Tumor grade,microvessel density,and activities of malate dehydrogenase,lactate dehydrogenase,and hexokinase in squamous cell carcinoma

Abstract

Squamous cell carcinomas were evaluated with respect to tumor differentiation (through use of hematoxylin and eosin stain), microvessel density (through use of CD-34 immunocytochemical stain), and magnitudes of malate dehydrogenase (MDH), hexokinase, and lactate dehydrogenase (LDH) enzyme activities. Direct correlations were found between tumor grade, MDH activity, and microves-sel density. Direct correlations were also found between hexokinase activity and MDH activity and microvessel density. Inverse correlations were found between LDH activity and both tumor grade and MDH activity. These results suggest that the high rate of glucose utilization (indicated by hexokinase activity) found in more poorly differentiated tumors has a higher component of aerobic oxidative metabolism (indicated by MDH activity) and a relatively lower contribution from anaerobic metabolism (indicated by LDH activity) than do the rates found in more differentiated tumors. It is also suggested that as the glycolytic rate increases, more pyruvate goes into the Krebs cycle than into lactate. The availability of glucose-derived pyruvate for oxidative metabolism would mean less of a dependency on glutamine as a carbon source in squamous cell carcinoma.

A number of different and independent approaches have been used to characterize cancerous tissue. One common method involves histopathologic evaluation of tumor differentiation through use of sections stained with hematoxylin and eosin. Tumor microvessel density has also been studied and has been correlated with metastasis and patient survival. 1 - 4 Enzymes and substrates of metabolism have been studied, particularly in animal models and cultured cells, and the unusual biochemistry of cancer tissue has been well documented. 5 - 7

Malate dehydrogenase (MDH) and lactate dehydrogenase (LDH) are key energy-related enzymes involved in aerobic and anaerobic metabolism, respectively. These enzymes are important reflections of oxidative and glycolytic reactions in brain, muscle, and other tissues. LDH activity has been measured in samples from animal and human tumors, tissue culture cells, and human serum. Activities of MDH and LDH were determined in samples of brain tumors, but little correlation was found between these enzymes among different types of tumors or among tumors of the same type. 8 However, it has been suggested that activities of MDH and LDH are useful in characterizing the metabolic biochemistry of squamous cell carcinoma tumors. 9

Relatively few studies have attempted to correlate several different parameters studied quantitatively in the same human tumor. This study correlates parameters in squamous cell carcinoma of the aerodigestive tract through use of 3 different and independent approaches: a cellular stain (reflecting degree of tumor differentiation), an immunocytochemical stain (CD-34, reflecting microvessel density), and enzyme activities (MDH, hexokinase, and LDH, reflecting energy metabolism). To facilitate this correlation, a numeric scale was developed to express the degree of tumor differentiation in a more quantitative manner than is usually done clinically. Not unexpectedly, it was found that more poorly differentiated tumors tend to have a higher blood vessel density and higher MDH and hexokinase activities. However, the reciprocal relationships between LDH and MDH activities and between LDH activity and tumor differentiation found in squamous cell carcinoma differ from what might have been expected on the basis of previous studies in which other tumor types were used.

METHODS AND MATERIAL

The study protocol was reviewed by the University of Oklahoma Health Sciences Center Institutional Review Board and given an exempt classification. Many of the methods used in this study have been previously published in detail.9 These descriptions have pertained to the obtaining, handling, and freezing of tumor tissue, cryostat sectioning, freeze-drying of frozen sections, microdissection and weighing of samples at room temperature with controlled humidity, fluorometric assays for MDH and LDH activities, and hematoxylin and eosin staining of sections thawed onto microscope slides.

Fig 1.

Correlations between tumor grade, or differentiation scale, and MDH activity (A), LDH activity (B), MDH/LDH activity ratio (C), and microvessel density (D). On x axis, 1 = well differentiated, 5 = moderately differentiated, and 9 = poorly differentiated. MDH and LDH activity units are same as those in Table 1.

Hexokinase activity was determined by measuring the rate of increase in reduced nicotinamide adenine dinucleotide phosphate fluorescence produced in a 1-mL reaction containing 2 mmol/L glucose, 5 mmol/L ATP, 5 mmol/L NADP+, 5 mmol/L MgCl₂, 0.05% bovine serum albumin, 0.5% Triton X-100, 0.3 U/mL glucose-6-phosphate dehydrogenase, and 100 mmol/L Tris buffer, pH 8.0. 10 Chemicals were obtained from Sigma Chemical Company (St Louis, MO).

