Cell-in-cell structures are more potent predictors of outcome than senescence or apoptosis in head and neck squamous cell carcinomas
© The Author(s). 2017
Received: 17 August 2016
Accepted: 21 December 2016
Published: 18 January 2017
This study sheds light on cell inactivating processes with focus on the phenomenon of cell-in-cell (CIC). Cell-in-cell describes a cell process where one cell is being engulfed by another non-professional phagocyte. We determined frequency and prognostic impact of CIC structures (CICs) as well as of senescent and apoptotic cells in head and neck squamous cell carcinomas (HNSCC).
These different forms of cell inactivation as well as the proportion of proliferating and tumor cells were assessed in 169 pre-radiochemotherapy biopsies and 32 post-therapy tumor resections by immunohistochemistry of tissue microarrays. Four consecutive cancer sections were stained with antibodies specific for E-cadherin for CIC detection, cleaved caspase-3 for apoptosis, H3K9Me for senescence and Ki67 as a proliferation marker. Positive events were quantified in corresponding tumor areas.
CICs were found in 55.5%, senescent cells in 67.1% and apoptotic cells in 93.3% of samples. While no prognostic impact of apoptotic and senescent cells was observed, CICs turned out to significantly influence overall-survival (p = 0.016) with a lack of CICs being prognostically beneficial. There was no correlation between CICs and apoptosis and 98.9% of CICs were negative for cleaved caspase-3.
CIC formation is a frequent event in HNSCC and a superior predictive marker compared to senescence and apoptosis. Independence of CIC and apoptosis and the adverse prognosis associated with numerous CICs lead to the assumption that CICs might take up necrotic rather than apoptotic cells preventing an adequate antitumoral immune response that would otherwise be initiated by necrotic cells through damage-associated molecular pattern molecules.
KeywordsHNSCC Cell death Cell-in-cell Senescence Apoptosis Proliferation
In multicellular organisms, precise control and coordination of cell proliferation, cell inactivation and cell death are vital to maintain homeostasis of cells in tissues and organs as slightest imbalances can lead to pathologies like tumors and autoimmune diseases. For deeper understanding of these processes and their meaning we investigated the cell processes of cell-in-cell, senescence and apoptosis. Much attention has been focused on the cell death mechanisms of apoptosis and necrosis and their significance in tumors as well as healthy tissue. However, the far less appreciated cell inactivating processes of CIC and senescence may be equally essential. To assess their role with regard to frequencies in tumor specimens and clinical outcome we investigated a cohort of HNSCC.
To best look into the subject of cell death stringent definitions are vital. Kroemer et al. suggest that cell death can be classified according to morphology, enzymological criteria, functional aspects or immunological characteristics . The Nomenclature Committee on Cell Death (NCCD) previously proposed three criteria for the identification of a dead cell: (1) permanent loss of the barrier function of the plasma membrane; (2) breakdown of cells into discrete fragments; or (3) engulfment of cells by professional phagocytes or other cells endowed with phagocytic activity .
According to the NCCD the phenomena of apoptosis and necrosis can be further defined as follows: apoptosis is characterized by cytoplasmic shrinkage, chromatin condensation (marginalization), nuclear fragmentation (karyorrhexis), so called blebbing and apoptotic bodies and is considered a “regulated cell death”, generally referred to as “programmed cell death”. Necrosis presents generalized swelling of the cytoplasm and organelles (oncosis), alteration of chromatin (condensation) and the nuclear membrane (dilatation) and is regarded as “accidental cell death” . Apart from these morphological features there was no reliable marker for detection of necrosis available.
