Autophagy in non-small cell lung carcinogenesis

Abstract

In a mouse model of non-small cell lung carcinogenesis, we recently found that the inactivation of the essential autophagy gene Atg5 causes an acceleration of the early phases of oncogenesis. Thus, hyperplastic lesions and adenomas are more frequent at early stages after adenoviral delivery of Cre recombinase via inhalation, when Cre—in addition to activating the KRas^G12D oncogene—inactivates both alleles of the Atg5 gene. The accelerated oncogenesis of autophagy-deficient tumors developing in KRas;Atg5^fl/fl mice (as compared with autophagy-competent KRas;Atg5^fl/+ control tumors) correlates with an increased infiltration by FOXP3⁺ regulatory T cells (Tregs). Depletion of such Tregs by means of specific monoclonal antibodies inhibits the accelerated oncogenesis of autophagy-deficient tumors down to the level observed in autophagy-competent controls. Subsequent analyses revealed that the combination of KRas activation and Atg5 inactivation favors the expression of ENTPD1/CD39, an ecto-ATPase that initiates the conversion of extracellular ATP, which is immunostimulatory, into adenosine, which is immunosuppressive. Pharmacological inhibition of ENTPD1 or blockade of adenosinergic receptors reduces the infiltration of KRas;Atg5^fl/fl tumors by Tregs and reverses accelerated oncogenesis. Altogether these data favor a model according to which autophagy deficiency favors oncogenesis via changes in the tumor microenvironment that ultimately entail the Treg-mediated inhibition of anticancer immunosurveillance.

Autophagy has a dual role in oncogenesis. Especially at late stages of tumor progression, cancer cells must be optimally fit to resist a hostile environment conditioned by hypoxia, as well as the absence of nutrients and trophic stimuli. However, in early oncogenesis, autophagy is often inhibited, leading to the speculation that autophagy might constitute an oncosuppressive mechanism.

Recently, we generated a model of non-small cell lung cancer, in which adenoviral delivery of the Cre recombinase (AdCre) can activate oncogenic KRAS (by excising a Lox-Stop-Lox cassette placed upstream of the KRas^G12D transgene) and inactivate either one or both alleles of floxed Atg5. Depending on the genetic background of the mice (KRas;Atg5^fl/fl or KRas;Atg5^fl/+), the resulting KRas-driven tumors hence are either autophagy-deficient or autophagy competent. In this model of KRas-driven lung carcinogenesis, autophagy clearly has a dual role in multistep oncogenesis. At early stages after inhalation of AdCre, the number of hyperplastic regions and adenomas and their volume are significantly increased in KRas;Atg5^fl/fl compared with control KRas;Atg5^fl/+ mice. However, at later stages (12–18 wk), autophagy is required for the progression of adenomas to adenocarcinomas, unless the Tp53 tumor suppressor gene is inactivated.

Given our interest in anticancer immunosurveillance, which describes the capacity of the immune system to eliminate or at least restrain the growth of cancers, we decided to characterize the impact of autophagy on the immune infiltrate in early hyperplastic region adenomas (2 wk post infection with AdCre). Although the number of myeloid cells and total CD3⁺ T lymphocytes infiltrating KRas;Atg5^fl/fl and KRas;Atg5^fl/+ adenomas is similar, the frequency of FOXP3⁺ regulatory T cells is enhanced in the context of autophagy deficiency. This result was obtained by using 2 different methods, namely i) dissociation of the tumors into monocellular suspensions followed by immunofluorescence staining and cytofluorometric analyses, and ii) quantitative in situ immunohistochemistry. Both methods led to the conclusion that the density of Tregs, as well as the proportion of Tregs among total T cells, is elevated in autophagy-deficient tumors. In a subsequent series of experiments, we depleted Tregs by injection of an antibody specific for IL2RA/CD25 (interleukin 2 receptor, α) in KRas;Atg5^fl/fl and KRas;Atg5^fl/+ mice. This manipulation eliminated the advantage of autophagy-deficient tumors in early oncogenesis. An alternative approach to inactivate Tregs, by injecting a monoclonal antibody specific for FOLR4/FR4 [folate receptor 4, delta (putative)], had very similar effects. These findings lead to the conclusion that KRAS-induced lung carcinogenesis is subjected to immunosurveillance if the tumors are autophagy-competent. However, autophagy-deficient cancer cells induce changes in the tumor microenvironment that favor infiltration by Tregs and consequent subversion of immunosurveillance.

