Introduction

Heterodimeric glycoprotein hormones, luteinizing hormone (LH) and human chorionic gonadotropin (hCG), are members of the cystine-knot growth factor families [1]. They are highly homologous and bind to the same receptor to initiate a series of signaling cascades for their endocrine, para/autocrine, and intracrine actions. LH/hCG receptor, a member of the G-protein coupled receptor superfamily, presents in gonads and a number of nongonadal tissues including thymus and lymphocytes [2–6]. Binding of LH and hCG to their cognate receptors results in activation of cyclic AMP-protein kinase A signaling. Protein kinase C and mitogen-activated protein kinase, with a potential crosstalk between them, also seem to be involved in mediating the LH and hCG actions [7].

LH is primarily produced by the pituitary gland and hCG by the placenta. Both hormones have been well recognized for many years in that a whole array of malignant tumors ectopically expresses dimer LH/hCG or their subunits as well as their receptors, implying that they might be involved in the malignant transformation [8]. The deliberate overexpression of LH β-subunit or hCG β-subunit induced tumorigenesis in ovaries and other organs [9–11]. Testicular tumors were found in mutant Lhr that was constitutively activated in the absence of LH [12]. Induction of adrenocortical tumors in gonadectomized mice was thought to be due to unopposed LH production by the pituitary [13]. Using antisense oligonucleotides to inhibit the expression of hCG-α subunit in human choriocarcinoma and lung cancer cells reduced their tumorigenic potential [14, 15]. LH-induced invasiveness of human primary endometrial cancer is found to correlate with the levels of LH/hCG receptors in these cells [16]. In contrast, the epidemiological and experimental data demonstrated that dimer LH and hCG protected against breast cancer [17]. LH and hCG reduced the risk of mutagen-induced rat mammary cancer and decreased proliferation and invasion of breast cancer cells by inhibiting NF-κB and AP-1 activation [18]. Previous studies reported that expression of LH/hCG receptor in ovarian cancer correlated with better survival and LH can inhibit FSH-induced ovarian epithelial cancer cells to grow in vitro [19]. Normal and malignant human prostates have been found to co-express LH/hCG and their corresponding receptors. hCG is capable of impeding the invasion of human prostate cancer cells [5, 20–22]. Several studies have demonstrated that hCG and its subunits have an inhibitory effect on the growth of Kaposi’s sarcoma [23, 24]. These findings reveal that LH/hCG signaling is closely linked to tumorigenesis in certain conditions, either promotive or repressive.

In this study, we investigated whether LH signaling deficiency affects the animal’s susceptibility to a simple alkylating carcinogen, N-methyl-N-nitrosourea (MNU), induced tumorigenesis. The results demonstrated that MNU induces a higher incidence and an earlier onset of aggressive lymphomas in LhrKO heterozygous and homozygous mice. This increase appears to be due to a reduction in apoptosis of thymocytes.

Materials and Methods

Animals and MNU Treatment

LhrKO mice were maintained on a C57BL/6 × 129/sv background. Adult male and female heterozygous mice were mated to obtain wild-type (Lhr+/+), heterozygous (Lhr+/−), and homozygous (Lhr−/−) animals. The genotype of LhrKO mice was determined by PCR with tail genomic DNA and Lhr and neomycin-resistant gene primers as previously described by Yuan et al. [25]. Seven-week-old female Lhr+/+, Lhr+/−, and Lhr−/− mice were used for MNU treatment. These animals were injected intraperitoneally with a single dose (50 mg/kg of body weight) of MNU (N-1517; Sigma, St Louis, MO, USA) dissolved in dimethyl sulfoxide and 0.01 M sodium phosphate buffer saline (PBS, pH 7.4) to a final concentration of 10 mg/ml and was used within 30 min. The animals were monitored for tumor development every other day. Animals exhibiting shortness of breath or other signs of mortal sickness were immediately sacrificed. All others were sacrificed 10 months after MNU exposure. Complete gross examination was performed for detection of tumor masses in thymus, lymph nodes (cervical and mesenteric), liver, kidney, and spleen. The normal thymuses were also harvested from 3-month-old untreated female mice. The tissues were either fixed in 10% formalin for histological analysis and immunohistochemical staining or stored at −80°C for later RNA isolation and protein lysate preparation. This study was approved by the Animal Care and Use Committee of the University of Louisville.

Histopathology

Normal and tumor tissues were fixed in 10% buffered formalin, processed through graded alcohols and xylene, and embedded in paraffin. Five-micrometer-thick sections were cut, stained with hematoxylin and eosin (H&E), and evaluated with an Olympus microscope.

