Abstract: Two techniques (surface photographic analysis [SPA] and microscopic tumour analysis [MTA]) are described for assessing the stage of tumour development induced by intraperitoneal 4-(methyl nitrosamino)- 1 -(3 pyridyl)-l-butanone in syngeneic A/J mice. Parameters used to assess the number and size of tumours were surface tumour count (STC), surface tumour area (STA), and relative tumour surface area (RTSA) - all using SPA; and mean tumour count (MTC), mean tumour area, and relative tumour area (RTA) - all using MTA. In a study examining the possible effects of Millipore chamber implantation on lung tumours, the stage of tumour development was assessed in 239 mice using both SPA and MTA. Statistically significant correlations (Spearman-Rank) were apparent between the derived parameters: STC vs. MTC (r = 0.6, P <0.0001); STA vs. mean tumour area (r = 0.5, P <0.0001); RTSA vs. RTA (r = 0.4, P <0.0001). Therefore, it is concluded that SPA - the simpler technique - is an appropriate method for assessing the stage of tumour development in this model.
Key words: Lung neoplasms. Models, biological. Mice. Microscopy.
Introduction
Carcinoma of the lung remains the most common cause of death in the UK. Although surgery offers the only potential cure, the five-year recurrence-free survival rate, even with early stages of the disease, is only 40-70%.1 Recurrences are caused by malignant cells that remain after surgery. If few in number, the immune system, which is known to contribute to tumour defence,2 should be able to eliminate them. Unfortunately, surgery has an adverse effect on many components of the immune response,3,4 and can be regarded as a `two-edged sword' - on the one hand being the only effective treatment, and on the other impairing the ability of the immune system to eliminate residual tumour cells.
Ideally, there should be a way of performing a surgical resection without depressing the immune response, and results of a study that looked at the effect of lung implanted subcutaneously in rats suggested a possible way to achieve this.5 It was shown that lung contained within a Millipore chamber (MPC) and implanted into an allogeneic (rather than syngeneic) host produced a prolific immune response around the implanted chamber.
The inference from this work is that necrosing lung within the chamber releases antigens that pass through the pores of the MPC and interact with local tissues. It is hypothesised that lung tumour implanted in a similar manner would result in the release of tumour antigen into the local tissues, causing an immune reaction, and that the MPC would prevent local spread.
A proposed clinical application is that, following surgical resection of a lung carcinoma, a segment of tumour contained within an MPC be re-implanted subcutaneously into the patient to augment the antitumour immune response, thereby reducing the likelihood of recurrent disease.
The mouse lung adenoma model6 is a long-established method for assessing potential treatments for lung carcinoma. 4-(methyl nitrosamino)-1-(3 pyridyl)1-butanone (NNK) is the most carcinogenic of the tobacco-specific nitrosamines.7 Lung tumours are induced effectively in syngeneic A/J mice by a single intraperitoneal dose of NNK, and multiple lung tumours develop within 16 weeks.8
Tumours appear as white, spherical lesions that are obvious to the naked eye, and counting these lesions has become a standard method for analysis.9,10 However, this procedure assumes that the tumour development apparent on the surface of the lung accurately reflects the situation within.
In order to assess tumour development within the lung, it is necessary to take multiple histological sections; and it has been demonstrated that sections (5 (mu)m thick) taken at 50 (mu)m intervals through the lung will identify all tumours (+/-2%) present.11 However, this process is considerably more laborious than surface assessment, and would be a major disadvantage in large-scale trials.
A study to examine the effect of subcutaneous implantation of MPCs containing lung tumour into mice with lung tumours, in order to assess a proposed method to reduce recurrence following surgical resection for lung carcinoma, is underway. An important part of this study is the analysis of the stage of tumour development in experimental subjects. Two methods are used: one is based on an examination of the lung surface (surface photographic analysis [SPA]); the other on examination of multiple histological sections of the lung (microscopic tumour analysis [MTA]).
Equivalent parameters are derived using these two techniques to assess tumour frequency and size, and a comparison made between them to determine whether the more simple surface analysis is a valid approximation of tumour development or if it is necessary to perform the more time-consuming microscopic analysis.
Here, the two techniques used to compare lung tumour development in the murine lung adenoma model are described and compared.