Senescence defines the process of a cell going into an irreversible cell-cycle arrest that is unresponsive to mitogenic or oncogenic stimulation. However, senescent cells are still viable and metabolically active without displaying the specific functions of their lineage . The CIC phenomenon describes a cell process where one cell is being phagocytized completely by another non-professional phagocytizing cell which has been observed in a variety of malignancies. “Cell-in-cell” is an umbrella term without further specification. Similar, yet different, processes giving rise to cell-in-cell structures have been introduced in literature: entosis, emperipolesis, cannibalism and phagocytosis . Entosis is the active invasion of a living cell into another cell’s cytoplasm . Emperipolesis defines the interaction of lymphocytes with other cells and has been observed in physiological and pathophysiological settings . Cannibalistic tumor cells are able to engulf other cells, including lymphocytes and erythrocytes, either dead or alive, with the main purpose to feed on them . These mechanisms all have in common that a living cell is taken up by another non-professional phagocytizing cell and can be broadly characterized as heterotypic or homotypic . Although it is shown that an engulfed cell is able to remain in the host cell, evade the host cell or undertake cell division, phagocytosis most likely leads to cell death of the incorporated cell and a subsequent decomposition [9, 10]. Although the CIC phenomenon has recently gained more and more attention, so far exact details of its characteristics and mechanisms remain unclear. Preliminary findings originate several questions: Which are role and condition of engulfing and inner cell? Is the inner cell viable or dead? Is cell-in-cell a kind of cell death or rather a cell interaction process?
Our main focus in this study was to examine the formation of CIC as a specific form of non-professional phagocytosis. We assessed CIC’s prognostic meaning to further examine its association with clinical pathogenesis as well as its relation to apoptosis and senescence to gain further knowledge about the identity of the interacting cells. We quantified the frequencies of each of these events in a HNSCC cohort pre- and post-radiochemotherapy (RCT) and investigated their impact on local failure-free-, metastasis-free- and overall-survival.
Materials and methods
Clinical characteristics of the head and neck squamous cell carcinoma cohort
169 patients with available pre-RCT biopsies
<56.4: 85 (50.3%) >56.4: 84 (49.7%)
Localization of tumor
Base of the mouth
Mean total dose
Mean dose per fraction
5-FU + Cisplatin
5-FU + Carboplatin
32 patients with available post-RCT biopsies
<51.4: 71 (46.4%) >51.4: 82 (53.6%)
Localization of tumor
Base of the mouth
Mean total dose
Mean dose per fraction
5-FU + Cisplatin
5-FU + Carboplatin
Antibodies and immunohistochemistry
The following antibodies were used: anti-E-cadherin (1:2000; clone 36/E-Cadherin; BD, Heidelberg, Germany) to visualize membranes of cells to detect CIC structures (Fig. 1a), anti-cleaved caspase-3 (1:200, polyclonal; Cell Signaling, Danvers, MA, USA) to display apoptotic cells (Fig. 1b), anti-H3K9Me (1:200; [BM1] Millipore, Darmstadt, Germany) to detect senescent cells by chromatin condensation (Fig. 1c) and anti-Ki67 (1:100, clone MIB1; Dako Cytomation, Hamburg, Germany) to highlight proliferating cells (Fig. 1d). Hematoxylin-eosin staining was used for nuclear counter-staining. H3K9Me was stained manually using a steam cooker and a commercially available target retrieval solution pH 6 (TRS6, Dako Cytomation, Hamburg, Germany) for antigen retrieval and an alkaline phosphatase-labeled polymer kit (Zytochem-Plus AP-PolymerKit, Zytomed Systems, Berlin, Germany) and Fast Red (Sigma-Aldrich, Deisenhofen, Germany) for detection. The other stainings were performed on a Ventana BenchMark ULTRA stainer (Roche) using an UltraView DAB IHC Detection Kit (Roche) for detection.
Imaging and image analysis
Stained TMAs were scanned with a high throughput scanner (Mirax Scan, Zeiss, Göttingen, Germany). The software Pannoramic Viewer (3D Histech, Budapest, Hungary) was used for further processing of the scans. Each TMA image was cut, exported and converted into TIF files to be saved separately. Cell events were counted in the same tumor area by precise alignment of each stained TMA using the image processing software Biomas (Erlangen, Germany). Each region of interest was defined in the E-cadherin slide and transferred to corresponding TMAs. Subsequently CICs, senescent cells, apoptotic cells, proliferating cells as well as tumor cells were counted partly manually partly half-automatically in each TMA by the use of the above-mentioned image processing software.
The following three criteria were defined for a coherent evaluation of CICs: complete encirclement of the inner cell by the host cell membrane, a round shape of the inner cell and a semilunar host cell nucleus displaced to the margin of the host cell (Fig. 1a i).
