Abstract

   Pyroptosis is a recently defined form of immunogenic cell death that
   shows great promise in cancer immunotherapy. However, almost all
   small‐molecule pyroptosis‐inducing agents (PyAs) reported to date
   indiscriminately induce pyroptosis in multiple cell types, leading to
   off‐target pyroptotic death of normal and immune cells. One promising
   approach to addressing this biosafety issue is the design of
   conditionally activatable PyAs that specifically respond to disease
   biomarkers. Herein, a general solution for facilely tailoring and
   synthesizing activatable PyAs based on a newly developed class of
   photoactive PyAs, termed PyPSs, is reported. The unique structurally
   encoded properties of PyPSs, including endo reticulum targeting,
   hypoxia tolerance, and sensing properties, excitingly meet a demanding
   set of performance requirements for constructing conditionally
   activatable PyAs for cancer immunotherapy. Based on PyPS‐1 scaffold,
   hypoxia‐activatable PyPS‐NF as a proof‐of‐concept example is prepared,
   demonstrating specific hypoxia‐activated pyroptotic cell death and
   favorable immunotherapeutic efficacy of solid tumors. Herein, a general
   design strategy for tailoring activatable PyAs to precisely control
   GSDME‐mediated cell pyroptosis is established, with great potential to
   advance cancer immunotherapy.

   Keywords: cell pyroptosis, immunogenic cell death, photosensitizers,
   stimuli-responsiveness, tumor immunotherapy
     __________________________________________________________________

   A unified design to prepare hypoxia‐tolerant and stimuli‐activatable
   pyroptosis‐inducing agents (PyAs) is established, opening an avenue to
   tailor and synthesize PyAs capable of inducing pyroptosis in a
   controlled manner, which is essential for advancing pyroptosis‐inducing
   cancer immunotherapy.

   graphic file with name SMSC-4-2300135-g007.jpg

1. Introduction

   Pyroptosis triggered by the inflammatory caspase‐mediated cleavage of
   gasdermin (GSDM) proteins is a recently defined immunogenic form of
   programmed cell death.^[ [34]^1 , [35]^2 , [36]^3 , [37]^4 , [38]^5 ^]
   It is characterized by plasma membrane pore formation, cell swelling
   with large bubbles, and massive leakage of cellular proinflammatory
   cytokines. Compared with traditional apoptosis and necrosis, pyroptosis
   is more liable to trigger cellular inflammation, showing great
   potential in cancer immunotherapy.^[ [39]^6 , [40]^7 , [41]^8 , [42]^9
   , [43]^10 ^] Key to this effort was the development of multiple
   pyroptosis‐inducing agents (PyAs) that have demonstrated the efficacy
   of inflammatory pyroptosis in cancer therapy.^[ [44]^11 , [45]^12 ,
   [46]^13 , [47]^14 , [48]^15 ^] Compared with conventional
   chemotherapeutic drug‐based PyAs, photoactive PyA possesses multiple
   significant advantages such as spatiotemporally controllable
   activation, no cross‐resistance with chemotherapeutic drugs, and
   minimal invasiveness. Canonical photoactive PyA requires oxygen to
   produce reactive oxygen species (ROS) to induce cell pyroptosis.
   However, severe O[2] deficiency prevalent in fast‐growing solid tumors,
   i.e., tumor hypoxia,^[ [49]^16 ^] substantially reduces the treatment
   outcome. Therefore, the development of photoactive PyAs with hypoxia
   tolerance is an urgent need.

   In addition, since the GSDM proteins are expressed in various normal
   tissues and immune cells in addition to cancer cells, the vast majority
   of photoactive PyAs reported so far induce pyroptosis indiscriminately,
   leading to off‐target pyroptotic death in normal tissues and cells.^[
   [50]^3 , [51]^10 ^] Therefore, the development of PyAs that can
   specifically induce pyroptosis in a controlled manner is expected to
   improve biosafety and is crucial for advancing pyroptosis‐inducing
   cancer immunotherapy.^[ [52]^17 , [53]^18 , [54]^19 ^] One promising
   approach is to engineer conditionally activatable PyAs that can
   specifically respond to disease biomarkers. Considering the complex and
   diverse tumor microenvironment, establishing a general platform to
   tailor oxygen‐tolerant and stimuli‐activatable PyAs in response to
   variable tumor biomarkers is a timely exercise.

