Immunohistochemical detection of aflatoxin B1-DNA adducts and hepatitis B virus antigens in hepatocellular carcinoma and nontumorous liver tissue.

Monoclonal antibodies recognizing the stable imidazole ring-opened form of the major N7-guanine aflatoxin B1-DNA adduct have been used in competitive enzyme-linked immunosorbent assays (ELISA) and indirect immunofluorescence assays to quantitate adduct levels in liver tissue. Methods were developed in AFB1-treated animals, then applied to paired tumor and nontumor liver tissues of hepatocellular carcinoma patients from Taiwan. An avidin-biotin complex staining method was also used for of the detection of hepatitis B surface (HBsAg) and X (HBxAg) antigens in liver sections. A total of 8 (30%) hepatocellular carcinoma (HCC) samples and 7 (26%) adjacent nontumor liver tissue samples from Taiwan were positive for AFB1-DNA adducts. For HBsAg, 10 (37%) HCC samples and 22 (81%) adjacent nontumorous liver samples were positive, and 9 (33%) HCC samples and 11 (41%) adjacent nontumor liver samples were HBxAg positive. No association with AFB1-DNA adducts was observed for HBsAg and HBxAg. These methods should be useful in determining the role of exposure in the induction of HCC in Taiwan. ImagesFIGURE 1. AFIGURE 1. B

MANY solid tumours have a poor blood supply and contain hypoxic cells that are resistant to treatment with radiation. Hypoxic and other poorly nourished cells may also limit the effectiveness of chemotherapy for a number of reasons. The concentration of anti-cancer drugs in such cells may be low because of their position relative to the blood supply, as demonstrated in tumours and spheroids by studying the fluorescence of ADR (Ozols et al., 1979;Sutherland et al., 1979). Also, poorly nourished tumour cells tend to be slowly proliferating (Tannock, 1968(Tannock, , 1970Hirst & Denekamp, 1979) whereas most anti-cancer drugs are more active against rapidly proliferating cells (Tannock, 1978). Finally, drug uptake or activity might be influenced by the nutritional state of the cells or by neighbouring tumour necrosis.
If hypoxic cells in solid tumours are resistant to some anti-cancer drugs, an improved therapeutic index (i.e. ratio of tumour damage to normal tissue damage) might be achieved by including in drug combinations agents with selective toxicitv for hypoxic cells. MISO and METRO are drugs that have been found to have greater toxicity for hypoxic cells than for aerobic cells in tissue culture (e.g. Mohindra & Rauth, 1976;Moore et al., 1976;Stratford & Adams, 1977;Taylor & Rauth, 1978) and to induce necrosis of hypoxic cells in the centre of multicellular spheroids (Sridhar et al., 1976). At high concentration, the same drugs have good penetration into hypoxic regions of tumours, and have been shown to kill hypoxic cells in some, but not all, mouse tumours (Foster et al., 1976;Brown, 1977;Denenkamp, 1978;Pederson et al., 1979). Also MISO may be metabolized in hypoxic tumour regions, with release of metabolites that can kill neighbouring better-oxygenated cells (Brown, 1977;Brown & Yu, 1979). The present series of experiments was designed to study the effect of combining MISO or METRO with conventional anticancer drugs on the response of two experimental tumours, and on host toxicity. Both single and multiple doses of MISO and METRO have been studied, because the half-lives of the drugs in mice (-1 h) are shorter than in man ( 10 h). Methotrexate, 5-fluorouracil, and Adriamycin were chosen for study bec4use they are used commonly to treat solid tumours in man, have selective toxicity for proliferating cells, and because ADR is known to have poor penetration from blood vessels. Experiments in which MISO or METRO were given after X-irradiation to the tumours are included in order to assess the in vivo toxicity of the drugs for hypoxic tumours.

