Prenatal stress enhances NNK-induced lung tumors in A/J mice

Abstract

Children born to women who experience stress during pregnancy have an increased risk of cancer in later life, but no previous animal studies have tested such a link. We questioned whether prenatal stress (PS) in A/J mice affected the development of lung tumors after postnatal response to tobacco-specific nitrosamine, 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK). Timed-bred A/J mice were randomly assigned on gestation day 12.5 to PS by restraint for 5 consecutive days or control (no restraint). Adult offspring of control and stressed pregnancies were all treated with three NNK injections (50 mg/kg every other day) and euthanized 16 weeks later to examine their lungs. Compared with controls, PS dams exhibited significantly increased levels of plasma corticosterone, increased adrenal weights and decreased fetus weights without fetal loss. Prenatally stressed litters had a significantly higher neonatal death rate within first week of life, and surviving male and female offspring developed lung epithelial proliferations with increase multiplicity, increased area and aggressive morphology. PS also induced more advanced atypical adenomatous hyperplasia lesions. We found no difference in lung NNK-derived methyl DNA adducts, but PS did significantly enhance CD3+ T cell and Foxp3+ T cell tumor infiltration. PS significantly increases multiplicity, area of NNK-induced lung tumors and advanced morphology. PS did not affect production of NNK-derived methyl DNA adducts but did increase lymphocytic infiltration of lung tumors. To our knowledge, this is the first animal model of PS with evaluation of cancer development in offspring.

PS significantly increases offspring NNK-induced lung tumor area, multiplicity, with more CD3+ and Foxp3+ T-cell infiltration compared with control mice. To our knowledge, this is the first animal model of PS with evaluation of carcinogenesis in offspring.

Introduction

Lung cancer is the leading cause of cancer-related death worldwide (1), and social stress in adult animals has been shown to promote growth of lung cancer (2). While maternal stress during pregnancy has a detrimental acute effect on both the mother and the fetus, epidemiology studies reveal that prenatal exposure to stress (glucocorticoids) as a risk factor for development in later life cardiovascular disease and neuropsychiatric illnesses, such as depression and schizophrenia (3,4). Furthermore, prenatal stress (PS) is also associated with increased risk of cancers, including testicular cancer, colon cancer and leukemia in later life of offspring (5). Yet, it is not known whether PS impacts susceptibility to lung cancer.

In humans, the proposed mechanisms of carcinogenic activity and developmental programming of disease susceptibility by PS include activation of fetal hypothalamic–pituitary–adrenal axis, epigenetic alteration of stress associated genes, telomere dysfunction or alteration of the immune system (6–10). Since human studies have limitations for understanding the mechanism of PS, animal models can be used to understand the dysregulation of normal prenatal processes that lead to disease later in life in offspring exposed to stress in utero.

Here, we test the hypothesis that PS enhances the sensitivity of offspring to the carcinogenic effects of environmental chemicals. To do this, we examined how exposure to restraint stress in mothers affects the development of pulmonary tumors in offspring using the well-established A/J mouse tumor model, where the tobacco-specific nitrosamine, 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK) produces lung epithelial hyperplasia and neoplasia. Restraint stress in pregnant animals, which is commonly employed to investigate the role of stress in disease development, is reproducible, relatively easy to perform, and induces stress pathways in predictable ways (11,12). NNK is a lung carcinogen in tobacco smoke, used to induce pulmonary epithelial hyperplasia and neoplasia (13,14), and has been classified as a human carcinogen (15). Reactive metabolites of NNK generate DNA adducts, and if not repaired, induce hyperplasia with K-RAS mutations (16,17), and eventually, lung tumor formation in A/J mice (18).

Persistence of O⁶-methylguanine (O⁶-mG) is correlated with the number of lung adenomas in this tumor model (19) and accordingly, we also measured the levels of NNK-derived methyl DNA adducts in our animal model to determine if PS influences NNK metabolism and levels of adducts after treatment. Since PS could also affect the developing immune system of the fetus, we measured CD3+ and Foxp3+ T cell infiltrates in proliferative lung lesions induced by NNK.

Materials and methods

Animals: A/J mice (stock #000646; Jackson Laboratories, Bar Harbor, ME) were housed in standard polycarbonate mouse cages in a temperature and humidity-controlled room under a 14 h:10 h light/dark cycle, lights on 07:00–21:00 h. The animals were allowed access to water and standard chow ad libitum. Female A/J mice were mated between 2 and 4 months of age, and if determined pregnant were randomly assigned to either a PS or control group who received no stress. We selected young A/J mice between the ages of 2–4 months for breeding as these mice do not have a tumor burden (20). Without NNK, the spontaneous tumorigenesis with A/J mice is low but by 18 months can reach 90% (21).