Southern Blotting

The EZ Tissue/Tail DNA isolation kit (EZ BioResearch LLC, St Louis, MO, USA) was used for isolation of genomic DNA from normal thymuses and thymic tumors according to the manufacturer’s instruction. Stu I-digested DNA was loaded on 1% agarose. After electrophoresis, the DNA was blotted on Hybond-N+ membrane (GE Healthcare Bio-Sciences Corp., Piscataway, NJ, USA) and hybridized with a 32P-labeled Lhr probe that detected an 8-kbp wild-type Lhr allele and a 10-kbp mutant allele as previously described [26]. Following washing under highly stringent conditions, the membrane was exposed overnight to X-ray film with an intensifying screen at −80°C.

Immunohistochemical Staining

The immunostaining procedures were conducted using an avidin–biotin complex immunoperoxidase (ABC) method (Vector Laboratories, Burlingame, CA, USA) as previously described [26]. Briefly, tissue sections were pretreated with an antigen unmasking solution (Vector Laboratories) at 100°C for 20 min. Endogenous peroxidase activity was blocked by incubation of 3% H2O2 in methanol for 10 min at room temperature. The slides were then incubated overnight at 4°C with a ready-to-use mouse monoclonal antibody against CD22 (AM439-5M; BioGenex Laboratories Inc., San Ramon, CA, USA), a 1:150 diluted goat polyclonal antibody against CD3 (sc-1127; Santa Cruz Biotech Inc., Santa Cruz, CA, USA), or a 1:100 diluted rabbit polyclonal antibody to proliferating cell nuclear antigen (PCNA, sc-7907; Santa Cruz Biotech). After rinsing in PBS, the slides were incubated with a biotin conjugated anti-mouse or goat or rabbit IgG secondary antibody then followed by ABC reagent and diaminobenzidine (DAB) chromogen (Santa Cruz Biotech). Omission of primary antibodies was used as a procedural control.

Thymocyte Apoptosis Assay

TdT-mediated dUTP Nick-End Label (TUNEL) staining was carried out on formalin-fixed and paraffin-embedded thymic sections with a DeadEnd colorimetric apoptosis detection system (Promega Corp., Madison, WI, USA) according to the manufacturer’s instruction. The brown-colored nuclei of apoptotic cells were visualized with a bright field microscope as previously described [25].

Semiquantitative RT–PCR

The total RNA of normal thymuses and thymic lymphomas was isolated using Trizol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instruction and reverse transcribed into cDNA using avian myeloblastosis virus reverse transcriptase (Promega) and random oligonucleotide hexamers (Invitrogen). PCR primers, listed in Table 1, were designed according to the sequences obtained from GenBank using the Vector NTI 9.0 computer program (InforMax, Frederick, MD, USA) and synthesized by Operon Technologies (Alameda, CA, USA). All the primers were designed to amplify products that spanned more than one exon. The housekeeping gene, ribosomal protein large subunit 19 (Rpl19), was co-amplified with Bcl2, Notch-1, and P53 as an internal control for both complementary DNA (cDNA) quantity and quality. The numbers indicated in Table 1 were the PCR cycles used, which were in the linear phase of the amplification for each target gene. The PCR-amplified products were electrophoresed in 1.5% agarose gels stained with ethidium bromide and analyzed using the TotalLab version 2.01 (Nonlinear USA, Durham, NC, USA) image analysis software. Then, the ratio of target gene to housekeeping gene was calculated.

Table 1 Oligonucleotide sequences used for semiquantitative RT–PCR
Full size table

Western Immunoblotting

The procedures were performed as previously described [27]. Briefly, the normal thymuses and thymic tumors were homogenized in 50 mM Tris–HCl (pH 8.0) containing 150 mM NaCl, 5 mM MgCl2, 1% Triton X-100, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate (SDS), 0.5 mM phenylmethyl sulfonyl fluoride, 5 μg/ml leupeptin, 5 μg/ml aprotinin, and 5 μg/ml pepstatin. Aliquots of the lysates (30 μg) were separated by 8% SDS–polyacrylamide gel electrophoresis and transferred to polyvinylidene fluoride membranes. After blocking with 3% nonfat dry milk, the membranes were incubated overnight at 4°C with 1:1,000 diluted rabbit polyclonal antibodies against un-cleaved (inactive) and cleaved (active) caspase-3 (9662 and 9661; Cell Signaling, Beverly, MA, USA), 1:500 diluted mouse monoclonal antibody against Bcl-2 (sc-509; Santa Cruz Biotech), and 1:1,000 diluted mouse monoclonal antibody against β-actin (A5441; Sigma), respectively. The blots were then incubated with a peroxidase-labeled secondary antibody. The immune complexes were detected using a chemiluminescence detection system according to the procedure recommended by the manufacturer (GE Healthcare Bio-Sciences). Omission of primary antibodies was used as a procedural control. TotalLab version 2.01 (Nonlinear USA) image analysis software was used for densitometric analysis of the specific bands, and the results were expressed as a ratio of active over inactive caspase-3.