Materials and methods
Tumour induction
Lung tumours were induced with a single intraperitoneal injection of 10 pmol/L NNK (Lancaster Synthesis Ltd, Morecambe, Lancs, UK) in 295 syngeneic A/J mice (female, six to nine weeks old). Sixteen weeks later, MPCs containing either normal lung or lung tumour harvested from appropriate donor mice were implanted in the induced tumour group. Between one and eight weeks after MPC implantation, subgroups of mice were sacrificed and their left lungs harvested for analysis of the stage of tumour development (211 subjects in 19 experimental groups). The right lungs were used for analysis of cellular content, which formed part of the overall study but not reported here. Uniform tumour distribution across both lungs has been reported previously in this model.12
Analysis of stage of tumour development
The left lung was harvested from each sacrificed experimental subject and placed in Bouin's solution for 4 h, then placed in 70% ethanol overnight. Each lung was photographed and histological processing completed to paraffin wax for subsequent sectioning.
Surface photographic analysis: Standardised black and white photographs were taken of the medial and lateral surfaces of each lung (Figure 1). Each photograph was analysed to derive the following parameters: surface tumour count (STC) - the total number of visible tumours identified on the medial and lateral surfaces of the lung; surface tumour area (STA) - the total surface area of the visible tumours (each tumour appeared as a circular lesion, and by taking the measured radius [r] of each visible tumour and summing all these to obtain R, the STA was derived from pi[R]^sup 2^); and relative tumour surface area (RTSA) - the ratio (%) of the STA to the lung surface area (LSA), which was derived by approximating the shape of the analysed lung to a right-angled triangle, and the product of the maximum length and width approximated to the LSA (medial and lateral surfaces).
Microscopic tumour analysis: Following embedding in paraffin wax, each lung was sectioned (5 (mu)m thick) every 50 (mu)m, two sections were picked up onto each microscope slide, and stained with haematoxylin and eosin (H&E). In the course of the study, over 4000 sections were produced and mounted on over 2000 slides. In order to make analysis feasible, a computerised image analysis system was used (Alpha Imager(TM) 2000 Documentation and Analysis System, Alpha Innotech Corporation, USA). Using a camera linked to the computer screen, the outer borders of each lung section and tumour were 'marked' using the computer mouse, and the relevant areas calculated automatically (Figure 2). From these measurements, the following parameters were derived: mean tumour count (MTC) - the total number of tumours identified for each lung, divided by the number of sections; mean tumour area - mean area of each section occupied by tumour; and relative tumour area (RTA) -the ratio (%) of the mean tumour area to the mean lung area.
Statistical analysis
The parameters derived by the two techniques assessed equivalent measurements. Both STC and MTC were measures of the number of lung tumours present. STA and mean tumour area were both measures of the absolute size of the tumours. RTSA and RTA were measures of the size of the tumours, relative to the lung.
These equivalent parameters derived from each experimental subject were compared using the Spearman-Rank correlation test to assess their similarity in assessing the degree of tumour development. A correlation coefficient based upon ranked rather than the original observed values was used so that the outcome would not be affected severely by extreme or influential observations.13
Results
The data obtained by the two methods (mean +/- SEM) are shown in Table 1. This was compared for each assessment of tumour development (tumour frequency, tumour area and RTA) between the values obtained by SPA and MTA. The resulting correlation coefficients and the associated statistical significance were as follows:
Assessment of tumour frequency, STC vs. MTC: r = 0.6, P < 0.0001
Assessment of tumour area, STA vs. mean tumour area: r = 0.5, P <0.0001
Assessment of relative tumour area, RTSA vs. RTA: r = 0.4, P <0.0001
Overall, no differences were apparent between the results of either SPA or NITA to assess the stage of tumour development. At one week post-implantation, there was significantly less tumour development in group 2, compared with groups I and 3. At eight weeks post-implantation, there was significantly greater tumour development in group 19, compared with groups 15, 16, 17 and 18. These differences between groups were demonstrated by both SPA and MTA.
Discussion
Prerequisites for use of an animal tumour model to assess different manipulations on tumour development are that measurements of the stage of tumour development are accurate, robust and reproducible. Moreover, it should be relatively simple to ensure that the study is viable. However, if simplicity results in a less accurate study, it would have little value.
Two techniques for analysis of stage of tumour development were used: SPA and MTA. SPA is based upon the surface appearance of the lungs, and attempts to increase the accuracy of the standard simple tumour count by using a permanent photographic record, and includes parameters to assess the size of the tumours (both absolute and relative to the lungs).
Whilst this approach provides more information than a simple tumour count, it assumes that the external appearance reflects tumour development within the lung parenchyma. An analysis based upon examination of histological sections does not rely upon such an assumption; however, the number of sections required to reflect tumour development accurately is unclear.
Previously, it has been reported that sections of lung taken every 50 ltm would identify all (+/-2%) tumours,11 and this protocol (MTA) was used as the second technique for assessing the stage of tumour development. On average, 20 sections were cut from each lung, and, for the whole study, this equated to over 4000 sections on over 2000 slides.
