Hematopoietic neoplasms in Prkar2a-deficient mice

Background Protein kinase A (PKA) is a holoenzyme that consists of a dimer of regulatory subunits and two inactive catalytic subunits that bind to the regulatory subunit dimer. Four regulatory subunits (RIα, RIβ, RIIα, RIIβ) and four catalytic subunits (Cα, Cβ, Cγ, Prkx) have been described in the human and mouse genomes. Previous studies showed that complete inactivation of the Prkar1a subunit (coding for RIα) in the germline leads to embryonic lethality, while Prkar1a–deficient mice are viable and develop schwannomas, thyroid, and bone neoplasms, and rarely lymphomas and sarcomas. Mice with inactivation of the Prkar2a and Prkar2b genes (coding for RIIα and RIIβ, respectively) are also viable but have not been studied for their susceptibility to any tumors. Methods Cohorts of Prkar1a+/−, Prkar2a+/−, Prkar2a−/−, Prkar2b+/− and wild type (WT) mice have been observed between 5 and 25 months of age for the development of hematologic malignancies. Tissues were studied by immunohistochemistry; tumor-specific markers were also used as indicated. Cell sorting and protein studies were also performed. Results Both Prkar2a−/− and Prkar2a+/− mice frequently developed hematopoietic neoplasms dominated by histiocytic sarcomas (HS) with rare diffuse large B cell lymphomas (DLBCL). Southern blot analysis confirmed that the tumors diagnosed histologically as DLBCL were clonal B cell neoplasms. Mice with other genotypes did not develop a significant number of similar neoplasms. Conclusions Prkar2a deficiency predisposes to hematopoietic malignancies in vivo. RIIα’s likely association with HS and DLBCL was hitherto unrecognized and may lead to better understanding of these rare neoplasms.


Background
Protein Kinase A (PKA) is the main mediator of cyclic adenosine mono-phosphate (cAMP)-dependent signaling [1,2]. In humans and rodents, PKA is a heterotetramer formed by two regulatory (RIα, RIβ, RIIα, RIIβ) and two catalytic subunits (Cα, Cβ, Cγ, Cx) [1,3]. The PKA subunits have distinct expression patterns: α-subunits are expressed ubiquitously, whereas β-subunits are expressed in a more tissue-specific fashion. There are two types of PKA holoenzymes, type I and type II, depending on the identity of the regulatory subunit [4]. The cAMP binds to the PKA regulatory subunits (two cAMP molecules per subunit), which leads to dissolution of the tetramer, allowing for the catalytic subunits to act as serine-threonine kinases, phosphorylating numerous targets [1,2].
Among the regulatory subunits, the PKA regulatory subunit type 1A (PRKAR1A or RIα) has the highest affinity for cAMP and is therefore more sensitive to cAMP, which makes RIα essential for maintaining regulated PKA activity [5]. The PRKAR1A gene on human chromosome 17q22-24 is the gene encoding for RIα; mutations in this gene are responsible for the multiple tumor syndrome Carney complex (CNC, Online Mendelian Inheritance in Man #160980) [6,7]. CNC patients develop myxomas, skin lesions, schwannomas, bone and endocrine tumors, and a variety of cancers but they are not known to be predisposed to hematologic malignancies. Mice haploinsufficient for a null allele of Prkar1a (Prkar1a +/− ) developed tumors, mainly bone tail and thyroid tumors and schwannomas, but the spectrum was somewhat different from that seen in humans [8]. Mice bearing a transgene expressing an anti-sense Prkar1a construct (AS-Prkar1a) developed more tumors than the Prkar1a +/− mice, including adrenal hyperplasia, histiocytic sarcomas (HS) and lymphomas [9]. Taken together, these studies clearly identified PRKAR1A as a tumor suppressor gene affecting several cell types in both humans and mice but left it unclear as to whether PKA defects might lead to a predisposition to hematologic malignancies.