IBM SPSS Statistics version 19 (IBM Corp, Armonk, NY, USA) was used for statistical analysis. Local failure-free survival, metastasis-free-survival and overall survival were determined according to Kaplan Meier. The log-rank test was used to compare survival curves between subgroups of patients. P-values below 0.05 were considered significant.
Frequency of cell events
Clinical characteristics of the study group are given in Table 1. The HNSCC cohort consisted of 169 patients with pre-RCT biopsies from all 169 patients and post-RCT tissue samples from 32 patients. The median follow-up was 4.5 years. Four consecutive cancer sections were stained with E-cadherin for CIC detection (Fig. 1a), anti-cleaved caspase-3 for detection of apoptosis (Fig. 1b), anti-H3K9Me for detection of senescence (Fig. 1c) and anti-Ki67 for identification of proliferating cells (Fig. 1d). Apoptosis, senescence and proliferation were assessed in positive cells morphologically consistent with carcinoma cells, whereas other cells like stromal or inflammatory cells were not included in the analysis.
We compared the frequency of CICs to senescent, apoptotic, proliferating and tumor cells (Fig. 1e-j). Center of the tumor pre-RCT (n = 169), invasive front pre-RCT (n = 103) and center of the tumor post-RCT (n = 32) derived 6 weeks after the end of RCT were separately evaluated and compared with each other. Significantly more CICs, senescent cells and apoptotic cells were identified in samples from tumor centers compared to invasive front samples (p = 0.003, p = 0.010, p = 0.043) (Fig. 1e-g). Average numbers of all positive events as well as the proportion of samples showing any positive event in all respective markers decreased distinctly post-RCT. The highest number of proliferating cells was found in the center of the tumor with distinct decrease post-RCT (Fig. 1h). Additionally, tumor cells in each TMA were counted (Fig. 1j).
Distribution and ratio of cell events
Taking all tissue samples into consideration, nearly every biopsy showed apoptotic events while only about half display CICs and senescent cells each. However 84.6% of tissues were positive for either CICs or senescent cells. When looking at cases with events, absolute numbers of cases with CICs and senescent cells were comparably frequent as cases with apoptotic cells. The overall mean of CICs was 8.2 CICs mm−2 and rising to 14.8 CICs mm−2 when considering only CIC-positive samples (Fig. 2b). Accordingly, overall mean of senescent cells was 12.6 mm−2 and increased to 18.8 mm−2 after exclusion of negative cases (Fig. 2c). The mean value of apoptotic cells was only slightly changed when considering positive cases only as apoptotic events were present in the great majority of cases from the beginning (Fig. 2d).
Correlation of cell events
Prognostic relevance of CIC structures, senescence and apoptosis
Our observations on frequency and prognostic impact of CICs, senescence and apoptosis evoke pivotal questions in view of previous findings and require further reflection. A remarking result is the fact that CICs and senescence occur in a comparable frequency range as apoptosis while not receiving equal consideration in literature. It is particularly noteworthy that CICs compared to apoptotic cell death, inactivation by senescence and proliferation is the only event with a prognostic relevance pre-RCT. Necrotic death could not be evaluated in this study due to the lack of a suitable immunohistochemical necrosis marker.
Another interesting point is that the prognostic value of CICs post-RCT was weaker than pre-RCT. Similarly, apoptosis showed no prognostic value pre-RCT whereas low apoptotic rates were associated with a favorable prognosis post-RCT. RCT-driven selection processes may lead to this change. The fraction of CIC-positive tissues post-RCT consisted of only 23%, while 50% of tissues were CIC-positive prior to RCT. Median apoptotic rates changed from 8.3 apoptotic cells per mm2 prior to RCT to 1.1 apoptotic cells per mm2 post-RCT.