   In this work, we have identified a unique class of photoactive PyAs,
   termed PyPSs, with fascinating structurally encoded properties that
   meet a demanding set of performance requirements for constructing
   conditionally activatable PyAs for cancer immunotherapy. First, the
   intrinsic endo reticulum (ER)‐targeting property of PyPSs enables
   significant GSDME‐mediated pyroptosis. Second, PyPSs‐induced pyroptosis
   occurs under both normoxic and hypoxic conditions, and this remarkable
   hypoxia tolerance allows immunotherapy of solid tumors characterized by
   hypoxia. Finally, and most strikingly, flexible phenol caging/uncaging
   chemistry allows one to tailor multiple conditionally activatable PyAs
   via installing PyPSs with various caging groups (CG) for precise cancer
   immunotherapy. Based on PyPS‐1 scaffold, we prepared a
   hypoxia‐activatable paradigm, PyPS‐NF, showing specific
   hypoxia‐activated pyroptotic cell death and antitumor efficacy. This
   study establishes a universal strategy, for the first time, to
   construct a variety of conditionally activatable PyAs, thereby
   providing a rich molecular toolbox for the precise control of cell
   pyroptosis.

2. Results and Discussion

2.1. Synthesis and Photophysical/Chemical Properties of PyPS Molecules

   Three putative PyAs (PyPS‐1, PyPS‐2, and PyPS‐3) with D‐π‐A structure
   (Figure [55]1a) were synthesized via Knoevenagel condensation (see the
   Supporting Information). Sulfur‐substituted
   dicyanomethylene‐4H‐chromene was chosen as a typical electron‐accepting
   group, in which the sulfur atom is expected to promote the intersystem
   crossing (ISC) process and thus enhance the generation of T[1] state.^[
   [56]^20 , [57]^21 , [58]^22 , [59]^23 ^] Phenolate groups substituted
   by halogen atoms at the ortho‐positions were adopted as
   electron‐donating moieties. As depicted in Scheme S1, Supporting
   Information, the phenolic hydroxyl group has to be ionized into an
   anionic phenate state to promote the intramolecular charge transfer
   (ICT) process,^[ [60]^24 , [61]^25 ^] which is the prerequisite for
   achieving photoactivity. Gratifyingly, as shown in Figure [62]1b and
   Figure S2, Supporting Information, all three PyPSs exhibited intense
   visible ICT bands and fluorescence emission at physiological pH 7.4.
   This is due to the pronounced electron‐withdrawing −I halogen effect of
   the iodine substituents, which significantly reduced the pK [a] value
   of the phenolic hydroxyl group (Figure [63]1c, Figure S3, Supporting
   Information, and Table [64]1 ).^[ [65]^26 , [66]^27 ^]

Figure 1.

   Figure 1
   [67]Open in a new tab

   Photophysical/chemical properties of PyPS molecules and their ability
   to generate superoxide anion radical (O[2] ^•−) and induce pyroptosis
   in living cells. a) Schematic illustration of chemical structures of
   PyPS molecules and their photosensitivity under both normoxic (21%
   O[2]) and hypoxic (2% O[2]) conditions. b) UV–vis absorption and
   fluorescence emission of PyPS‐1 in aqueous medium at pH 7.4. c)
   pH‐dependent changes in the absorbance of the ICT‐bands of PyPS
   molecules. d) Time‐dependent fluorescence emission spectra of DHE
   within 2.5 μm aqueous solution (pH 7.4, 40 μg mL^−1 DNA‐ct) of PyPS‐1
   upon light irradiation. e) Fluorescence enhancement of DHE induced by
   the photogenerated O[2] ^•− of PyPS molecules (n = 3). f) EPR spectra
   of PyPS molecules in the presence of DMPO upon light irradiation. g)
   Confocal fluorescence imaging of photogenerated ROSs by PyPS‐1 in MCF‐7
   cells under (left) normoxic and (right) hypoxic conditions by using
   DCF‐DA and DHE as the general ROS and O[2] ^•− fluorescence indicators,
   respectively. Scale bar: 50 μm. h) IC[50] values of different PyPS
   molecules under normoxia and hypoxia (n = 4). i,j) Representative phase
   contrast images of MCF‐7 cells incubated with different PyPS molecules
   in dark or upon light irradiation under normoxic (i) and hypoxic (j)
   conditions. The red arrows indicate the characteristic large bubbles
   accompanying pyroptosis. Scale bar: 20 μm. The light source was a Xe
   lamp (490–700 nm) with the power fixed at 5 and 10 mW cm^−2 for
   solution and cell experiments, respectively.