MATERIALS AND METHODS
Animals and Tumours.-C3H male mice (Flow Laboratories) at least 10 weeks old were used in all experiments. Experimental tumours were the KHT fibrosarcoma and the 16/C mammary adenocarcinoma. The KHT tumour has been serially transplanted in our laboratory, and is known to respond to some anti-cancer drugs and to contain hypoxic cells (Lin & Bruce, 1972;Hill & Bush, 1977). The 16/C tumour was obtained from the NCI tumour bank at Mason Research Laboratories, Worcester, Massachusetts, and subsequently has been serially transplanted; it is known to respond to several anti-cancer drugs including ADR and FU (Corbett et al., 1978). For implantation of tumours a single cell suspension was obtained by a method described previously (Thomson & Rauth, 1974) and 2 x 105 cells were injected into the left hind leg of each animal. Both types of tumour will grow progressively from implants of 100 cells.
Growth curves for the tumours were generated as follows. Tumour diameter was recorded to the nearest 0 5 mm by passing the leg through a series of graded holes drilled in perspex, and the tumour weight was estimated from this measurement by a previously defined calibration curve. Animals were coded with numbered ear tags prior to treatment, and randomized groups of 6-8 mice received various treatments. Treatment was given when tumours had a mean diameter of 8-9 mm (weight 0 3 g). Tumour diameter and animal weight were then recorded at 1-3-day intervals by an observer who was unaware of the treatment history. Mean tumour weight and its standard error were plotted against time after treatment. Most experiments were repeated to check reproducibility.
Drugs and Radiation.-Methotrexate (MTX Lederle), 5-fluorouracil (FU, Roche) and Adriamycin (ADR, Adria Laboratories) were standard parenteral formulation obtained from our hospital pharmacy. Metronidazole (METRO) was donated by Poulenc (Montreal) and Misonidazole (MISO) by Roche (Welwyn Garden City, England). All drugs were diluted to an appropriate concentration with physiological saline shortly before use.
All the anti-cancer drugs were given in a fluid volume of 0 01 ml/g body weight by i.p. injection. Because of their limited solubility, MISO and METRO were injected i.p. in volumes of 0-02 or 0 05 ml/g body weight. Control animals received equal volumes of saline.
Serum concentration of MISO or METRO was measured by high-pressure liquid chromatography (HPLC, Gudaskas et at., 1978;Workman et al., 1978). Groups of 3 mice were killed at various times after injection of MISO or METRO, and heparin was injected shortly before death to prevent clotting. Blood samples were collected from the inferior vena cava, and the pooled blood was centrifuged. Plasma was mixed with methanol to precipitate protein, filtered and injected on to the HPLC column. Drug concentration was obtained from a previous calibration curve, derived from samples of known concentration.
Tumours in some experiments were irradiated locally using a double-headed lOOkVp X-ray unit at a dose rate of 11 Gy/min (Siemann et al., 1975). Mice were not anaesthetized for irradiation.
Assessment of haematoloyical toxicity.-Haemoglobin levels (Hb) and white blood cell counts (WBC) were measured f0f individual mice. A small incision was made in the tail and 446 ,ul of blood was collected in a heparinized pipette. The blood was diluted in 10 ml saline, and analysed using a Coulter Counter Model SSr. Blood smears were also prepared and differential counts of polymorphs and other WBCs were recorded for some samples. The method allowed serial estimation of blood counts without killing the animals.

MISO or METRO alone
A single i.p. injection of MISO at a dose of 1 mg/g was given in many experiments. This dose was well tolerated and usually produced only a transient weight loss, < 5 %. The single-dose LD50 for MISO is -1-4 mg/g. Multiple i.p. injections of MISO or METRO were given in other experiments in an attempt to sustain plasma levels over 36 h. Nine doses of 0-2 mg/g (MISO) or 0 4 mg/g (METRO) were given at 4h intervals; these doses were 30-50% of LD50 but caused no deaths, and weight loss < 5%. Any   alone have negligible effects on tumour growth. Haemoglobin and WBC count were measured at various times from 1-7 days after treatment with MISO (1 mg/g single dose) and compared with values for mice receiving saline. Some of these data are included in Table I. There was a tendency for Hb to fall and WBC to increase in mice that were bled serially. There were no differences in mean Hb level between mice receiving MISO or saline, but in each of 5 experiments mean WBC count was slightly lower in animals that had received MISO. However, MISO did not suppress values of WBC and polymorph counts below the range found in control animals.