As only very few of the cells engulfed in CICs were apoptotic and there was no significant correlation between CICs, apoptotic and senescent cells, the question rises whether CICs are formed by phagocytized necrotic cells. In our previous study 99.5% of CICs were negative for cleaved caspase-3 going in line with the present finding that apoptotic cells are very infrequently involved in the formation of CICs. There was also no correlation between these events, indicating that high numbers of apoptotic cells do not promote CIC formation . Additionally we demonstrated in our previous studies that phagocytosis by nonprofessional phagocytes including primary fibroblasts, pancreatic adenocarcinoma, glioblastoma and head and neck cancer cells serves as a mechanism to remove primary necrotic cells . While our data are consistent with a non-apoptotic cell-in-cell process they do not rule out other scenarios previously described in literature. For instance, Huang et al. identified four subtypes of CICs which could be formed homotypically among tumor cells or heterotypically between tumor cells and leukocytes. They also stated that macrophages could target live tumor cells to generate CICs . As mentioned above, entosis is described a nonapoptotic cell death program in matrix-detached cells initiated by the invasion of one living cell into another . Wang et al. report “emperitosis”, an apoptotic cell-in-cell death process that occurs in heterotypic immune killer cells expressing granzyme B inside tumor cells which may serve as an in-cell danger sensation model to prevent the killing of target cells from inside, implying a unique mechanism for tumor cells to escape from immune surveillance .
In our study, CICs turned out highly significant for overall-survival whereby high CIC occurrences were associated with poor survival rates. Going in line with other published data, this may support the possibility of CIC formation stimulating tumor survival and progression by inducing aneuploidy and promoting nutrient scavenging [5, 12]. Concomitantly, differing assumptions on this issue have been provided and Huang et al. point out that CIC profiles may vary from tumor to tumor which may indicate different malignant stages and/or inflammatory conditions [12, 14]. On the ground of our previous findings we can support the proposition that each tumor type may harbor its own unique CIC pattern [10, 12].
Kepp et al. coined the term “immunogenic cell death” where multiple stimuli can trigger cell death that does not go unnoticed by the immune system . Accordingly, immunogenic cell death is accompanied by the production of a series of immunostimulatory damage-associated molecular patterns (DAMPs), i.e. molecules that are released or exposed during cytoprotective stress responses or upon cell death . Emission of DAMPs promotes the recruitment of antigen presenting cells to sites of ongoing immunogenic cell death and assists to their ability to take up dead cell-derived particulate material, as well as their capacity to prime an adaptive immune response . Assuming that CICs engulf necrotic cells which would under normal circumstances release DAMPs, an immunologic reaction to the tumor could thereby be inhibited. Accordingly, occurrence of CICs as an adverse prognosticator by attenuating anti-tumoral immune responses would be coherent.
The versatile and partly antithetic observations made on CICs convey that CIC represents a highly complex form of intercellular interaction. Conceivably, various forms of CICs coexist affecting tissues and tumor progression in different ways. Elucidation of the mechanisms underlying these observations is essential to further examine the significance of cell-in-cell in healthy tissue and pathological processes.
Our immunohistochemical analysis shows that CICs and senescence, taken together, occur with similar frequency in head and neck squamous cell carcinomas as apoptosis. We demonstrated that CICs have a significant impact on clinical outcome pre-RCT as opposed to senescent and apoptotic cells. We speculate that CICs impair the production of DAMPs by taking up necrotic cells inhibiting an adequate anti-tumoral immune response which is reflected by the adverse prognosis of patients with high CIC occurrences. CIC frequency might therefore be a valuable independent prognostic factor which could be taken into consideration for diagnostics and treatment strategies.
Damage-associated molecular pattern molecules
Head and neck squamous cell carcinoma
Nomenclature Committee on Cell Death
No evidence of disease
We thank Rudolf Jung and Christa Winkelmann for excellent technical assistance. We thank the Tumor Center at the Friedrich-Alexander University of Erlangen-Nürnberg, Erlangen, Germany for providing us with patient data. The present work was performed in fulfillment of the requirements for obtaining the degree “Dr. med.”.
Availability of data and materials
Clinical data were obtained from the Erlangen Tumor Center Database and extracted from the patients’ records. Tumor tissue samples were obtained from the Department of Pathology of the University of Erlangen-Nuremberg.
HMS analyzed and interpreted the patients’ data, performed the histological examination of the tumor tissue samples and the statistical analysis. MBH performed the stainings of the tissue samples. RF designed the study and provided materials and analysis tools. LVD designed the study, analyzed the data and provided the statistical analysis tools. All authors read and approved the final manuscript.
The authors declare that they have no competing interests.
