ANTIGENIC AND INFLAMMATORY PROPERTIES OF RET/PTC3 ONCOGENE

 Significance

Overview   RET/PTCs are a group of oncogenic fusion proteins derived from the proto-oncogene c-RET, structurally related to a family of receptor tyrosine kinases (1-3). RET/PTCs result from joining the carboxy-terminus of fusion partners with the amino-terminus of c-RET, leading to constitutively active kinase. Of the 11 different fusion genes reported, RET/PTC1 or RP1 and RET/PTC3 or RP3 are the most prevalent (1). In the case of RP3, the amino terminus is derived from the androgen receptor-associated protein, ARA-70 (Fig. 1) (4). RP3 drives three different pathways that strongly influence biological properties of the tumor. First, the constitutively active c-terminal RET kinase domain activates RAS/BRAF/MEK/ERK, PI3K/AKT, p38MAPK, and JNK pathways leading to thyrocyte transformation (5). Second, kinase activity leads to precocious phosphorylation of RP3 itself and other intracellular proteins that provide tumor-specific targets for the adaptive immune system (6). Finally, RP3 also activates NF-κB leading to the production of various inflammatory cytokines and chemokines such as interleukin-1 (IL-1), granulocyte monocytes-colony stimulating factor (GM-CSF), tumor necrosis factor-alpha (TNF-α), (7-10), thereby providing key co-stimulatory signals that license adaptive immune responses. 

 Kinase activity, Phosphopeptides and Adaptive Immunity Dysregulated signaling pathways, thought to arise from aberrant kinase activity, are considered to be a prominent outcome of oncogenic transformation (11-15). Aberrant kinase activity can lead to errant phosphorylation of proteins, and the presentation of tumor-specific phosphopeptides by MHC class I and class II molecules (16-18). It has been demonstrated that CD4 T cells can readily discriminate between a phosphopeptide and its unphosphorylated counterpart (19). This strong specificity to discriminate between peptide and tumor-specific phosphopeptide gives credence to the proposition that peptide-based immunization need not lead to autoimmune manifestations (20).This further justifies interest in identifying phosphopeptides in RP3-expressing and other transformed cells. Indeed, our previous publication demonstrated a CD4 T-cell response to the kinase domain of RP3, suggesting phosphopeptide-specific recognition (6). Indeed, thus far, we have detected only MHC-II-restricted responses to RP3-expressing cells. Although most focus in T cell mediated cancer immunotherapy has been on CD8 T cells, due to their cytolytic activity, CD4 T-cell participation is increasingly appreciated as being critical for effective and long lasting anti-tumor immunity (21-24). Thus, for this proposal, we will focus on MHC-II-restricted presentation of RP3-derived peptides. 

According to convention, MHC class II-bound peptides are derived from proteins that are taken up from the extracellular space by the antigen-presenting cell. However, most phosphopeptides are derived from cytosolic proteins (25), suggesting presentation from endogenous sources of antigen. Data coming from our lab, especially in relation to presentation of viral proteins, demonstrates far more MHC class II-restricted epitope presentation via endogenous pathways than from the conventional, exogenous pathway. In recent work (submitted) we have shown that live virus immunization led to the generation of 13 immunogenic peptides in C57BL/6 mice compared to just 3 using much larger amounts (>106-fold) of inactivated virus. Additional experiments showed that these peptides are generated by true endogenous pathways rather than by transfer from one cell to another and processing by the classical pathway (class II “cross presentation”). The kinase activity of RP3 could thus lead to the generation of far more immunogenic phosphopeptides which can be used in vaccination strategies against RP3 mediated tumor. 