Mice were bred overnight and the male removed the next morning. All mice were weighed before breeding and 12.5 days after mating. Any mouse with a 2.3 g increase of body weight was considered pregnant based on preliminary studies. On gestation day 12.5 (G12.5), 50% of pregnant A/J mice were assigned to the PS group and were given restraint stress for 2 h G12.5 through G16.5—see scheme (Figure 1A). For restraint stress, pregnant mice were placed in a 45 ml pill container constructed with 20 holes for heat and air exchange (Figure 1B). Chow and water were provided ad libitum except during the 2 h restraint stress. Pregnant mice in PS group were housed by themselves for remainder of pregnancy, which is also considered a stressor for mice (22). The control mice were returned to their cages after weighing at G12.5, housed with one or two other female littermates, and were not stressed during their pregnancies. All procedures were approved by the Animal Care and Use Committee of the Johns Hopkins Medical Institutions.

Figure 1.

Experiment design: (A) A/J mice were determined pregnant and randomly assigned to control or PS group. Restraint stress was used for 2 h a day from G12.5 through G16.5 day. Pregnant mice in MC1 were euthanized at G16.5 day to collect plasma samples. MC2 group was euthanized at G19.5. Whereas, pups (OC1) from MC3 dams were weaned after 28 days after birth and at 11 weeks were given NNK injections every other day for 3 days (50 mg/kg). Mice were euthanized at 27 weeks of age. Pups in OC2 were weaned after 28 days after birth and at 7 weeks were given NNK injections every other day for 3 days (50 mg/kg). Mice in OC2 were euthanized 3 days after the last NNK injection to measure DNA adducts. (B) Restrainer for mice. The restrainer is a 45 cc pill case with holes for heat and air exchange. A pregnant A/J mouse is in the restrainer on right.

Experimental design—cohorts assignment

To achieve all of goals of the study, the animals were divided into three maternal cohorts and two offspring cohorts, as described below. All maternal cohorts were subjected to the same scheme of PS (Figure 1).

Maternal Cohort 1 (MC1), pregnant dams were euthanized at G16.5 to measure plasma corticosterone levels. Maternal Cohort 2 (MC2), pregnant dams were followed through pregnancy for body weight and food intake measurements and euthanized at G19.5 to count the number of fetuses and placentas. Maternal Cohort 3 (MC3), pregnant dams which gave birth were monitored for neonatal death rates and their offspring were used for the offspring cohorts.

Offspring Cohort 1 (OC1)—offspring of MC3 were administered NNK at 11 weeks of age, and euthanized at 27 weeks for lung proliferation counting, histological evaluation and plasma corticosterone levels measurements. Offspring Cohort 2 (OC2)—offspring of MC3 were administered NNK and euthanized 3 days later for DNA adducts measurements.

Measurement of stress glucocorticoid hormones in pregnant dams (MC1) and adult offspring (OC1)

To determine if stressors increase plasma glucocorticoid levels, hormone was measured in plasma of pregnant females (n = 5 mice/group). Blood samples were collected into heparinized tubes through a small nick on tail between 8:00 am and 9:00 am (animal room) on the last day of the PS period (G16.5). The stress group was placed under restraint stress for 2 h, and all pregnant mice were euthanized between 10:00 am and 11:00 am. Since cutting the tail of mice was considered a stress, we did not raise pups from these experiments and collected blood at euthanasia to measure poststress glucocorticoid levels. In separate experiments, corticosterone levels were also measured in 27-week-old adult offspring. Plasma samples were separated from whole blood with centrifugation (3000g for 5 min). Plasma corticosterone levels were measured by RIA kits for corticosterone (MP Biomedicals, Costa Mesa, CA) according to the manufacturer’s instructions.

Evaluation of maternal body weight and the number of fetuses and placentas (MC2)

To determine how PS affects food intake and body weight, body weight trends of pregnant mice (n = 6–7 mice per group) during pregnancy were evaluated at G12.5, G16.5 and G19.5. In addition, total amount of food intake was measured between G12.5 and G16.5, and also between G12.5 and G19.5. One day before expecting birth, pregnant females were euthanized at G19.5 in order to evaluate dams and fetuses. Body weight of the stressed pregnant mice was compared with that of control pregnant mice and a ‘pregnant body weight’ was calculated with the formula, (Pregnant body weight) = (Whole body weight) − (Whole uterus weight). Numbers of fetuses and placentas were counted.

Neonatal loss induced by PS (MC3)

To determine whether PS increased incidence of neonatal death, 19 female control mice and 27 PS mice were observed daily for a week after giving birth to monitor the behavior of the dams toward the pups and to count the number of live pups. The number of pups remaining in both groups was compared 7 days after birth.