Whilst a computerised image analysis system was used, the time commitment required for analysis of the sections was considerable (over 200 man-hours); in addition to which there was the time required to produce the slides. It is likely that MTA would be unfeasible in a significantly larger study.
The question that this report seeks to answer is whether or not a less time-consuming technique, based upon the surface appearance of the lung, would provide sufficiently comparable results from which to derive valid conclusions. The results presented here show statistically significant correlations (P <0.0001) between equivalent parameters measured by the two techniques (SPA and MTA); however, the correlation coefficients varied between the two techniques (0.4-0.6), indicating that 'similar' estimates are not identical measurements.
To a certain extent, the apparent highly significant P value is a reflection of the large overall size of the study group. The differences in measurements obtained by the two techniques may be due either to one of them providing a less accurate estimate, or to the possibility that they are measuring different aspects of that parameter. This study does not differentiate between these two possibilities. It is likely that MTA provides a more accurate measurement, and SPA is an approximation to this measurement. However, in any experimental study, the relevance of the accuracy of a measurement is whether or not it is sufficient to determine true differences if they exist between experimental groups.
Both MTA and SPA identified differences between the same experimental groups; therefore, whilst it has not been possible to demonstrate whether or not SPA is as accurate as MTA, both provided 'similar' measures and identical conclusions in the overall study.
Conclusions
Surface photographic analysis provides a valid assessment of tumour development in the mouse tumour model used, and is particularly appropriate for largescale studies. Microscopic tumour analysis, whilst potentially the more accurate assessment, should be restricted to small-scale studies because it requires a considerable amount of time to perform.
[Reference]
References
[Reference]
I Takita H. Postoperative therapy for locoregional non-small cell lung cancer. J Surg Oncol 1996; 62: 65-74.
2 Kripke ML. Immunoregulation of carcinogenesis: past, present and future. J Natl Cancer Inst 1988; 80: 722-7.
3 Perttila J, Liliu EM, Salo M. Effects of anaesthesia and surgery on serum opsonic capacity. Acta Anaesthesiol Scand 1986;30: 173-6.
4 Stephan RN, Saizawa M, Conrad PJ, Dean RE, Geha AS, Chaudry IH. Depressed antigen presentation, function and membrane interleukin-1 activity of peritoneal macrophages after laparotomy. Surgery 1987; 102: 147-54.
5 John LCH. Subcutaneously implanted tumour: a hypothesised novel approach to individualised immunotherapy for lung carcinoma. Med Hypoth 1996; 47: 315-9.
6 Shimkin MB, Stoner GD. Lung tumours in mice: application to carcinogenesis bioassay. Adv Cancer Res 1975: 21: 1-58. 7 Hoffmann D, Adams JD, Brunnemann KD, Hecht SS. Assessment of tobacco-specific N-nitrosamines in tobacco products. Cancer Res 1979; 39: 2505-9.
8 Belinsky SA, Devereux TR, Foley JF, Maronpot RR, Anderson MW. The role of the alveolar type II cell in the development and progression of pulmonary tumours induced by NNK in the A/J mouse. Cancer Res 1992; 52: 3164-73.
9 Hecht SS, Morse MA, Eklind KI, Chung FL. A/J mouse lung tumorigenesis by the tobacco-specific nitrosamine NNK, and its inhibition by aryl alkyl isothiocyanates. Exp Lung Res 1991; 17:501-11
[Reference]
10 Morse MA, Amin SG, Hecht SS, Chung FL. Effects of aromatic isothiocyanates on tumorigenicity, O'-methylguanine formation and metabolism of the tobacco-specific nitrosamine NNK in A/J mouse lung. Cancer Res 1989; 49: 2894-7.
11 Rehm S, Ward JM. Quantitative analysis of alveolar type II cell tumours in mice by whole lung serial and step sections. Toxicol Pathol 1989; 17: 737-42.
12 Rehm S, Ward JM, Anderson LM, Riggs CW, Rice JM. Transplacental induction of mouse lung tumours: stage of foetal organogenesis in relation to frequency, morphology, size and neoplastic progression of N-nitrosoethyl urea-induced tumours. Toxicol Pathol 1991; 19: 35-46.
13 Devlin SJ, Gnanadesikan R, Ketterring JR. Robust estimation and outlier detection with correlation coefficients. Biometrics 1975; 62:531-45.
[Author Affiliation]
LINDSAY C. H. JOHN
Department of Cardiothoracic Surgery, Kings College Hospital, Denmark Hill, London SE5 9RS, UK
(Accepted 10 April 2001)
No comments:
Post a Comment