Inflammation and Transformation pathways in RP3  While NF-κB-driven inflammation has been linked to tumorigenesis and tumor growth, in certain cases it can be suppressive (26), and its inhibition can aid tumorigenesis (27-29).  Similarly, chronic inflammation due to mononuclear cell infiltration along with the presence of lymphoid follicles has been associated with a better prognostic outcome and significantly fewer tumor recurrences (30-33). Furthermore, modification of the tumor microenvironment in ways that elicit inflammatory cytokines such as TNF-α and interleukin-2 (IL-2) can stimulate infiltration of lymphocytes leading to tumor rejection (34-37). With an eye toward therapeutic application, our lab has explored the mechanistic basis of transformation and inflammation in RP3 oncogene. The result was a clear demonstration that the inflammatory pathway (NF-κB) is functionally and spatially distinct from RAS/BRAF/ERK/MEK transformation (Fig. 2). TRAF2 and TRAF6, members of TNF receptor associated factor (TRAF) family of E3 ubiquitin ligases, initiate either the classical or alternative NF-κB pathway. The binding sequences for TRAF6 and TRAF2 on RP3 (38) were mutated using site-directed mutagenesis to make TRAF6 (RP3T6mut) and TRAF2 (RP3T2mut) mutants. Both mutations significantly reduced induction of the inflammatory pathway but had no effect on the transformation of cells, with maintenance of AKT and ERK activity and growth in soft agar. The ability of RP3 to signal mitogenesis is dependent upon phosphorylation of tyrosine 588 (8). Mutation of this tyrosine to phenylalanine abrogated binding of src-homology (SHC) adaptor proteins and downstream signaling pathways while the NF-κB-driven pathways remained intact (8). Consequently, the cells lost transformation activity, evident by lack of colony formation in soft agar, but the inflammatory cytokine secretion was intact. RP3 thus drives TRAF-mediated proinflammatory cytokine production which is independent of RAS/BRAF/MEK/ERK transforming pathway (38). These novel findings mechanistically unlink inflammation and transformation and provide an exceptional opportunity to explore the impact of inflammation on tumor malignancy in vivo.

RP3 and association with HT   PTC and HT are strongly co-incidental, and both are associated with an inflammatory environment. However, they are clearly mechanistically distinct. HT is a T-cell mediated autoimmunity with known autoantigens, namely thyroglobulin (Tg) and thyroid peroxidase (TPO). The mouse model of HT, EAT has validated the role of Tg as an autoantigen by induction of disease using human Tg or mouse Tg (mTg) (39,40). The incidence of HT is very high at 8/1000 in the US population (41). Moreover, 4.6 % of the US population suffers from subclinical thyroiditis (42,43) and focal thyroiditis lesions have been identified in ~41% of Caucasian females (US population) post mortem (44). Expression of HLA-DR3 and HLA-DR5 are major risk factors (42,45,46). HT is considered to be a pre-cancerous condition that induces infiltration of lymphocytes, release of various pro-inflammatory cytokines and, consequently, oxidative DNA damage and development of PTC. Yet very few cases of HT progress to PTC. 

Indeed, PTC is relatively rare with an incidence of 3 of 100,000 (6). And while the co-incidence of PTC and HT is high (30-60%) (47), not every case of PTC is associated with HT. The consequence of continued presence of chronic inflammation on PTC progression is also still debated.  Further, presence of HLA-DQ4, not HLA-DR3 nor HLA-DR5, is a risk factor for PTC (48). Thus, formation of the RP3 fusion protein may create an inflammatory environment that eventually leads to HT, as proposed (6). The association of HT with PTC due to RP3 augurs well for PTC as per the clinical data with lower incidence of central lymph node metastasis and overall lower rates of PTC (49-51). It was therefore hypothesized that presence of HT is a better prognostic marker leading to immune-mediated elimination of PTC (49).        

Aim 3. Develop a transgenic mouse model of inducible RP3 expression and assess its inflammatory and immunogenic impact on HT