Radiation
Tumour cells which survive moderate or large doses of radiation in vivo are usually hypoxic. Therefore MISO or METRO was given to tumour-bearing mice after 15 Gy tumour radiation to seek evidence for killing of hypoxic cells in situ. The drugs were given after radiation to avoid radiosensitization.
Regression and regrowth curves after irradiation of KHT and 16/C tumours are shown in Fig. 3. Single dose MISO (1 mg/g) after radiation led to a small increase in effect against the 16/C tumour but in only 1 of 2 experiments (shown in Fig. 3B) was the separation of regrowth curves significant. The same dose of MISO had no effect when given after radiation to the KHT tumour (Fig. 3A) but a higher dose (1.2 mg/g) gave short prolongation of growth delay of about 2 days (not shown).
Methotrexate MTX given as 3 injections of 25 mg/kg at 4h intervals had no effect on growth of the 16/C tumour, and the addition of MISO was also without effect. The same dose and schedule of MTX led to a delay in growth of 2-3 days for the KHT tumour, and a higher dose of 40 mg/kg/injection  given in the same schedule caused some regression and delayed tumour growth by about 4 days. The above dose schedules of MTX were combined with single dose MISO (given with the second of 3 MTX injections), or with the multiple-dose schedule of MISO or METRO (injections of MTX given with the 4th-6th injections of the nitroimidazole). Results of one of the latter experiments are shown in Fig. 4. MISO or METRO did not increase the anti-tumour effects of MTX in any experiment.
MISO and METRO added considerable toxicity when combined with MTX. There were more deaths (usually on Days 4-6 after treatment) and more weight loss in mnice receiving combined treatment (Table   II). Also, the fall in WBC count was lower and more prolonged when Misonidazole was added to treatment with MTX (Table  I) though there was no effect on Hb level.

5-Fluorouracil
A single dose of 100 mg/kg FU delayed growth of the KHT tumour by -3 days and of 16/C by 5 days. The same total dose was more effective against the KHT tumour when given as a single injection than as 3 equal doses at 4h intervals. When a single dose of FU was given in the middle of a course of multiple injections of MISO or METRO to animals bearing the KHT tumour, there was no increase in growth delay over those mice receiving FU alone. A single injection of MISO (1 mg/g) given simultaneously with FU led to a small increase in growth delay of both tumours (Fig. 5).
The addition of a nitromidazole to treatment with FU increased the toxicity as measured by death and loss of body weight (Table III), though the effect was less than for MTX. The fall in WBC count was also lower and more prolonged after combined treatment than after treatment with FU alone.

Adriantycin
Single doses of ADR up to 15 mg/kg i.p. or 20 mg/kg i.v. were ineffective against the KHT tumour, and higher doses killed the animals. In contrast, the 16/C tumour is known to be sensitive to ADR (Corbett et al., 1978) and single doses of 10 or 15 mg/kg i.p. caused complete regression of some tumours (but no cures) and delay to regrowth of 11-13 days (Fig. 6). Simultaneous injection of single-dose MISO (1 mg/g) with ADR prolonged the delay to regrowth by 3-5 days (Fig. 6). Mean loss of body weight after combined treatment with ADR and MISO (8 7%O, range 4-6-12-2%) was no greater than after treatment with ADR alone (9.4%, range 6-8-14-3%). There were no treatment-related deaths in either group, and MISO could be given with 15 mg/kg i.p. of ADR (0/12 deaths) even though a slightly higher dose of ADR alone (20 mg/kg i.p.) was lethal (7/8 deaths). Myelosuppression after ADR was minimal, and although WBC count tended to be slightly lower in mice that also received MISO, the effect was not significant (Table I)