Statistical Analyses

Kaplan–Meier analysis with logrank test was applied to analyze the data presented in Fig. 1. Comparison of lymphoma metastasis among genotypes was carried out by chi-square test. One-way analysis of variance and Tukey multiple comparison post test were used to evaluate the results shown in Figs. 5 and 6. The data presented are the means ± SEM. All statistical analyses were performed by using a Prism (version 5) program (Graphpad Software Inc., San Diego, CA, USA). A p value <0.05 was considered significant.

Fig. 1
figure 1

MNU-induced lymphoma in LhrKO female mice. a The gross appearance of enlarged thymus (yellow arrow) and cervical lymphoids (white arrow) in an LhrKO homozygous female mouse sacrificed 3 months after a single dose of MNU injection. b The percent of tumor-free animals after MNU injection

Full size image

Results

MNU-Induced Tumorigenesis in LhrKO Mice

We monitored the animal’s health and appearance of tumors in Lhr +/+ (n = 14), Lhr +/− (n = 15), and Lhr −/− (n = 13) mice for a 10-month period from the day of MNU injection unless they displayed signs of impending death, in which case they were sacrificed immediately. By the end of 10 months, in contrast to 35.7% (5/14) with a median latency of 118 days in Lhr +/+ mice, lymphomas in the thymus and lymphoids were found to be 70.6% (12/17) with a median latency of 125 days, and 100% (13/13) with a median latency of 86 days in Lhr +/− and Lhr −/− animals, respectively (Fig. 1b). Comparing Lhr +/− and Lhr −/− to Lhr +/+ mice, the incidence of lymphomas was significantly increased (p < 0.001). This increase was more profound in Lhr −/− than Lhr +/− siblings (p < 0.01). The development of tumors in Lhr −/− but not Lhr +/− mice was much earlier than that found in Lhr +/+ mice (p < 0.002). Autopsy results revealed that lymphomas were typically accumulated in the thymus and cervical lymph nodes (Fig. 1a). Lymphomagenesis in Lhr −/− mice displayed a more aggressive phenotype that metastasized to spleen, liver, or kidney. The invasion of lymphomas to one or more of these organs in Lhr +/− (23.5%) and Lhr −/− (38.5%) mice were substantially greater than in Lhr +/+ (7.1%) animals (p < 0.01).

MNU-Induced Non-Hodgkin’s Lymphoma in all Three Genotypes

All the tumor samples collected from Lhr +/+, Lhr +/−, and Lhr −/− animals, including primary and metastasized tumor tissues, were non-Hodgkin’s lymphomas according to their histopathological characteristics. Morphological analysis of the tumor tissues from thymus showed that most of these tumors were medium-poor differentiation (Fig. 2). To ascertain the cellular origin of the lymphomas, immunohistochemical staining for CD3 (a T-cell-specific antibody) and for CD22 (a marker for B cell) was performed which demonstrated that the lymphomas, including the metastatic lesions of lymph nodes, liver, spleen, and kidney, originated from T lymphocytes (Fig. 2). Thus lymphomagenesis appears to be the malignant transformation of T cells.

Fig. 2
figure 2

MNU induces non-Hodgkin’s lymphomas. Lymphomas in the thymus (a–c), lymph node (d–f), spleen (g–i), liver (j–l), and kidney (m–o) are all stained positive for a T-cell marker, CD3, but negative for a B-cell marker, CD22. Magnification: a and f, ×300; b–e and g–o, ×150

Full size image

Southern blotting on the genomic DNA isolated from the normal and thymic lymphomas indicated that there was no rearrangement in the Lhr locus (Fig. 3).

Fig. 3
figure 3

Southern blotting performed on EcoRI digested genomic DNA isolated from normal and thymus tumors showed that wild-type and mutant alleles in the tumor tissues were the same as in the normal tissues

Full size image

Effect of Lhr Deficiency on Proliferation and Apoptosis in Thymus and MNU-Induced Thymic Lymphoma

To determine whether the increased susceptibility of MNU-induced lymphomagenesis was associated with increased cell proliferation or reduced apoptosis in the thymuses, immunostaining of PCNA and TUNEL assay were performed. Immunostaining showed that there was no difference in cell proliferation among the genotypes (Fig. 4a–c). TUNEL assay, on the other hand, demonstrated that apoptotic thymocytes were detected in clusters throughout the entire thymus of Lhr +/+ animals and their numbers of clusters containing apoptotic cells clearly decreased in Lhr +/− and Lhr −/− mice (Fig. 4d–f).