NNK-induced lung pathology (OC1)

Mouse pups from dams in each group (n = 13–18/group) were weaned 28 days after birth and transferred to cages with up to 5 littermates per cage. At 11 weeks of age, all offspring were given three injections of tobacco-specific carcinogen, nitrosamine 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK) (Toronto Research Chemicals, Toronto, Ontario; 50 mg/kg body weight, given intraperitoneally in vehicle) with 48 h between injections, as described previously (23). Mice were euthanized at 27 weeks of age (16 weeks post-NNK injections) and blood samples were collected through inferior vena cava after euthanasia between 9:00 and 10:30 in the morning for measurement of plasma corticosterone levels (see above). The lungs were perfused in situ through the trachea with 10% formalin and fixed for 48 h. Surface lung proliferations (Figure 2) were counted and photographed under an optical microscope (World Precision Instruments, PZMTIV) and lung lobes were separated to prepare for standard histology. Three different levels of each lung lobe, 250 µm apart, were stained by hematoxylin and eosin (H&E). Proliferative lesions were counted, and after largest radius (r) of each lesion was measured using a Nikon Eclipse 50i microscope (Nikon Corp., Tokyo, Japan), individual proliferation area (cm²) was calculated as 3.14 × r². According to the International Harmonization of Nomenclature and Diagnostic Criteria (INHAND) (24) and the Mouse Models of Human Cancers Consortium (MMHCC) (25), pulmonary proliferative lesions were classified as hyperplasia, atypical adenomatous hyperplasia (AAH) or adenoma. Two pathologists (MD and DVM pathologist) reviewed literature and established criteria on NNK-induced lung tumors in A/J mice and devised a classification scheme to characterize lung proliferation in the current study. Representative lesions are shown in Figure 3. C1 lesions, designated pulmonary epithelial hyperplasia, are characterized by a single layer of cuboidal alveolar cells without nuclear atypia with bronchiolo-alveolar structure maintained (A and B). C2 lesions, designated AAH, are characterized by multiple cell layers of cuboidal heterogeneously enlarged hyperplastic alveolar cells (C and D) with oval to round nuclei. C3 lesions, designated AAH with a central adenoma, are characterized by less clear bronchiolo-alveolar structure papillary growth of cuboidal cells with large round nuclei (E and F). C4 lesions are characterized by distinct adenoma features, with large cuboidal cells, increased mitoses, hyperchromatic cell nuclei and distinct nucleoli (G and H). The tumor cell nuclear membrane is not smooth. Some lesions in this category have expansive growth greater than 1–3 mm in diameter.

Figure 2.

Macroscopic comparison of control and PS group lungs. PS group had significantly more proliferations than control group (males and females). (A) Representative lung of control male group. Scale bar, 2 mm. (B) Representative lung of PS male group. (C) The number of lung proliferations in control male group and PS male group. Values are mean ± standard error (SE). PS male group had significantly more lung proliferations than control male group (n = 18–22, control male group versus PS male group, 9.6 ± 1.01 versus 12.6 ± 0.96, P = 0.0211). (D) Representative lung of control female group. (E) Representative lung of PS female group. (F) The number of lung proliferations in control female group and PS female group. Values are mean ± SE. PS female group also had significantly more lung proliferations than control female group (n = 16–22, control female group versus PS female group, 10.3 ± 0.54 versus 16.9 ± 1.08, P < 0.0001). Mann–Whitney U-test. (G–R) Representative lung sections from three mice (H&E staining) from each group at low magnification where PS groups had more lung epithelial proliferations than control groups. (G, K and O) Representative lungs in control male group. (H, L and P) Representative lungs in PS male group. (I, M and Q) Representative lungs in control female group. (J, N and R) Representative lungs in PS female group.

Figure 3.

Representative examples of classification of four categories for lung proliferation stained by hematoxylin and eosin (H&E). (A) C1 (hyperplasia) for (single layer of hyperplastic alveolar Type II cells lined along alveoli), (B) high power of (A), (C) C2 (AAH) for (multiple layer of hyperplastic alveolar Type II cells lined along alveoli). (D) High power of (C), (E) C3 for (adenoma within AAH), (F) high power of (E), (G) C4 for (adenoma), (H) high power of (G), scale bar for (A, C, E and G); 250 µm, scale bar for (B, D, F and H); 50 µm.

NNK DNA adduct analysis (OC2)

As in OC1, pups were weaned on postnatal day 28 and given three injections of NNK at 7 weeks of age. Mice were euthanized 3 days after the last NNK injection and lungs were excised and frozen for DNA adduct measurements. DNA was isolated from lung tissue using Qiagen PureGene reagents (Germantown, MD) according to the manufacturer’s protocol with the modification that all incubations were performed at room temperature. Levels of 7-methylguanine (7-mG) and O⁶-mG were measured using an established liquid chromatography/tandem mass spectrometry method (26). Guanine concentrations were determined as described previously (27). The amount of DNA adducts was normalized to amount of the guanine in each sample.

