**A Small Number of HER2 Redirected CAR T Cells Significantly Improves Immune Response of Adoptively Transferred Mouse Lymphocytes against Human Breast Cancer Xenografts**

#### **Gábor Tóth 1, János Szöll ˝osi 1,2, Hinrich Abken 3, György Vereb 1,2,4,\* and Árpád Szö ˝or 1,\***


Received: 30 November 2019; Accepted: 31 January 2020; Published: 4 February 2020

**Abstract:** HER2 positive JIMT-1 breast tumors are resistant to trastuzumab treatment in vitro and develop resistance to trastuzumab in vivo in SCID mice. We explored whether these resistant tumors could still be eliminated by T cells redirected by a second-generation chimeric antigen receptor (CAR) containing a CD28 costimulatory domain and targeting HER2 with a trastuzumab-derived scFv. In vitro, T cells engineered with this HER2 specific CAR recognized HER2 positive target cells as judged by cytokine production and cytolytic activity. In vivo, the administration of trastuzumab twice weekly had no effect on the growth of JIMT-1 xenografts in SCID mice. At the same time, a single dose of 2.5 million T cells from congenic mice exhibited a moderate xenoimmune response and even stable disease in some cases. In contrast, when the same dose contained 7% (175,000) CAR T cells, complete remission was achieved in 57 days. Even a reduced dose of 250,000 T cells, including only 17,500 CAR T cells, yielded complete remission, although it needed nearly twice the time. We conclude that even a small number of CAR T lymphocytes can evoke a robust anti-tumor response against an antibody resistant xenograft by focusing the activity of xenogenic T cells. This observation may have significance for optimizing the dose of CAR T cells in the therapy of solid tumors.

**Keywords:** breast cancer; trastuzumab; chimeric antigen receptor; immunotherapy; cell therapy

#### **1. Introduction**

Human epidermal growth factor receptor 2 (HER2) is overexpressed in 20–25% of breast cancer tumors [1]. HER2 expression is associated with an aggressive disease with a high recurrence rate and increased mortality [2]. Specific monoclonal antibody therapy has revolutionized the treatment of HER2 positive breast cancer since the FDA (U.S. Food and Drug Administration) approval of trastuzumab (Herceptin®) in 1998 [3]. The addition of trastuzumab to chemotherapy results in a lower rate of death after one year (22 percent vs. 33 percent, *P* = 0.008), longer survival (median survival, 25.1 vs. 20.3 months; *P* = 0.046), and a 20 percent reduction in the risk of death [4]. Despite the success, resistance to therapeutic antibodies is a clinical reality that affects the outcome of 60–80% of HER2+ breast cancer patients [5]. One of the underlying mechanisms is epitope masking by components of the tumor microenvironment (TME) such as the MUC4 (mucin 4) or the CD44/Hyaluronan complex [6–11].

The JIMT-1 cell line was established from the pleural metastasis of a breast cancer patient and has recapitulated the trastuzumab resistance of the original tumor in vitro and also in vivo if treatment of JIMT-1 xenografts SCID mice was initiated at a few hundred mm<sup>3</sup> tumor volumes [12–14]. Our recent data indicate that simultaneous targeting of two epitopes on the HER2 molecule with clinical doses of trastuzumab and pertuzumab additionally improves the efficacy of antibody-dependent cellular cytotoxicity and thereby also the anti-tumor response; however, eventually all JIMT-1 xenografts become resistant to antibody treatment at a certain tumor size [15].

In such cases of antibody resistance, Chimeric Antigen Receptor (CAR) engineered T cells [16] represent an appealing option for improving the outcome for patients with advanced breast cancer [17–20]. Several tumor-associated membrane proteins are targeted in clinical trials by CAR T cells, including HER2 (NCT02547961, NCT02713984), CEA (NCT02349724) and mesothelin (NCT02792114). While no results have been disclosed of current HER2 targeting trials, the first reported clinical use of HER2 specific CAR T cells resulted in a serious adverse event following CAR T cell infusion [21]. In this trial, a HER2 positive colon cancer patient was treated with a large number (1010) of CD28-41BB costimulatory (3rd-generation) CAR T cells, which derived their antigen specificity from trastuzumab. The patient developed respiratory distress, followed by multiple cardiac arrests over the course of 5 days, leading to death. The death of this patient may have occurred due to the result of HER2 recognition of highly active and numerous anti-HER2 CAR T cells in the normal lung tissue that caused pulmonary toxicity and edema followed by a cytokine release storm causing multiorgan failure. The immune-mediated recognition of tumor antigens in normal tissues is referred to as "on-target, off-tumor" toxicity. It is thus clear from both preclinical experiments and clinical trials that while CAR T cell-based immune therapy has great potential to improve the outcome for patients with HER2 positive tumors, it still needs plentiful optimization.

Here, we report the generation of mouse T cells that are genetically modified to express a chimeric antigen receptor that consists of a HER2 specific single-chain variable fragment (scFv) derived from trastuzumab, a CD28 costimulatory endodomain, and a CD3z intracellular signaling domain. We demonstrate that these T cells recognize and kill HER2+ tumor cells in vitro and significantly improve the xenogenic immune response against human breast cancer even at very low numbers (17,500), resulting in complete tumor regression, and significant survival advantage.

#### **2. Results**

#### *2.1. Generation of Murine HER2 Specific CAR T Cells*

To genetically modify mouse T cells (Figure 1), first, we generated VSVG-pseudotyped retroviral particles encoding HER2 specific chimeric antigen receptors (Figure 1A). T cells were isolated from the freshly dissected spleen of congenic Balb/c mice and activated by anti-mouse CD3e and anti-mouse CD28 antibodies. After 24 h, the medium was supplemented with mouse interleukin 2. Finally, activated mouse T cells were retrovirally transduced on RetroNectin-coated plates (Figure 1B).

The CAR contains an scFv obtained from trastuzumab, an IgG1 CH2-CH3 extracellular stalk, a CD28 costimulatory endodomain and a CD3z effector domain (Figure 2A). Using trastuzumab as a recognition domain allowed us to compare the impact of CAR T cells as living drugs with the impact of antibodies. The mean transduction efficiency was 8.7% (range: 5.58–11.84%; *n* = 8) as judged by flow cytometric analysis of the HER2 specific scFv (Figure 2B,C). We confirmed that CARs are stably expressed and re-confirmed CAR expression on day 10.

**Figure 1.** Genetic modification of mouse T cells with chimeric antigen receptors: (**A**) Scheme of retrovirus production. (**B**) Scheme of mouse T cell separation and activation.

**Figure 2.** Generation of HER2 specific mouse CAR T cells: (**A**) Schematic diagram illustrating the modular composition of the retroviral vector encoding HER2 specific CAR. (**B**,**C**) Representative flow cytometry dot-plots and summary data (HER2 CAR mouse T cells (*n* = 8) and non-transduced (NT) mouse T cells (*n* = 8)).

#### *2.2. HER2 Specific CARs Redirect Mouse T Cells to HER2 Positive Trastuzumab Resistant Tumor Cells*

To demonstrate that the HER2 specific CAR redirects mouse T cells to HER2 positive target cells, we co-cultured HER2 CAR T cells with JIMT-1 cells in various effector to target ratios (from 2.5:1 to 0.01:1). HER2 specific CAR T cells recognized the HER2 positive tumor cells indicated by a significant increase in IFNg secretion (*p* < 0.001). Unmodified (NT) T cells did not induce mIFNg release (Figure 3A). While HER2 specific CAR induced T cells killed JIMT-1 tumor cells, no killing was observed in co-cultures with unmodified T cells (*p* < 0.001; Figure 3B). Taken together, the HER2 specific CAR activates mouse T cells in an antigen-dependent manner and induces antigen-dependent tumor cell killing.