What then are the mechanisms that link autophagy deficiency in tumor cells to increased Treg infiltration? Microarray and protein expression analyses of AdCre-infected Atg5^fl/fl, Atg5^fl/+, KRas;Atg5^fl/fl and KRas;Atg5^fl/+ pneumocytes revealed that the combination of Kras expression plus Atg5 knockout stimulates HIF1A/Hif1α [hypoxia inducible factor 1, α subunit (basic helix-loop-helix transcription factor)]-dependent activation of ENTPD1, an ecto-enzyme that is exposed on the cell surface and catalyzes the conversion of ATP into ADP and AMP. ENTPD1 hence initiates the degradation of extracellular ATP, which has immunostimulatory properties, into adenosine (generated from AMP through the action of another ecto-enzyme, NTSE/CD73), which has immunosuppressive effects (Fig. 1). While ATP stimulates purinergic receptors to attract antigen-presenting cells into the proximity of stressed tumor cells, adenosine binds adenosinergic receptors to facilitate the recruitment and function of Tregs. Driven by this knowledge, we performed experiments in which ENTPD1 was inhibited pharmacologically by injection of polyoxometalate-1, or in which adenosinergic receptors were blocked by means of PSB1115. Both manipulations had a similar effect; they eliminate the difference in Treg numbers in the early hyperplastic tumor regions of KRas;Atg5^fl/fl and KRas;Atg5^fl/+ mice. Moreover, both treatments reduce the frequency and sizes of KRas;Atg5^fl/fl tumors down to the level of KRas;Atg5^fl/+ controls. Altogether, these results suggest the existence of a pathway that links autophagy deficiency to Treg infiltration via ENTPD1-mediated conversion of immunostimulatory ATP into immunosuppressive adenosine (Fig. 1). Of note, additional mechanisms for how autophagy deficiency couples to increased intra-tumoral Tregs and hence reduced anticancer immunity most likely exist and need to be explored in future experiments.

Figure 1. Consequences of autophagy deficiency on the immunosurveillance system and possible corrective measures. As compared with autophagy-competent tumors, autophagy-deficient KRas-driven tumos overexpress ENTPD1, an ectoenzyme that initiates the conversion of immunostimulatory ATP into immunosuppressive adenosine (A). Via its action on ADORA1/2 adenosine receptors, adenosine favors the infiltration of tumor by immunosuppressive regulatory T cells (Tregs). This cascade offers a range of targets for therapeutic intervention (B). ENTPD1 can be inhibited by pharmacological agents exemplified by polyoxometalate-1 (POM1). NTSE inhibitors are being developed by the pharmaceutical industry. Adenosinergic receptors can be blocked by specific antagonists including PSB1115. Finally, Tregs can be depleted by monoclonal antibodies targeting IL2RA or FOLR4.

We have found previously that autophagy deficiency induced by knockdown of Atg5 or Atg7, as well as transfection-enforced overexpression of ENTPD1 in transplantable tumors subverted chemotherapy-induced immunosurveillance. Our present findings that knockout of Atg5 upregulates ENTPD1 and subverts natural immunosurveillance in oncogene-induced cancers are compatible with these former findings. It will be interesting to know whether these observations, which have been obtained in the realm of tumor immunology, may be applicable to infectious diseases as well. If so, one may postulate that viruses or intracellular bacteria that subvert the autophagic machinery will cause the infected cells to emit signals inducing the inopportune recruitment of Tregs, thereby suppressing the pathogen-specific immune response.

Rao S, Tortola L, Perlot T, Wirnsberger G, Navatchkova M, Nitsch R, Sykacek P, Frank L, Schramek D, Komnenovic V, et al. A dual role for autophagy in a murine model of lung cancer. Nat Commun. 2014 doi: 10.1038/ncomms4056.

10.4161/auto.27643