Lymphocyte cellular infiltrate in lung proliferative lesions analysis (OC1)

Immunohistochemical staining for CD3+ and Foxp3+ cells was performed on three levels of all lobes of lungs for the four treatment groups. Paraffin sections 5 µm thick were stained using anti-CD3 antibody (clone RM0027-3B19, Santa Cruz Biotechnology) and anti-FOXP3 monoclonal antibody (clone FJK-16s, Thermo Fisher Scientific) and antibody reactivity was detected using Alkaline phosphatase ImPRESS-AP kit, Vector Laboratories and Vector Red (ImPACT Vector Red Alkaline Phosphatase substrate, Vector Laboratories with Methyl Green (Vector) counterstain. For quantification, one lung level was selected per mouse (n = 10/group) and each lung proliferative lesion was identified and photographed by high power magnification (×400) and numbers of CD3+ (cell membrane) and Foxp3+ (nuclei) positive lymphoid cells were quantified using Image J software and checked with manual counts.

Statistics analysis

Data were analyzed by JMP® 12.2.0 (SAS Institute, Cary, NC). The results were expressed as means ± standard error. Measurement data were analyzed by a Unpaired Student’s t-test or Mann–Whitney U-test and multiple comparisons were performed using the Tukey method. Categorical data were analyzed using Fisher’s exact test. P < 0.05 was considered to indicate a statistically significant result.

Results

Restraint stress causes increase in plasma corticosterone levels in pregnant mice

We verified induction of significant physiological stress by restraint with measurements of glucocorticoids in pregnant mice G16.5 in stressed compared with control (non-stressed) pregnant mice. Corticosterone levels were measured in mornings before stress and immediately after 2 h periods of restraint stress in all mice (restraint or no restraint), and as expected, plasma corticosterone levels showed significantly greater increases in mice after stress (Tables 1 and 2). We also noted a trend toward increased morning baseline corticosterone levels in the PS group before stress, but these differences did not meet criteria for statistical significance.

Table 1.

Dam’s features

Cohort	Control	PS	P-value
MC1
Plasma corticosterone levels at prestress (ng/ml)	1388.0 ± 186.8	1979.1 ± 295.5	NS
	(n = 5)	(n = 5)
Plasma corticosterone levels at poststress (ng/ml)	1785.9 ± 261.4	3739.0 ± 346.5	0.0002
MC2
Body weight (g)	34.9 ± 0.86	30.9 ± 0.48	0.0082
	(n = 6)	(n = 7)
Whole uterus weight (g)	11.9 ± 0.67	9.0 ± 0.50	0.0124
Pregnant body weight (g)	23.0 ± 0.60	21.8 ± 0.45	NS
Dams with undeveloped fetuses	2	2	NS
The number of fetuses*	8.0 ± 0.70	7.0 ± 0.63	NS
Placenta and fetus weight/fetuses	1.44 ± 0.03	1.29 ± 0.05	0.0373
Adrenal glands (g)/BW (g) × 103	0.08 ± 0.01	0.11 ± 0.01	0.0096
Kidneys (g)/BW (g) × 103	11.1 ± 0.41	12.2 ± 0.32	NS
Spleen (g)/BW (g) × 103	3.3 ± 0.22	3.2 ± 0.13	NS
MC3
The number of females that lost litters 7 days post birth	3	14	0.0156
	(n = 19)	(n = 27)
MC3
Age at breeding date (weeks)	12.2 ± 0.55	11.1 ± 0.76	NS
	(n = 10)	(n = 11)
Number of pups 3 days after birth	5.4 ± 0.54	6.7 ± 0.42	NS
Number of total pups at weaning	4.9 ± 0.55	6.1 ± 0.39	NS
Weaning offspring sex ratio (male/female)	1.14 ± 0.32	1.44 ± 0.22	NS

BW, body weight; NS not significant

Table 2.

Offspring’s features

Cohort	Male			Female
	Control (N = 18)	PS (N = 22)	P-value	Control (N = 16)	PS (N = 21)	P-value
OC1
The number of offspring mice in a same cage after weaning	3.0 ± 0.24	3.7 ± 0.18	NS	4.1 ± 0.18	4.2 ± 0.18	NS
Body weight (g)	28.7 ± 0.72	29.95 ± 0.53	NS	23.3 ± 0.33	24.3 ± 0.40	0.0142
Heart (g)/BW (g) × 103	4.4 ± 0.09	4.7 ± 0.16	NS	4.6 ± 0.13	4.4 ± 0.10	NS
Right adrenal gland (g)/BW (g) × 103 (n = 8–9)	0.02 ± 0.001	0.03 ± 0.02	0.0184	0.07 ± 0.07	0.09 ± 0.04	0.0313
Right kidney (g)/BW (g) × 103	7.3 ± 1.66	6.1 ± 0.13	NS	5.4 ± 0.16	5.7 ± 0.16	NS
Spleen (g)/BW (g) × 103	2.4 ± 0.01	2.4 ± 0.01	NS	5.1 ± 2.15	3.2 ± 0.12	NS
Plasma corticosterone levels (ng/ml) (n = 5)	410.5 ± 110.3	305.1 ± 43.1	NS	382.5 ± 34.9	407.9 ± 79.6	NS
OC2
7-mG (fmol/nmol G) (n = 5)	50.0 ± 4.20	46.5 ± 1.77	NS	54.6 ± 2.87	52.1 ± 1.23	NS
O6-mG (fmol/nmol G) (n = 5)	16.7 ± 1.00	14.9 ± 0.61	NS	18.2 ± 0.37	18.6 ± 0.79	NS