**Figure 3.** In vitro anti-tumor function of HER2 specific mouse CAR T cells: (**A**) HER2 CAR or non-transduced (NT) mouse T cells were co-cultured with HER2+ JIMT-1 cells at various (2.5:1–0.01:1) T cell to tumor cell ratios. After 48 h, IFNγ release was determined by ELISA (*n* = 2, assay performed in duplicates; HER2 CAR versus non-transduced (NT) T cells: \* *p* < 0.05, \*\*\* *p* < 0.001). (**B**) XTT-based cytotoxicity assay using HER2 CAR T cells or non-modified mouse T cells and HER2 positive JIMT-1 cells as target at various (2.5:1–0.04:1) T cell to tumor cell ratios (*n* = 2; assay was performed in duplicates; HER2 CAR versus NT T cells: \*\*\* *p* < 0.001).

#### *2.3. HER2 Specific CAR T Cells Have Antitumor Activity In Vivo against HER2*+ *Trastuzumab-Resistant Tumor Xenografts*

To compare the anti-tumor function of antibody treatment and HER2-redirected CAR T cells, we established subcutaneous JIMT-1 xenografts (3 <sup>×</sup> 106 cells) in SCID mice (day <sup>−</sup>35, Figure 4). Mice were injected with 100 μg trastuzumab intraperitoneally twice weekly from day 0 (35 days after tumor cell injection), when average tumor size reached 800 mm3. Control animals were injected with PBS (Figure 4) and did not show delayed tumor growth and consequently their overall survival was not improved (*p* = 0.79, Figure 5). These data are in line with our previous observations [13].

**Figure 4.** Outline of in vivo animal treatment schedule: Mice were s.c. injected with 3 <sup>×</sup> 10<sup>6</sup> JIMT-1 cells. 35 days later (on day 0), mice received 100 μL PBS buffer twice weekly (untreated, *n* = 5), 100 μg trastuzumab in 100 <sup>μ</sup>L PBS buffer twice weekly (trastuzumab, *<sup>n</sup>* <sup>=</sup> 5), or an i.v. dose of 2.5 <sup>×</sup> 105 HER2 CAR mouse T cells (low-dose HER2 CAR, *<sup>n</sup>* <sup>=</sup> 5), or an i.v. dose of 2.5 <sup>×</sup> <sup>10</sup><sup>6</sup> HER2 CAR mouse T cells (HER2 CAR, *n* = 5). Tumor growth was followed by caliper and was derived as the product of the length, width and height.

**Figure 5.** Antitumor activity of HER2-CAR mouse T cells in a xenograft model: (**A**) Quantitative measurement of tumor volumes (volume = mm3; low dose HER2-CAR versus HER2-CAR groups: \*\*\* *p* < 0.001; NT T cell versus HER2-CAR groups \*\*\* *p* < 0.001). (**B**) Kaplan-Meier survival curve (NT T cell versus both HER2 CAR groups \*\*\* *p* < 0.001). (**C**) Representative images of animals. (**D**) Representative images for the detection of HER2-CAR mouse T cells in JIMT-1 xenografts (field of view: 92 μm × 92 μm). HER2-CAR mouse T cells were visualized by AlexaFluor647 conjugated anti-mouse CD3e and HER2-GFP co-staining.

To evaluate the in vivo efficacy of HER2-CAR T cells against these trastuzumab resistant xenografts, 35 days after JIMT-1 inoculation mice were injected iv. with a single dose of 2.5 <sup>×</sup> 106 (HER2-CAR T cell group), or 2.5 <sup>×</sup> 105 (Low Dose HER2-CAR T cell group) congenic mouse T cells, 7% of them expressing the HER2-CAR. Control mice were treated with 2.5 <sup>×</sup> <sup>10</sup><sup>6</sup> unmodified T cells (NT T cell group) (Figure 4). In this group, the high number of mouse T cells, in some cases, delayed the progression of the human xenografts resulting in better overall survival in comparison to the untreated group (Figure 5A–C). In contrast, the same dose of mouse T cells, when transduced with the CAR at 7% efficiency, completely eradicated the tumors in 57 days and resulted in long-term tumor-free survival. Moreover, complete tumor regression was also observed, by day 105, in the low dose HER2-CAR T cell group in which animals received only 250,000 T cells, among them 17,500 CAR T cells (Figure 5A–C). There is a cell dose dependence of the rate of tumor regression (*p* < 0.001). Despite the difference in time to complete regression, all CAR T cell treated mice remained tumor free until the termination of the experiment (day 150). To assess on-target off-tumor toxicity, formaldehyde-fixed paraffin-embedded tissue section weremade from the heart and lungs of each sacrificed animal. HE-stained sections were characterized based on morphology by an expert histopathologist and showed no signs of mononuclear infiltration (Figure S1A,B). Visual inspection upon dissection also did not show signs of inflammation.

To demonstrate that HER2 specific CAR T cells penetrated the JIMT-1 xenografts, tumor samples from week 3 after CAR T cell injection were immunostained and analyzed by confocal microscopy. We found T cells positive for the CAR and mouse-CD3e confirming the presence of CAR T cells in tumor xenografts (Figure 5D). Taken together, we conclude that HER2 specific CAR T cells have potent in vivo antitumor activity and penetrate HER2 positive xenografts, which are not eliminated by trastuzumab treatment.

#### **3. Discussion**

In this study, we described the generation of HER2 specific CAR-modified mouse T cells that obtained their antigen specificity from trastuzumab, a HER2 specific monoclonal antibody applied in clinical practice. We demonstrated that these cells specifically recognize trastuzumab resistant HER2 positive JIMT-1 target cells in vitro. Moreover, in vivo a low dose of HER2 CAR T cell expressed potent anti-tumor activity in a trastuzumab-resistant mouse xenograft model.

Although HER2 specific CAR T cells are effective against breast cancer cells [19,22], the tumor cells preferentially used as targets were trastuzumab sensitive cell lines (SKBR3 or BT474) leaving the question unanswered whether trastuzumab-resistant xenografts can be successfully treated with trastuzumab-derived CAR T cells.

Incorporation of the trastuzumab scFv into the CAR backbone allowed us to compare the efficacy in tumor eradication by cytolytic CAR T cells versus antibody mediated cytotoxicity.

SCID mice exhibit natural killer cell (NK) activity [23], through which trastuzumab treatment induces antibody dependent cellular cytotocixity against therapy sensitive xenografts (MCF7; BT-474) [24], which makes a direct comparison possible. In vivo, we confirmed that trastuzumab treatment has no potential to delay or revert the growth of established JIMT-1 tumors. which is in line with previous observations [11–14].

Although unmodified mouse T cells in co-cultures with JIMT-1 cells did not release cytokine and did not exhibit cytotoxicity, we wanted to confirm that xenogenic immune response does not reject human tumor xenografts in vivo [23]. A single injection of 2.5 <sup>×</sup> 106 unmodified mouse T cells on day 35 following JIMT-1 implantation could not eradicate the human tumor xenografts; however, in some cases, it caused tumor regression and resulted in stable disease.

