Background
The incidence of non-Hodgkin's lymphoma has increased over recent years, an increase that cannot be totally explained by improvements in diagnosis or reporting. The exact etiology of lymphoma remains unknown but viral infection, chronic antigen stimulation, and/or immunosuppression, either primary or acquired immunodeficiency, may all contribute to the occurrence of lymphoma [
1]. Some reports suggest that exposure to the UV radiation in sunlight may play a role in the development of lymphoma in humans [
2]. This conclusion is based primarily on epidemiological data showing a geographic correlation between sunlight exposure and lymphoma incidence (i.e., a latitude gradient) [
3‐
7]. However, not all the reports supported a link between sunlight exposure and lymphoma development [
8‐
11], some findings indicate an inverse association between the solar UVB exposure and the occurrence of non-Hodgkin lymphoma [
12‐
15]. The rise in lymphoma incidence parallels the dramatic rise in melanoma incidence, and patients with non-Hodgkin's lymphoma or chronic lymphocytic leukemia are at a higher risk of developing skin cancer [
16,
17].
Studies with experimental animals suggest a strong correlation between UV exposure and lymphoma development [
18,
19]. Animal experiments offer the distinct advantage of controlled irradiation with defined UV light sources, without the complications that arise from exposure to any other environmental carcinogens or toxins, nor the complication of recall bias. We previously reported that UV irradiation augments lymphoid malignancies in mice with one functional copy of wild-type p53. In that study we demonstrated that UV irradiated p53 heterozygous mice developed lymphoid tumors at a much higher rate (88% of irradiated mice developed tumors) than found in un-irradiated animals (6% spontaneous tumor rate). Sequencing data indicated that the UV-irradiated p53
+/- mice retained the non-mutated p53 allele, suggesting loss of heterozygosity did not play a role in the induction of this tumor [
19].
The biological effects of UV exposure are well known. UVB, wavelengths in the 280-320 nm range of the solar spectrum, can induce a wide variety of adverse effects. Chief among them are sunburn, inflammation, premature ageing of the skin, the induction of non-melanoma skin cancer and the induction of immune suppression (reviewed in [
20]). UVA (320-400 nm) has been suggested to be important in melanoma induction [
21]. It is clear that the adverse effects of UV exposure are not solely limited to the skin; how UV exposure influences lymphoma development in humans, however, is still far from being completely understood. Here we analyzed the UV-induced lymphoid malignancies that arise in mice with one functional copy of p53. We have previously analyzed UV-induced lymphoid malignancies by histopathological examination of formalin-fixed tissues. To be able to study the biology of the tumors we needed to generate tumor cell lines from UV-irradiated
p53
+/- mice. We exposed
p53
+/- mice to solar simulated UV radiation for a period of 30 weeks and then injected spleen cells into immune deficient Rag2
-/- mice. We generated a number of transplantable tumor cells lines, some of which were further adapted to grow in vitro. We found that UV-irradiation of
p53
+/- mice resulted in the induction of B-cell tumors (CD5
+, CD19
+, B220
+, IgM
+) that grew progressively in immune deficient mice, but not in immune competent animals, and grew at low rate in UV-irradiated C57Bl/6 mice. We suggest the induction of B cell lymphomas by UV radiation in
p53
+/- mice may serve as a useful animal model for lymphoma development.
Methods
Mice
Specific pathogen-free
p53 knockout mice were backcrossed onto a C57Bl/6 background [
19]. Specific pathogen-free Rag2
-/- mice were obtained from Dr. Fred Alt (Howard Hughes Medical Institute, Children's Hospital, Boston, MA). All mice were housed in a pathogen-free barrier facility. The Institutional Animal Care and Use Committee approved all procedures.
UV irradiation
A 1000 W Xenon UV solar simulator equipped with a Schott WG-320 atmospheric attenuation filter and a visible/infrared band-pass filter (Oriel, Stratford, CT) was used to irradiate the mice. The intensity and spectral output was measured using an Optronics Model OL-754 scanning spectrophotometer (Optronics Lab, Orlando, FL) [
22]. Seventeen p53 heterozygous mice were divided in two groups of 8 and 9 mice each for UV irradiation. The mice were shaved on the dorsal skin and were kept individually in a Plexiglas container; then, they were irradiated three times a week during 10-15 minutes for a period of 30 weeks. During this procedure the mice remained conscious all the time.
Transplantation
Cells (2.5 × 106) from enlarged spleens of the irradiated mice were injected subcutaneously into each flank of a Rag2-/-, wild type and/or UV irradiated (10 kJ/m2, 3 times per week for 12 weeks) C57Bl/6 mouse. Mice were monitored daily for tumor development. The tumors were excised and analyzed by flow cytometry and histopathology.
