solid pimonidazole hydrochloride is very stable being unchanged after storage for 2 years at room temperature in subdued light. Saline solutions of pimonidazole hydrochloride used for infusion (34 mmol/L in 0.9% saline, pH 3.9 ± 0.1) are also extremely stable being unchanged after 1.5 years at 4°C in subdued light as determined by high performance liquid chromatography and ultraviolet spectroscopy. In addition to chemical stability, pimonidazole hydrochloride has high water solubility (400 mmol/L; 116 grams per 100 mL) that facilitates intravenous marker infusion and produces a short plasma half-life of 5.1 ± 0.8 h. In spite of the water solubility of its hydrochloride salt (pKa 8.7), pimonidazole itself has an octanol-water partition coefficient of 8.5 (38) and diffuses readily into tumors and normal tissues including brain (39). Consistent with a large, 155-L volume of distribution, pimonidazole concentrates approximately threefold above plasma levels in tumors and normal tissues (39) thereby increasing the sensitivity of hypoxia marking. At the dose of 0.5 g/m² used in hypoxia marking, pimonidazole hydrochloride causes neither central nervous system toxicity nor sensation (e.g., flushing) (39). Central nervous system toxicity was of particular interest because this was the dose limiting toxicity for pimonidazole hydrochloride at the higher, multiple doses used in radiosensitizer trials. In addition to the absence of central nervous system effects, the overall procedure from pimonidazole hydrochloride infusion to tumor biopsy is well tolerated in both inpatient and outpatient settings.

Protein adducts of reductively activated pimonidazole are effective immunogens for the production of both polyclonal and monoclonal antibodies. The antibodies can be used for immunoperoxidase analysis of formalin fixed, paraffin embedded sections 22, 24, 25; for immunofluorescence analysis of frozen fixed sections (37); and, for flow cytometry with directly labeled or secondary fluorescent antibodies (35). The antibodies have also been used in enzyme-linked immunosorbent assays 7, 21. As is the case for pimonidazole hydrochloride itself, the antibodies to pimonidazole adducts are very robust. For example, aqueous solutions of the IgG₁ monoclonal antibody against pimonidazole adducts (clone 4.3.11.3) is stable indefinitely when stored at −20°C and is stable for at least 4 months at 4°C when supplemented with 10 mg/mL of bovine serum albumin and 10 mmol/L sodium azide. One final attractive feature of pimonidazole is the fact that pimonidazole adducts in vivo are long-lived (21). This provides flexibility in the timing of biopsy taking, which is an advantage in a clinical setting. In summary, pimonidazole hydrochloride and associated antibodies form a very attractive basis on which to develop a low tech, low cost, semiquantitative assay for human tumor hypoxia.