In contrast to the control and trastuzumab-treated groups, 2.5 million mouse T cells with a proportion of 7% CAR-transduced cells (total ~175,000) caused a complete remission in 57. Even a tenth of this dose, 250,000 T cells including ~17,500 CAR T cells, was fully curative, although only in 105 days. The number of CAR T cells yielding complete remission in the latter case is only 0.2–0.3% of the usual 5 to 10 million CAR T cells used in successful mouse CAR-T therapy models [25]. Thus, it is likely that this small fraction of specifically redirected T lymphocytes successfully penetrates the tumor mass and evokes, in addition to direct tumor killing, a focusing effect that concentrates the xenogenic response of non-transduced mouse lymphocytes against the human tumor. By extrapolation, it is possible that in the case of HER2 positive solid tumors, a reduced number of CAR T cells could still maintain therapeutic efficacy through actively penetrating the tumor and enhancing the activity of tumor infiltrating lymphocytes (TILS, [26]). At the same time, while "on target off tumor" toxicity (mainly in the cardiopulmonary system), could be avoided owing to the lower expression of the HER2 target and the lack of TILs in healthy tissues.

Overall, we conclude that even a small number of CAR T lymphocytes can evoke a robust anti-tumor response against an antibody resistant xenograft by focusing the activity of xenogenic T cells. This observation may have significance for optimizing the dose of CAR T cells in the therapy of solid tumors.

#### **4. Materials and Methods**

All materials were from Sigma-Aldrich (St. Louis, MO, USA) unless otherwise indicated.

#### *4.1. Cells and Culture Conditions*

HEK 293T packaging cells were purchased from the American Type Culture Collection (ATCC, Manassas, VA, USA). Cells were cultured in Dulbecco's Modified Eagle Medium (DMEM) supplemented with 2 mmol/L Glutamax and 10% Fetal Calf Serum (FCS) and antibiotics. The JIMT-1 human breast cancer cell line was established in the laboratory of Cancer Biology, University of Tampere, Finland [12]. These cells were cultured in 1:1 ratio of Ham's F-12 and DMEM supplemented with 20% FCS, 300 U/L

insulin, 2 mmol/L GlutaMAX and antibiotics. Primary mouse splenocytes, T cells and CAR T cells were cultured in RPMI 1640 supplemented with 2 mmol/L GlutaMAX, 10% FCS and antibiotics. All of the above-listed cells and cell lines were maintained in a humidified atmosphere containing 5% CO2 at 37 ◦C and were routinely checked for the absence of mycoplasma contamination.

#### *4.2. Retrovirus Production and Transduction of T Cells*

Retroviral particles were generated by transient transfection of HEK 293T cells with the MSGV retroviral vector containing a trastuzumab derived HER2 specific CAR [17], a Peg-Pam-e plasmid containing the sequence for MoMLV gag-pol, and a pMEVSVg plasmid containing the sequence for VSVg, using jetPrime transfection reagent (Polyplus, Illkirch, France). Supernatants containing the retrovirus were collected after 48 h (Figure 1A).

To generate HER2 specific CAR T cells, T cells of syngenic Balb/c mice were isolated from a freshly dissected spleen by using a mouse-specific T cell isolation MACS sorting kit (130-095-130; Miltenyi Biotech; Bergisch Gladbach, Germany). MACS sorted mouse T cells were plated on non-tissue culture treated 24-well plates (5 <sup>×</sup> 106 cells/well), which were pre-coated with 1μg/manti-mouse CD3e (ThermoFischer, Waltham, MA, USA) and anti-mouse CD28 (R&D Systems, Min L neapolis, MN, USA) antibodies. After 24 h, mouse interleukin 2 (mIL2; 700 U/ml) was added to cultures. T cells were then transduced with the previously described retroviral particles on RetroNectin-coated (Takara, Kusatsu, Japan) plates on day 3 in the presence of mIL2 (200 U/m L). The expansion of T cells was subsequently supported with mIL2. Anti-mouse CD3e/CD28 activated non-transduced (NT) T cells were expanded in parallel with mIL2. Following 48-72h incubation, cells were collected and used for further experiments (Figure 1B).

#### *4.3. Flow Cytometry*

HER2 specific CAR expression was confirmed by a HER2-GFP recombinant protein. T cell purity was determined by Alexa Fluor 647 conjugated anti-mouse CD3 antibody (BD Biosciences, San Jose, CA, USA) staining. Both molecules were used at 10 μg/mL final concentration for 10 minutes on ice. Analysis was performed on at least 10,000 cells per sample using a FACS Calibur (Becton Dickinson, Franklin Lakes, NJ, USA) instrument and FCS Express 6 software (De Novo Software, Glendale, CA, USA).

#### *4.4. CAR-Mediated T Cell Activation*

Mouse CAR T cells and non-transduced controls were cultivated in indicator-free RPMI 1640 medium and 10% (*v*/*v*) FCS, without additional stimuli for 24 h, washed and incubated on 96-well round-bottomed plates in the presence of JIMT-1 target cells for 48 h. Culture supernatants were analyzed for IFN-γ by ELISA (BD Biosciences). IFN-γ was bound to a solid-phase mAb R46A2 and detected by a biotinylated mAb XMG1.2. The reaction product was visualized by a peroxidase-streptavidin conjugate (1:10,000) and ABTS as substrate. To monitor the cytolytic activity, genetically modified and control T cells were co-cultured with JIMT-1 target cells with increasing T cell numbers for 48 h in 96-well round-bottomed plates. Specific cytotoxicity was monitored by a 2,3-bis[2-methoxy-4-nitro-5-sulphophenyl]-2H-tetrazolium-5-carboxanilide salt (XTT)-based colorimetric assay ('Cell Proliferation Kit II', Roche Diagnostics, Risch, Switzerland). Reduction of XTT was determined as OD at 480 nm for treated tumor cells (Tu), for untreated tumor cells (Max) and for T cells only (T). Background (Bg) was measured in complete medium with XTT but no cells. Measurements were run with minimally 3 technical replicates in three independent experiments. Cytotoxicity was calculated as (1).

$$\text{Cytotoability} \left( \% \right) = \left( 1 - \frac{T\mu - T}{Max - Bg} \right) \cdot 100\%. \tag{1}$$

#### *4.5. Xenograft Tumors and In Vivo Antibody Treatment*

SCID (C.B-17/Icr-Prkdcscid/IcrIcoCrl, Fox-Chase) mice were purchased from Charles River Laboratories, and housed in a specific-pathogen-free environment. All animal experiments were performed in accordance with FELASA guidelines and recommendations and DIN EN ISO 9001 standards. Only non-leaky SCID mice with murine IgG levels below 100 ng/mL were used. Each seven-week-old female SCID mouse participating in the study was given a subcutaneous injection in both flanks, each containing 3 <sup>×</sup> 106 JIMT-1 cells suspended in 100 <sup>μ</sup>L PBS buffer and mixed with an equal volume of Matrigel (BD Biosciences, San Jose, CA, USA). Tumor volumes were derived as the product of the length, width and height measured with a caliper.

The trastuzumab group was treated with 100 μg trastuzumab intraperitoneally in 100 μL PBS twice weekly from day 35 post tumor cell injection. The untreated control group was injected with 100 μL PBS buffer i.p. twice weekly. In the HER2 CAR T cell and unmodified mouse T cell groups mice received a single dose of 2.5 <sup>×</sup> 106 <sup>e</sup>ffector cells i.v. on day 35 post JIMT-1 inoculation. In the low dose HER2 CAR T cell group mice received 2.5 <sup>×</sup> <sup>10</sup><sup>5</sup> cells i.v. (Figure 3A). At the end of the experiment, the animals were euthanized. Experiments were approved by the National Ethical Committee for Animal Research (# 5-1/2018/DEMÁB).