Table 1.

   Properties of three PyPS molecules
   Sample[68]^a) λ [abs] [nm] ε [M^−1 cm^−1] λ [em] [nm] Φ [fl] [%] pK [a]
   PyPS‐1 603 38 070 730 2.03 5.23
   PyPS‐2 585 35 440 724 2.19 5.76
   PyPS‐3 573 44 180 685 14.1 5.78
   [69]Open in a new tab
   ^a)

   The experiments for determining fluorescence quantum yields and pK [a]
   values were performed in citric acid/Na[2]HPO[4] buffer solutions; the
   other data were collected in 10 mM PBS (pH 7.4). The results represent
   the average of three independent experiments.

   The photochemical properties of PyPS molecules were then studied. QDPBF
   (a water‐soluble 1,3‐diphenylisobenzofuran derivative) was used to
   assess the photosensitivity of PyPSs in generating ROS (Scheme S2,
   Supporting Information),^[ [70]^28 ^] and the obvious reduction of
   QDPBF absorption indicated their high efficiency in generating ROS
   (Figure S4, Supporting Information). Notably, the photosensitivity of
   PyPS‐1 was estimated to be approximately 2.4 times greater than that of
   methylene blue (MB), a widely used commercial photosensitizer, by
   recording the absorbance decrease of QDPBF at 415 nm (Figure S4f,
   Supporting Information). By further using ABDA as the specific ^1O[2]
   indicator (Scheme S3, Supporting Information), and only a small amount
   of ^1O[2] was detected during photosensitizing process of PyPSs (Figure
   S5, Supporting Information). In sharp contrast, as shown in Figure
   [71]1d and Figure S6, Supporting Information, PyPSs induced a drastic
   increase in the fluorescence emission of dihydroethidium (DHE), a
   fluorogenic probe of O[2] ^•− (Scheme S4, Supporting Information), and
   the fluorescence increased in the order of PyPS‐3 < PyPS‐2 < PyPS‐1
   (Figure [72]1e). EPR (electron paramagnetic resonance) spectroscopy
   analysis further identified the photogenerated O[2] ^•− (Scheme S5,
   Supporting Information), as evidenced by the characteristic 1:1:1:1
   quartet signal attributed to the DMPO‐O[2] ^•− adduct, i.e., DMPO‐OOH
   (Figure [73]1f).^[ [74]^29 ^] Similar to the DHE assay, the EPR signals
   of DMPO‐OOH increased in the order of PyPS‐3 < PyPS‐2 < PyPS‐1. The
   earlier results indicated that PyPSs mainly underwent type‐I pathway,
   which is less oxygen‐dependent compared to conversional type‐II
   pathway.^[ [75]^30 , [76]^31 ^]

   The difference in photosensitivity between the three PyPSs should be
   attributed to the diverse heavy atom effects of their substituted
   halogens (where I > Br > Cl).^[ [77]^32 ^] Time‐dependent density
   functional theory (TD‐DFT) calculations was then performed to elucidate
   the type‐I mechanism of PyPS‐1. As shown in Figure S7a, Supporting
   Information, the lowest energy gap (ΔE [ST]) between the singlet
   excited state (S) and triplet excited state (T) of PyPS‐1 was
   calculated to be 0.09 eV, which is much less than 0.3 eV, facilitating
   the singlet‐triplet ISC process to form ^3PyPS‐1* upon
   photoexcitation.^[ [78]^33 ^] On the other hand, the electron‐rich
   nature of anionic PyPS‐1 promotes the formation of PyPS‐1^•− from
   ^3PyPS‐1*.^[ [79]^34 , [80]^35 , [81]^36 ^] The formed PyPS‐1^•− then
   transfers the electron to oxygen to generate O[2] ^•−.^[ [82]^30 ,
   [83]^31 ^] The Gibbs free energy (ΔG [1]) of the electron transfer
   process was calculated to be −13.88 kcal mol^−1, indicating an energy
   favorable reaction (Figure S7b, Supporting Information). Given the
   remarkable type‐I photosensitivity, we then assessed the photodynamic
   therapy (PDT) performance of PyPS‐1 in living cells.