Flow cytometric analysis
Cell surface molecule expression was analyzed by flow cytometry using a BD FACS-Calibur (BD Biosciences, San Jose, CA). Fluorescently-labeled monoclonal antibodies to mouse CD1d, CD3, CD4, CD5, CD8a, CD11b, CD16/32, CD19, CD23, CD25, CD38, CD40, CD43, CD44, CD45R/B220, CD49b, CD62L, CD80, CD86, CD90, CD117, CD123, CD127, CD138, IgM and IgD, GR-1 were purchased from commercial sources (Becton-Dickinson, Mountain View, California, and eBioscience, San Diego, California).
In vitro class switch recombination assay
The cells were stimulated with lipopolysaccharide (LPS) (Sigma, St. Louis, MO), IL-4 (10ng/ml) plus CD40L (10 ng/mL), and a combination of CD40L/TGFβ (10 ng/mL each, all from PeproTech Rocky Hill, NJ). The cells were analyzed 96 hours later by flow cytometry using a BD FACS-Calibur or BD LSR II (Becton-Dickinson, Mountain View, California).
Southern blotting analysis
High molecular weight DNA samples were prepared from the tumor, digested with EcoR1 and hybridized with a JH probe using Gene images AlkPhos Direct Labeling and detection System (Amersham Biosciences, Buckinghamshire, UK).
Cytogenetic analysis
Tumor cells were cultured in the presence of Colcemid. Metaphases were analyzed with trypsin-Giemsa stain and with whole chromosome painting probes (Carlsbad, CA, USA). To analyze potential chromosomal translocations, the metaphases were also used for spectral karyotyping analysis (SKY; Applied Spectral Imaging, Inc. Vista, CA).
Cell culture and lentiviral vector transduction
Tumor cells (1 × 105) were incubated with lentiviral vectors. The number of green fluorescent protein (GFP) positive cells was analyzed by flow cytometry at multiple time points after transduction. Highly expressing GFP bearing cells were isolated by cell sorting (BD FACS-Aria, BD Biosciences, San Jose, CA).
Cytokine determination
Cytokine secretion of the tumor cells into the supernatant fluid was measured by the multiplexed mouse cytokine Ten-Plex Antibody Bead system (BIOSOURCE, Camarillo, CA) and the Luminex 100 fluorescence detection system (Austin, TX). Cytokines analyzed by this system include: IL-1beta, IL-2, IL-4, IL-5, IL-6, IL-10, IL-12, GM-CSF, Interferon-gamma and TNF-alpha.
Discussion
We characterized the lymphoid tumors that arise in UV-irradiated p53
+/- mice. No evidence of T cell tumors was found in this study. This may reflect the different way in which this study and our previous work was carried out. Our observation that UV exposure induced T cell tumors was based on a retrospective study in which histopathological analyses of paraffin-embedded tissue was used to determine the phenotype of the tumor [
19]. Here, we generated tumors by injecting spleen cells isolated from UV-irradiated p53 heterozygous mice into Rag2-/- mice. The tumor cells were subjected to several passages in vivo until we were able to generate stable tumor cell lines that could grow in Rag2
-/- mice and in culture. From these experiments we obtained three cell lines that grew in vivo but only one was also able to grow in vitro. It is possible that our "culture" favored the growth of B cells but not T cells. In one series of experiments, we phenotyped each successive in vivo passage, and in one of the earlier passages, we noted the presence of some T cell receptor positive cells (which by definition must be of tumor origin since the Rag2
-/- mice cannot rearrange the T cell receptor). These cells were purified by cell sorting and then injected into Rag2
-/- mice. No T cell tumors developed. This suggests that abnormal cells may have been present, but were not capable of growing progressively upon further passage in vivo.
It is interesting to compare our results to those previously published by MacPherson and colleagues [
28]. In their study, transgenic mice were constructed with a point mutation in the murine p53 gene (serine 23 was replaced with alanine). These mice, which displayed defective apoptosis, had a reduced life span and most developed tumors. However, unlike the tumor spectrum found in p53
-/- mice (predominately thymic lymphomas and sarcomas), the most common tumors found in homozygous p53 Ser23 mice were of B cell origin. The authors suggest that the long latency period to tumorigenesis in the p53 Ser23 mice provides the time required to acquire additional genetic changes to develop B cell tumors. We also note a long latency period for tumor development in our system, and as mentioned below, we suggest UV-induced inflammation may play a significant role.