Fig. 4
figure 4

Immunohistochemical staining for a cell proliferation marker, PCNA, in normal thymuses (a–c) showed no genotype differences among the genotypes (a–c). TUNEL assay in normal thymuses (d–f) revealed a decrease in clusters of apoptotic cells in heterozygous (+/−) and homozygous (−/−) animals compared with wild-type siblings (+/+). Magnification: a–f, ×150; insets in a–c, ×600

Full size image

Semiquantitative RT–PCR and Western immunoblotting revealed while mRNA and protein levels of Bcl-2 in the thymuses of Lhr +/− and Lhr −/− animals were dramatically elevated compared with Lhr +/+ siblings (Fig. 5), the P53 and Notch-1 were similar (data not shown). In MNU-induced thymic lymphomas, a substantial increase in mRNA and protein levels of Bcl-2 in Lhr +/+ animals was observed despite no further increase in Bcl-2 expression in Lhr +/− and Lhr −/− mice (Fig. 5).

Fig. 5
figure 5

Semiquantitative RT–PCR and Western blotting revealed that mRNA (a) and protein (b) levels of Bcl-2, an anti-apoptotic factor, in both normal thymuses and thymic tumors of heterozygous (+/−) and homozygous (−/−) mice were significantly higher than in wild-type (+/+) siblings. n = 6; a p < 0.001, compared with normal thymuses of +/− and −/− animals and tumors of all genotypes; b p < 0.05, compared with normal thymuses of wild-type animals

Full size image

To further characterize apoptotic potential in thymuses and thymic lymphomas, Western blotting was performed to assess the caspase-3 activation. The results showed a significant decrease in activated caspase-3 levels in the thymuses of Lhr +/− and Lhr −/− mice compared to Lhr +/+ siblings with a more dramatic reduction in Lhr −/− mice (Fig. 6). In MNU-induced thymic lymphomas, the activated caspase-3 levels were lower than the normal thymus, especially the 17-kDa activated caspase-3. However, there were no genotype differences (Fig. 6).

Fig. 6
figure 6

Western blotting demonstrated that inactive caspase-3 levels are not significantly different among the genotypes in either normal thymus or thymic tumor. The active caspase-3, on the other hand, markedly decreased in the normal thymuses of heterozygous and homozygous mice compared to wild-type siblings. Although active caspase-3 levels in thymic tumors were not different among the genotypes, they were all significantly lower than in normal thymuses of wild-type animals. n = 6; a p < 0.05, b p < 0.01, compared with normal thymuses of wild-type siblings

Full size image

Discussion

LH/hCG receptors have been demonstrated in the thymus and lymphocytes from a number of species [2, 28, 29]. A great deal of data indicates that pituitary hormones directly and indirectly play a role in the control of thymus development and functions [29–32]. Serum LH levels were shown to be highly correlated with the changes in the thymic hormone, thymosin beta 4, in heifers during maturation [33]. Gonadotropin-releasing hormone (GnRH) is a hypothalamic decapeptide hormone that modulates LH and FSH synthesis and secretion. The number of mature thymocytes in hypogonadal mice, due to deficiency of GnRH and as a result have very low LH, FSH, and gonad steroid levels, was significantly reduced [34]. Several other studies have shown that treatment of GnRH analogues in rats and mice influences thymus weight, lymphocyte counts, and their functions [35–38]. The present study demonstrated that MNU-induced malignant lymphomas were remarkably increased in LhrKO mice, suggesting deficiency of LH signaling in the thymus may predispose the thymus to greater vulnerability to lymphomagenesis.

MNU is a monofunctional alkylating agent that modifies guanine in DNA, resulting in mutation to adenine after DNA replication [39]. It is often used to induce carcinogenesis in rat and mouse. It has a relative broad spectrum of target organs including lymphohematopoietic system, skin, mammary gland, prostate, bladder, and lung [40–44]. Interestingly, thymic and lymphoidal lymphomas were the only tumors found in this study. The plausible explanation may be that the thymuses of LhrKO mice were very sensitive to MNU in the absence of LH signaling and lymphomas developed early, rapidly, and aggressively so that these animals had to be terminated before other tumors appeared.