   The weak but visible fluorescence of PyPSs enabled the fluorescence
   monitoring of its internalization by MCF‐7 cells. Taken PyPS‐1 as
   example, as the molecule was uptaken by MCF‐7 cells, its fluorescence
   increased with time and reached a plateau at about 2 h (Figure S8,
   Supporting Information). Therefore, in subsequent cell experiments, the
   incubation time of PyPSs was fixed at 2 h. After confirming the low
   dark cytotoxicity of PyPS‐1 (Figure S9a, Supporting Information), we
   then assessed its photosensitivity in MCF‐7 cells. Dichlorofluorescein
   diacetate (DCF‐DA) and DHE were employed as universal ROS and specific
   O[2] ^•− fluorescent indicator, respectively.^[ [84]^20 , [85]^34 ^] As
   expected, cells treated with light alone or PyPS‐1 alone showed
   negligible DCF or DHE fluorescence (Figure [86]1g). In contrast, in the
   group treated with PyPS‐1/light, intense intracellular DCF and DHE
   fluorescence were observed, indicating significant photoactivity of
   PyPS‐1 in generation of O[2] ^•−. Encouragingly, PyPS‐1 remained strong
   photoactivity in cells incubated under hypoxic condition (2% O[2]).
   These observations of hypoxia‐tolerant photoactivity of PyPS‐1 were in
   good agreement with the results of ROS studies in solution.

   Emboldened by the earlier findings, we subsequently assessed the
   potential of PyPS‐1 to ablate cancer cells. Live/dead cell staining
   assay with Calcein AM and propidium iodide (PI) was carried out. As
   shown in Figure S10, Supporting Information, under both normoxic and
   hypoxic conditions, MCF‐7 cells treated with 1 μm PyPS‐1 were stained
   by PI after light irradiation, demonstrating the remarkable PDT
   efficiency of PyPS‐1. MTT assay of PyPS‐1 further demonstrated its
   effective PDT outcome under both normoxic and hypoxic conditions
   (Figure S9a, Supporting Information), with IC[50] values as low as
   0.083 and 0.106 μm, respectively (Table S1, Supporting Information).
   Although type‐I photosensitizers have a certain tolerance to hypoxia,
   their ability to generate ROS under hypoxic conditions is still
   slightly weaker than under normoxic conditions.^[ [87]^20 , [88]^30 ^]
   Remarkably, under hypoxic condition, the IC[50] value of PyPS‐1 was
   80‐fold lower than that of MB (Figure S9 and Table S1, Supporting
   Information), owing to its superb hypoxia‐tolerant type‐I
   photoactivity. Notably, MTT assay of PyPS‐2 and PyPS‐3 also
   demonstrated their good biocompatibility and effective PDT outcome
   (Figure S9, Table S1, Supporting Information). Not surprisingly, PyPS‐1
   displayed the strongest photocytotoxicity among all PyPSs no matter in
   normoxia or hypoxia (Figure [89]1h, Table S1, Supporting Information),
   which is consistent with the photosensitivity results in solution.

