Cell-in-cell (CIC) is a structure form wherein one viable cell is internalized into another cell, i.e., the internal cell (referred to as the target cell) is enclosed within a large vesicle in the external cell (referred to as the host cell) (1–3). The presence of lymphocytes within intestinal epithelial cells was first reported in 1864 (4), which led to the discovery of cell-in-cell structures (CICs). Despite being identified over a century ago, CIC has been traditionally viewed as a pathological feature that is primarily investigated by researchers in the field of pathology. With the advancements in microscopy techniques and cellular microscopic examination methods, an increasing number of CICs have been observed in both pathological and physiological conditions, with their most frequent occurrence being in cancer tissues (5). Thus far, numerous studies have explored the significance and mechanisms of CIC in diverse biological processes. CIC plays a crucial role in embryonic development, viral infection, immune homeostasis in the liver, and tumor development. Given the escalating rate of cancer-related mortality in recent years, coupled with the intricate underlying mechanisms of tumorigenesis, our focus is on the classification and significance of CICs in cancer. Although the formation of cell–cell interactions results in CIC structures with similar appearances, the classification of CICs into homotypic and heterotypic CICs is determined by the target cell types involved ( Figure 1 ). Homotypic CICs are formed through internalization among tumor cells, where one tumor cell is internalized by another, whereas heterotypic CICs are formed by the internalization of immune cells into tumor cells whereby an immune cell resides within a tumor cell. There are distinct molecular mechanisms involved in the initiation of target-cell internalization into host cells among different subtypes of CIC. Most target cells within CICs are degraded to facilitate tumor progression through, for example, energy reprogramming (6) and immune escape (7). Therefore, it can be inferred that CICs are closely related to tumor diagnosis and prognosis.

Chemoradiotherapy, immunotherapy, and molecular-targeted drugs have emerged as promising therapeutic modalities for preventing tumor recurrence. However, despite these advancements in tumor treatment, a significant number of patients still exhibit non-response and drug resistance to these therapies (5, 8, 9). The mechanisms of non-response to tumor therapy and drug resistance, including the effects of CIC on anticancer agents, have been extensively studied. To date, however, no review has been conducted on the implications of CIC in cancer treatment. In the present article, we review the impact of CICs on anticancer therapy and aim to provide evidence for future research and the clinical treatment of CIC in tumors.

CICs were first observed in tumors by Steinhaus in 1891 (10). In the following year, the phenomenon of tumor cells engulfing other tumor cells was also reported by H. Stroebe (11). Because this phenomenon was identical to phagocytosis, it was referred to as “cell phagocytosis” (12). Subsequently, Leyden described the CIC phenomenon in the tumor as a “bird’s eye cell” and termed it “cell cannibalism,” wherein a tumor cell with a large crescent-shaped nucleus contains a small tumor cell in its cytoplasm (13). In 1925, Lewis first experimentally demonstrated the viability of internalized cells in CIC; this enabled the differentiation of CIC from phagocytosis, in which the phagocytes internalize only dead cells or dying cells (3). Later, Ugini et al. found that metastatic melanoma cells could engulf lymphocytes. This phenomenon was described as “heterotypic cell cannibalism,” whereas the cell cannibalism process occurring between two tumor cells was termed “homotypic cell cannibalism” (14). In 1956, Humbel observed that lymphocytes could invade tumor cells and yet remain intact; the author termed this phenomenon “emperipolesis”, which is a Greek word meaning “inside round about wandering” (15). Presently, emperipolesis is often used as a general term to describe the movement of a living cell into another cell; this is because a variety of specific names have been used according to the different mechanisms involved in CIC formation (5). In 2007, a non-apoptotic mechanism of programmed cell death was discovered; this process was termed “entosis,” a Greek word meaning “inside” or “into”. In this process, once the cells detach from the matrix, they actively invade the neighboring host cells, thereby inducing the host cells to undergo degradation or release from the matrix (16). In 2013, Wang et al. demonstrated that live cytotoxic immune cells were engulfed by tumor cells, and the phenomenon of internalized cell death occurred through apoptosis; the authors termed this phenomenon “emperitosis,” which was derived from “emperipolesis” and “apoptosis” (17).