#### *4.6. Tumor Xenograft Sections*

At termination, mice were dissected, and fresh tumors were embedded in cryomatrix (Thermo Fischer Scientific, Waltham, MA, USA) and snap-frozen in isopentane submerged in liquid nitrogen. Serial 14 μm thick cryosections were made with a Shandon Cryotome (Thermo Fischer Scientific, Waltham, MA, USA) at −24 ◦C and air-dried. Staining was carried out at room temperature and all labeling molecules were diluted in PBS buffer supplemented with 1% BSA. After 5 min of rehydration in PBS buffer containing 1% BSA and 0.01% TritonX-100 (Thermo Fischer Scientific, Waltham, MA, USA) HER2 CAR mouse T cells were stained with HER2-GFP recombinant protein and Alexa Fluor 647 conjugated anti-mouse CD3e antibodies. Both molecules were used at 2 μg/mL concentration at 4 ◦C for 10 h. Sections were washed three times, for 5, 20, and 60 minutes, fixed in formaldehyde, and mounted in Mowiol antifade.

#### *4.7. Haematoxylin and Eosin Stained Sections*

The heart and lung were resected from sacrificed mice, fixed in 4% paraformaldehyde for 4–6 h, dehydrated in ethyl alcohol, and embedded in paraffin. Serial sections of 6 μm were cut with a microtome. Deparaffinized sections were HE stained using standard procedures and imaged using a Pannoramic digital histopathology scanner with a 20× objective in transmission mode. A trained histopathologist has examined the digital slides.

#### *4.8. Confocal Laser Scanning Microscopy*

Immunofluorescence-labeled tissue sections were analyzed with a confocal laser scanning microscope (LSM 510, Carl Zeiss GmbH, Jena, Germany) using a 40× C-Apochromat water immersion objective (NA = 1.2). GFP was excited at 488 nm and Alexa Fluor 647 at 633 nm. Corresponding fluorescence emissions were separated with an appropriate quad-band dichroic mirror, and detected through 505 to 550 nm bandpass and 650 nm longpass filters, respectively. Pinhole was set for 4 μm thick optical sections.

#### *4.9. Statistical Analysis*

GraphPad Prism 5 software (GraphPad software, Inc., La Jolla, CA) was used for statistical analysis. Data were presented as mean ± SD or ± SEM. For comparison between two groups, a two-tailed *t*-test was used. For comparisons of three or more groups, one-way ANOVA with Bonferroni's post-test was used. For the mouse experiments, survival, determined from the time of tumor cell

injection, was analyzed by the Kaplan-Meier method and log-rank test. *p*-values < 0.05 were considered statistically significant.

**Supplementary Materials:** Supplementary materials can be found at http://www.mdpi.com/1422-0067/21/3/1039/s1.

**Author Contributions:** G.V. and Á.S. designed the study. G.T and Á.S. performed the experiments. H.A. provided reagents. G.T., J.S., H.A., G.V. and Á.S. analysed data. Á.S. wrote the manuscript, G.V., H.A. and Á.S. revised the manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** We acknowledge the financial support from OTKA K119690 and FK132773 (the National Research, Development and Innovation Office, Hungary), GINOP-2.3.2-15-2016-00050 (co-financed by the European Union and the European Regional Development Fund), and Deutsche Krebshilfe, Bonn, FRG. A.S. was supported by the János Bolyai Research Scholarship of the Hungarian Academy of Sciences and by the UNKP-19- 4-DE-167 New National Excellence Program of the Ministry for Innovation and Technology.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **Abbreviations**


#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Flow Cytometry Reveals the Nature of Oncotic Cells**

**Anna Vossenkamper <sup>1</sup> and Gary Warnes 2,\***


Received: 30 July 2019; Accepted: 4 September 2019; Published: 6 September 2019

**Abstract:** The term necrosis is commonly applied to cells that have died via a non-specific pathway or mechanism but strictly is the description of the degradation processes involved once the plasma membrane of the cell has lost integrity. The signalling pathways potentially involved in accidental cell death (ACD) or oncosis are under-studied. In this study, the flow cytometric analysis of the intracellular antigens involved in regulated cell death (RCD) revealed the phenotypic nature of cells undergoing oncosis or necrosis. Sodium azide induced oncosis but also classic apoptosis, which was blocked by zVAD (z-Vla-Ala-Asp(OMe)-fluoromethylketone). Oncotic cells were found to be viability+ve/caspase-3–ve/RIP3+ve/–ve (Receptor-interacting serine/threonine protein kinase 3). These two cell populations also displayed a DNA damage response (DDR) phenotype pH2AX+ve/PARP–ve, cleaved PARP induced caspase independent apoptosis H2AX–ve/PARP+ve and hyper-activation or parthanatos H2AX+ve/PARP<sup>+</sup>ve. Oncotic cells with phenotype cell viability<sup>+</sup>ve/RIP3–ve/caspase-3–ve showed increased DDR and parthanatos. Necrostatin-1 down-regulated DDR in oncotic cells and increased sodium azide induced apoptosis. This flow cytometric approach to cell death research highlights the link between ACD and the RCD processes of programmed apoptosis and necrosis.

**Keywords:** accidental cell death; oncosis; DDR; parthanatos; flow cytometry

#### **1. Introduction**

The recent re-definition of cell death from Type I (programmed cell death by apoptosis), Type II (autophagic cell death), and Type III (programmed necrosis) to programmed cell death (PCD, homeostatic and embryonic), accidental cell death (ACD, oncosis), and regulated cell death (RCD), which includes apoptosis, necroptosis, autophagy, parthanatos or hyper-activation of Poly (ADP-ribose) polymerase (PARP) caused by an excessive DNA damage response (DDR, e.g., by pH2AX phospho H2AX histone) has been advantageous in understanding the complexity of cell death. RIP1- (Receptor-interacting serine/threonine protein kinase 1) dependent apoptosis or RIP1/RIP3/caspase-3 cells not being included highlights that the cell death nomenclature should be reviewed on a regular basis [1–5]. Programmed or regulated necrosis now includes necroptosis and parthanatos, amongst other forms of reported programmed cell death. However, necrosis is also the term commonly used to indicate the presence of dead cells that have lost plasma membrane integrity by any cell death pathway, but is strictly a reference to the degradation of the cell contents and plasma membrane after death [6–10]. The term oncosis or ACD is a better description of cell death induced by mechanical, chemical, and environmental factors that cause a rapid decrease in intracellular ATP leading to the deactivation of Na<sup>+</sup> and K+-ATPase, resulting in an influx of Na<sup>+</sup>, Cl–, and Ca2<sup>+</sup> ions. The cell then undergoes osmosis and swelling with a bursting of cell organelles and the plasma membrane [11]. Uncoupling protein 2 (UCP-2), located in the inner mitochondrial membrane where protons are pumped by UCP-2 into the mitochondrial matrix or the intermembrane space, which then regulates ATP and superoxide

production is modestly up-regulated by oncosis, resulting in a rapid depolarization of the mitochondrial membrane, which has been measured by flow cytometry [12,13]. The signal pathways involved in oncosis are under studied, so little information is known about these mechanisms compared to the knowledge of pathways in RCD [1,2,6].

