Direct evidence of cellular transformation by prion-like p53 amyloid infection

p53 has been extensively studied in the field of cancer biology due to its central role in the regulation of anti-proliferative and tumor suppressive programs by activating or repressing key effector genes (Lane and Crawford, 1979; Levine, 1997; Mello and Attardi, 2018; Vousden, 2002). Many studies have established the mechanism of loss of tumor suppressive functions by p53 and the consequent initiation of cancer (Kim et al., 2009; Moll and Schramm, 1998). The loss of p53 function is mostly associated with its mutations, wherein p53 either misfolds (structural mutation) or is unable to bind to its cognate DNA sequence (contact mutation) (Joerger et al., 2006; Olivier et al., 2010). Mutant p53 proteins not only lose their native tumor suppressive functions but often gain additional oncogenic functions and essentially behave as an oncogene (Freed-Pastor and Prives, 2012; Oren and Rotter, 2010; Rivlin et al., 2011). p53 mutations enhance the proliferative and tumorigenic potential of the cells contributing to various stages of tumor progression and increased anti-cancer drug resistance (Bargonetti and Prives, 2019; Stein et al., 2019). Hotspot mutations in p53 show a common tendency to destabilize/misfold the native conformation of the protein (Wilcken et al., 2012), leading to the formation of aggregates of inactive p53, which are sequestered and targeted by the proteasomal degradation machinery (Ano Bom et al., 2012; Levy et al., 2011; Ano Bom et al., 2012; Levy et al., 2011). Interestingly, wild-type p53 has also been reported to aggregate in vitro (Bullock et al., 1997; Wang and Fersht, 2012, 2015), and also in the cytoplasm and/or nucleus of cancer tissues (De Smet et al., 2017; Moll et al., 1996; Ostermeyer et al., 1996). Recently, many studies have suggested that p53 aggregation and amyloid formation might be the causative factors of p53 loss/gain-of-function in cancers (Ghosh et al., 2017; Rangel et al., 2014; Silva et al., 2014; Lasagna-Reeves et al., 2013). This is further supported by the detection of p53 amyloids in cancer cells and tumor tissue biopsies (Ghosh et al., 2017; Ano Bom et al., 2012; Levy et al., 2011). Moreover, similar to many other amyloids associated with neurodegenerative disorders, p53 amyloids also show prion-like properties in cells (Ghosh et al., 2017).

Although previous studies have shown the link between p53 amyloid formation and cancer (Silva et al., 2014; Navalkar et al., 2020), the direct evidence that endogenous p53 amyloids can induce tumor formation is lacking. Here, we investigate whether the wild-type p53 amyloid formation can directly induce the cancerous transformation of normal cells leading to tumor initiation. We use a unique cell model (Ghosh et al., 2017) to induce aggregation of the endogenous wild-type p53 via exogenous addition of in vitro synthesized p53 core fibrils (seeds) to normal cells. These cells with p53 amyloids exhibit enhanced cancer-reminiscent properties, such as increased survival, colony formation and proliferation rate, and enhanced migration with resistance to apoptosis. More importantly, p53 amyloid-containing cells are tumorigenic and can establish a tumor xenograft in immunocompromised mice. Disaggregation of cellular p53 aggregates using a peptide inhibitor rescues p53 function and reduces the tumorigenic properties of these cells. Overall, our study reveals that wild-type p53 amyloid formation can induce oncogenic traits and lead to the gain of tumorigenic potential in non-cancerous cells. These results propose that endogenous p53 amyloid formation can be a plausible cause of cancer initiation.

To examine whether wild-type p53 amyloid formation can lead to the transformation of normal cells to a cancerous phenotype, we chose two non-cancerous cells of different origins – MCF 10A cells (non-tumorigenic human breast epithelial cells) and HFF cells (human foreskin fibroblast). These cells have been used to identify oncogenic agents that can cause transformation in cells (Yusuf and Frenkel, 2010; Imbalzano et al., 2009; Scott et al., 2004). As the exogenous addition of p53 core amyloid fibrils (seeds) leads to intracellular p53 amyloid formation (Ghosh et al., 2017), we have used this method to induce p53 amyloid formation in the two non-cancerous cells. For amyloid seed preparation, p53 core protein was incubated and its aggregation and amyloid formation was confirmed using electron microscopic imaging, circular dichroism and Thioflavin T (ThT) fluorescence (Fig. 1A, left panel; Fig. S1A-C). Further, MCF 10A and HFF cells were treated with sonicated core p53 fibrils for 48 h to induce aggregation of p53, and untreated cells were kept as a control (Fig. 1A, right panel; Fig. S1D). Immunofluorescence using p53 DO-1 antibody and confocal microscopy confirmed p53 puncta formation in the cells (Fig. 1B; Fig. S2A). Interestingly, the majority of the cell population showed the presence of nuclear aggregates of p53, both for MCF 10A (∼79%) and HFF cells (∼83%) (Fig. 1C). Untreated cells and α-synuclein amyloid fibril-treated cells did not show significant p53 stabilization (Fig. S2B), indicating that p53 aggregation in cells is specific to the treatment of p53 core-derived amyloids. To confirm the internalization and templating ability of these p53 core fibrils, we exogenously added rhodamine-labeled p53 core fibrils seeds (Rhod-Core) to MCF 10A cells and immunostained with the DO-1 p53 antibody (green). Rhodamine-labeled core seeds were internalized, showing cytoplasmic and nuclear localization (pseudocolored magenta). We observed significant colocalization of magenta and green signal in the nucleus (Fig. S2C). This indicates the templating ability of internalized p53 core fibril seeds, leading to amyloid aggregation of endogenous p53.

To confirm the amyloid nature of p53 in cells, p53 was immunoprecipitated using p53 DO-1 antibody from core fibril-treated and untreated cells (MCF 10A and HFF), followed by dot blot with p53 DO-1 and amyloid fibril-specific OC antibody. The dot blot confirmed the amyloid nature of p53 aggregates (OC positive) in the fibril-treated MCF 10A and HFF cells, whereas p53 was not detected in untreated cells (Fig. 1D).

For its tumor suppressive function, p53 modulates the cellular expression of several hundreds of genes by binding to its specific DNA response elements (Kastenhuber and Lowe, 2017). As p53 formed aggregates inside the nucleus of cells, we first examined p53-specific DNA binding and downstream transcriptional activity of p53 in MCF 10A cells using a luciferase reporter assay (El-Deiry et al., 1993). Core fibril-treated cells showed significant loss of p53 transcriptional activity (luciferase signal) compared to the untreated cells, suggesting that amyloid formation of p53 leads to loss of its specific DNA binding (Fig. 1E). Considering the loss of transcriptional activity of p53, we hypothesized that p53 amyloid formation may also lead to the loss of key p53 functions, such as the triggering of apoptosis under stress conditions. To examine this, core fibril-treated MCF 10A and HFF cells were subjected to actinomycin D (ActD) treatment as a stress inducer (Fig. 1F). ActD-treated cells without fibril treatment were used as a control. In both the cell lines (MCF 10A and HFF), the core fibril-treated cells showed resistance to apoptosis in the presence of stressor compared to untreated cells (Fig. 1F; Fig. S3A). This suggests that the loss of apoptotic function of p53 is due to amyloid formation.