Both hexokinase and MDH have mitochondrial and soluble forms, and although LDH is soluble, it exists in different isoenzymes. In cancer cells, the mitochondrial isoenzyme of hexokinase is considered to predominate, 6 , 7 , 11 and this is probably the case for MDH as well.8 In this study, activities of isoenzymes were not separately determined; reported values represent total enzyme activities.

To determine whether correlations existed between histo-pathologic tumor grade and assayed enzyme activities, an attempt was made to express the grade in a more quantitative manner than is usually done for clinical purposes. Hematoxylin-and-eosin-stained sections of each tumor, in the same series as those freeze-dried for biochemical assay, were examined independently by two pathologists (E.G., R.J.) who at the time of analysis did not know any of the biochemical results. In some cases the biochemical assays were done after the histologic examination; in other cases, before. Histopathologic evaluation was based on criteria similar to those used clinically, including squamous differentiation, cytologic atypia, degree of stromal and lymphovascular invasion, tumor necrosis, mitotic activity, and infiltrative margin in the tumor section. Each pathologist assigned a numeric grade to each tumor through use of a scale from 0 to 10, 0 being normal-appearing, 1 being well differentiated, 5 being moderately differentiated, 9 being poorly differentiated, and 10 being an undifferentiated malignancy. Each pathologist gave consistent grades to sections of the same tumor examined at different times. The grades assigned by the two pathologists for the same tumor were very similar, sometimes identical. The average of the two assigned grades was calculated for each tumor.

Fig 2.

Correlations between microvessel density and MDH activity (A) and hexokinase activity (B). Microvessel density is mean count from five 100× fields in CD-34-stained sections. MDH and hexokinase activity units are same as those in Table 1.

Table 1.

Numeric grade of tumor differentiation, enzyme activities, and microvessel density

Tumor ∗	Grade	MDH activity †	LDH activity †	MDH/LDH	Hexokinase activity †	Microvessel density ‡
A	9.0	4.72 ± 0.23 (11)	8.51 ± 0.41 (7)	0.555	3.16 ± 0.20 (8)	674
B	7.5	3.15 ± 0.22 (8)	9.51 ± 0.24 (7)	0.331	0.83 ± 0.06 (8)	242
C	3.8	2.33 ± 0.07 (34)	19.9 ± 0.92 (33)	0.117	1.09 ± 0.05 (4)	286
D	6.0	2.28 ± 0.16 (24)	9.92 ± 0.75 (29)	0.230	1.47 ± 0.08 (4)	205
E	2.0	1.72 ± 0.13 (14)	12.0 ± 0.41 (4)	0.143	0.82 ± 0.18 (4)	97
F	3.0	1.88 ± 0.10 (14)	15.7 ± 1.40 (5)	0.120	1.09 ± 0.06 (5)	106
G	1.5	1.28 ± 0.11 (21)	20.5 ± 0.85 (9)	0.062	1.29 ± 0.04 (10)	94
H	4.0	1.95 ± 0.07 (14)	11.5 ± 1.02 (7)	0.169	1.17 ± 0.05 (8)	163
I	3.0	2.21 ± 0.18 (14)	24.1 ± 3.13 (8)	0.092	1.15 ± 0.15 (6)	199
J	8.0	2.01 ± 0.14 (17)	11.4 ± 1.42 (8)	0.177	1.54 ± 0.22 (9)	189
K	0.5	1.64 ± 0.15 (13)	12.4 ± 0.85 (4)	0.133	0.70 ± 0.06 (5)	235
L	8.5	2.10 ± 0.18 (9)	11.3 ± 0.60 (4)	0.186	1.16 ± 0.07 (4)	479
M	3.9	2.44 ± 0.15 (9)	13.8 ± 1.32 (7)	0.177	1.60 ± 0.11 (7)	271
N	7.5	1.93 ± 0.06 (8)	11.7 ± 0.57 (4)	0.165	1.09 ± 0.11 (4)	245
O	6.0	1.80 ± 0.08 (10)	19.1 ± 1.18 (7)	0.094	1.02 ± 0.05 (7)	107
P	6.0	3.62 ± 0.20 (10)	9.55 ± 1.12 (10)	0.379	1.82 ± 0.18 (3)	131
Q	8.0	2.46 ± 0.02 (3)	8.41 ± 0.65 (3)	0.293	1.04 (1)	217
R	8.0	3.39 ± 0.37 (5)	7.23 ± 0.64 (5)	0.469	1.34 ± 0.12 (3)	271

∗Tumor letters correspond to those used in Figs 1–3. Locations were oral (C, F, G, I, K), hypopharynx (A, B, D, J, L), and larynx (E, H, M, N, O, P, Q, R).