In contrast to oncosis, classic apoptosis is caspase-3 dependent and these cells form blebs on the surface of the plasma membrane with a gradual loss in mitochondrial membrane potential and hence a gradual lowering of intracellular ATP. The generation of reactive oxygen species (ROS) and loss of cytochrome *c* from the mitochondria to the cytoplasm results in activation of caspases and generation of apoptosomes and DNA fragmentation, accompanied with cell shrinkage and the formation of apoptotic bodies [1,2,8,14].

Until recently, necrotic or oncotic cells were measured via flow cytometry using the Annexin V assay in which such cells are gated as cell viability+ve/AnnexinV–ve and numerous researchers have attempted to understand this dead cell population with varying success [13,15,16]. In our laboratory, mitochondrial and plasma membrane dysfunction were detected by the multiplexing of mitochondrial and plasma membrane probes into the Annexin V assay, leading to a better understanding of the biological processes occurring in this oncotic population [13,15,16]. The recent development of a polychromatic flow cytometric assay in this laboratory, which identifies most RCDs simultaneously and demonstrates pathways affected by use of pan-caspase and RIP1 protein blockers zVAD and necrostatin-1 (Nec-1), led us to re-investigate oncotic cell death for potential pathways by comparison with apoptosis. The markers measured by flow cytometry included a fixable cell viability marker, activated caspase-3 (apoptosis), up-regulated RIP3 (necroptosis, or resting when not), pH2AX (DDR), cleaved PARP (apoptosis), parthanatos, or hyper-activation of cleaved PARP (pH2AX/cleaved PARP; Table 1, Figure S1). Potential modulation of the oncotic response to sodium azide was further investigated by the use of zVAD and necrostatin-1 to evaluate if the oncotic signalling pathways can be modified by these inhibitors before the cell loses plasma membrane permeability and the cell undergoes oncosis [3,4,13,17–19]. This approach may indicate the nature of the oncotic cell phenotype and highlight potential mechanisms that can modify the oncotic cellular response and the ACD connection to RCD processes. This may increase the potential for the use of therapeutic drugs to target the ACD process in the treatment of cancer.


**Table 1.** Cell description and phenotypes; Figure S1 provides a diagrammatical representation.

#### **2. Results**

#### *2.1. Induction of Oncosis*

NaN3 induced early apoptosis (28%, Figure 1B, lower right quadrant) and lower levels of late apoptotic (13%, Figure 1B, upper right quadrant) and oncotic cells (17%, Figure 1B, upper left quadrant) compared with untreated cells after 24 h (Figure 1A,B, see Materials and Methods section for details

of cell phenotype and gating strategy, Table 1, Figure S1). A lower incidence of live resting cells was observed (RIP3+ve/caspase-3–ve, 44%, Figure 2C, upper left quadrant) but with more early apoptosis after treatment (RIP3–ve/caspase-3<sup>+</sup>ve, 25%, Figure 2A,C, lower right quadrant). Dead cells arising from NaN3 treatment showed less late apoptosis (25%, Figure 2D, lower right quadrant) than untreated cells (Figure 2B).

**Figure 1.** Cell death and caspase-3 activation assay. Cells were (**A**) untreated; (**B**) treated with 0.25% sodium azide (NaN3) for 24 h; (**C**) pre-treated with 20 μM zVAD for 2 h, then with 0.25% NaN3; (**D**) pre-treated with 60 μM necrostatin-1 (Nec-1) for 2 h, then with 0.25% NaN3; (**E**) pre-treated with 20 μM zVAD and 60 μM Nec-1 for 2 h, then with 0.25% NaN3; (**F**) treated with 1 μM Etoposide (Etop) for 24 h; (**G**) pre-treated with 20 μM zVAD for 2 h, then with 1 μM Etop; (**H**) pre-treated with 60 μM Nec-1 for 2 h, then with 1 μM Etop; and (**I**) pre-treated with 20 μM zVAD and 60 μM Nec-1 for 2 h, then with 1 μM Etop. *n* = 3, % Mean ± % SEM; Student's *t*-test: NS (not significant), \* *p* < 0.05, \*\* *p* < 0.01\*\*, \*\*\* *p* < 0.001; arrows indicate change compared with untreated cells.

**Figure 2.** RIP3 and caspase-3 activation analysis of oncosis. After gating on live and dead cells from a Zombie NIR vs. caspase3-BV650 dot-plot (**A**) untreated live and (**B**) dead Jurkat cells were analysed on a RIP3-PE vs. caspase-3-BV650 dot-plot with resting phenotype indicated by RIP3+ve/caspase-3–ve, apoptosis by RIP3–ve/caspase-3<sup>+</sup>ve, RIP1-dependent apoptosis RIP3<sup>+</sup>ve/caspase-3<sup>+</sup>ve, and double negative RIP3–ve/caspase-3–ve. Live and dead cells treated with (**C,D**) 0.25% NaN3 for 24 h; (**E**,**F**) pre-treated with 20 μM zVAD for 2 h, then treated with 0.25% NaN3; (**G**,**H**)pre-treated with 60 μM Nec-1 for 2 h, then treated with 0.25% NaN3; and (**I**,**J**) pre-treated with 20 μM zVAD and 60 μM Nec-1 for 2 h, then treated with 0.25% NaN3, respectively. *n* = 3, % Mean ± % SEM, Student's *t*–test: NS (not significant), \* *p* < 0.05, \*\* *p* < 0.01\*\*, \*\*\* *p* < 0.001; arrows indicate change compared with untreated cells.

The oncotic cells resulting from NaN3 treatment were mainly double negative (55%) for RIP3 and caspase-3 expression (dead resting oncotic cells, <10% caspase-3–ve/RIP3<sup>+</sup>ve, Figure 2D).

After NaN3 treatment, the two live and dead apoptotic populations showed increased levels of pH2AX hyper-activation of cleaved PARP and a lower degree of apoptosis via cleaved PARP and DDR than untreated cells (Figure 3A–C, Figure S1A,B,E,F,I,J,M,N). Whereas late apoptotic cells showed increased DDR (Figure 3C, Figure S2B,E,M,J). The dead resting oncotic cells (Zombie+ve/caspase-3–ve/RIP3<sup>+</sup>ve) were, 31% negative for both H2AX and PARP, whereas the dead oncotic DN (Zombie+ve/caspase-3–ve/RIP3–ve) cells were 57% negative for both markers (Figure 3A–C, Figure S2O,P). The live and dead DN populations showed increased levels of parthanatos and DDR (Figure 3A–C, Figure S2D,H,L,P).

**Figure 3.** Parthanatos/hyper-activation of cleaved PARP, apoptosis via cleaved PARP, and DDR analysis of oncosis. Untreated Jurkat cells, treated with 0.25% NaN3 for 24 h, or pre-treated with zVAD (20 μM) and/or Nec-1 (60 μM) for 2 h, then incubated with 0.25% NaN3. Gating live and dead cells from a Zombie NIR vs. caspase-3-BV650 plot then both were analysed on a RIP3-PE vs. caspase-3-BV650 plot. Next, early and late apoptotic, necroptotic/resting, RIP1-dependent apoptotic, and double negative (DN) populations were analysed for pH2AX and cleaved PARP (Figures S2, S3). The incidence of (**A**) parthanatos/hyper-activation of cleaved PARP, (**B**) apoptosis via cleaved PARP, and (**C**) DDR were determined for all populations listed above. *n* = 3, % Mean, error bars % SEM, Student's *t*-test; NS (not significant), \* *p* < 0.05, \*\* *p* < 0.01\*\*, \*\*\* *p* < 0.001 compared with untreated cells.