Next, we proposed that similar to mutant p53 (Stein et al., 2019), p53 amyloids may also act as an oncoprotein. Previously, we demonstrated that p53 amyloids can enhance the existing transformative properties of SH-SY5Y cells (transformed neuroblastoma cell line) (Ghosh et al., 2017). We hypothesized that similar changes might also induce the transformation of normal cells. Cells acquiring such transformative capabilities should exhibit various properties similar to transformed cells, such as colony formation in soft agar, hyperproliferation and enhanced migration (Dong et al., 2013; Muller et al., 2011; Oren and Rotter, 2010; Brosh and Rotter, 2009). Indeed, core fibril-treated MCF 10A and HFF cells displayed an increased ability to form colonies in the soft agar assay (cell transformation assay) compared to the untreated controls (Fig. 1G; Fig. S3B). This transformative potential is further assisted by their enhanced proliferation, as evidenced by the higher number of cells expressing Ki67 (a cellular proliferation marker; Scholzen and Gerdes, 2000) compared to untreated cells (Fig. 1H; Fig. S3C). Further, the cell migration study, using a wound healing assay (Liang et al., 2007), showed higher migration of core fibril-treated MCF 10A cells (migration rate of ∼13.61 μm/h) compared to untreated cells (migration rate of ∼3.44 μm/h) (Fig. S3D,E). A similar observation was also seen in HFF cells (Fig. S3D,E). This property of enhanced migration, analogous to cancerous cells (Yamaguchi et al., 2005), indicates a gain of transformative properties by these cells.

The role of p53 in pro-survival activities of a cell is tightly regulated and is integrated with its tumor suppressive function to prevent cancer induction (Reisman and Loging, 1998). Loss of regulation over these pathways provides the cells with a survival advantage. Several yeast prion proteins, like Mod 5, are known to transmit from mother-to-daughter cells and transfer this survival advantage to the next generation (Suzuki and Tanaka, 2013; Suzuki et al., 2012). We aimed to investigate whether the intracellular p53 amyloid aggregates can be maintained through the mitotic divisions and transmit from mother-to-daughter cells, which might also affect the survival or number of cell generations (passage number). To evaluate this, both the cells (MCF 10A and HFF) were treated with core fibrils and were passaged continuously for many generations (Fig. 2A,B; Fig. S4A,B). Untreated MCF 10A and HFF cells were used as controls. It is important to note that only the first generation of cells was treated with core fibril seeds. All subsequent generations did not receive any exogenous fibril treatment. Each cell generation after core fibril treatment was termed T1, T2 and so on (Fig. S4A,B). An Annexin-V PI assay coupled with fluorescence-activated cell sorting (FACS) analysis of MCF 10A and HFF cells showed increased survival of core fibril-treated cells (reduced apoptosis) over the generations compared to corresponding untreated cells (Fig. 2A,B). At each passage, MCF 10A cells were immunostained using p53 DO-1 antibody for the detection of p53 in the cells. For core-treated cells, p53 aggregates (predominantly nuclear puncta) were detected throughout all the cell generations (Fig. 2C), indicating that p53 amyloids propagate through multiple generations. Untreated cells did not show the presence of p53 aggregates (Fig. 2C). Furthermore, to confirm the p53 stabilization, we quantified the p53 signal in untreated and core-treated cells at passage T5. The core-treated cells showed a significantly higher p53 signal compared to the untreated cells (Fig. 2D). We performed co-immunostaining using the amyloid fibril-specific OC antibody and p53 antibody at the T5 passage of MCF 10A cells. The p53 signal colocalizes with the OC antibody signal, predominantly in the nucleus (Fig. 2E), confirming the amyloid nature of p53 aggregates across passages. Further, the presence of p53 amyloid fibrils was confirmed by immunoprecipitation of p53 followed by dot blot assay using OC antibody. The presence of OC positive aggregates in the nucleus at the later generations for both MCF 10A and HFF cells (Fig. 2F) suggests that p53 amyloid aggregates can be passed from one cell generation to the next, contributing to enhanced survival of the cells. Interestingly, the functional inactivation of p53 also persists through cell generations due to the propagation of p53 amyloids. This is evident as the luciferase reporter assay using MCF 10A core fibril-treated cells showed loss of p53 transcriptional activity at T5 in contrast to untreated cells, which is further consistent with the extent of p53 amyloid formation in cells (OC immunoreactivity, Fig. 3A, Fig. 2F). This indicates that the amyloid form of p53 is no longer able to participate in cellular apoptotic mechanisms, allowing core fibril-treated cells to overcome apoptosis for higher survival, whereas untreated cells undergo apoptosis in the continuous culture.

p53 is known to regulate senescence and aging of cells through its tumor suppressive pathways (Itahana et al., 2001), and inactivation of p53 has been shown to induce reentry of cells into the cell cycle and rapid tumor progression (Kim et al., 2009). Therefore, p53 amyloid formation can dysregulate the pathways involved in activating senescence, thereby changing the fate of the cells. We examined senescence-associated beta-galactosidase (SA-β-gal) expression with chromogenic substrate X-Gal using a cytochemical assay (Gary and Kindell, 2005). Core fibril-treated cells (T5 generation) showed a reduction in the percentage of SA-β-gal⁺ cells compared to untreated control (Fig. 3B). This indicates that fibril treatment plays a role in the inactivation of senescence/aging pathways of cells, which further contributes to the cancer-like phenotype. Furthermore, loss of p53 function has been established to play a critical role in conferring a high level of drug resistance to cancer cells against doxorubicin, cisplatin, alkylating agents, gemcitabine and microtubule targeting drugs (Zhou et al., 2019; Hientz et al., 2017). p53 aggregation has also been shown to enhance platinum resistance in ovarian cancer cells (Yang-Hartwich et al., 2015). To examine the possible drug resistance of MCF 10A cells due to p53 amyloid formation, the viability of fibril-transformed cells (at T5) were analyzed using a 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay in the presence and absence of cisplatin, doxorubicin and paclitaxel, which are well-known drugs used against breast cancer (Onda et al., 2004). Core fibril-treated cells showed significantly higher EC₅₀ values of these drugs compared to the untreated control (Fig. 3C). The survival trend of untreated cells and core fibril-treated cells was plotted for each of the drug concentrations (Fig. S4C). These data indicate that p53 amyloid formation can induce drug resistance in cells.

Extracellular matrix (ECM) majorly regulates the tumor microenvironment and similar signals cannot be mimicked when cells are cultured on two-dimensional substrates (Nelson and Bissell, 2006). As MCF 10A cells acquire a transformative phenotype after p53 amyloid formation, they might also show gain-of-oncogenic traits, characteristic of tumor cells, such as the formation of spheroids in three-dimensional culture (Tevis et al., 2017). ‘Matrigel’ is a naturally derived ECM extracted from the Engelbreth–Holm–Swarm (EHS) mouse tumor, and is used as a gold standard scaffold for the development of three-dimensional cell cultures (Benton et al., 2014; Kleinman and Martin, 2005). It facilitates the cells to express key features reflecting their inner malignancy, such as invasion capability (Kenny et al., 2007). Hence, for establishing the transforming capability and tumor-like nature of p53 amyloid-transformed cells, we cultured the fibril-treated and untreated MCF 10A cells in Matrigel (Fig. 3D). Core fibril-treated MCF 10A cells at T5 passage showed three-dimensional spheroid formation (Fig. 3E) with an average diameter of ∼100 µm on day 2 (Fig. S5A,B). The volume of spheroids was calculated from the diameter and was found to increase gradually over 10 days (Fig. 3F), indicating proliferation and migration of core fibril-treated cells. However, untreated cells did not show significant cellular aggregation to form distinct three-dimensional spheroids (Fig. 3F; Fig. S5A). Further, spheroids showed positive staining with Calcein-AM dye, indicating cell viability (Fig. S5C). As p53 maintains a transcriptional program to prevent epithelial-to-mesenchymal transition (EMT) of cells (Chang et al., 2011), the loss of this p53 activity may also induce an EMT-like phenotype in cells (Muller et al., 2011). In this context, the amyloid-induced three-dimensional spheroids showed high expression of EMT markers, such as β-catenin, Slug, Vimentin and N-cadherin, as evident by immunofluorescence on day 10 (Fig. 3G). This suggests that cells with p53 amyloid aggregates promote EMT potential, which is known to contribute to cancer progression (Roche, 2018).