CIC exists in various cancers, thus indicating its relevance in cancer development. Under nutrient-replete conditions, such as glucose starvation and amino acid starvation, the rapid proliferation and metastasis of tumor cells require more energy support; consequently, the tumor cells degrade and digest their sibling or heterotypic cells by engulfing these cells and reusing their metabolites as their own nutrients to promote their survival and proliferation (14, 32, 37). This feature may be related to the reprogramming of tumor metabolism—a hallmark of cancer. Meanwhile, CIC enhances glucose metabolism through the transfer of mitochondria (6), which may contribute to the reprogramming of tumor metabolism—which is also a hallmark of cancer—and promote tumor malignancy. In heterotypic CIC, immune cells are engulfed and eliminated by tumor cells, thereby resulting in the immune escape and survival of tumor cells (38). During CIC formation, the target cells enter the host cells through myosin contraction and are killed. The stiffness of a cell is determined by the structure of its cytoskeleton, with cells that are softer exhibiting greater deformability, a property that determines whether they can traverse through the interstitial space between the fibrous stroma and the narrow endothelium of blood vessels. Therefore, RhoA and actomyosin can determine the deformability of the cell. The host cells have a stronger deformability, whereas the target cells are more rigid. This promotes competition between tumor cells, and the survival of highly deformable cells is associated with a higher metastatic capacity (39). This finding indicates that tumor cells with stronger deformability are more likely to survive and that CIC can facilitate clonal selection during tumor development. The internalized cells can persist in the outer cells and even undergo mitosis; this persistence can interrupt the mitosis of the outer cells, leading to the destruction of the contraction ring and the production of binuclear tumor cells (16). Ultimately, these cell behaviors result in genomic instability and the accumulation of mutations in tumor cells. Genomic instability is a hallmark of tumor development and promotes tumor progression and invasion. Entosis can also eliminate tumor cells detached from the stroma and inhibit their metastasis, which affects tumor suppression. A recent study suggested that genetic transfer, such as cell fusion and phagocytosis of apoptotic bodies, render the acquisition of phenotypes such as tumor aggressiveness and drug resistance. CIC formation between tumor cells results in cell clones associated with the genetic transfer and gain of malignancy by allowing the assimilation of aggressive phenotypes from distinct coexisting subpopulations; this indicates a new mechanism that allows rapid tumor development (40). As mentioned earlier, the promotive or suppressive effects of CIC may be related to the diagnosis and prognosis of tumors.

Immunotherapy, chemoradiotherapy, and targeted therapy have been widely used in tumor management strategies, including both systemic and local treatment methods. Cancer recurrence and the resistance of cancer cells to anticancer drugs are important causes of treatment failure and remain major challenges for anticancer treatment (9). Overcoming these challenges is essential to achieve long-lasting and effective cancer treatment and to prevent cancer recurrence. The mechanisms of anticancer-drug resistance and cancer recurrence have been investigated but have not yet been fully characterized.

The internalized cells in homotypic CIC can escape in an intact state after entering the cells; this feature may afford protection to the internalized cells in an unfavorable therapeutic microenvironment, such as exposure to a chemotoxic drug and immune surveillance. Because immune checkpoint inhibitors have shown limited effects in most mouse models, Kellie Smith et al. (60) conducted in-vivo experiments where human breast, colon, and melanoma tumor cells were incubated with pre-activated allogeneic T cells. The results demonstrated that the majority of individual cells underwent cell death whereas surviving tumor cells formed organized CIC structures. In-vitro studies on melanoma tumors treated with autologous reactive T-infiltrating lymphocytes revealed a high frequency of CIC formation in relapse tumors. This is due to the induction of phosphorylation in tumor cells by cytotoxic T-cell immunotherapy, which specifically targets the transcription factors signal transducer and activator of transcription 3 (STAT3) and early growth response-1 (EGR-1) and subsequently promotes the formation of CIC. Therefore, to mitigate tumor resistance toward subsequent drugs, it is recommended that these factors are inhibited prior to immunotherapy. Similarly, in the targeted treatment of prostate cancer with nintedanib (61), prostate cancer cells acquired resistance to nintedanib and survived by activating entosis formed by the upregulation of E-cadherin and ROCK1/2 through the PI3K/CDC42 signaling pathway. Thus, during treatment with anticancer drugs, cancer cells hide inside each other to avoid destruction. Radiotherapy and chemotherapy can destroy tumors by inducing apoptosis and necrosis in tumor cells. Radiotherapy and chemotherapy can also induce tumor cells to enter a senescence state, that is, move from cell cycle arrest into cellular quiescence (62). Doxorubicin induces senescence in p53 wild-type breast cancer cells, leading to the acquisition of a senescent phenotype. These senescent cells exhibit a significant enrichment for genes related to macrophage and phagocytosis, and efficiently engulf neighboring senescent or non-senescent tumor cells at a high frequency, forming numerous CICs. The target cells that undergo processing in the lysosome and subsequent breakdown confer an advantage to senescent host cells, which promotes the viability of senescent cancer cells. This ultimately leads to tumor persistence and poor patient survival (63). With this perspective in mind, Gottwald et al. (64) discovered a significant increase in CICs in rectal cancer tissues after neoadjuvant radiochemotherapy (RCT) compared with pre-RCT conditions. In addition, a notably high presence of senescent cells and CICs is associated with unfavorable clinical outcomes following RCT. It is speculated that CICs formed by senescent cells enhance cell viability and re-enter the cell cycle during antitumor treatment, thus providing additional energy and material for tumor recurrence and growth. This may be the cause of tumor recurrence and drug resistance during radiotherapy and chemotherapy.