#### *2.2. Induction of Apoptosis*

Induction of apoptosis with Etop showed an increase in early and late apoptosis as well as oncotic cells compared with untreated cells (Figure 1A,F). Live and early apoptotic cells showed increased levels of both types of apoptosis and the DN cells, whereas dead cells showed no such change (Figure 2A,B and Figure 4A,B).

**Figure 4.** RIP3 and caspase-3 activation analysis of apoptosis. Gating on live and dead cells from a Zombie NIR vs. caspase-3-BV650 plot followed by analysis on a RIP3-PE vs. caspase-3-BV650 plot with resting phenotype indicated by RIP3+ve/caspase-3–ve, apoptosis by RIP3–ve/caspase-3<sup>+</sup>ve, RIP1-dependent apoptosis by RIP3+ve/caspase-3+ve and double negative by RIP3–ve/caspase-3–ve. (**A,B**) Treated with 1 μM Etop for 24 h; (**C**,**D**), pre-treated with 20 μM zVAD for 2 h, then treated with 1 μM Etop; (**E**,**F**) pre-treated with 60 μM Nec-1 for 2 h, then treated with 1 μM Etop; and (**G**,**H**) pre-treated with 20 μM zVAD and 60 μM Nec-1 for 2 h, then treated with 1 μM Etop. *N* = 3, % Mean ± % SEM, Student's *t*–test: NS (not significant), \* *p* < 0.05, \*\* *p* < 0.01\*\*, \*\*\* *p* < 0.001; arrows indicate change compared with untreated cells.

Early and late apoptotic cells after Etop treatment showed increased pH2AX hyper-activation of cleaved PARP with a decrease in apoptosis via cleaved PARP compared with untreated cells (Figure 5A,B and Figure S2Q,U). Live RIP1-dependent cells, however, showed increased pH2AX hyper-activation of cleaved PARP and caspase-3-dependent apoptosis via cleaved PARP with decreased DDR (Figure 5A–C and Figure S2R). However, live and dead oncotic resting and DN phenotypes also showed increased parthanatos and caspase-3-independent apoptosis via cleaved PARP with no increase in DDR, except for an increase observed in the live resting population (Figure 5A–C and Figure S2S,T,V,W,X).

**Figure 5.** Parthanatos/hyper-activation of cleaved PARP, apoptosis via cleaved PARP, and DDR analysis of apoptosis. Untreated Jurkat, treated with 1 μM Etop or pre-treated with zVAD (20 μM) and/or Nec-1 (60 μM) for 2 h, then incubated with 1 μM Etop for 24 h. Gating on live and dead cells from a Zombie NIR vs. caspase-3-BV650 plot then both were analysed on a RIP3-PE vs. caspase-3-BV650 plot. Early and late apoptotic, necroptotic/resting, RIP1-dependent apoptotic, and double negative (DN) populations were analysed for pH2AX and cleaved PARP (Figures S2,S4 for detailed information). The incidence of (**A**) parthanatos/hyper-activation of cleaved PARP, (**B**) apoptosis via cleaved PARP, and (**C**) DDR were determined for all populations listed above. Mean, error bars % SEM, Student's *t*-test: NS (not significant), \* *p* < 0.05, \*\* *p* < 0.01\*\*, \*\*\* *p* < 0.001 compared with untreated cells.

#### *2.3. Blockade of Caspases*

Pre-treatment with zVAD to block the activation of caspases by NaN3 and Etop resulted in lower levels of early apoptosis (<20%) and late apoptosis (<10%) with no change in the incidence of oncotic cells (Figure 1C,G, Figure 2E,F and Figure 4C,D). The proportion of live DN (RIP3–ve/caspase-3–ve) and dead oncotic cells (RIP3–ve/caspase-3–ve) increased after zVAD blockade of both drugs (Figure 2E,F and Figure 4C,D).

After zVAD caspase blockade of NaN3 and Etop treatments, all live populations showed increased H2AX hyper-activation of cleaved PARP or parthanatos compared with untreated cells but decreased compared to drugs alone (Figure 3A, Figure 5A and Figure S3A–D, S4A–D). The live RIP1-dependent apoptotic, resting, and DN cells from both treatments also showed increased levels of apoptosis via cleaved PARP compared with untreated cells (early apoptosis showed a decrease, Figure 3B, Figure 5B and Figure S3A–D, S4A–D). The live DN population after zVAD/NaN3 or Etop treatment showed a decrease or increase of DDR compared with drugs alone, respectively (Figure 3C, Figure 5C and Figure S3D, S4D). Live RIP1-dependent apoptotic and resting phenotypes, after both treatments with zVAD, showed increased DDR compared with drugs alone (Figure 3C, Figure 5C and Figure S3B,C, S4B,C). Dead cells from zVAD blockade of NaN3/Etop treatments returned pH2AX and cleaved PARP expression to that of untreated dead cells (Figure 3, Figure 5 and Figure S3E–H, S4E–H). Except after NaN3/zVAD treatment, an increase in DDR was observed in the dead RIP1-dependent apoptotic and oncotic DN phenotypes (Figure 5 and Figure S3F,H).

#### *2.4. Blockade with Necrostatin-1*

Blockade of NaN3 with Nec-1 resulted in very high levels of early apoptosis compared with NaN3, but was lower with Etop treatment (Figure 1D,H). Cell death was lower with Nec-1/NaN3 but not changed with Etop treatment (Figure 1D,H). The live cells showed a higher incidence of early and RIP1-dependent apoptosis compared with NaN3 treatment, with no change observed with Etop or in the incidence of dead cells (Figure 2A–D,G,H and Figure 4A,E,F).

pHA2X hyper-activation of cleaved PARP in live and dead cells after Nec-1 showed increased values similar to that observed with only drugs, except for the lower levels in both the DN populations and live resting cells after Nec-1/Etop treatment (Figure 3A, Figure 5A and Figure S3I–P, S4I–P). Apoptosis via cleaved PARP was increased in live and dead cells after Nec-1 blockade of NaN3 treatment and decreased with Nec-1/Etop treatment compared with drugs alone, except for a decrease in live resting cells (Nec-1/NaN3) and no change in RIP1-dependent apoptosis after Nec-1/Etop treatment (Figure 3B, Figure 5B and Figure S3I–P, S4I–P). Very low levels of DDR were observed in early and live RIP1-dependent apoptotic cells, but increased in live resting and DN cells (Figure 3C, Figure 5C and Figure S3I–L, S4I–L). Dead cell phenotypes after Nec-1 blockade of NaN3 showed no increase in DDR compared with untreated cells, but was increased with Nec-1/Etop treatment (Figure 3C, Figure 5C and Figure S3, S4M–P).

#### *2.5. Blockade with zVAD and Necrostatin-1*

Pre-treatment with zVAD and Nec-1 to block the activation of caspases and RIP proteins resulted in lower levels of early and late apoptosis (<20%), although oncosis (caspase-3–ve/Zombie<sup>+</sup>ve) was still maintained (Figure 1E,I, Figure 2I,J and Figure 4G,H). Blocked NaN3 treated cells had a higher level of live DN cells (and Etop), whereas dead cells showed higher levels of RIP1-dependent apoptosis, which indicated that zVAD did not block caspases in the RIP1-dependent apoptotic pathway in the presence of Nec-1 (Figure 2I,J and Figure 4G,H).