Finally, to confirm the p53 amyloid-mediated gain of tumorigenicity by the MCF 10A cells, core fibril-treated cells at T5 were injected into severe combined immunodeficiency (SCID) mice to examine tumor induction in a mouse xenograft model (Fig. 4A). All animals injected with core-treated cells at T5 showed tumor formation within 2-3 weeks (n=8). The untreated MCF 10A (at T5) and MCF 7 cells were also injected as negative and positive controls, respectively. The injection of fibril-treated cells in the mammary fat pad of female mice showed gradual tumor generation after 2 weeks, whereas no tumor was observed for untreated MCF 10A cells. MCF 7 cells, the positive control for the xenograft, also showed tumor development in the animals. The increase in tumor volume was calculated up to the fifth week, after which the animals were sacrificed and tumors were isolated (Fig. 4B,C). Further, positron emission tomography and computed tomography (PET-CT) scans were used as a non-invasive technique to visualize the tumor anatomy and assay the metabolic activity of the tumor (Mulero et al., 2011). For this, the tumors were first identified with small animal CT followed by a small animal PET scan for the evaluation of ¹⁸F-FDG (fluorodeoxyglucose) accumulation. The tumors induced by cellular xenograft showed an accumulation of ¹⁸F-FDG, indicating active metabolic activity (Fig. 4D). Increased standardized uptake values (SUVs) of ¹⁸F-FDG for MCF 7-injected animals (mean±s.e.m value of 1366.7±225.6) and core fibril-treated MCF 10A-injected animals (mean±s.e.m value of 1051.6±145.5) reflected the viability and metabolic activity of cells in the tumor (Fig. 4E). Further, the histology of excised tissue (tumor/normal) was examined using Hematoxylin and Eosin staining (H&E) (Fig. 4F). Tumor tissue from animals injected with core fibril-treated MCF 10A showed loss of normal ductal-lobular architecture with a high degree of nuclear pleomorphism. The extensive nuclear staining by Hematoxylin indicated hyperproliferative cells in the tumors with loss of normal tissue histology, similar to positive control MCF 7 tumor (Fig. 4F). In contrast, normal tissue morphology was observed in untreated MCF 10A-injected animals (Fig. S5D, upper panel). The immunohistochemistry of tumor tissues using p53 DO-1 antibody and amyloid specific antibody (OC) demonstrated the presence of p53 aggregates in the tumor tissue (Fig. 4G), whereas negligible p53 expression was observed in control tissue from MCF 10A only-injected animals (Fig. S5D, lower panel). Tumors induced by MCF 7 cells showed stabilized expression of p53 but without colocalization with OC, indicating the absence of a significant level of p53 amyloid aggregates (Fig. 4H). This suggests that p53 amyloid formation leads to the cellular transformation of non-cancerous MCF 10A cells. However, this mechanism may not be exclusively involved in the cancerous nature of cells and/or tissues as other stimuli, such as p53 misfolding (without amyloid formation), loss of other tumor suppressors proteins and overexpression of additional pro-oncogene(s), which can also contribute to the transformation of cells.

If the observed transformed phenotype of the cells is initiated due to core fibrils treatment (i.e. endogenous p53 aggregation), then rescuing the p53 aggregation using inhibitor should reduce the oncogenic traits. In this context, Soragni et al. (2016) have developed a peptide inhibitor based on the aggregation-prone region of p53 termed as rescue peptide, which specifically targets aggregated p53 (Soragni et al., 2016). We used in vitro synthesized rescue peptide and treated the cells containing aggregated p53 (at T5). Immunofluorescence staining of p53 in cells showed reduced p53 aggregates in the nucleus compared to the rescue peptide-untreated cells (Fig. 5A). We further evaluated the restoration of the transcriptional function of p53 after rescue peptide treatment. p53 disaggregation led to the rescue of p53 sequence-specific binding and transcriptional function, as demonstrated by higher luciferase activity of rescue peptide-treated cells (+RP) compared to rescue peptide-untreated cells (−RP) (Fig. 5B). Further, to check whether the apoptotic function of p53 is restored due to p53 disaggregation by inhibitor peptide, an Annexin-V PI assay followed by FACS analysis was performed. Rescue peptide-treated cells showed loss of viability with an increase in apoptosis, as seen in the FACS data (Fig. 5C; Fig. S6A). As these cells were at T5 generation, when ideally the MCF 10A cells without p53 aggregation (untreated) do not maintain their viability (Fig. 2A), high apoptotic death was observed even in absence of external stressor (due to rescue peptide treatment). Moreover, the rescue peptide-treated cells showed decreased colony formation potential in soft agar assay (Fig. 5D; Fig. S6B) and reduced migratory potential (Fig. 5E), which is comparable to untreated MCF 10A cells without p53 aggregation. Furthermore, when the rescue peptide-treated cells were injected in mice (Fig. 6A), unlike fibril-treated cells, these cells did not induce tumor formation (Fig. 6B). All six mice injected with rescue peptide-treated cells did not show tumor formation. This suggests that rescue peptide treatment of cells mostly reverts them to the non-cancerous state (similar to untreated MCF 10A cells). CT imaging further confirmed the tumor anatomy in core fibril-treated MCF 10A-injected animals with increased ¹⁸F-FDG uptake, whereas no signal was detected for animals injected with the same cells but treated with rescue peptide (Fig. 6B). Further, core fibril-treated MCF 10A cells xenograft tumor showed neoplastic cells with extensive Hematoxylin staining of the nuclei, indicating the presence of hyperproliferative cells (Fig. 6C). However, rescue peptide-treated cells neither showed tumor formation in mice nor hyperproliferative cells in the tissue of the injected area (Fig. 6C). This suggests that rescue peptide treatment results in p53 disaggregation and cells are unable to establish a tumor xenograft in mice.

Protein aggregation and amyloid formation are linked with several human neurodegenerative diseases, such as Alzheimer's, Parkinson's and prion disease (Chiti and Dobson, 2017; Soto, 2003). In these diseases, the cellular toxicity is associated with (1) loss of native functions of the protein and/or (2) gain of toxic functions due to protein aggregates (Eisenberg and Jucker, 2012; Winklhofer et al., 2008). In vitro aggregation of amyloidogenic proteins suggest that disease-associated protein aggregation proceeds through nucleation-dependent polymerization reaction, which is greatly accelerated by the addition of preformed seeds (Chiti and Dobson, 2017). Indeed, this is the mechanism by which infectious prion PrP^sc infects cells containing normal PrP (PrP^c) and converts them to the pathological form of PrP^Sc (Prusiner et al., 1998). Multiple studies suggest that prion-like amplification and disease spread is common in various neurodegenerative disorders, including Alzheimer's and Parkinson's disease (Aguzzi and Rajendran, 2009; Eisenberg and Jucker, 2012; Jucker and Walker, 2013). The latest addition to the family of amyloid diseases is p53, implicated in cancer pathogenesis, wherein loss of p53 native function is suggested to be linked with its amyloid formation (Costa et al., 2016; Ghosh et al., 2017; Lasagna-Reeves et al., 2013).

p53 contains unstructured domains (Wang et al., 1994) and an aggregation-prone sequence in the core domain of the protein (Ghosh et al., 2014). Thus, in vivo, several hotspot mutations and post-translational modifications leading to exposure of p53 aggregation prone-sequence are linked with p53 aggregation (Bouaoun et al., 2016; Navalkar et al., 2020). A study on p53 aggregation in six different human cancer tissue types demonstrated the poor clinical outcome of tumors with p53 aggregates (De Smet et al., 2017). Similarly, p53 aggregation was also observed in bladder carcinoma, astrocytoma, osteosarcoma, colon carcinoma and pharynx carcinoma (Xu et al., 2011). Furthermore, our group also demonstrated the presence of p53 amyloids in human breast and lung cancer tissues (Ghosh et al., 2017). Taken together, this suggests that p53 amyloids are present in cancer tissues of several origins. Further, p53 fibril uptake by cells (Forget et al., 2013) and cell-to-cell transmission of p53 aggregates (Ghosh et al., 2017) also raises the possibility that p53 amyloids can act as a transmissible prion-like protein (Navalkar et al., 2020).