Rationale:   The large prevalence rate of HT without PTC due to RP3 indicates a possibility, that only a subset of HT is involved with PTC. HT is triggered by various environmental factors which coupled with genetic predisposition factors could lead to development of the disease (42,92). As shown in the schematic, expression of immunogenic fusion oncogene in thyroid could be one of the trigger factors leading to the development of HT (Fig. 16). HT therefore might not a pre-cancerous lesion leading to the development of RP3-mediated PTC.  Presence of HT also might not enhance PTC progression as per the lack of increased incidence of PTC due to RP3 in HT susceptible population. In order to investigate our hypothesis and clarify the interrelationship between HT and PTC, we propose to develop a transgenic mouse model with inducible expression of RP3 in thyroid and overlay EAT induction. An earlier attempt was made to model RP3-driven PTC, using constitutive expression of RP3 in thyroids but this did not lead to transformation (93) unless combined with p53 KO (94).   Constitutive expression of RP3 could however lead to development of T-cell tolerance against RP3 hindering inflammatory reaction and thereby preventing the development of EAT which is unlike the clinical scenarios.  The T-cell tolerance has been demonstrated in other mouse models of constitutive expression of oncogene (95,96). Moreover, RP3 which is fusion oncogene is transiently expressed in thyroid for the development of PTC, therefore constitutive expression is not a clinically relevant model.  The inducible expression system has been demonstrated to induce tumors in a transgenic mouse system using other PTC oncogenes, such as BRAF and RET/PTC1 (97,98). Our system of inducible expression would lead to the stable expression of RP3 once turned on which will be unlike the conditional expression system reported previously where oncogene expression was turned off without continued feeding of doxycycline. Inducible expression better models the natural situation in any event. RP3 could be turned on at different stages of life and one can examine the ease of generation of thyroid lesions. Therefore use of an inducible system of RP3 expression in thyroid is appropriate.  

  Immunotherapy of cancer generally involves breaking tolerance against self-antigens or altered-self antigens. This perturbation of host’s immune system however also leads to undesirable autoimmune sequelae along with beneficial anti-tumor immune responses (99-101). However, mechanistic detail of this interaction remains poorly understood. We will use our combined model to examine the implications of enhanced anti-PTC immune responses on development of EAT. Immunogenic phosphopeptides generated from aim 1 will be used to assess their influence on PTC regression and EAT development. As previously discussed about the phosphopeptides , host’s immune system has the capability of discerning between phosphorylated and unphosphorylated CD4 T-cell epitopes. This potentially can be utilized to immunize against RP3-mediated tumors with no or little cross reaction to autoantigens, and thus limiting autoimmunity.  

Our efforts to develop this combined model of PTC and EAT is to study this interaction, along with development of immunotherapy protocols for PTC, without accompanying EAT. Our mouse model will therefore clarify naturally occurring relationship between a chronic inflammatory condition (HT) and a solid organ tumor (PTC) and help see other tumor systems in a new light for mechanistic insights. 

Preliminary results: Our lab has already developed conditional expression of antigen in the thyroid using a transgenic mouse model. A cell-surface specific glycoprotein (human interleukin-2 receptor alpha chain, Tac) was engineered to contain CD4, CD8 and B cell specific epitopes. This antigen was targeted for single- transgene insertion into hypoxanthine phosphoribosyltransferase (hprt) locus (Fig. 17). Transgene expression was targeted to the thyroid via the bovine thyroglobulin promoter and a Cre/loxP-based strategy was used for conditional expression. The loxP-flanked STOP cassette was inserted between the tissue-specific promoter and the transgene open reading frame. Upon removal of the STOP cassette by Cre-mediated recombination, the thyroglobulin promoter drives expression of the modified Tac antigen exclusively in thyroid epithelial cells. In order to achieve this, TacOFF transgenics were further crossed with mice expressing a Cre-estrogen receptor fusion protein (CreER) driven by the actin promoter. Upon tamoxifen administration, these mice demonstrate thyroid-specific expression of the transgene and presentation of the CD8 epitope (Fig. 18).

Research Plan: 

a) Develop Cre-transgenic mice with inducible expression of RP3 that will be combined or not with experimental autoimmune thyroiditis (EAT), subsequently assessing PTC progression. 