Animal protocol
Mice deficient for Prkar1a and Prkar2a alleles that have been described previously [8,12], were maintained on a mixed genetic background (C57BL/6/129Sv) and were crossed to generate Prkar2a +/− and Prkar2a −/− mice. We also studied Prkar2b +/− mice on the same mixed genetic background. All studies were performed under animal protocol 12-033 and were approved by and conducted in accordance with the Eunice Kennedy Shriver National Institute for Child Health and Human Development Institutional Animal Care and Use Committee. Mice have been observed between the ages of 5 and 25 months and were necropsied when found to have splenomegaly, lymphadenopathy, hepatomegaly, labored breathing, ruffled fur, or other signs of significant morbidity. At necropsy, harvested tissues were fixed in 10 % neutral buffered formalin for histopathologic study and other samples were frozen for subsequent DNA, RNA and protein studies.

Southern blot hybridization
DNA prepared from frozen samples of spleen obtained at necropsy was processed for Southern blot hybridization using established procedures and hybridized with an IgH J H probe [15,[17][18][19].
Prkar2a-deficienct mice developed hematopoietic tumors with higher frequency than Prkar1a-deficient and WT mice ( Fig. 1). Remarkably, a total of 16 cases (51.6 %), including 11/19 cases from the Prkar2a +/− mice (58 %), were diagnosed as histiocytic sarcomas (HS) [11] (Table 1). Seven others cases were tumors of B cell lineage origin, including diffuse large B cell lymphomas (DLBCL), a plasmocytoma and marginal zone lymphomas (MZL) ( Table 1) [13,[20][21][22].  Overall, the most frequent neoplasms were hematopoietic in origin and included diffuse large B cell lymphomas (DLBCL) with 3 of 4 of these cases (Table 1, mice 3209, 3217, 2537, 3150) being histiocyte-rich. In one of these cases, the mouse appeared to have a coexisting DLBCL and a HS (Fig. 2, Table 1 mouse #3217). The spleen of this mouse was infiltrated by DLBCL that stained intensely with antibody to the B cell marker, B220 (Fig. 2a) and had metastasized to the liver. The liver mass stained uniformly with B220 (Fig. 2b) but was negative for the macrophage marker, Mac-2 (data not shown). In another part, the same spleen was essentially replaced with a HS associated with the presence of large multinucleate cells (Fig. 2c) that are often seen in mouse as well as human HS [23].
We identified a case of high-grade splenic marginal zone lymphoma [24] (Table 1, mouse #3345). At low power, sheets of pale staining cells can be seen, surrounding the follicles and invading both the follicles and the red pulp (Fig. 3c). At high power, the marginal zone is seen to be populated by centroblasts with one or more prominent nucleoli at the nuclear membrane and some immunoblasts with a large magenta bar-shaped nucleolus appended to one side of the nuclear membrane (Fig. 3d).

DNA studies
To determine if the B cell lineage tumors were clonal, we performed DNA extraction from 4 cases of DLBCL (Table 1, mice # 3217, 3150, 3209, 2537), 1 HS (Table 1, mouse #2622) and a normal spleen and analyzed them by Southern blot hybridization for organization of the IgH locus (Fig. 4). All the DLBCL cases exhibited clonal or oligoclonal rearrangements of IgH while the HS and normal spleen, as expected, had only a 6.8 kb band characteristic of the germline sequence. It is noteworthy that the intensity of the germline band was reduced in only one of the cases of DLBCL indicating that the tumors were not the dominant cell types in the other cases. We concluded from these studies, that the tumors diagnosed histologically as DLBCL were clonal B cell neoplasms.