Although these previous studies speculated the role of p53 amyloids in cancer pathogenesis (Ghosh et al., 2017), p53 amyloids leading to cellular transformation and tumor formation have not been directly demonstrated yet due to the lack of an appropriate cellular model (controlled p53 amyloid induction). In this context, our work suggests that similar to many other amyloids (Eisenberg and Jucker, 2012; Jucker and Walker, 2013), exogenous addition of p53-derived amyloid seeds of the p53 core domain (Ghosh et al., 2017) or the amyloid seeds of the high aggregation-prone sequence of p53 (Ghosh et al., 2014) can induce the amyloid formation of endogenous p53 in cells. Using this technique, we show here the cellular transformation and tumor induction by normal cells upon endogenous p53 amyloid formation. When a minute amount of p53 core amyloid seeds were exogenously added, mostly nuclear p53 aggregates (puncta) were observed in MCF 10A and HFF cells (Fig. 2E). This suggests that the core p53 fibrils seed aggregation of otherwise soluble p53, possibly by a direct transport to the nucleus. Such direct transport of amyloid seeds into the nucleus and subsequent nuclear aggregation is also observed for Huntington disease-associated amyloids (Deguire et al., 2018; Yang et al., 2002). Interestingly, localization of p53 aggregates might vary with different cell types; for example, with similar experiments, mostly cytoplasmic p53 aggregates were observed in transformed neuroblastoma cell line SH-SY5Y (Ghosh et al., 2017). This difference in cellular localization (cytoplasmic or nuclear) of p53 aggregates could be due to differences in the nuclear barrier and also might be dependent on the specific cell types. However, both cells (MCF 10A and HFF) with nuclear p53 aggregates also showed loss of transcriptional and apoptotic activities of wild-type p53 (Fig. 1B,E; Fig. S2A). Similar to cytoplasmic p53 aggregates, which are known to inhibit p53 translocation into the nucleus along with sequestration of other proteins and tumor suppressors (p63, p73) (Xu et al., 2011), the nuclear p53 aggregates cannot bind p53-specific cognate DNA sequence for transcriptional activities (Fig. 1E, Fig. 3A). Instead, they might bind non-specifically to alternative sequence elements and initiate aberrant transcription of genes resulting in a gain-of-function phenotype, as observed for mutant p53 (Muller and Vousden, 2014). Indeed, MCF 10A and HFF cells containing p53 amyloids showed resistance to apoptosis when exposed to stress conditions, and also enhanced the pro-metastatic properties (such as transformation potential and increased motility), resembling cancer cells (Fig. 1F-H). Therefore, the cellular activities of p53 amyloids have much more resemblance to functional amyloids (Fowler et al., 2007; Maji et al., 2009), in contrast to amyloids associated with neurodegenerative diseases.

Furthermore, prion-like functional amyloids in yeast show horizontal transmission (cell-to-cell transmission) and propagation through generations from mother-to-daughter cells (Halfmann et al., 2012; Liebman and Chernoff, 2012; Wickner et al., 2015). This transfer of misfolded protein aggregates allows the spread of loss-of-function/gain-of-function phenotype to the recipient cells. Similarly, p53 amyloids are also propagated from one generation to the next (mother-to-daughter cells), leading to an increase in survival in terms of the number of generations (passages) (Fig. 2). When cells containing aggregates undergo division, nuclear aggregates might pass from mother-to-daughter cells similar to cytosolic prion transmission in yeast (Uptain and Lindquist, 2002; Wickner, 1994). The enhanced survival capacity of cells containing p53 amyloids further supports that p53 prions must provide fitness benefit to the cells under stress conditions (providing additional survival advantage), similar to yeast prion (Serio and Lindquist, 2000; Shorter and Lindquist, 2005; Suzuki et al., 2012; Wickner et al., 2000). Moreover, the protein homeostasis machinery, including chaperones, must influence the inheritance and persistence of p53 amyloids. The survival benefit is further evident as cells containing p53 amyloids even at the fifth generation show decreased senescence and increased resistance against cytotoxic drugs (Fig. 3B,C). The direct evidence that p53 aggregates confer tumor-like properties to the cells is the observation that cells at the late passage (T5) with p53 amyloid aggregates induce tumorigenesis upon xenograft in mice (Fig. 4). In addition, the xenograft tumor tissue from the mice demonstrated the presence of p53 aggregates, establishing a direct link between p53 amyloid formation and tumorigenesis. In this regard, the accumulation and subsequent aggregation of p53 have also been observed in clinical biopsy tissues of neuroblastoma, retinoblastoma, breast and colon cancers (Levy et al., 2011; Moll et al., 1995). Importantly, disaggregation of p53 partially restores the native function of p53 and rescued p53 re-engages in cell death and diminishes the tumorigenic propensity of cells (Figs 5, 6).

Therefore, the present study suggests that the altered cellular properties due to p53 amyloid formation might arise through a multitude of cellular pathways, which are likely to contribute to the transformation of cells. The consequence of p53 aggregation can be dictated by the in vivo cell and tissue-specific condition, gain-of-specific oncogenic pathways, as well as the dysregulation of proteostasis machinery, which needs to be explored further. The partial degradation of p53 amyloids by proteasomal machinery in cells can further generate secondary infectious protein aggregates (seeds), which can aid in cell-to-cell transmission and the acquisition of inheritable p53 prion transmission (Halfmann et al., 2012; Liebman and Chernoff, 2012; Wickner et al., 2015; Winkler et al., 2012). Hence, systematic investigations are required to establish how pro/anti-aggregating conditions can govern the clinical outcome of p53 amyloid formation associated with cancer in a cell/tissue-specific manner. Overall, our study demonstrates that p53 protein misfolding/aggregation, as well as an infection of preformed p53 amyloids, contribute to cellular transformation and cancer pathogenesis in a subset of cancers (if not all), which can be targeted by using p53 aggregation inhibitors.

All reagents and chemicals used were of the highest quality and were purchased either from Sigma-Aldrich (St Louis, MO, USA) or Merck (Darmstadt, Germany). Milli-Q system (Millipore Corp., Bedford, MA, USA) was used for obtaining double-distilled de-ionized water, followed by autoclaving before use. Rescue peptide was custom synthesized by USV Limited (Mumbai, India) using the solid-state peptide synthesis method with >95% purity. Peptide mass and purity were confirmed using high performance liquid chromatography (HPLC) and matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) mass spectrometry.

p53 core protein was purified according to our previously published protocol (Ghosh et al., 2017), using p53 core domain plasmid (Addgene, 24866). Briefly, p53 core domain plasmid was transformed in Escherichiacoli BL21 (DE3) for purification of p53 core protein with N-terminal His6 tag. Isopropyl β-D-1-thiogalactopyranoside (IPTG, Himedia, Mumbai, India) (1 mM) was used to induce p53 core expression at 25°C for 8 h. After IPTG induction, the cells were pelleted down and lysed in 50 mM sodium phosphate (pH 8.0) with 0.3 M NaCl and 1 mM phenylmethylsulfonyl fluoride (Sigma-Aldrich, St Louis, MO, USA). The cell pellet was sonicated at 40% amplitude for 15 mins with 3 s on and 3 s off cycle. The cell lysate was centrifuged at 10,000 g at 4°C for 45 min. The supernatant was collected and added to the Ni²⁺ sepharose affinity chromatography column to purify His-tagged p53 core protein from the cell lysate. The elution of the protein was performed using the imidazole gradient from 100 to 250 mM. The purity of protein fractions was confirmed using 12% SDS-PAGE. UV absorption spectra were taken for measuring the concentration of the core protein. The molar absorptivity of protein (ε) was taken as 17,420 M⁻¹ cm⁻¹

p53 core monomer protein (500 μl) at a concentration of ∼50 μM in 50 mM sodium phosphate buffer (pH 7.4), 0.3 M NaCl and 0.01% sodium azide with chondroitin sulphate A (Sigma-Aldrich, St Louis, MO, USA) at equimolar concentration dissolved in the same buffer were mixed. The protein solution was then used for setting up in vitro aggregation of p53 core domain according to the previously published protocol (Ghosh et al., 2017). Briefly, the solution was incubated at 37°C for 96 h with slight agitation. After incubation, the protein solution was centrifuged at 200,000 g for 45 min at 4°C. The fibril pellet was resuspended in sterile PBS (pH 7.4).