RP3 transgenic mice will be developed under the Cre/loxP system using the bovine-thyroglobulin promoter to target thyroid-specific expression. RP3-loxP transgene will be introduced in embryos. The mice will then be bred to Cre-ER mice. Unlike the doxycycline conditional expression, once the RP3 is turned on it cannot be turned off again, as would be the case for natural RP3-expressing tumors.  RP3 will be turned on at different time points, from 6 weeks to 2 yrs (to mimic human age) to examine the ease of development of lesions. One aspect of the hypothesis is that HT can develop secondary to RP3 expression. Therefore RP3 will be expressed and the development of EAT will be examined along with PTC progression (Fig. 19). Along with the wild type RP3, we will also use RP3 mutants ablated in either of the pathways as controls. The other hypothesis to be tested is that HT will impede PTC progression. To answer this question, EAT will be induced prior to RP3-expression and its influence on PTC progression will be assessed (Fig. 20). Various grades of EAT using mTg and varying doses of IL-1 (5,000, 10,000 or 20,000U IL-1) could be induced to examine their relative influence on PTC progression (102). In a therapeutic model, EAT will be induced after RP3 expression to assess influence of enhanced inflammation on ongoing PTC progression (Fig. 21). 

In all cases, thyroids will be grossly examined for PTC (enlargement). Histopathological sections of the thyroid (50-60 vertical sections at 10-15 step levels) will be examined by hematoxylin and eosin staining of formalin fixed tissue. Mononuclear cell infiltration and inflammation for EAT and presence of various markers of transformation such as hyperplasia, papillary carcinoma lesions for PTC progression will be examined. Immunofluorescense staining will be performed on the thyroid for RP3 expression along with cellular infiltration to assess tolerogenic (MDSCs, inhibitory macrophages) or non-tolerogenic (CD4, CD8 T-cells, less or no infiltration of MDSCs) tumor microenvironment. To measure cell mediated immune responses, proliferation assays and ELISPOT assays to measure T-cell reactivity will be assessed for mTg, along with RP3. Humoral responses will be examined by analyzing anti-mTg and anti-RP3 antibody using enzyme linked immunosorbent assay (ELISA). The thyroid extracts will be a source of mTg whereas for RP3, recombinant protein will be made and quantified according to the given protocol (6).

b) Prophylactic and therapeutic immunization against RP3 using in vivo model of PTC and HT. RP3 transgenic mice will be immunized with phosphopeptides identified from aim 1 or recombinant RP3-vaccinia virus (RP3-VACV) to examine the protection afforded against PTC in a prophylactic (Fig. 22) and/or therapeutic setting (Fig. 23). The influence of robust anti-tumor immunity on EAT development will be assessed. Our readout therefore will be thyroid histology for identifying PTC progression or regression and mononuclear cell infiltration for EAT. ELISPOT assays will be used to measure T-cell reactivity against RP3 and mTg and antibody levels against both antigens.   Thyroid sections will be stained with immunofluorescent antibodies to assess the cellular infiltration and to identify other key components of anti-tumor immunity. Cell depletion studies along with cell transfer studies (T-cells, MDSCs, Tregs) will be performed to characterize the anti-tumor immune responses. Presence or absence of EAT should not change the outcome of immunization.  

 Discussion:  

Subaim a: If the hypothesis is correct, induced expression of immunogenic and inflammatory RP3 will lead to immune infiltration in thyroids. This low grade thyroiditis could be one of the triggering factors leading to the development of EAT, only under appropriate conditions such as presence of large pool of primed autoreactive T-cells against Tg due to breakdown in immune tolerance against mTg following infection (Fig. 11). C57BL/6 mice are resistant to EAT induction. In that scenario, regulatory T cell depletion (Treg) which is required for peripheral tolerance against different antigens including mTg should lead to robust development of EAT (103-106). Since development of EAT might be contingent on the presence of inflammatory signaling of RP3, transforming only RP3 mutant control (TRAF2 and 6 mutants) should lead to lower or no development of EAT. Presence of pre-existing EAT should not enhance PTC as per our hypothesis but rather reduce or slow down PTC progression or cause regression of RP3 or transformation only RP3. In this scenario, low grade EAT should have a reduced effect on PTC progression whereas high grade EAT pathology should lead to increased regression or impeded tumor growth. Similarly, induction of EAT after the RP3 expression also should reduce the PTC progression. The readout will be primed T-cell response against RP3 along with antibody response. As discussed, conditional expression of RP3 is expected to cause transformation since this was the outcome with two other PTC oncogenes.  If necessary, we can always combine it with knockdown of tumor suppressor genes such as p53 or PTEN to induce tumors (94,107).