DISCUSSION
The present experiments were designed to provide information relevant to two related questions: (i) Are there important interactions between anti-cancer drugs and MISO or METRO that would encourage or discourage their combination in patients? (ii) Is there evidence for sparing of hypoxic cells by conventional chemotherapy, and for selective killing of hypoxic cells by MISO or METRO?
The first of the above questions is of current and practical importance because clinical trials of nitroimidazoles used alone or in combination with anti-cancer drugs have been proposed. Current trials of MISO as a radiation sensitizer are also including some patients receiving chemotherapy. METRO alone was inactive against colo-rectal carcinoma (Frytak et al., 1978) but its role in combination chemotherapy remains to be defined. However, the results of this and the succeeding paper show that MISO and METRO may add considerable toxicity to conventional anti-cancer drugs. Introduction of such drug combinations in clinical medicine should be made with caution, and in Phase I clinical trials to study toxicity.
The data reported do not answer conclusively the second question about the importance of hypoxic cells in chemotherapy. Limited drug delivery to poorly nourished tumour cells (Ozols et at., 1979), and their documented low rate of proliferation (Tannock, 1968(Tannock, , 1970Hirst & Denekamp, 1979) should convey resistance to many drugs. However, there is no direct evidence that the chronically malnourished cells retain clonogenicity and the ability to re-establish the tumour. There is contrary evidence in some tumours that the hypoxic cells which convey resistance to radiation may be acutely hypoxic because of changes in blood flow (Brown 1979;Yamaura & Matsuzawa, 1979). Such transient hypoxia conveys resistance to radiation because exposure to radiation is short, but might have less influence on cell proliferation and 1.0 : O.5 K8 on drug concentration, unless serum halflife were also very short. There have been few direct studies of the response to chemotherapy of hypoxic and aerobic cells in solid tumours. Cyclophosphamide was reported to spare hypoxic cells in a rat carcinoma but to have no specificity for B16 melanoma, while BCNU and nitrogen mustard were reported to spare hypoxic cells in the B16 melanoma and the KHT sarcoma respectively (Hill & Stanley, 1975;Hill & Bush, 1977;Dixon et al., 1978). I am unaware of data for MTX, FU or ADR.
If chronically hypoxic cells in solid tumours are both clonogenic and resistant to conventional chemotherapy, it remains uncertain whether MISO or METRO can be given in adequate concentration to kill surviving hypoxic cells. Experiments in which these drugs have been given after radiation have demonstrated hypoxic cell toxicity in about half the reported studies (Denekamp, 1978) and in current experiments MISO at 1 mg/g increased the antitumour effect when given after irradiation to the 16/C but not to the KHT tumour. In vitro studies of the toxicity of MISO and METRO for hypoxic cells suggest that drug-cell contact time may be more important than peak concentration (Hall et al., 1978;A. M. Rauth, personal communication). Sustained but lower serum concentration of the drugs is more easily achieved in man where the serum halflives of the drugs are much longer than in mice. Multiple injections were used in current experiments to try to increase drug-cell contact time, but also had little or no effect in prolonging growth delay of the irradiated KHT tumour. Other factors that might limit the contact time between drug and hypoxic cells include a relatively short life of even "chronically hypoxic" cells in murine tumours (Tannock, 1968) and rapid reoxygenation after radiation or drugs (Hill & Bush, 1977). These factors, together with a lower temperature in peripheral murine tumours, and hence lower sensitivity of hypoxic cells to MISO or METRO, could lead to a lower prob-ability of detecting specific toxicity for hypoxic cells in mice than in man (Stratford & Adams, 1978).
In summary, my results show that MISO may increase the effectiveness of FU and ADR against some experimental tumours, though the observed effects are small and the mechanism uncertain. Toxicity was increased when MISO or METRO were combined with MTX or FU, and the combination with MTX led to a decrease in Therapeutic Index. Increased toxicity was not demonstrated when MISO was added to ADR but further experiments to assess a range of normal-tissue toxicity would be essential before concluding that this combination might convey therapeutic benefit.