†Mean in moles product per kg dry weight per hour at 25°C ± SEM (number of samples).

‡Mean of counts in five 100× fields.

Cryostat sections of each tumor were thawed onto Probe-On Plus slides (Fisher Scientific, Pittsburgh, PA) for determination of blood vessel density. Sections were fixed for 10 minutes in phosphate-buffered 4% paraformaldehyde, rinsed in buffer, incubated for 30 minutes in 0.3% hydrogen peroxide, and rinsed again in buffer. After incubation for 30 minutes in normal blocking serum, sections were incubated for 30 minutes through use of the avidin/biotin blocking kit before an overnight incubation with CD-34 primary antibody at 1:25 dilution. After several rinses, staining was visualized through use of the Vectastain ABC and DAB substrate kits with nickel intensification. Fixation, rinsing, and staining procedures were done at room temperature. Blocking, antibody, and visualization steps were all done with materials from Novocastra (Vector Laboratories, Burlingame, CA). After air-drying, mounting, and cover-slipping, microvessels were counted by means of a microscope with bright-field illumination. Micro-vessels were counted in five 100× fields in tumor areas containing blood vessels, and means were determined. A stained endothelial cell or cell cluster separate from other endothelial cells, tumor cells, or other connective tissue elements was considered a countable microvessel. Visualization of vessel lumen was not necessary. All counts were performed in a consistent manner by one investigator (M.G.).

Fig 3.

A, Inverse correlation between MDH and LDH activities. B, Direct correlation between MDH and hexokinase activities. Enzyme activity units are same as those in Table 1.

RESULTS

Histologic and biochemical examinations were done on tumors from 18 patients (Table 1). Tumors were located in the oral cavity, hypopharynx, or larynx. MDH activity was correlated (P < 0.05) with tumor stage (T1–T4) in laryngeal tumors only. No other correlation between experimentally measured parameters and tumor stage or size was found, perhaps because of the relatively small number of tumors. The degree of differentiation varied among tumors from almost normal-appearing (grade 0.5) to poorly differentiated (grade 9.0). Microvessel density also varied greatly among tumors, with a greater than 7-fold difference between the lowest and highest. Enzyme activities varied between different tumors, with almost a 4-fold difference between lowest and highest values of MDH activity, a greater than 3-fold difference in LDH activity, a 6-fold range in MDH/LDH ratio, and a greater than 4-fold range in hexokinase activity (Table 1).

The differentiation scale of tumor grade correlated significantly with MDH and LDH activities; higher MDH but lower LDH activity was associated with more poorly differentiated tumors (Fig 1 A and B). The ratio of MDH to LDH activity correlated with tumor grade slightly better than did MDH alone (Fig 1 C). Hexokinase activity tended to be higher in more poorly differentiated tumors, but the relationship was not significant. Microvessel density was higher in more poorly differentiated tumors (Fig 1 D). Microvessel density also correlated directly with MDH and hexokinase activities (Fig 2 A and B) and with the MDH/LDH ratio (r = 0.582; P < 0.05). The inverse correlation between microvessel density and LDH activity was not significant.

A reciprocal relationship was found between MDH and LDH activities: tumors with higher MDH activity had, on average, lower LDH activity (Fig 3 A). Hexokinase activity directly correlated with MDH activity (Fig 3B) and also with the MDH/LDH ratio (r = 0.656; P < 0.01). An inverse correlation between hexokinase activity and LDH activity was not significant.

Pronounced spontaneous necrosis within tumors is generally considered to be of significance in characterizing a tumor as poorly differentiated and aggressive. This observation is supported by our biochemical analyses. The tumor with the largest necrotic areas (Fig 4) had the highest microvessel density, the highest activities of MDH and hexokinase, and the highest MDH/LDH ratio (tumor A; Table 1), suggesting a very high aerobic metabolic rate of the tumor cells.

DISCUSSION

It is well known that cancer cells have a high rate of metabolism. 6 , 7 It has been shown that poorly differentiated and rapidly growing tumors have an increased glycolytic capacity, which accounts for a significant proportion of their energy needs. 6 , 7 , 11 Glucose taken up by cells is phosphorylated by hexokinase to produce glu-cose-6-phosphate. Our finding that hexokinase activity is generally higher in more poorly differentiated squamous cell carcinomas is consistent with previous evidence that hexokinase activity is high in rapidly growing, highly glycolytic tumors. 6 , 7 , 11 One isoform of hexokinase, type II, particularly expressed in cancer cells, is mitochondr-ial-bound, has a high affinity for glucose, and has a reduced sensitivity to feedback inhibition by glucose-6-phosphate. 11 It is possible that this isoform predominates in the squamous cell carcinoma tumors in our study.