   Significantly, the dying MCF‐7 cells evoked by PDT exhibited typical
   features of pyroptosis with distinct cell swelling and large bubbles
   from the plasma membrane (Figure [90]1i). It was observed that the size
   and number of bubbles generated on the plasma membrane of cells treated
   with different PyPS molecules were positively correlated with their
   photosensitivity. When the concentrations of PyPS‐2 and PyPS‐3 were
   reduced to the same 1 μM as that of PyPS‐1, the cells treated with them
   produced significantly fewer and smaller bubbles on the plasma membrane
   than the PyPS‐1 group (Figure S11, Supporting Information). Annexin
   V/PI staining for MCF‐7 cells was then performed. As shown in Figure
   S12, Supporting Information, MCF‐7 cells after PyPS‐1/light treatment
   proceeded directly to the annexin V and PI double‐positive stage, in
   line with the previously reported results of cell pyroptosis.^[ [91]^8
   , [92]^11 ^] PyPS‐1‐triggered cell pyroptosis remained prominent under
   hypoxic conditions owing to its type‐I photoactivity (Figure [93]1j).
   Notably, light irradiation alone had no obvious effect on cells under
   both normoxic and hypoxic conditions (Figure S13, Supporting
   Information). Moreover, this PDT‐induced pyroptotic cell death is
   prevalent in other cancer cell lines, such as A549, HeLa, A375, and
   HepG2 cells (Figure S14, Supporting Information).

2.2. The Mechanism Underlying Pyroptosis Activation

   In view of the unusually drastic cell pyroptosis, we moved to explore
   the mechanism underlying pyroptosis activation (Figure [94]2a). Growing
   evidence suggests that organelle stress can induce pyroptosis,^[
   [95]^18 , [96]^37 ^] so we speculated that our PyPSs might specifically
   target and attack certain organelles. Through confocal laser scanning
   microscope, we were surprised to find that the three PyPSs were mainly
   localized in ER of MCF‐7 cells with Pearson's correlation coefficients
   above 0.85 (Figure [97]2b and S15, Supporting Information).
   Colocalization imaging of PyPS‐1 in different z‐axis sections was then
   performed, and the result showed that the spatial distribution of
   PyPS‐1 is highly consistent with the specific ER probe (ERTG; Figure
   S16, Supporting Information), further demonstrating that PyPS‐1 is
   specifically localized to the ER. This specific ER‐targeting property
   of PyPSs was also observed in other cancer cell lines, such as HeLa and
   HepG2 cells (Figure S17, Supporting Information). Notably,
   photosensitizers that can achieve specific ER targeting without
   chemically attaching additional targeting groups have been rare
   reported, and this structurally encoded ER targeting property can
   effectively reduce tedious chemical synthesis and molecular size.

Figure 2.

   Figure 2
   [98]Open in a new tab

   The mechanism underlying pyroptosis activation. a) Schematic
   illustration of the mechanism of the cell pyroptosis triggered by
   PyPS‐1. b) Colocalization studies of PyPS‐1 toward ER in MCF‐7 cells;
   cells were incubated with PyPS‐1 (2.0 μm) for 1 h. Pearson's
   correlation coefficients (PCCs) for ER and Mito are 0.95 and 0.69,
   respectively. c) Fluorescence imaging of Ca^2+ ions in MCF‐7 cells
   after different treatments using Fluo‐3 AM as a fluorescent Ca^2+
   indicator. d) Evaluation of the mitochondrial damage of MCF‐7 cells
   treated with PyPS‐1 (1 μm) under light irradiation in normoxic
   conditions by using JC‐1 as the fluorescent indicator. e) Western
   blotting of the cleavage of GSDME by caspase‐3 in MCF‐7 cells after
   treatment of PyPS‐1/light under normoxic conditions. c‐Caspase‐3,
   active Caspase‐3 after cleavage of pro‐Caspase‐3; GSDME‐FL, full‐length
   GSDME; GSDME‐N, the N‐terminal cleavage product of GSDME. f–h)
   Investigation of LDH (n = 4) (f), ATP (n = 4) (g), and HMGB‐1 (h)
   release in MCF‐7 cells after PyPS‐1/light treatment. i) Western
   blotting of the cleavage of GSDME in MCF‐7 cells after treatment of
   PyPS‐1 and light under hypoxic conditions. j) Representative phase
   contrast images of healthy 293 T or NIH‐3T3 cells after treatment of
   PyPS‐1/light. Scale bar in (c): 50 μm. Scale bars in other cell images:
   20 μm. Light source: Xe lamp (490–700 nm, 10 mW cm^−2).