Chromosomal analysis revealed a t(14:19) translocation, which may play a role in UV-induced lymphomagenesis. However, to the best of our knowledge, a similar translocation has not been reported in lymphoma in the mouse. We analyzed the translocated region of chromosome 19 and found that this region contains the apoptotic genes Caspase 7 and programmed cell death 4, and the DNA cross-link repair 1a gene (SNM1a) which is involved in G1 cell cycle checkpoint downstream of ATM[
29]. We suspect the t(14;19) translocation contributes to the deregulation of these and other genes and probably plays an important role in the oncogenic process in these cells. In addition, trisomy of chromosome 6 has been reported in chemically induced malignant skin tumors in TG.AC mice, which carry the v-Ha-ras transgene[
30], and a similar imbalance has been reported in both benign and malignant skin tumors in the mouse [
31]. Therefore, it is likely that the trisomy of chromosome 6 observed in our model is the result of the chronic exposure to UVB and probably it is also a contributing factor to the transformation of the UV-induced tumor B cells. We are currently exploring the possible deregulated oncogenes/tumor suppressors in these tumor cell lines.
What are the signals involved in the maintenance and proliferation of these tumors? Simple loss of heterozygosity does not appear to be an adequate explanation because when the B cell lymphomas were sequenced, we found they retained a normal wild-type p53 allele (data not shown). We favor the view that the immune suppression and chronic inflammation induced by UV exposure are involved. As noted above, the B cell lymphomas are highly antigenic. From our previous study we know that p53
+/- mice are immune competent and UV exposure will induce immune suppression in these mice [
19], so it is reasonable to assume that the immune suppression induced by chronic UV exposure provides a window that will allow these antigenic cells to grow in the UV-irradiated p53
+/- mice. While it has been argued that an inverse association of sun exposure with NHL risk is related to the protective effects of vitamin D synthesis by UV radiation [
15,
32‐
36]; we know that the induction of systemic immune suppression following UV exposure of the skin is associated with the release of immune modulatory factors and immune suppressive cytokines. These may have direct relevance to the results reported here. The first is IL-10. UV-irradiation of the skin induces keratinocytes to secrete IL-10 [
37,
38] and IL-10 can be found in the serum of UV-irradiated mice [
39]. Moreover, IL-10 promotes the growth and differentiation of human B cells [
40] and has been shown to promote the growth of B cell lymphomas [
41,
42]. Serum IL-10 is also considered to be a negative prognostic indicator for the survival of patients with non-Hodgkin's lymphoma [
43]. The second is platelet-activating factor (PAF). PAF is a phospholipid that plays a primary role in the induction of inflammation. As the name implies PAF activates a wide variety of cells, including platelets, monocytes, mast cells and polymorphonuclear leukocytes [
44]. PAF is secreted by epidermal cells almost immediately following UV exposure [
45] and serves as a transcriptional activator for a variety of genes including cyclooxygenase (COX)-2, IL-6 and IL-10 [
46,
47]. PAF also plays an essential role in UV-induced immune suppression. Blocking the binding of PAF to its receptor, with a series of selective PAF receptor antagonists, blocks UV-induced transcription of COX-2 and IL-10 and blocks UV-induced immune suppression [
47]. Normal and leukemic B cells express the PAF receptor [
48,
49] and PAF has been shown rescue B cells from apoptosis [
50].
Our results suggest the following scenario: the loss of a single p53 allele predisposes the B cells for transformation. Long-term chronic UV exposure then induces the release of inflammatory mediators by irradiated keratinocytes that promote tumor development. For example, UV-irradiated keratinocytes secrete macrophage migration inhibitory factor (MIF) [
51], which can inhibit p53 tumor suppressor activity [
52], thus resulting in a functional "loss of heterozygosity" in the UV-irradiated p53
+/- mice. Other UV-induced mediators of inflammation, such as PAF, may also play important roles in the process. First, PAF induces chronic inflammation, which appears to be associated with the induction of non-Hodgkin's lymphoma [
53]. Second, PAF binds to receptors on B cells and protects them from apoptosis. PAF up-regulates the transcription of COX-2 and IL-10. This is important because the expression of either prostaglandin E
2 or IL-10 has been associated with a poor prognosis in lymphoma patients [
43,
54]. In addition, both are immune suppressive, and we suggest UV-induced immune suppression (mediated by PAF, prostaglandin and IL-10), protects the developing antigenic B cell lymphoma from immune destruction. Our data provide a link between chronic UV exposure and the incidence of internal tumors in the 53 heterozygous background. Although there is no development of skin tumors in p53 heterozygous mice, somehow chronic UV exposure develops cellular transformation away form the site of exposure by a mechanism still under investigation.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
NP: designed and performed experiments, interpreted results, drafted and revised the manuscript; YM, CS, AYL performed experiments, interpreted results, drafted the manuscript; NK performed experiments using animals; SEU and CZ designed and supervised the research, analyzed the data and wrote and modified manuscript.
All authors read and approved the final manuscript.