Fig 4.

A and B, Hematoxylineosin–stained section of tumor A shows two different areas of necrosis. This tumor had highest activities of MDH and hexokinase, highest MDH/LDH ratio, and highest microvessel density (Table 1). (A and B, Scale bar = 1.0 mm.)

The magnitude of hexokinase activity in these tumors is assumed to be proportional to the rate of glucose uptake and the rate of glycolysis, including, at least approximately, the rate of pyruvate production. We examined the activities of LDH and MDH to indicate the relative amounts of pyruvate that might be converted to lactate or enter the Krebs cycle for oxidative metabolism.

The magnitudes of MDH and LDH activities have been studied in nontumor tissues to indicate the degree to which oxidative or glycolytic metabolism contributes to energy production in each tissue. Neural areas with high mitochondrial densities have high contributions from oxidative metabolism and high MDH activities. 12 , 13 In skeletal muscle, MDH is high in “slow-oxidative” muscle fibers whereas LDH is high in “fast-glycolytic” muscle fibers. 14 , 15 In nervous tissue, MDH/LDH ratios are relatively constant in different areas, major exceptions being the retina 13 and the organ of Corti, 15 where relatively high LDH values reflect a poorer blood supply and more energy production through anaerobic metabolism. 16 , 17

Therefore, the correlations between tumor grade and increasing MDH and decreasing LDH activities found in these squamous cell carcinomas suggest that more poorly differentiated tumors tend to have a relatively higher contribution of aerobic metabolism, and a correspondingly lower contribution of anaerobic metabolism, to energy production. Oxidative metabolism requires oxygen, but it produces many more molecules of ATP per glucose molecule than does glycolytic metabolism. The correlation between microvessel density and MDH activity is consistent with a greater tumor blood supply and presumably with higher oxygen availability, associated as it is with a greater component of oxidative metabolism and more energy production. This higher production of ATP in more oxidative tumor tissue might be associated with cellular energy-requiring functions, such as a high rate of proliferation or the active drug efflux pump. 18

Glucose uptake in normal cells under anaerobic conditions decreases when oxygen is made available. 6 , 7 In most cancer cells, rather than glucose utilization being reduced with increased oxygen availability, the glucose is energy-inefficiently converted to more lactate through LDH. 6 , 7 In these cancers, glutamine provides a carbon source for the Krebs cycle. 6 , 7 , 19 – 22 However, the decrease in LDH activity with increased MDH found in our study of squamous cell carcinoma implies that with increased glycolysis, additional pyruvate may be channeled to the oxidative Krebs cycle reactions rather than into lactate. The availability of glucose-derived pyruvate for oxidative metabolism would mean less of a dependency on glutamine as a carbon source in squamous cell carcinoma. This suggestion presumes a relation between the magnitudes of LDH and MDH activities and substrate/product concentrations and turnover.

The significant correlations found in this study between tumor grade, microvessel density, and enzyme activities reflect general trends seen in a population of tumors. However, large quantitative differences, particularly in enzyme activities, occur among individual tumors (Figs 1–3). Obviously, much more understanding is needed concerning the biochemistry of squamous cell carcinomas before one tumor sample from a patient could be examined and the information from the analysis used with high confidence for a prognosis or for prediction of treatment response.

CONCLUSION

These results indicate that more poorly differentiated squamous cell carcinomas generally have higher micro-vessel density and use more aerobic mechanisms for energy production, as indicated by their higher MDH activity and lower LDH activity. The reciprocal relationship between MDH and LDH is actually a logical pattern based on the activities of these enzymes in non-tumor tissues. The difference in MDH/LDH ratio between well and poorly differentiated tumors implies that on the average, well differentiated tumors are more anaerobic and poorly differentiated tumors more aerobic in their energy production, with more glucose-derived pyruvate possibly entering the Krebs cycle. The fate of glucose-derived pyruvate is important because this may reflect the extent that glutamine is used as an energy source in squamous cell carcinoma. Energy-metabolic reactions may fuel processes that are important to the tumor, such as angiogenesis, proliferation, metastasis, and mechanisms of resistance to treatment. These processes are also important considerations in the design of effective cancer therapies. Therefore, further study on the metabolic biochemistry of squamous cell carcinoma is warranted.

