A new mechanism for autophagy regulation of anti-tumor immune responses

ABSTRACT

In this commentary I discuss a recent paper that describes a new mechanism for how macroautophagy/autophagy regulates the immune response to cancer, and relate it to other recent studies in this area. These recent developments may allow more effective strategies to manipulate autophagy to improve cancer therapy.

The autophagy literature often uses the term “double-edged sword” as a metaphor to convey the idea that autophagy can affect biological processes in competing ways [1]. This complexity can make it hard to predict how to manipulate autophagy for therapeutic purposes because an autophagy-focused intervention might inadvertently promote the opposite of the desired effect. The problem is important in cancer – it has been recognized for many years that autophagy can both promote and inhibit cancer [2,3] – yet, cancer is the disease where the application of therapeutics intended to manipulate autophagy has been most widely attempted [4]. However, double-edged swords are arguably more useful than single-edged ones; as has been pointed out in talks by Ravi Amaravadi, in cancer treatment we are usually trying to kill things (cancer cells). And, if your killing tool is a sword, one that is sharp on both sides and potentially lethal irrespective of which way you swing it could be a better tool than a sword with just one sharp side. To treat disease by applying our understanding of autophagy, we need to know when to swing our autophagy sword leftwards and when to swing it to the right.

A recent study [5] from Lorenzo Galluzzi’s lab exemplifies how we might do this. Yamazaki et al. set out to understand the role of autophagy in an important phenomenon in radiation oncology called the “abscopal response” [6]. During radiation therapy, an oncologist has an opportunity to direct therapy very precisely by targeting a specific region of the body where a tumor is present. (This is quite different from treatment with drugs, which are usually widely distributed throughout the body.) When such targeted radiation therapy is done, one sometimes sees eradication not just of the targeted tumor whose cells are directly killed by the radiation, but also a systemic response whereby nonirradiated tumor cells in a different part of the body are also killed. This effect where tumor cells that were not themselves subjected to radiation treatment are still killed is the abscopal response and is associated with better clinical responses and enhanced overall survival [6]. In animal models, abscopal responses can be studied by irradiating a tumor on one side of a mouse while monitoring the effect on an unirradiated tumor on the other side of the animal. This outcome allows the design of scientifically rigorous experiments (e.g., by having the irradiated and unirradiated tumors differ in their autophagy status). The biological basis for this effect is through targeting of the nonirradiated tumor cells by the adaptive immune system and, by combining other ways of enhancing immune control of cancer, it may be possible to boost the abscopal effect [7].

Unfortunately, abscopal responses are rare and it is unclear how to increase the likelihood of one. Thus, it is exciting that the new paper from Yamazaki et al. reports that inhibiting autophagy, by inactivation of the Atg5 or Atg7 genes in the irradiated target cells, increases abscopal responses in cancer cells that were never irradiated. This occurs when the unirradiated tumor cells are autophagy competent, indicating that the autophagy machinery is regulating the ability of an irradiated dying cancer cell in one location to control how effectively the immune system can kill tumor cells elsewhere in the body. Importantly, the authors provide mechanistic insight into how this happens. They found that in irradiated tumor cells, autophagy inhibits type I IFN secretion. Thus, when autophagy is inhibited they see more IFN secretion, which activates a better abscopal anti-tumor immune response. This occurs due to release of mitochondrial DNA (mtDNA) into the cytoplasm, which activates the CGAS-STING1 pathway. The conclusion is that autophagy suppresses the response by degrading permeabilized mitochondria, thus limiting cytosolic release of mtDNA and reducing activity of the STING1 pathway, leading to IFN secretion. Additionally, the authors provide translational evidence suggesting that the mechanism may apply in patients by demonstrating that an autophagy gene signature in the tumor inversely correlates with the apparent mitochondrial abundance and type I IFN signaling, and that this is correlated with prognosis in breast cancer patients.

The Yamazaki et al. paper [5] adds to an accumulating body of work suggesting that autophagy is critical in how the immune system targets cancer. Another recent paper [8] from the Perera and Kimmelman laboratories reports that in pancreas cancer, NBR1-mediated selective autophagy degrades MHC-I, reducing tumor antigen presentation and limiting the anti-tumor immune response. And, as with the study of abscopal responses to radiation, inhibition of autophagy by either genetic of pharmacological approaches can combine with immune checkpoint inhibitors to enhance tumor control and, sometimes, completely eradicate tumors. In these examples, autophagy is critical in the tumor cells themselves. Autophagy is also critical in immune cells. For example, a recent study from Julian Lum’s lab [9] demonstrated that inhibiting autophagy can enhance tumor control through regulation of T cell metabolism. While Doug Green’s lab [10] demonstrated that inhibiting LC3-associated phagocytosis/LAP leads to enhanced tumor control via STING1-mediated control of IFN release in macrophages. These, and other studies, suggest that we should swing our autophagy sword in the direction of inhibiting autophagy in order to enhance the way that the immune system can target cancer cells. Indeed, in the studies from Kimmelman and Perera [8], they found that systemic (though incomplete and nonspecific) pharmacological inhibition of autophagy with chloroquine is better at inhibiting tumor growth and combining with immune checkpoint inhibitors than more complete genetic inactivation of autophagy in the tumor cells alone. This result suggests that targeting autophagy in both tumor cells and the rest of the body, including immune cells is more effective because it potentially takes advantage of multiple mechanisms.

However, our double-edged sword metaphor still applies because there is also an extensive body of literature over several years showing that autophagy can be required during cell death in order for a dying tumor cell to activate the immune system [11]. Indeed, this mechanism can be enhanced when one swings the autophagy sword the other way, in the enhancing autophagy direction, using caloric restriction mimetics [12]. This raises another critical issue; it is important not just whether or not we kill cancer cells; the way that they are killed is also important. One important aspect of this point is that the most common way that cells die is through caspase-dependent apoptosis. However, apoptosis tends to suppress rather than enhance immune responses [13], and has other oncogenic effects as well [14]. Consistent with this observation, caspases inhibit abscopal responses [15,16]. Autophagy, and the autophagy machinery, regulates the apoptotic sensitivity of cancer cells by controlling expression of BBC3/PUMA, which is a key protein involved in permeabilization of mitochondria [17], and by autophagic membranes serving as scaffolds for efficient caspase activation [18,19]. The autophagy machinery can also serve as a scaffold for a different type of cell death called necroptosis [20], which is much more effective at inducing an anti-tumor immune response than apoptosis [21,22]. And, the necroptosis pathway can also be enhanced through BBC3/PUMA-dependent mtDNA release and activation of CGAS-STING1 [23].

All these mechanisms presumably work together, and this complexity means that delivering on the idea of manipulating autophagy to enhance the results we want is a difficult problem. However, the new paper by Yamazaki et al., along with the other studies, also points a direction forward by suggesting ways to select patients and design treatment combinations to take better advantage of our double-edged autophagy sword. One such study might be to test for enhanced abscopal responses when radiation therapy is combined with autophagy inhibition and immune checkpoint inhibition in a tumor type, such as pancreatic cancer, where other mechanisms by which autophagy inhibition might enhance immune responses also apply. Ultimately such studies need to be tested in people, but the key is to not just randomly swing the sword when moving into the clinic but instead to design the study so that a defined swing in a specific direction is done under circumstances when it is most likely to make lethal contact with the target [24]. By understanding the biological mechanisms through which autophagy has its biological effects, we are better able to do that.

Funding Statement

This work was supported by the National Cancer Institute [R01 CA150925].

Disclosure statement

All authors declare no conflict of interest.