BW, body weight

Restraint stress causes decreased food intake and decreased body weight gain in pregnant mice during the 5 day stress period

The trend of gained body weight for cohort pregnant mice (MC2) and food intake during the stress period is shown in Supplementary Figure 1, available at Carcinogenesis Online. Food intake was less in PS mice between G12.5 and G16.5 compared with control mice (P = 0.0065) (Supplementary Figure 1B, available at Carcinogenesis Online). Also, pregnant mice in the PS group showed a lower percentage in gain of body weight during the stress period (Supplementary Figure 1A, available at Carcinogenesis Online). After the stress period, there were no differences in food intake in the two groups between G12.5 and G19.5 days (P = 0.8303) (Supplementary Figure 1C, available at Carcinogenesis Online), and the rate of body weight gain was similar among stressed and non-stressed mice.

Restraint stress causes decreased fetal and placental weight

At day G19.5, groups of stressed and non-stressed mice were euthanized and uteruses examined by necropsy (Tables 1 and 2). Uterus weight in the PS group was lower than that in control group, but there was no difference between the two groups in incidence of dams with undeveloped fetuses. However, the average weight for a set of fetuses and placenta was lower in PS group than in the control group. Larger adrenal glands were observed in stressed dams compared with control dams. There was no difference in body weights (without uterus) or weights of kidneys and spleen in control and stressed dams.

Neonatal deaths are higher in litters from restraint stressed pregnancies

Nineteen control pregnant mice and twenty-seven stressed pregnant mice (MC3) delivered at G20.5 and all were first-time mothers. Numbers of live pups delivered per litter in the control and PS groups were not statistically different (P = 0.0824), but 3 of 19 females in control group and 14 of 27 females in PS group lost their entire litters within first week after birth (Tables 1 and 2), with a significant difference in death incidence between the two groups (P = 0.0156). Offspring sex ratios in control and PS groups were not different. Litters that were lost after birth could not be analyzed due to cannibalism.

Prenatal restraint stress in pregnant mice causes increased body weight of adult female offspring

Control and PS adult offspring (both sexes) were compared for body weight, organ weights and plasma corticosterone levels (Tables 1 and 2). Body weights of adult female offspring mice in PS group (n = 21) were increased significantly (P = 0.0142) compared with adult female controls (n = 16). However, there was no difference in weight between control male (n = 18) and PS male (n = 22) groups. Weights of adrenal glands in the PS males (n = 9) and females (n = 8) were significantly increased compared with control males (n = 8) and females (n = 8). There were no significant differences in the weights of heart, kidney and spleen between control and PS mice. In addition, there were no differences in baseline (no stress) plasma corticosterone levels between the control and PS groups (Tables 1 and 2) in samples collected at euthanasia.

Macroscopic evaluation—NNK-induced lung proliferative lesions are increased in number per mouse in male and female mice that were stressed in utero

To determine whether PS causes increased susceptibility to carcinogens when offspring mice reach adulthood, offspring of both sexes in control (no stress) and PS groups were exposed to three doses of NNK at 11 weeks of age and euthanized 4 months later to evaluate lung gross pathology. First, we found that there were significantly more macroscopic lesions (gross pathology) in the PS groups as compared with control groups for both sexes (Figure 2). Note the white circular lesions in the representative photographs of lungs from each group (Figure 2A and B for males, D and E for females). The number of proliferative lesions (gross) was counted and is graphed in (Figure 2C for males and F for females). Offspring in the PS female group had significantly more proliferations (or increased multiplicity) in lungs compared with control female group, 9.5 ± 1.60 versus 14.3 ± 1.14, P = 0.0137. In males, the differences in proliferations were borderline significant but showed the same trend, (control male versus PS male group, 7.8 ± 1.16 versus 10.5 ± 1.05, P = 0.0544). Figure 2G–R is representative photograph of slides from the four treatment groups with proliferative lesions that stain dark purple with the H&E staining.