Cell sorting
To determine if altered PRKAR2A expression might affect early events in hematopoietic differentiation in the BM or the development and distribution of cell subsets in peripheral lymphoid tissues, we performed extensive multiparameter FACS analyses of single cell suspensions prepared from bone marrow (BM), spleen, lymph nodes and thymi from 3-6 month old Prkar2a +/− and Prkar2a −/− mice. The analyses were for B cells, T cells and their subsets, macrophages, dendritic cells and granulocytes. We found no significant changes in any cell subset in any tissue from data obtained with WT mice (data not shown). These results indicate that if aberrant PRKAR2A expression affected hematopoietic differentiation, the effects only became evident well into adult life.

Non-hematopoietic tumors
We identified mice with tumors of the lung (Table 1, mice # 2378, 3436, 3069). These included a pulmonary adenoma (Fig. 5a, b mouse #2378) and a pulmonary fibrosarcoma [21] (Fig. 5c, mouse #3069). The adenoma, which presented as a round mass lesion, was comprised of monomorphic well-differentiated cells with complete loss of normal alveolar architecture. In contrast, the fibrosarcoma presented as a large mass lesion displacing all normal lung elements and was associated with a large necrotic area. It is not known if the lung was where this tumor originated from, but sarcomatous lesions were not identified in other tissues of this mouse.
One mouse (#3211) developed an uncommon liver vascular lesion readily evident at necropsy (Fig. 6a). Histologic sections of liver revealed features consistent with a diagnosis of malignant hemangiosarcoma but that also resembled peliosis hepatis, a non-malignant lesion in which multiple blood filled cavities can be found throughout the liver [25] (Fig. 6b, c).

Discussion
Our findings in this study identify Pkrar2a as a gene whose deficiency predisposes mice to a spectrum of mostly hematopoietic neoplasms. This phenotype has some similarities but also differences from the spectrum of tumors exhibited by humans and mice with RIα defects [6,9,26]. The major difference is that Prkar2a +/− and Prkar2a −/− mice develop with a high frequency hematopoietic neoplasms, particularly macrophage-derived HS tumors, as well as neoplasms derived from mature cells of the B cell lineage. Lymphomas or HS have not been reported in patients with CNC but were among the tumors found relatively rarely in Prkar1a-deficient mice [8,9]. This suggests that alterations in the RIα and RIIα signaling in different cell types may preferentially predispose them to transformation. It has been shown, for example, that PRKAR1A inactivation promotes the proliferation of human B cells in association with reduced apoptosis [27], both of which may enhance B cell sensitivity to malignant transformation.
HS is an uncommon lesion in humans and mice and the etiology in either species remains unknown [16,23]. Studies of Prkar2a-deficient mice revealed that many animals without clear histologic evidence of HS exhibited histiocytic proliferation in association with DLBCL, a condition we refer to as histiocyte-rich DLBCL [13,28] or, isolated, in the absence of other tumors. It is not clear if such proliferation might be the precursor to tumor development. However, such progression appears possible since in a few animals distinct areas of the spleen were occupied by HS with other areas showing classic features of histiocyte-rich DLBCL. Rarely, HS have been found to harbor IgH or TCR rearrangements, likely as the result of trans-differentiation [29], but this was not found to be the case with the single HS tested here for IgH organization.
It remains to be determined how dysregulation of PRKAR2A expression contributes to transformation of these cells. There is a report of a single in-frame expressed fusion gene resulting from a chromosomal translocation in breast cancer between PRKAR2A and SLC26A6, which encodes an anion transporter [30], but the oncogenetic significance of this event has not been determined. A comprehensive review of the cancer genome atlas (TCGA) did not reveal PRKAR2A mutations in any human tumors but several studies show copy number changes of PRKAR2A's chromosomal locus at 3p21 which, however, is gene-rich and is frequently lost in a variety of neoplasms [31][32][33][34][35].
In conclusion, the present study supports the notion that alterations in PKA signaling in a variety of cell types, including those of the hematologic lineage, may lead to neoplastic transformation. Prkar2adeficiency, in particular, predisposes mice to a spectrum of mostly hematopoietic neoplasms, which may lead to better understanding of the molecular pathology of these lesions in both mice and humans.