Heterotypic CIC is formed by tumor cells and immune cells. In the tumor microenvironment, although NK cells and CD8⁺ T cells can kill tumor cells, they can also become engulfed by tumor cells, leading to their degradation and immune escape. Yun-Jeong et al. (25) have demonstrated a positive correlation between anticancer drug resistance and heterotypic cancer-initiating cell structures in EGFR-overexpressing lung cancer cells. The heterotypic CICs formed by non-small-cell lung cancer (NSCLC) cells and NK cells exhibited a higher proliferative capacity and more malignant phenotype than the non-CIC NSCLC cells. Following immunotherapy with NK-cell therapy and chemotherapy with cisplatin, doxorubicin, gemcitabine, etoposide, and docetaxel treatment, heterotypic CICs displayed lower sensitivity to NK cytotoxicity and greater resistance to anticancer drugs than non-CIC cancer cells.

Taken as a whole, the CIC phenomenon can be considered a mechanism of drug resistance and used to predict whether a patient would respond to anticancer therapies or show resistance to anticancer drugs.

Various studies have reported and highlighted the phenomenon and effect of CIC in tumors. Here, we discussed the mechanism of occurrence of two subtypes of CIC and the role of CIC in tumors. CIC has been frequently observed in various human tumor tissues (40). Thus, it can be considered a common feature in the tumor development process. We propose that CIC could be a new tumor marker that affects the prognosis and anticancer therapy of tumors. CIC formation can provide nutritional support to tumors, induce genomic instability to further promote a malignant phenotype, select populations with more survival advantages, suppress the immune microenvironment, evade immune cell attack, and promote tumor development and metastasis. Entosis can also act as a mechanism to eliminate abnormal cells, such as those with abnormal mitosis and loss of adhesion; thus, it could play a role in tumor suppression. We predict that entosis, similar to autophagy, has a dichotomous role; it plays contrasting roles at different times and in different contexts of tumors (65). Entosis can perform an inhibitory function by eliminating abnormal cells in tumorigenesis and the early stage of tumor development. As the tumor progresses, tumor cells use entosis to adapt to different environments, such as nutrient-replete conditions or the tumor immune microenvironment, thus promoting further tumor development.

Although we have a preliminary understanding of the mechanism of CIC formation, several questions that have not yet been answered remain. Deciphering these questions will enable us to further understand the role of CIC in clinical treatment. First, before the initiation of homotypic CIC formation, the mechanism for determining the identity of host and target cells remains unclear. It is currently believed that the degree of stiffness determines the tumor cells that will engulf or become engulfed (66). KRAS (39), c-Myc (67), mutant P53 overexpression (68), loss of CDKN2 (69), and CD44 downregulation (54) confer a “host” status to the tumor cells by making them softer. However, current knowledge of mechanisms of engulfment appears to be insufficient. Studies on CIC formation indicate that the stiffness of cells might not be the only factor leading to the invasion of soft cells and that other sophisticated mechanisms manipulating this process are present in tumor cells. For heterotypic CIC, the mechanisms responsible for the inversion of the host–target relationship between tumor cells and immune cells remain unknown. Second, based on the formation process, CIC can be classified into two types: (1) target cell-initiated and (2) host cell-initiated. The former entails the active invasion of host cells by target cells, and the latter entails the active engulfment of target cells by host cells. We predict that the formation of CIC requires the combined efforts of both the host and target cells, that is, the active engulfment of host cells and the cell entry momentum provided by target cells. Moreover, under normal conditions, only specialized phagocytes have the ability to phagocytose. The mechanism of acquisition of phagocytic capability by tumor cells remains undefined. Cell fusion can lead to gene transfer, and tumor cells may acquire the ability to phagocytose through fusion with macrophages. The most critical limitation is the lack of specific markers for differentiating homotypic cell cannibalism and entosis from heterotypic cell cannibalism and emperitosis in tumor tissues; this limitation should be promptly addressed for a greater understanding of the CIC phenomenon. Finally, due to the lack of a unified and specific counting method, CIC counting currently relies on manual counting, which may cause bias in the results. Tang et al. (70) proposed an automated identification method for specific CICs. Uniform high-throughput counting methods such as this are needed to improve the reliability of research results and provide more precise data support for clinical treatment.

Based on current studies, CIC can be considered a potential biomarker for tumor prognosis and resistance to anticancer agents. The anticancer treatment process can induce CIC formation and cause treatment recurrence and resistance. Thus, targeting CIC could open new avenues for cancer therapy. In the future, more large-scale studies on a variety of cancers are required to provide a more reliable basis for CIC formation and its role in tumor development.

Conceptualization: XL and JY; writing and original draft preparation: XL; writing, review, and editing: JY. All authors contributed to the article and approved the submitted version.