Microscopic evaluation—NNK-induced lung proliferative lesions are increased in area and more histologically advanced in male and female mice that were stressed in utero

To next validate macroscopic findings with microscopic findings, we discovered that not all white macroscopic lesions on the lung surface are adenomas. We examined three different H&E sections (three levels) of each lung lobe and counted the number of proliferations, measured the area of proliferations and classified the type of proliferations as hyperplasia versus AAH versus adenoma with examples of each category given in (Figure 3). In Figure 4A and B, prenatally stressed female mice have significantly increased number of lung proliferations in the three histological sections examined compared with the control mice, where the prenatally stress male mice showed a trend toward increased numbers of proliferative lesions. In Figure 4C and D, prenatally stressed male and female mice have significantly increased area of proliferative lesions compared with the control mice. The microscopic proliferative lesions were further characterized histologically, details described in methods and examples of categories provided in Figure 2. Figure 4E and F summarizes the frequency of lesion categories in males and females. The control male group had a significantly higher percentage of simpler C1 lesions (hyperplasia) compared with the PS male group with 67.9 ± 5.59% versus 50.4 ± 5.22%, P = 0.0412, whereas control male mice had a lower percentage of C2 (more complex) lesions compared with PS mice, 28.6 ± 5.53% versus 46.6 ± 4.63%, P = 0.0280. The same pattern occurred in PS female group with a higher percentage of C2 proliferation 43.5 ± 5.16% versus 28.8 ± 5.27%, P = 0.0683 and a lower percentage of C1 compared with control group, 50.4 ± 5.11% versus 70.4 ± 5.40%, P = 0.0411. In both males and females, PS offspring had more aggressive lung proliferation lesions, AAH, indicating increased susceptibility to NNK lung carcinogenesis.

Figure 4.

Microscopic comparison of control and PS group lungs. (A and B) The number of epithelial proliferations at three section levels of H&E stained lung tissue by microscopic evaluation (mean ± SE). PS female group had significantly more proliferations in three level sections of lung compared with control female group (n = 13–16, control female group versus PS female group, 9.5 ± 1.60 versus 14.3 ± 1.14, P = 0.0137). In males, there was a marginal difference (n = 16–18, control male group versus PS male group, 7.8 ± 1.16 versus 10.5 ± 1.05, P = 0.0544). M, male; F, female. Mann–Whitney U-test. (C and D) The sum of area (mm²) for all proliferations at three level sections of H&E stained lung tissue by microscopic evaluation (mean ± SE). PS group had larger area of all proliferations at three sections of lung than control group in both genders significantly (n = 13–18, control male group versus PS male group, 1.19 ± 0.27 versus 1.67 ± 0.19, P = 0.0324; control female group versus PS female group, 1.21 ± 0.26 versus 1.94 ± 0.17, P = 0.0168). (E and F) Frequency of each categorized proliferations according to our defined classification microscopic evaluation (mean ± SE). Control male group had higher frequency of C1 proliferations than PS male group significantly (n = 13–18, control male group versus PS male group, C1, 67.9 ± 5.59 versus 50.4 ± 5.22, P = 0.0412, C2, 28.6 ± 5.53 versus 46.6 ± 4.63, P = 0.0280). As for females, PS female group had higher frequency of C2 and lower frequency of C1 than control group marginally (n = 13–18, control female group versus PS female group, C1, 70.4 ± 5.40 versus 50.4 ± 5.11, P = 0.0411, C2, 28.75 ± 5.27 versus 43.53 ± 5.16, P = 0.0683). M, male; F, female. Mann–Whitney U-test.

NNK induces similar levels of DNA adducts in lungs in control and PS offspring groups

Since there were more lung cancer precursor lesions in animals exposed in utero to PS, we questioned whether PS increases the formation of NNK-induced DNA adducts in lung after NNK exposure. Formation of DNA adducts are important for initiation of lung neoplasia in this model (18), and NNK metabolites can both methylate and pyridyloxobutylate DNA with methylation pathway dominating in the A/J mouse lung (28). To evaluate whether PS causes increased levels of DNA adducts, levels of NNK-derived methyl DNA adducts, 7-mG and O⁶-mG were measured in sets of NNK-exposed mice euthanized 3 days after the last NNK dose. No significant difference in levels of adducts was noted between mice that had been stressed in utero compared with control mice in either sex (Tables 1 and 2).

Increased numbers of CD3+ T cells and Foxp3+ T cells are within proliferative lesions in prenatally stressed male and female mice

Next, we evaluated whether PS could affect the immune response to the pulmonary proliferative lesions induced by NNK, using immunohistochemistry to measure CD3+ T cells in lung proliferative lesions in mice from all treatment groups (Figure 5 and Supplementary Figure 2, available at Carcinogenesis Online). Numbers of T-regulatory cells (Tregs, Foxp3+ cells) were also evaluated. We counted CD3 and Tregs cells in lung proliferations and found that prenatally stress exposed adult male (P = 0.0288) and female (P = 0.0001) mice had significantly more CD3 cells within the lung lesions compared with controls (Figure 5). Previously, Foxp3+ T cells (Tregs) have been described in lung tumors of NNK-treated mice, and since Tregs are thought to provide a permissive environment for the development of K-RAS-driven lung tumors (29), we measured Foxp3+ staining lymphocytes in tumor tissues from our experiments. We found that prenatally stress exposed adult male (P = 0.0369) and female mice (P = 0.0007) had significantly more T-regulatory cells within the lung proliferative lesions compared with controls, suggesting that PS induces changes in the tumor immune microenvironment in our experimental model.