All cell phenotypes after dual blockade showed the same reduced levels of pH2AX hyper-activation of cleaved PARP as that observed after zVAD blockade (no change in Etop DN cells, Figure 3A, Figure 5A and Figure S3Q–X, S4Q–X). Apoptosis via cleaved PARP was lower in the early apoptotic and DN cells after dual blockade of NaN3 and increased in live resting and RIP1-dependent apoptotic cells (Figure 3B

and Figure S3Q,T). In contrast, all live cell (dual-blocked Etop) populations showed no change in apoptosis via cleaved PARP compared with drug treatment, except for the lower levels observed in live resting cells (Figure 5B and Figure S4Q,T). All dead cell populations after dual blockade of both treatments showed lower levels of apoptosis via cleaved PARP compared to any treatment protocol, except no change was observed in Etop-induced late and RIP1-dependent apoptotic cells (Figure 3B, Figure 5B and Figure S3U–X, S4U–X). Lastly, the DDR levels after both treatment protocols showed that all cell populations had higher levels than untreated cells (Figures 3C and 5C), except for the lower levels found in the live RIP1-dependent apoptotic cells compared with both treatments (Figures 3C and 5C).

#### **3. Discussion**

The use of oncosis and apoptosis-inducing drugs NaN3 and Etop, together with caspase and RIP protein blockers, zVAD and Nec-1, has allowed the tracking of the cell death processes involved in ACD and apoptosis a form of RCD using flow cytometry (Figure 6) [17–19]. Induction of oncosis or apoptosis resulted in measurable oncosis (Zombie+ve/caspase-3–ve, commonly termed necrotic) but also early and late apoptosis, which was reduced by zVAD blockade without affecting the degree of oncosis induced by both drugs. Blockade of both drugs with Nec-1 resulted in increased NaN3-induced early apoptosis while increasing Etop-induced oncosis. Induction of ACD and early apoptosis (by NaN3) showed that live resting Jurkat cells move the RIP3<sup>+</sup>ve/caspase-3–ve/Zombie–ve phenotype to the early apoptotic phenotype of RIP3–ve/caspase-3+ve/Zombie–ve, then later to the RIP1-dependent phenotype RIP3+ve/caspase-3+ve/Zombie–ve (Figure 6B), even in the presence of RIP1 inhibitor, Nec-1. This effect has been previously reported [17,19], where it was shown that although Nec-1 inhibited necroptosis by abrogation of the up-regulation of RIP3 (RIP3+ve/caspase-3–ve/Zombie–ve), it did not inhibit cells from undergoing apparent RIP1-dependent apoptosis. The limitation of the current assay is highlighted by the use of RIP3 and caspase-3 antibodies to indirectly identify RIP1-dependent apoptosis due to the lack availability of a fluorescenated RIP1 antibody [17,19]. The interactions of RIP1, RIP3, TRADD (TNFR1-associated death domain), FADD (Fas associated via death domain), and caspase-8 in apoptotic Complex IIa and IIb pathways are not completely understood, so another explanation of the apparent presence (indirectly via RIP3) of RIP1-dependent apoptosis in the presence of Nec-1 is required, which may be elucidated by the use of a fluorescenated RIP1 antibody [5,20,21].

Induction of apoptosis by Etop showed that a high proportion of live resting cells become DN (losing their RIP3), as well as another population expressing caspase-3, possibly indicating that the cells first lose RIP3, become DN, and then express caspase-3 (Figure 6B), but also that early apoptotic cells can also later express RIP3 (Figure 6B) [5,17,19–21]. The route to RIP1-dependent apoptosis may be that these resting cells, rather than lose their RIP3, also start to express caspase-3 (Figure 6B).

Once the cells lose plasma membrane integrity and become Zombie<sup>+</sup>ve, the cells presumably maintain the late apoptotic phenotype RIP3–ve/caspase-3+ve/Zombie+ve before further degradation resulting in the cells becoming DN (Figure 6C). Oncotic cells (caspase-3–ve/Zombie<sup>+</sup>ve) induced by NaN3/Etop, however, can also be divided into those with RIP3+ve/caspase-3–ve/Zombie+ve or the DN phenotype RIP3–ve/caspase-3–ve/Zombie+ve (Figure 6C).

**Figure 6.** ACD and RCD pathways. (**A**) Live cells can undergo either early apoptosis (EAPO) or oncosis after drug treatment, with early apoptotic cells moving to late apoptotic (LAPO), then later with cell degradation, to the oncotic or necrotic phenotype. (**B**) Live cells may express RIP3+ve/caspase-3–ve when resting, or be RIP3high+ve/caspase-3–ve when undergoing necroptosis, or be double negative (DN). EAPO cells lose RIP3 or, if retained, undergo RIP1-dependent apoptosis (RIP1-APO). EAPO cells can also become RIP1<sup>+</sup>ve. (**C**) Loss of plasma membrane integrity or cell death results in cell phenotypes mirrored in (**B**), with degradation of cells resulting in the DN population. (**B**,**C**) Live and dead cell phenotypes can also express pH2AX (DDR) or cleaved PARP (apoptosis), both of which can ultimately express both proteins, (**D**,**E**) resulting in pH2AX hyper-activation of cleaved PARP in the presence of active caspase-3 or parthanatos in the absence of caspase-3. Arrows indicate movement of cell populations.

The expression of pH2AX and cleaved PARP in the four identified phenotypes in live and dead cells are resting, early or late apoptotic, RIP1-dependent apoptosis, and DN, the incidence of which is modified by the action of the two drugs used in this study and can be further manipulated by blockade of ACD and RCD processes by zVAD and Nec-1 (Figure 6D,E) [22–25]. In the first instance live resting cells, the main phenotype of untreated Jurkat cells express RIP3 with a high degree of DDR (37%), whereas resting DN cells showed little DDR (2%) but a high degree of cleaved PARP (39%) in the absence of caspase-3 [25]. Induction of ACD resulted in enhanced levels of pHA2X hyper-activation of cleaved PARP in all live cell phenotypes with consequent reduced levels of DDR in live RIP1-dependent apoptotic cells, but with increased levels in the live DN phenotype (Figure 6D).

Once the cells undergo death, all phenotypes still showed increased pHA2X hyper-activation of cleaved PARP above the levels observed with dead untreated Jurkat cells (Figure 6E), whereas late apoptotic and dead oncotic DN cells showed increased DDR with no change in dead resting cells. Similar results were observed when cells undergo Etop-induced apoptosis, except no increase in DDR was observed in the dead oncotic or resting cells. So, NaN3-derived oncotic cells can express more DDR than Etop-induced oncotic cells, which expressed higher levels of cleaved PARP (in the absence of caspase-3) and parthanatos/pHA2X hyper-activation of cleaved PARP, again in the absence of caspase-3 (Figure 6E).

The main effect of blockade with zVAD was an increased incidence of live and dead DN cells with both treatments, which displayed no change in DDR, increased cleaved PARP and reductions in pHA2X hyper-activation of cleaved PARP, as was the case with most cell phenotypes (Figure 6D, E). This was with the notable exception of no change in levels of cleaved PARP in dead oncotic resting and DN cells compared to NaN3 treatment.

In contrast, blockade of NaN3 with Nec-1 resulted in no change in the incidence of dead DN and fewer live DN cells (as untreated cells), whereas Etop-induced levels were similar to that observed by drug alone. However, all blocked NaN3 cell phenotypes showed increased cleaved PARP and pH2AX activation of cleaved PARP with reductions in DDR; this observation was especially noteworthy in the dead oncotic resting and DN cells, whereas the opposite was observed in these populations after Nec-1 blockade of Etop.