Circular dichroism spectroscopy of the p53 core monomeric protein (0 h) and p53 core amyloid fibrils (up to 96 h) was conducted to confirm the structural transition and amyloid formation of the p53 core. For this, 200 μl p53 core monomeric protein or resuspended p53 core fibrils (10 μM) were taken in washed quartz cell (Hellma, Forest Hills, New York, NY, USA) with a path length of 0.1 cm. Spectra were obtained by circular dichroism instrument (JASCO-1500, JASCO, Hachioji, Tokyo, Japan) between the wavelength range of 200-260 nm at room temperature. For each sample, three accumulations were acquired and the average was taken. Smoothing and buffer subtraction was carried out for the processing of raw data.

To evaluate the amyloid nature of p53 core fibrils, the ThT binding was measured for p53 core monomeric protein (0 h) and p53 core amyloid fibrils (96 h). For this, 200 μl p53 core monomeric protein or resuspended p53 core fibrils (10 μM) were used, and 2 μl of 1 mM of ThT prepared in Tris-HCl buffer (pH 8.0) containing 0.01% sodium azide was added. ThT fluorescence was measured with an excitation wavelength of 450 nm and an emission range of 460-500 nm (slit width of 5 nm) using a Horiba Jobin Yvon (Fluomax4) instrument. Fluorescence intensity at 480 nm for all the samples was noted for ThT data analysis. Three independent sets of experiments were performed for each of the samples.

The fibril morphology of the aggregated p53 core domain was confirmed using transmission electron microscopy. After p53 aggregation in vitro, the protein solution was centrifuged at 200,000 g for 45 min at 4°C. The fibril pellet was resuspended in PBS (pH 7.4). An aliquot of 10 μl of the fibril solution (10 μM) was spotted on carbon-coated Formvar EM grids and incubated for 5 min. The grid was washed gently with autoclaved milli-Q water. Freshly prepared and filtered 1% (v/v) uranyl formate solution (5 μl) was used for staining for 5 min. The grids with samples were air-dried and imaged using a transmission electron microscope (Philips CM-200, The Netherlands) with a magnification range of 6600× to 12,000× at 200 kV. Images were recorded digitally with the aid of the Keen View Soft imaging system (Olympus, Japan). Further, for the preparation of fibril seeds, the resuspended core fibrils were sonicated in a water bath for 3 min at room temperature.

p53 core fibrils were labeled with Rhodamine dye (10 M excess) as per the manufacturer's instructions (Invitrogen, Carlsbad, CA, USA). Rhodamine-labeled core fibrils were centrifuged at 200,000 g for 45 min. Milli-Q water was used to wash the fibrils twice. For fibril seed preparation, the labeled core fibrils were resuspended in sterile PBS and were sonicated for 3 min at room temperature in a water bath.

MCF 10A cells (non-tumorigenic human breast epithelial) were procured from the cell repository at the National Centre for Cell Science, Pune, India. HFF cells (normal human foreskin fibroblasts) were obtained from the American Type Culture Collection (ATCC). Both MCF 10A and HFF cells were used for experimentation at early passages. For culturing of MCF 10A, Dulbecco's modified Eagle medium (DMEM) growth medium (Gibco, Waltham, MA, USA) supplemented with 5% horse serum (Gibco, Waltham, MA, USA), 0.5 μg/ml hydrocortisone (Himedia, Mumbai, India), 20 ng/ml hEGF (Invitrogen, Carlsbad, CA, USA), 10 μg/ml insulin (Himedia, Mumbai, India) and 1× Pen-Strep (Himedia, Mumbai, India) antimicrobial agent was used. Cells were not allowed to reach more than 80% confluency at each passage. For all experiments with MCF 10A, assay medium was used with the following composition: DMEM growth medium supplemented with 2% horse serum; 0.5 μg/ml hydrocortisone; 10 μg/ml insulin; and 1× penicillin-streptomycin antimicrobial agent. For culturing HFF cells, DMEM growth medium supplemented with 5% fetal bovine serum (Gibco, Waltham, MA, USA), 10 mM HEPES buffer (Himedia, Mumbai, India) and 20 mg/ml gentamycin (Gibco, Waltham, MA, USA) as an antimicrobial agent was used.

For core fibril treatment of MCF 10A, the assay medium was used for all the experiments. For this, 30 μM of core fibril seeds (or Rhodamine-labeled core fibril seeds) were added to 1 ml of DMEM medium supplemented with 2% horse serum, 0.5 μg/ml hydrocortisone, 10 μg/ml insulin and 1× penicillin-streptomycin antimicrobial agent. For fibril treatment of HFF, 30 μM of core fibril seeds were added to DMEM medium supplemented with 5% fetal bovine serum, 10 mM HEPES buffer and 20 mg/ml gentamycin as antimicrobial agent. Cells were incubated for 48 h in a 5% CO₂ incubator at 37°C.

Cells were treated with p53 core fibrils as discussed above, and immunofluorescence was performed to study the state of p53 using p53 antibody. For this, 4% paraformaldehyde (Himedia, Mumbai, India) was used for fixing cells on the coverslips for 20 min at 25°C. Further, the fixed cells were washed with PBS (pH 7.4), and 0.2% Triton X-100 (Sigma-Aldrich, St Louis, MO, USA) in PBS was used to permeabilize cells for 10 min. The cells were then washed with PBS and incubated with 2% bovine serum albumin (BSA, Himedia, Mumbai, India) in PBS for 1 h to block non-specific epitopes. This was followed by incubation at 4°C overnight with a 1:200 dilution of mouse monoclonal anti-human IgG p53 primary antibody DO-1 (sc-126, RRID: AB_628082, Santa Cruz Biotechnology, Dallas, TX, USA). After incubation, cells were washed three times with PBST (0.1% Tween 20 containing PBS, pH 7.4). Then the coverslips were incubated with 1:500 dilution of goat anti-mouse IgG Alexa Fluor 488-conjugated secondary antibody (A28175, RRID: AB_2536161, Invitrogen, USA) for 2 h at 25°C. Cells were rinsed three times with PBST (pH 7.4). Mounting medium containing 1% 1,4-diazabicyclo[2.2.2]octane (DABCO, Sigma-Aldrich, St Louis, MO, USA) in 90% glycerol and 10% PBS was used to mount the coverslips on slides. The coverslips were observed using a Zeiss Observer Z1 microscope fitted with a high-speed microlens-enhanced Nipkow spinning disc. Images were pseudocolored red for the p53 signal. Cells showing cytoplasmic and/or nuclear punctate were counted, and the percentage of cells was plotted in each fraction [n=3 independent experiments (Total cells counted, N>200)].