Figure 5.

Effect of PS on immune infiltrate in lung epithelial tumor growths (hyperplasia to neoplasia) of NNK-treated mice. Three consecutive tissue sections with lymphocytic infiltration within an alveolar epithelial proliferation were compared by H&E staining (A) and immunohistochemistry for CD3+ T cells (B) and T-regulatory cells (C) in a female prenatally stress mouse. (D) To compare all treatment groups (n = 10/group), CD3+ T cells were stained and counted within lung proliferative lesions of male and female control and prenatally stressed mice. PS induces significantly more T-cell infiltration into lung epithelial proliferations (tumors). (E) The next consecutive section was evaluated in lung sections for Foxp3+ T-cell infiltration. PS induces significantly more Foxp3+ T-cell infiltration into lung epithelial proliferations (tumors). Positive cells for CD3 or Foxp3 stain pink by alkaline phosphatase immunohistochemistry staining. CD3 is positive for cytoplasmic membrane of T cell and Foxp3 is positive for nucleus of T-regulatory cells in the lung lesions. Unpaired Student’s t-test comparison within each sex.

Discussion

Using a novel mouse model to examine cancer susceptibility, we demonstrated that the growth of NNK-induced neoplastic lung lesions (hyperplasia, AAH and adenomas) are enhanced in offspring of mothers that experienced stress while pregnant. These effects were not due to differences in NNK metabolism since PS had no impact on the levels of pulmonary NNK-derived DNA damage. However, PS did influence the immune system as T-cell infiltration of the lung tumor proliferative lesions was greater in the prenatally stressed mice, aligning with the more aggressive carcinogenesis pathway observed in the in utero stressed mice.

Previously, there were no animal models to study the influence of in utero maternal stress on cancer susceptibility. This association has been observed in large human populations exposed to environmental hardships caused by genocide, famine and natural disasters. Since the Developmental Origins of Health and Disease (DOHaD) hypothesis links in utero exposures to undernutrition, over-nutrition, chemicals, hormones and stress to increased risk of diseases in adulthood (30,31), we set out to create a model to test this hypothesis.

A number of models of stress and timing of stress were considered (32,33). Since we did not want the stress to cause fetal death, we selected 5 days of restraint stress that had previously been successfully used for mice (34). To validate that our approach caused physiologically significant stress in pregnant mice, we demonstrated that plasma corticosterone levels were about three times higher in the restrained pregnant mice than in the control animals. Normally maternal corticosterone levels increase with gestational stage, which is important for fetal lung maturation (35).

Stressed pregnant mice gained less weight compared with controls partly due to reduced food consumption during the restraint period, consistent with previous reports on PS in rats (36). Stressed pregnant mice had smaller uteruses at G19.5, and the average weight for fetuses and placentas in PS group was less than that in the control group. The number of fetuses in the stressed group was not different from that in control group, indicating that PS did not induce fetal death. These results suggest that stress during pregnancy affected size and weight of fetuses at birth. This finding is consistent with previous human studies linking maternal undernutrition during pregnancy with intrauterine growth retardation (37), while undernutrition was associated with increased breast cancer risk in adult offspring of the Dutch Famine (38).

Even with our conservative approach to PS, we observed increased (first week of life) neonatal loss in the PS group. Remarkably, 14 of 27 litters (51.9%) died the first week after birth in dams that were exposed to PS compared with 3 of 19 litters (15.8%) in the control group (P = 0.0156). We were not able to determine the cause of offspring death, but a possible reason could be the lack of adequate mothering in prenatally stressed dams.

Exposure to PS in humans is associated with diabetes in offspring and weight gain in adults (38). In the current study, body weights of prenatally stressed female offspring mice at 27 weeks were significantly increased compared with body weights in control females. Conversely, PS in pregnant rats induced a lower body weight in the adult offspring at 12 weeks of age (39). Yet, similar to our mouse findings, in humans, young adults whose mothers experienced life stress during pregnancy had higher body mass index, percent body fat and insulin resistance (40). Taken together, the body weight of offspring from stressed mothers is partially dependent on the mother, genetic background and how the mother responds to stress during pregnancy.

We chose A/J mice, since this strain is prone to developing lung neoplasia with age as a result of their genetic susceptibility to a higher incidence of activating mutations in the K-RAS oncogene (41,42). NNK, a tobacco-specific carcinogen, induces lung hyperplasia, adenomas and adenocarcinomas in this strain (14). Activated NNK induces mutations in oncogenes such as K-RAS by forming DNA adducts, such as O⁶-mG and 7-mG (43). Consistently, K-RAS mutations are found in hyperplastic lesions in mice after NNK treatment (16).