Dual blockade of both treatments again showed a high incidence of DN cells with reduced levels of pHA2X hyper-activation of cleaved PARP and cleaved PARP expression in the various populations of cells, whereas DDR was generally increased after Nec-1 blockade of both drugs in most populations of cells. So, the blockade of apoptotic and/or necroptotic pathways when the cells are undergoing ACD and RCD processes radically increased the incidence of oncotic cells, especially in the case of blockade of apoptosis alone and necroptosis, which resulted in increased levels of DDR, whereas Nec-1 blockade reduced ACD-related DDR but not in the case of RCD.

For decades, oncotic cells have been an undetermined population of dead cells that have been overlooked due to the difficult in their characterisation rather than being a point of interest. They have been detected by use of the Annexin V assay and classed as cell viability+ve/Annexin V–ve [13,15,16]. Using an active caspase-3 and RIP3 antibodies in tandem with a fixable live/dead dye, we showed that these oncotic cells have two main phenotypes, both of which are cell viability+ve/caspase-3–ve/RIP3+/–ve, which can be further divided into DDR, hyper-activation of PARP or parthanatos, apoptosis via PARP or cells undergoing programmed necrosis, as well as double negative oncotic cells. All of them are expressed to different degrees compared with cells derived from untreated cultures. This perhaps reflects their different origins, with differences in expression of these markers when the cells are derived from an ACD or RCD induction process. These differences in origin are further highlighted by their differing oncotic responses to blockade by zVAD and Nec-1 or both.

#### **4. Materials and Methods**

#### *4.1. Induction of Oncosis and Apoptosis*

Jurkat cells (human acute T cell leukaemia cell line, ECACC, Salisbury, UK) were grown in RPMI (Roswell Park Memorial Institute 1640 Medium) 1640 with 10% FBS (Fetal Bovine Serum, Invitrogen, Paisley, UK) at 37 ◦C and 5% CO2, either untreated or treated with 0.25% sodium azide or 1 μM etoposide (NaN3, Etop, Sigma, Poole, UK) for 24 h. Cells were pre-treated with pan-caspase blocker zVAD (20 μM, Enzo Life Sciences, Exeter, UK) and/or necroptosis blocker necrostatin-1 (Nec-1, 60 μM Cambridge Bioscience, Cambridge, UK) for 2 h before induction of oncosis with 0.25% sodium azide or apoptosis with 1 μM Etop for 24 h.

#### *4.2. Flow Cytometry Assay*

Harvested cells were labelled with fixable live dead stain, Zombie near infra red (NIR; Biolegend, San Diego, CA, USA) at room temperature (RT) for 15 min. Washed cell were fixed in Solution A (CalTag, Little Balmer, UK) then 0.25% Triton X-100 (Sigma, Poole, UK) for 15 min each at RT. Jurkat cells (1 <sup>×</sup> 106) were incubated for 20 min at RT with anti-RIP3-PE (phycoerythrin clone B-2, Cat. No. sc-374639, Santa Cruz, Dallas, Tx, USA), cleaved PARP-PE-CF-595 (clone F21-852, Becton Dickinson, San Jose, CA, USA), H2A.X-Phospho (ser139)-PE-Cy7 (clone 2F3, Biolegend, San Diego, CA, USA) and anti- active caspase-3-BV650 (clone C92-605, Becton Dickinson, San Jose, CA, USA) for 20 min at RT. Washed cells were resuspended in 400 μL PBS (Phosphate Buffered Saline) and analysed on a ACEA Bioscience Novocyte 3000 flow cytometer (100,000 events, San Diego, CA, USA). Zombie NIR was excited by a 633 nm laser and collected with a 780/60 nm detector. Caspase-3-BV650 was excited by a 405 nm laser and collected at 675/30 nm. RIP3-PE, cleaved PARP-PE-CF-595, and pH2AX-PE-Cy7 were

excited by a 488 nm laser and collected at 572/28, 615/20, and 780/60 nm, respectively. Single colour controls were used to determine the colour compensation using the pre-set voltages on the instrument using Novo Express software (ver 1.2.5, ACEA Biosciences, San Diego, CA, USA). Cells were gated on FSC (Forward Scatter) vs. SSC (Side Scatter) with single cells being gated on a FSC-A (Area) vs. FSC-H (Height) plot. Cells were then gated on a plot of caspase-3-BV650 vs. Zombie NIR, with a quadrant placed marking off live cells in the double negative quadrant (Figure S1A, lower left quadrant), with caspase-3-BV650+ve/Zombie NIR–ve (Figure S1A, lower right quadrant) indicating early apoptotic cells (EAPO), and lastly with caspase-3-BV650+ve/Zombie NIR+ve and caspase-3-BV650–ve/Zombie NIR+ve (Figure S1A, upper quadrants) indicating dead late apoptotic (LAPO) (Figure S1A, upper right quadrant) and oncotic cells (Figure S1A, upper left quadrant). Further labelling with RIP3 allowed identification within the live resting and dead oncotic populations of RIP3+ve/caspase-3–ve or necroptotic cells when RIP3 is up-regulated, early or late apoptotic (RIP3–ve/caspase-3<sup>+</sup>ve), RIP1-dependent apoptosis (RIP3+ve/caspase-3<sup>+</sup>ve, RIP1APO), and live double negative (DN) or dead oncotic DN cells (Figure S1B,D). Further gating on each of these eight populations for pH2AX and cleaved PARP allowed the identification of DDR (H2AX+ve/PARP–ve), pH2AX hyper-activation of cleaved PARP or parthanatos (H2AX+ve/PARP<sup>+</sup>ve), apoptotic cell death via cleaved PARP (H2AX–ve/PARP<sup>+</sup>ve), and DN cells (H2AX–ve/PARP–ve) (Figure S1C,E) [19]. In particular, dead resting oncotic cells (Zombie+ve/caspase-3–ve/RIP3<sup>+</sup>ve) and dead oncotic DN cells (Zombie<sup>+</sup>ve/caspase-3–ve/RIP3–ve) were gated for pH2AX and cleaved PARP, revealing the phenotypic nature of these two types of oncotic cells (Figure S1F,G).

#### *4.3. Statistics*

For all experiments, *n* = 3 and data are reported as mean ± SEM for percentage positive. Student's *t-*tests were performed in GraphPad software Inc. (San Diego, CA, USA) with *p* ≥ 0.05 considered not significant (NS). \* denotes *p* ≤ 0.05, *\*\** denotes *p* ≤ 0.01, and \*\*\* denotes *p* ≤ 0.001 when treated cells were compared to untreated cells.

**Supplementary Materials:** Supplementary materials can be found at http://www.mdpi.com/1422-0067/20/18/ 4379/s1.

**Author Contributions:** Conceptualization, G.W.; methodology, G.W.; software, G.W., validation, G.W.; formal analysis, G.W. and A.V.; investigation, G.W.; resources, G.W.; data curation, G.W.; writing—original draft preparation, G.W.; writing—review and editing, G.W. and A.V.; visualization, G.W.; supervision, G.W.; project administration, G.W., funding acquisition, G.W.

**Funding:** APC was funded by Queen Mary University Internal funding.

**Conflicts of Interest:** A.V. received salary funding from GSK. GSK had no input or role in the conception or undertaking of this study in any manner. Other authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