Consent for publication
Ethics approval and consent to participate
All patients signed a “front door” informed consent allowing collection of their tissue and clinical data. The study was approved by the Ethics Review Committee of the University Hospital Erlangen-Nürnberg, Erlangen, Germany.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
- Kroemer G, Galluzzi L, Vandenabeele P, Abrams J, Alnemri ES, Baehrecke EH, Blagosklonny MV, El-Deiry WS, Golstein P, Green DR, et al. Classification of cell death: recommendations of the Nomenclature Committee on Cell Death 2009. Cell Death Differ. 2009;16:3–11.View ArticlePubMedGoogle Scholar
- Galluzzi L, Bravo-San Pedro JM, Vitale I, Aaronson SA, Abrams JM, Adam D, Alnemri ES, Altucci L, Andrews D, Annicchiarico-Petruzzelli M, et al. Essential versus accessory aspects of cell death: recommendations of the NCCD 2015. Cell Death Differ. 2015;22:58–73.View ArticlePubMedGoogle Scholar
- Yeh CK. Cellular senescence and aging. Oral Dis. 2016;22;587-90.
- Kroemer G, Perfettini JL. Entosis, a key player in cancer cell competition. Cell Res. 2014;24:1280–1.View ArticlePubMedPubMed CentralGoogle Scholar
- Overholtzer M, Mailleux AA, Mouneimne G, Normand G, Schnitt SJ, King RW, Cibas ES, Brugge JS. A nonapoptotic cell death process, entosis, that occurs by cell-in-cell invasion. Cell. 2007;131:966–79.View ArticlePubMedGoogle Scholar
- Wang S, Guo Z, Xia P, Liu T, Wang J, Li S, Sun L, Lu J, Wen Q, Zhou M, et al. Internalization of NK cells into tumor cells requires ezrin and leads to programmed cell-in-cell death. Cell Res. 2009;19:1350–62.View ArticlePubMedGoogle Scholar
- Lozupone F, Fais S. Cancer cell cannibalism: a primeval option to survive. Curr Mol Med. 2015;15:836–41.View ArticlePubMedGoogle Scholar
- Krishna S, Overholtzer M. Mechanisms and consequences of entosis. Cell Mol Life Sci. 2016;73:2379-86.
- Overholtzer M, Brugge JS. The cell biology of cell-in-cell structures. Nat Rev Mol Cell Biol. 2008;9:796–809.View ArticlePubMedGoogle Scholar
- Schwegler M, Wirsing AM, Schenker HM, Ott L, Ries JM, Buttner-Herold M, Fietkau R, Putz F, Distel LV. Prognostic value of homotypic cell internalization by nonprofessional phagocytic cancer cells. Biomed Res Int. 2015;2015:359392.View ArticlePubMedPubMed CentralGoogle Scholar
- Schwegler M, Wirsing AM, Dollinger AJ, Abendroth B, Putz F, Fietkau R, Distel LV. Clearance of primary necrotic cells by non-professional phagocytes. Biol Cell. 2015;107:372–87.View ArticlePubMedGoogle Scholar
- Huang H, Chen A, Wang T, Wang M, Ning X, He M, Hu Y, Yuan L, Li S, Wang Q, et al. Detecting cell-in-cell structures in human tumor samples by E-cadherin/CD68/CD45 triple staining. Oncotarget. 2015;6:20278–87.View ArticlePubMedPubMed CentralGoogle Scholar
- Wang S, He MF, Chen YH, Wang MY, Yu XM, Bai J, Zhu HY, Wang YY, Zhao H, Mei Q, et al. Rapid reuptake of granzyme B leads to emperitosis: an apoptotic cell-in-cell death of immune killer cells inside tumor cells. Cell Death Dis. 2013;4:e856.View ArticlePubMedPubMed CentralGoogle Scholar
- Gupta K, Dey P. Cell cannibalism: diagnostic marker of malignancy. Diagn Cytopathol. 2003;28:86–7.View ArticlePubMedGoogle Scholar
- Kepp O, Senovilla L, Vitale I, Vacchelli E, Adjemian S, Agostinis P, Apetoh L, Aranda F, Barnaba V, Bloy N, et al. Consensus guidelines for the detection of immunogenic cell death. Oncoimmunology. 2014;3:e955691.View ArticlePubMedPubMed CentralGoogle Scholar