We thank Dr Donald Godfrey, Department of Otolaryngology–Head & Neck Surgery, Medical College of Ohio, for helpful comments on the manuscript.

References

1. Heimann

Ferguson

Powers

, Angiogenesis as a predictor of long-term survival for patients with node-negative breast cancer. J Natl Cancer Inst 1966; 88: 1764–9.

2. Obermair

Wanner

Bilgi

, Tumor angiogenesis in stage IB cervical cancer: correlation of microvessel density with survival. Am J Obstet Gynecol 1998; 178: 314–9.

3. Shpitzer

Chaimoff

Gal

, Tumor angiogenesis as a prognostic factor in early oral tongue cancer. Arch Otolaryngol Head Neck Surg 1996; 122: 865–8.

4. Roychowdury

Tseng

Jr Fu

, New prognostic factors in nasopharyngeal carcinoma: tumor angiogenesis and C-erbB2 expression. Cancer 1996; 77: 1419–26.

5. Weber

. Biochemical strategy of cancer cells and the design of chemotherapy: GHA Clowes Memorial Lecture. Cancer Res 1983; 43: 3466–92.

6. Eigenbrodt

Fister

Reinacher

. New perspectives on carbohydrate metabolism in tumor cells. In: Beitner

, editor. Regulation of carbohydrate metabolism. Vol 2. Boca Raton: CRC Press, Inc; 1985. p 141–79.

7. Baggetto

. Deviant energetic metabolism of glycolytic cancer cells. Biochimie 1992:74:959–74.

8. Lowry

Berger

Carter

, Diversity of metabolic patterns in brain tumors: enzymes of energy metabolism and related metabolites and cofactors. J Neurochem 1983; 41: 994–1010.

9. Ross

Medina

. Quantitative biochemical analysis of samples from squamous cell carcinoma. Otolaryngol Head Neck Surg 1998; 119: 455–9.

10.

10. Passonneau

Lowry

. Enzymatic analysis: a practical guide. Totowa, NJ: Humana Press; 1993.

11.

11. Mathupala

Rempel

Pederson

. Aberrant glycolytic metabolism of cancer cells: a remarkable coordination of genetic, transcriptional, post-translational, and mutational events that lead to a critical role for Type II hexokinase. J Bioenerg Biomembr 1997; 29: 339–43.

12.

12. Ross

Godfrey

. Distributions of aspartate aminotransferase and malate dehydrogenase activities in rat retinal layers. J Histochem Cytochem 1985; 33: 624–30.

13.

13. Ross

Godfrey

Parli

. Amino acid concentrations and selected enzyme activities in rat auditory, olfactory and visual systems. Neurochem Res 1995; 12: 1483–90.

14.

14. Peter

Barnard

Edgerton

, Metabolic profiles of three fiber types of skeletal muscle in guinea pigs and rabbits. Biochemistry 1972; 11: 2727–33.

15.

15. Godfrey

Wiet

Ross

. Quantitative histochemistry of the cochlea. In: Altschuler

Hoffman

Bobbin

, editors. Neurobiology of hearing: the cochlea. New York: Raven Press; 1986. p 149–60.

16.

16. Lowry

Roberts

Schulz Dw

, Quantitative histochemistry of retina, II: enzymes of glucose metabolism. J Biol Chem 1961; 236: 2813–20.

17.

17. Matschinsky

Thalmann

. Energy metabolism of the cochlear duct. In: Paparella

, editor. Biochemical mechanisms in hearing and deafness. Springfield: Charles C. Thomas; 1970. p 265–94.

18.

18. Baggetto

. Biochemical, genetic, and metabolic adaptations of tumor cells that express the typical multidrug-resistance phenotype: reversion by new therapies. J Bioenerg Biomembr 1997; 29: 401–13.

19.

19. Kovacevic

Morris

. The role of glutamine in the oxidative metabolism of malignant cells. Cancer Res 1972; 32: 326–33.

20.

20. Reitzer

Wice

Kennell

. Evidence that glutamine, not sugar, is the major energy source for cultured HeLa cells. J Biol Chem 1979; 254: 2669–76.

21.

21. Moreadith

Lehninger

. The pathways of glutamate and glutamine oxidation by tumor cell mitochondria. J Biol Chem 1984; 259: 6215–21.

22.

22. Matsuno

. Bioenergetics of tumor cells: glutamine metabolism in tumor cell mitochondria. Int J Biochem 1987; 19: 303–7.