The sensitivity to MNU-induced carcinogenesis can be dramatically modified by numerous other factors, which include sex steroid hormones, DNA repair proteins, cell cycle regulators, immune surveillant molecules, oncogenes, tumor suppressors, and apoptotic modulators [45–51]. For example, MNU-induced mammary carcinomas were largely estrogen dependent [45]. The susceptibility of MNU-mediated thymic lymphomas in P53 haploinsufficient mice was three-fold higher than in wild-type siblings [46]. MNU characteristically induced an early and higher incidence of malignant lymphomas in both heterozygous and homozygous LhrKO mice. The tumors behaved much more aggressively with metastases in multiple organs. Thus, we postulate that LH signaling may function as a modifier to influence tumorigenesis. It is possible that certain cancer promotions resulting from LH signaling deficiency in the thymus may potentiate the tumor development and progression upon carcinogen insults. Estrogen seems to be an unlikely promoting factor to this enhanced sensitivity to MNU-induced lymphomagenesis because estrogen levels in heterozygous LhrKO mice were comparable to wild-type siblings. Also, estrogen levels in homozygous LhrKO mice were in fact decreased [26], whereas the incidence of lymphomas in both genotypes was significantly increased as compared with wild-type mice.

To assess whether Lhr deficiency in the thymus affected thymocyte proliferation and apoptosis, thymic tissues were immunostained for PCNA and subjected to TUNEL evaluation. The cell proliferation in the thymuses was indistinguishable between wild-type and LhrKO animals. Surprisingly, thymocyte apoptosis was quite common in the thymuses of wild-type animals. The reason for this fairly high rate of apoptosis in normal adult thymus is unknown. One possibility may be relevant to the process of age-related thymic involution which is characterized by loss of thymocytes and a weight decrease [31]. The thymuses of LhrKO heterozygous and homozygous mice, in contrast, showed a clear reduction in thymocyte apoptosis. The inhibition of apoptosis could lead to an accumulation of cells that contain mutant DNA, favoring aberrant cellular progression through mitosis and ultimately neoplastic transformation. Hence, the inability of Lhr-deficient thymocytes to undergo apoptosis may underlie the observed differences in susceptibility to MNU-mediated lymphomagenesis.

Caspases are well-established cysteine proteases in the execution of apoptosis. The activation of caspase-3 requires proapoptotic molecules such as cytochrome c released from mitochondria. The antiapoptotic effect of Bcl-2 is achieved by its capability to inhibit cytochrome c release [52]. Thus, it could be rationalized that overexpression of Bcl-2 in the thymus of Lhr-deficient mice may be responsible for the decrease in caspase-3 activation and subsequently the reduction of thymocyte apoptosis. Abnormal activation of transcription factor NF-κB in lymphoma is known to induce a number of survival factors including Bcl-2 family members [53]. Interestingly, the LH/hCG signaling has been shown to repress the activation of NF-κB and AP-1 in several cell types [18, 54], raising the possibility that overexpression of Bcl-2 may result from dysregulated activation of NF-κB and AP-1 in the thymus due to Lhr deficiency. Notch-1 and p53 are known to play important roles in the regulation of thymocyte apoptosis [55–57]. However, they do not appear to be involved in the change of thymocyte apoptosis in Lhr-deficient mice since the expression of these two molecules was unaltered.

MNU is a common genotoxic chemical. Our data did not detect any change in the Lhr locus in MNU-induced lymphomas. The possibility that MNU may have caused deleterious point mutations in the Lhr gene and/or others that may alter thymocyte apoptotic cascades remains to be established. Although elevation of Bcl-2 is associated with MNU-induced lymphogenesis, it should be stressed that perhaps other unidentified factors may contribute to the different propensity to thymocyte apoptosis between wild-type and LhrKO mice, which requires further investigations.

In summary, our study reveals that a genotoxic agent, MNU, induced a higher incidence, an earlier onset, and a more aggressive phenotype of thymic lymphomas in LhrKO animals, suggesting that deficiency of Lhr may render the thymus more susceptible to a carcinogen challenge. As such, LH signaling may function as a tumor modifier in inhibiting mutagen-induced tumorigenesis. Simple alkylating agents are widely presented as environmental carcinogens and also endogenously formed by metabolic processes [58, 59]. Moreover, they are frequently used in tumor chemotherapy. Lymphomas are one of the most common cancers. While the incidence of most cancers is decreasing, lymphoma is one of two tumors increasing in frequency [60]. Therefore, a better understanding of the mechanistic link between LH/hCG signaling and lymphomagenesis in future studies may be useful in developing LH/hCG as a potential preventive agent to deter alkylating agent-induced tumorigenesis.