In the current study, we observed more macroscopic and microscopic lung proliferations in the PS offspring group compared with controls. We selected the 16 week endpoint because NNK-treated mice developed adenomas at 14 weeks after NNK injection (17,44), with hyperplasia progressing to adenomas between 16 and 20 weeks post-NNK injection (17). The relatively early time point (16 weeks) selected in our study allowed us to observed novel differences in the degree and extent of hyperplasia caused by PS. This prenatal exposure impacted the frequency and progression of these lesions from simple hyperplasia (C1) to AAH (C2). These pulmonary lesions differences may not have been observed in these mice if a later time point had been selected.

AAH is more frequently used in human literature but has been used to describe proliferative lesions in mice lungs (25). Interestingly, some AAH lesions in human lung tissue have K-RAS mutations (45). Our work offers a novel observation that the PS group has more C2 (AAH) proliferative lesions compared with control group suggesting PS induces a lesion progression (C1 > C2) resulting in more advanced C2 lesions. We do not know if the number of K-RAS mutations increase in proliferative lesions progressing from C1 to C2. Additionally, we did not evaluate NNK DNA adducts specifically in K-RAS or in tumor suppressor genes. It is possible that there could be a significant presence of NNK DNA adducts (7-mG and O⁶-mG) at the K-RAS locus, yet not contribute to an overall difference between the stress groups when measured globally. As far is we could determine, NNK lung hyperplastic proliferations have never been morphologically subcategorized and mapping the mutation profiles of these two different lesions could be very informative in the pairing of mutational events in carcinogenesis, cellular morphology and cancer progression.

One mechanism by which PS could increase susceptibility to the carcinogenic effects of NNK is by altering NNK metabolism to cause more pulmonary DNA adducts. To evaluate this possibility, O⁶-mG and 7-mG adduct levels were measured in lungs of mice. There was no difference in the levels of DNA adducts in the lungs of the control and PS groups 72 h after the last dose of NNK. This observation rules out heightened metabolism of NNK as a mechanism explaining the increased sensitivity of PS exposed mice to NNK.

Another possible mechanism is that PS alters the immune system and, consequently, the carcinogenesis process. Previously, it was shown that NNK-induced Foxp3+ T cells are thought to provide a permissive environment for the development of K-RAS-driven lung tumors (29). We found that PS does induce significantly more CD3+ T cell and T-regulatory cell infiltrates in lung proliferative lesions in males and in females early in the carcinogenesis process, and this may be due to a more aggressive (AAH) proliferations observed in the prenatally stressed mice.

In conclusion, PS in mice increases the lung carcinogenesis of NNK in A/J mice. This is not acutely associated with more adducts 3 days subsequent to dosing. Significantly more T cells and specifically Treg cells infiltrate these more aggressive lesions observed in the mice exposed to PS. This is the first animal model that demonstrates that PS predisposes offspring to cancer. This enhanced carcinogenesis effect was seen in both male and female offspring. Multiple epidemiological studies in humans have shown an association between PS and increased cancer outcomes in offspring (5,46,47); in addition, body length at birth and/or the size of placenta was associated with lung cancer occurrence in adult life (48). This new mouse model can be studied to further the understanding of mechanisms behind carcinogenesis enhancement by PS.

Supplementary material

Supplementary data are available at Carcinogenesis online.

Supplementary Figure 1. The trend of body weight and the comparison of food intake during pregnancy. (A) Trend of relative gained body weight for pregnant mice and amount of food intake between G12.5 and G19.5 days. Pregnant mice in PS group gained less than control mice significantly. Blue dots and line, control group; red dots and line, PS group (n = 6–7). *P = 0.0053, **P = 0.0034. (B) The comparison of food intake between G12.5 and G16.5. There was significant difference in the food intake between the control and PS groups (P = 0.0065). (C) The comparison of food intake between G12.5 and G19.5. There was no difference between the amount of food intake between G12.5 and G19.5 days in control group and that in PS group (n = 6–7, mean ± standard error, P = 0.8303). PS, prenatal stress group.

Supplementary Figure 2. Immunohistochemistry examples from each treatment group. Immunohistochemistry for T cells and T-regulatory cells in lung epithelial proliferations. (A, C, E, G) CD3 staining (A: control male, C: PS male, E: control female, G: PS female). (B, D, F, H) Foxp3 staining (B: control male, D: PS male, F: control female, H: PS female). Positive cells for CD3 or Foxp3 stain pink by alkaline phosphatase staining. CD3 is positive for cytoplasmic membrane of T cell and Foxp3 is positive for nucleus of T-regulatory cells in the lung lesions.

Glossary

Abbreviations

AAH

atypical adenomatous hyperplasia

O6-mG

O6-methylguanine

PS

prenatal stress

Funding

This study was supported in part by a Grant-in-Aid from MEXT-Supported Program for the Strategic Research Foundation at Private Universities, 2015–19. Tomoaki Ito is currently receiving a grant from Toray Medical Co., Ltd. DNA adduct measurements were supported by National Institutes of Health (R01 CA184987) and a generous gift from Minnesota Masonic Charities to the Masonic Cancer Center, University of Minnesota.