For colocalization studies with p53 and OC antibody, the paraformaldehyde-fixed cells were permeabilized for 15 min. Further, after blocking with 2.5% BSA, cells were incubated with primary antibody p53 DO-1 at a dilution of 1:200, and rabbit amyloid fibril-specific OC antibody (Abcam, Cambridge, UK) at a 1:500 dilution, at 4°C overnight. Secondary antibodies goat anti-mouse IgG FITC (against p53) and goat anti-rabbit IgG Alexa Fluor-555 (against OC) (A27039, RRID: AB_2536100, Invitrogen, Carlsbad, CA, USA) at 1:500 dilution were incubated with the cells for 2 h at 25°C. Images were acquired using a Nipkow CSU-X1 spinning disc confocal microscope (Yokogawa, Japan). Images were pseudocolored (magenta) for the OC signal.

For the detection of EMT markers, immunostaining of spheroids derived from MCF 10A cells (at T5 generation) was performed. MCF 10A spheroids were grown on coverslips (details discussed in the relevant section below), fixed and permeabilized, followed by blocking non-specific epitopes as mentioned above. MCF 10A spheroids were incubated with 1:500 dilution of rabbit monoclonal EMT antibodies, β-catenin, Slug, Vimentin, N-cadherin (9782, Cell Signaling Technology, Danvers, MA, USA), overnight at 4°C in a humidified chamber. Coverslips were washed three times with PBST (pH 7.4) and were further incubated with goat anti-rabbit IgG Alexa Fluor 555-conjugated secondary antibody (1:1000) for 2 h at room temperature in a humidified chamber. MCF 10A spheroids were rinsed three times with PBST (pH 7.4) and stained with 1 μg/ml DAPI for 5 min, and were washed with PBS twice. The coverslips were mounted and observed under a Leica DMi8 fluorescence microscope, and images were analyzed using ImageJ software.

MCF 10A and HFF cells were grown and treated with core fibrils for 48 h. Immunoprecipitation in non-denaturing conditions was carried out as per the manufacturer's protocol (Abcam). Briefly, cells were trypsinized and lysed under cold conditions using the lysis buffer containing 20 mM Tris-HCl (pH 8.0), 137 mM NaCl, 1% NP-40, 2 mM EDTA and protease inhibitor cocktail (Roche, Basel, Switzerland) under constant agitation for 20 min at 4°C. The mixture was centrifuged at 13,000 g for 20 min at 4°C. The supernatant was incubated with anti-p53 antibody (DO-1) under rotation at 4°C for 8-10 h. Further, 120 μl of Sepharose G beads were added to the solution and incubated at 4°C under constant agitation for 4 h. This was centrifuged at 900 g for 30 s to collect the beads. Elution was performed using 100 μl of 0.2 M glycine buffer (pH 2.6) with a 10 min incubation with constant mixing and centrifugation at 900 g for 2 min at 4°C. The eluted fractions were pooled and neutralized by adding an equal volume of Tris HCl (pH 8.0). The eluted protein was quantified using Bradford assay to ensure equal loading in the dot blot. For the immunoprecipitation of p53 from cell generations, cell pellets from sequential cell generations were collected for both fibril-treated and untreated cells, and immunoprecipitation of p53 was carried out using the same method.

The immunoprecipitated p53 protein from MCF 10A and HFF cells was used for dot blot assay. For the detection of p53 amyloid aggregates in cells, 5 μg protein was used. For the passaging experiments, 3 μg of protein was used for the blots to prevent over-saturation of signals. For dot blot, a nitrocellulose membrane was used to spot the samples and was allowed to air dry. The membranes were then treated with a blocking solution of 5% non-fat skimmed milk powder (Himedia, Mumbai, India) in TBST (TBS with 0.1% Tween 20) for 1 h at 25°C. The membranes were washed with TBS and incubated with primary antibody (anti-p53, 1:200 dilution; or OC, 1:500 dilution) overnight at 4°C on a rocker. Further, the blots were washed with TBST three times (each wash for 10 min). This was followed by a 1 h incubation with anti-mouse IgG (for p53) (401253, RRID: AB_437779) and anti-rabbit IgG (for OC) (401353, RRID: AB_437794) horseradish peroxidase (HRP)-conjugated secondary antibody (Sigma-Aldrich, St Louis, MO, USA) with 1:5000 dilution at room temperature. Three washes of TBST, each for 10 min, were performed to remove non-specific binding. Immobilon Western chemiluminescent HRP substrate was used for the detection of chemiluminescence (Merck Millipore, USA).

Both MCF 10A and HFF cells were seeded (10⁴ cells per well) and incubated for 48 h in 5% CO₂ incubator at 37°C. These cells were then treated with fibrils as mentioned above and then further ActD (Sigma-Aldrich, St Louis, MO, USA) treatment. ActD (2 μg/ml) was added for another 12 h to induce apoptotic stress. The cells were trypsinized and stained using an Annexin-V PI apoptosis kit (556547, BD Biosciences, USA) following the manufacturer's protocol. For quantifying the apoptotic population, FACS analysis was performed using a BD FACS Aria SORP instrument (BD Biosciences, USA) (n=3 independent experiments).

A soft agar assay was performed to study the transformation potential of core fibril-treated MCF 10A and HFF cells, with corresponding untreated cells as controls. For the soft agar assay, the 2× cell culture medium was prepared by dissolving 1 g of DMEM powder medium and 0.2 g of sodium bicarbonate (Himedia, Mumbai, India) in 50 ml autoclaved de-ionized water, and sterilized using a 0.22 μm filter. Agarose (1.8%) was prepared by adding 1 g of noble agar to 100 ml of distilled water. Further, agarose was diluted to 0.3% and 0.6% using pre-warmed 2× culture DMEM medium for the bottom and top layers, respectively. The 12-well plates were covered with 0.3% agarose mixture (bottom layer), which was allowed to solidify at room temperature for 30 min. For the upper layer, fibril-treated and untreated cells were harvested by trypsinization, and were then centrifuged and resuspended in complete medium. Agarose solution (0.6%) was melted in a microwave and mixed with cell suspension in a 1:1 ratio (at 42°C), such that there were 10⁶ cells per well. This mixture was then layered on top of the bottom layer. The cell/agar mixture was allowed to solidify at room temperature before incubating at 37°C. The plates were observed daily for the formation of colonies. The colonies were imaged under a bright-field microscope (Leica DMi1, Wetzlar, Germany) and counted. Three independent experiments were performed.

MCF 10A and HFF cells (fibril treated for 48 h and untreated) were seeded at a cell density of 10⁴ cells per well on a glass coverslip. Cells were fixed and permeabilized, and non-specific epitopes were blocked as mentioned above. Cells were then incubated overnight at 4°C with 1:1000 dilution of anti-Ki67 antibody-Alexa Fluor 488 tagged (ab92742, EPR360, Abcam). Then cells were washed three times with PBST (0.1% Tween 20 containing PBS) (pH 7.4). The coverslips were mounted with mounting medium containing 1% DABCO. The slides were observed using a Zeiss Observer Z1 fitted with a high-speed microlens-enhanced Nipkow spinning disc, and the cells positive for the Ki67 marker were counted for each set of samples (N>200; n=3 independent experiments).

To examine the effect of p53 amyloid formation on cell migration capability, a wound-healing assay was performed (Rodriguez et al., 2005). To do that, MCF 10A or HFF cells were cultured on a 12-well plate and were treated with core fibril seeds for 48 h, as described previously. Untreated cells were used as controls. Cells were gently washed with PBS (pH 7.4) and a scratch was made with a plastic tip (sterile). The layer of peeled-off cells was removed with PBS washes and medium with mitomycin C (0.5 μg/ml) (Sigma-Aldrich, St Louis, MO, USA) was added to the cells. Mitomycin C acts as a cell proliferation inhibitor to calculate the migration rate accurately. Further, the cells were incubated for 48 h to allow wound healing. Images at zero and final time-points were acquired using a bright-field microscope (Leica DMi1). The wound width covered was calculated using ImageJ software. Cell migration was estimated using the formula: rate of cell migration (μm/h)=[(0) time wound width-final wound width)/time taken (in h)] (n=3 independent experiments).

MCF 10A and HFF cells were seeded in a six-well plate (cell density of 5×10⁴ cells per well). The cells were then treated with 30 μM p53 core fibril seeds for 48 h, as described previously. Untreated MCF 10A and HFF cells were kept as controls. The cells were allowed to reach confluency (∼80%) before passaging them further. To do that, ∼1% cells from the previous generation mixed with fresh medium were added to a fresh well to get the next generation. Untreated cells were passaged similarly. For estimating the cell viability at each generation, Annexin-V PI staining coupled with FACS analysis was performed. For each sample, 10,000 cells were analyzed. The passaging of cells to obtain sequential generations was continued until the cells lost their adherence potential (along with low viability). During this passaging, at each generation, cells were also seeded on glass coverslips for immunostaining with p53 antibody as mentioned before to check the p53 status (n=2 independent experiments). The cell pellet was collected at each generation for immunoprecipitation of p53 as discussed.

MCF 10A cells were cultured in 24-well plates and treated with p53 core fibrils for 48 h and passaged, as discussed. The corresponding untreated cells were used as a control. Senescence was evaluated at T5 generation. For this, the cell culture medium from 24-well plates was aspirated and the cells were washed with PBS (500 μl per well) twice. After washing, 250 μl of 4% paraformaldehyde solution was used to fix the cells for 5 min at room temperature, followed by washing the cells two times for 5 min each with PBS. Then, to evaluate the senescent cell population, senescence-associated β-galactosidase (SA-beta-gal) staining was performed according to the established protocol (Gary and Kindell, 2005). Briefly, the SA-β-gal staining solution (pH 6.0) was prepared by mixing 0.1% X-gal (Thermo Fisher Scientific, USA), 5 mM potassium ferrocyanide, 5 mM potassium ferricyanide, 150 mM sodium chloride and 2 mM magnesium chloride in 40 mM citric acid/sodium phosphate solution. SA-β-gal (250 μl) staining solution was added to each of the wells containing fixed cells, and the plate was incubated in the dark in a 37°C incubator. After 72 h, the cells were washed with distilled water. The senescent cells contain galactosidase enzyme, which cleaves the chromogenic substrate X-Gal present in staining solution to develop a blue precipitate, which stains the senescent cells blue. Each well was observed under an inverted bright-field microscope Leica DMi1 (Leica, Wetzlar, Germany) using a 10× objective, and blue-stained cells were counted. The SA-β-gal⁺ cells are represented as a percentage of the total cell number (total cells, N>200) with three independent experiments (n=3).

A luciferase assay was used for evaluating the p53 transcriptional activity in MCF 10A core fibril-treated and untreated cells. For this assay, PG-13 vector containing 13 repeats of p53 binding sites cloned upstream of firefly luciferase was used (Addgene, 16442), along with a plasmid containing Renilla luciferase (Addgene, 74444) as a control to normalize for translational variations across the samples. Cells were seeded (50,000 cells/well) in a 24-well plate and transfection was performed at 70% confluency. To do that, 1 µg of plasmids (500 ng of PG-13 plasmid and 500 ng of Renilla plasmid) were transfected per well using Lipofectamine 3000 (Invitrogen, Carlsbad, CA, USA). Cells were harvested after 48 h of transfection and luciferase levels were quantified using a dual luciferase assay according to the manufacturer's protocol (E1910, Promega, Madison, WI, USA) in a Tecan microplate reader (Tecan, Switzerland). For rescue of p53 aggregation experiments, rescue peptide (10 µM) was added 16 h before harvesting the cells, and the luciferase levels were measured as mentioned above. The ratio of firefly and Renilla (relative luciferase activity) was used to indirectly evaluate the p53 transcriptional activity, and the graphs were analyzed and plotted in GraphPad Prism 6.0 (n=2 independent experiments).

An MTT assay was performed to evaluate the cytotoxicity of the MCF 10A cells (fibril treated and untreated at the T5 generation) in the presence and absence of various drugs. The drugs cisplatin, doxorubicin and paclitaxel (Sigma-Aldrich, St Louis, MO, USA) were dissolved in DMSO to make stock concentrations of 1 mM and further dilutions of each drug were made in sterile PBS (pH 7.4). MCF 10A cells at T5 (untreated and core fibril treated) were independently seeded in a 96-well plate at a density of 10⁴ cells per well (the same number of viable cells were kept for untreated and core-treated samples while setting up the assays) and incubated overnight. DMEM assay medium (100 μl) containing the drugs of increasing concentration (0.5, 1, 2.5, 5, 10 and 20 µM for cisplatin as well as paclitaxel; 0.05, 0.10, 0.50, 1, 1.5, 2, 5, 10 and 20 µM for doxorubicin) was added to the wells. The samples were performed in triplicates. PBS buffer with DMSO was added to the medium with cells as a buffer control (negative control). Triton X-100 added to the medium with cells was used as a positive control. The medium with cells (100% viability) and cell-free medium alone (for background absorbance) were also used. The plate containing cells and drugs were incubated for 24 h. After incubation, 10 μl of MTT dye (Sigma-Aldrich, St Louis, MO, USA) (5 mg/ml) in PBS was added to the wells and incubated for 4 h. Then, overnight incubation was peformed with 100 μl of N,N-dimethylformamide-SDS solution. Further, the absorbance at 560 nm and background scattering at 690 nm was measured using a SpectraMax M2 (Molecular Devices, USA) plate reader. The background scattering values at 690 nm were subtracted from the absorbance values at 560 nm to calculate the viability of cells both in the presence and absence of various compounds at each drug concentration. The viability of untreated and fibril-treated cells was normalized (100% viability at 0 µM drug concentration) during calculations. EC₅₀ of each drug was calculated using GraphPad Prism 6 (n=3 independent experiments).

MCF 10A (core fibril-treated and untreated cells) passaged cells at T5 generation were evaluated for their ability to form spheroids in Matrigel (Corning, USA). Matrigel (10 µl) was used to layer the surface of clean glass coverslips placed in a 24 well plate and allowed to solidify. Fibril-treated and untreated MCF 10A cells (10³ cells per well) were mixed individually with 50 µl Matrigel and layered on the top of the Matrigel bed on the coverslip. The Matrigel was allowed to solidify before the addition of fresh medium. The cells were incubated for 10 days and imaged periodically along with the untreated control to detect spheroid formation. The diameter of spheroids (D) was measured from the images of larger spheroids. Spheroid volume was calculated using the formula: volume of spheroid=4/3×3.14(r)³, where r is the radius and r=D/2 (n=2 independent experiments). Following spheroid formation, the spheroids were stained with a mixture of two dyes: 1 μM Calcein-AM (Thermo Fisher Scientific, USA) and 1 μM of ethidium homodimer-1 (Thermo Fisher Scientific, USA). Dyes were prepared freshly in sterile PBS and added directly to the medium without aspiration and incubated for 2 h. The dye solution was washed out with PBS slowly to avoid spheroid loss or displacement. Spheroids were imaged using a DMi8 microscope (Leica, Wetzlar, Germany) with a color HD camera

Female SCID mice (6-8 weeks old) weighing ∼20-25 g each were obtained from the Advanced Centre for Treatment, Research and Education in Cancer (ACTREC), Mumbai, India and acclimatized for ∼1 week. All animals were kept in standard conditions with a 12 h dark/light cycle in micro-isolator cages with autoclaved bedding, with 50% humidity, a temperature of 22±2°C and free access to water and food (standard pellet feed). All animal experiments were performed in accordance with the relevant guidelines and regulations. The study was approved by the Institutional Animal Ethics Committee, Agharkar Research Institute, Pune (approval certificate: ARI/IAEC/2019/03).

MCF 10A cells were treated with core fibrils and passaged until the T5 generation as discussed. p53 aggregates in the passaged cells at T5 were confirmed by immunofluorescence. In parallel, untreated MCF 10A cells (non-tumorigenic) were passaged until T5 and did not show p53 aggregation. Both core fibril-treated and untreated cells at T5 generation were used to analyze the tumorigenicity by injection in immunocompromised SCID mice. The mice were divided into groups of four for each experimental set. The groups were as follows: MCF 10A untreated cells, MCF 10A core-treated cells and MCF 7 cells. The viability of cells was ensured before injection with Trypan Blue stain. Each mouse received an injection subcutaneously in the mammary fat pad with 2×10⁷ core-transformed cells in sterile PBS (150 μl each injection). Untreated passaged cells (1×10⁷) were injected as a control. MCF 7 cells (1×10⁷ cells) were used as a positive control for tumor induction. Estradiol (Estrabet tablets, 2 mg, Abbott India) treatment was given orally to mice via their drinking water (1000 nM, i.e. 0.054 µg/25 g mouse/day). All animals that received injections of MCF 10A core-treated cells and injections of MCF 7 developed tumors. The mice were weighed once a week and the area of injection was observed. After tumor induction, the tumor volume was periodically measured using width (a) and length (b) measurements (a²b/2, where a<b) to assess the tumor growth. PET and CT scan was performed to study tumor morphology and metabolism. At the end of the experiments, the tumors were excised and embedded in paraffin for sectioning for characterization. The experiment was repeated twice with mice in groups of four each time (combined n=8).

The experimental mice injected with MCF 10A (core fibril treated and untreated) were kept for overnight fasting with access to drinking water before ¹⁸F-FDG (2-deoxy-2-[fluorine-18] fluoro-D-glucose) injection. Animals were anesthetized by isoflurane and 25-30 MBq ¹⁸F-FDG (∼800 microcuries) was injected intravenously through the tail vein. Whole-body PET scan integrated with CT scan was performed after ¹⁸F-FDG administration (within 1 h of ¹⁸F-FDG administration). Both CT and PET scans were performed with the tri-modality gamma imaging system (Triumph, Gamma Medica Ideas, Northridge, CA, USA). Reconstructed PET/CT images were analyzed using PMOD software, version 3.2 (PMOD Technologies, Zurich, Switzerland). Regions of interest were drawn for the tumor-bearing animals in the micro-PET image using the aligned micro-CT image as a guide. The standard uptake values (SUVs) of ¹⁸F-FDG in the tumors were calculated, taking into consideration the body weight and injected a dose of ¹⁸F-FDG, using the images acquired. The SUV was calculated using the following formula: SUV=(activity in region of interest)/(injected dose/body weight).

To compare the histology of tumor tissue (MCF 10A core fibril treated and MCF 7) and normal tissue (only MCF 10A and rescue peptide-treated core-transformed MCF 10A), H&E staining was performed. Tissues were formaldehyde-fixed and paraffin embedded. A microtome was used to prepare thin sections of the paraffin-embedded tissues. Deparaffinization of tissue sections was performed using decreasing concentrations of the clearing agent xylene (100% xylene, followed by 50% xylene in 50% ethanol). The tissue was further rehydrated in decreasing concentrations of ethanol (100%, 95%, 70% and 50%) and ultimately with distilled water. The sections were then washed thoroughly with distilled water for 10 min and then stained with 0.5% Hematoxylin prepared in distilled water for 2 min. The unbound stain was washed using 70% and 90% ethanol for 10 min each. Next, 0.8% eosin in 95% ethanol was used to stain the sections for 1 min and kept in xylene for 1 h. The sections were then sealed with a coverslip using DPX mounting solution (a mixture of polystyrene, tricresyl phosphate and xylene) and observed using a DMi8 microscope (Leica, Wetzlar, Germany) with a color HD camera.

Paraffin-embedded tissues, tumor tissue (MCF 10A core fibril treated and MCF 7) or normal tissue (only MCF 10A and rescue peptide-treated core-transformed MCF 10A) were used for immunohistochemistry. Paraffin-embedded tissues were sectioned using a microtome. Further, the slides with tissue sections were deparaffinized using a xylene gradient (100% xylene, followed by 50% xylene in 50% ethanol) and eventually rehydrated in decreasing concentrations of graded ethanol (100%, 95%, 70% and 50%), and finally with distilled water. Sections were treated with 0.05% trypsin at 37°C for 2 min for enzymatic antigen retrieval. Further, the tissue sections were washed with TBST at pH 7.4 and permeabilized with 0.2% Triton X-100 in TBST for 10 min. TBST containing 2% BSA was used for the blocking of non-specific antigenic sites for 1 h. First, the sections were incubated with mouse p53 antibody (DO-1) (1:200 dilution) and OC (1:500 dilution) overnight at 4°C in a humidified chamber. After incubation, TBST (pH 7.4) was used to wash the sections three times. For secondary staining, the slides were then incubated with goat anti-mouse IgG Alexa-Fluor 488-conjugated secondary antibody (1:1000) and goat anti-rabbit IgG Alexa-Fluor 555-conjugated secondary antibody (1:1000) for 2 h at room temperature. Sections were then rinsed three times with TBST (pH 7.4) and stained with 1 μg/ml DAPI for 3 min, followed by washing with TBS, twice. The sections were mounted with 1% DABCO in 90% glycerol and 10% PBS. Control tissue sections were stained using the procedure mentioned above. Sections were observed using a Leica Mi8 fluorescence microscope, and images were analyzed using ImageJ.

A cell-penetrating peptide, called rescue peptide, previously established by Soragni et al. (2016), was used to disaggregate p53 amyloids. The rescue peptide was custom synthesized by USV Limited (Mumbai, India) using the solid-state peptide synthesis method. Lyophilized rescue peptide was dissolved in DMSO to make a stock concentration of 10 mM. MCF 10A cells were treated with core fibrils in a 12-well plate, as discussed, to induce p53 aggregation. These cells were passaged until the T5 generation. In parallel, untreated cells were passaged as a control. A 10 μM working concentration of the rescue peptide was used to treat MCF 10A core fibril-transformed cells for 12 h. Fibril-transformed cells (at T5) without rescue peptide treatment and untreated MCF 10A cells (T5) were kept as control. The rescue peptide-treated cells were used for soft agar (n=3 independent experiments), apoptosis (n=3 independent experiments) and p53 luciferase reporter assays for understanding the effect of p53 disaggregation (n=2 independent experiments), as described previously. To check the tumorigenic potential of rescue peptide-treated MCF 10A core fibril-transformed cells (at T5), the mice were divided into groups of three for each experimental set. The viability of cells was evaluated before injection with Trypan Blue stain. The mice were injected subcutaneously in the mammary fat pad with 2×10⁷ cells [fibril-transformed cells (T5) with/without rescue peptide treatment and untreated MCF 10A cells (T5)] in sterile PBS (150 μl each injection). The mice were maintained and tumor induction was monitored as described previously. PET and CT scans were carried out similarly to the previous experiment. Tissue sectioning along with H&E staining for tumor and normal tissue was performed as described previously. The animal experiment was repeated twice.

The statistical significance [*P<0.05, **P<0.01, ***P<0.001, non-significant (NS P>0.05)] was calculated by one-way ANOVA followed by Bonferroni multiple comparison post-hoc test with 95% confidence interval. KaleidaGraph version 4.1 software was used for calculating the statistical significance. No outliers were excluded from the analysis. No assessment was made for outliers and the normality of data.

We thank the central facilities of the Indian Institute of Technology Bombay for assistance with transmission electron microscopy, FACS and confocal microscopy. We also thank Prof. Charles Glabe (University of California Irvine, CA, USA) for the kind gift of the OC antibody. We are grateful to Professor Roop Mallik for reading the manuscript and making suggestions.