   Cellular uptake pathway of PyPS‐1 was then investigated. As shown in
   Figure S18, Supporting Information, the cellular uptake of PyPS‐1
   significantly decreased by 85.5% when incubating at 4 °C, suggesting an
   energy‐dependent pathway of the cellular uptake. In addition,
   endocytosis inhibitor assays were employed, including genistein (GEN,
   an inhibitor of caveolae‐mediated endocytosis), methyl‐β‐cyclodextrin
   (Mβ‐CD, an inhibitor of lipid raft‐mediated endocytosis),
   chlorpromazine (an inhibitor of clathrin‐mediated endocytosis),
   5‐(N‐ethyl‐N‐isopropyl)‐amiloride (an inhibitor of macropinocytosis),
   and cytochalasin B (CytB, an inhibitor of phagocytosis and/or
   macropinocytosis). Among these inhibitors, GEN and Mβ‐CD resulted in a
   69.9% and 66.6% decrease in cellular uptake, respectively, indicating
   that PyPS‐1 was uptaken mainly via the caveolae‐mediated endocytosis
   pathway (Figure S18, Supporting Information). The earlier finding
   echoed previous reports that caveolae‐mediated endocytic pathways are
   capable of delivering cargos to the ER.^[ [99]^38 , [100]^39 ^]

   The ER‐targeting property of PyPS‐1 should enable concentrated ROS
   attacks, leading to severe ER stress, as demonstrated by the
   observation of a transfer of free Ca^2+ ions from the ER to the
   cytoplasm using a fluorescent Ca^2+ indicator Fluo‐3 AM (Figure [101]2c
   and S19, Supporting Information).^[ [102]^40 ^] It has been reported
   that Ca^2+ release caused by severe ER stress can promote
   ER‐mitochondria Ca^2+ flux and thus mitochondrial Ca^2+ overload,
   leading to decreased mitochondrial membrane potential and caspase‐3
   activation.^[ [103]^40 , [104]^41 , [105]^42 ^] JC‐1 staining
   demonstrated that PyPS‐1 lowered mitochondrial membrane potential upon
   light irradiation (Figure [106]2d),^[ [107]^43 ^] and Western blotting
   showed that caspase‐3 was activated in these cells and triggered
   GSDME‐mediated pyroptosis (Figure [108]2e). In contrast, a caspase‐3
   inhibitor Z‐DEVD‐FMK attenuated caspase‐3‐triggered GSDME cleavage,
   thereby inhibiting cell pyroptosis (Figure S20a,b, Supporting
   Information).^[ [109]^8 , [110]^44 ^] It echoed the previous reports
   that GSDME can convert caspase‐3‐mediated apoptosis to pyroptosis.^[
   [111]^7 , [112]^8 ^] Importantly, leakage of cellular contents
   including lactate dehydrogenase (LDH), ATP, and nuclear high‐mobility
   group box 1 protein (HMGB‐1) was also demonstrated (Figure
   [113]2f–h).^[ [114]^45 , [115]^46 ^] GSDMD is another well‐studied
   executor of pyroptosis, and it has been reported that ER‐targeted PDT
   could induce GSDMD‐mediated cell pyroptosis.^[ [116]^47 ^] However, no
   GSDMD cleavage was observed after PyPS‐1/light treatment via Western
   blotting (Figure S20c, Supporting Information). Collectively, this work
   demonstrates, to our knowledge, the first example of significant
   GSDME‐mediated pyroptosis initiated by ER‐targeted PDT.

   In addition, under hypoxic conditions, PyPS‐1‐triggered cell pyroptosis
   via the same mechanism as under normoxic conditions (Figure [117]2i,
   S21 and S22, Supporting Information). It is nevertheless noteworthy
   that PyPS‐1‐induced pyroptosis was indiscriminate, as the GSDM proteins
   are expressed in various cell types.^[ [118]^3 , [119]^10 ^] Distinct
   pyroptosis occurred in healthy 293 T and NIH 3T3 cell lines following
   PDT treatment with PyPS‐1 (Figure [120]2j). Likewise, PyPS‐2 and PyPS‐3
   also induced significant pyroptosis in these healthy cells (Figure S23,
   Supporting Information). Therefore, from the perspective of cancer
   treatment, adequate derivatizing PyPS‐1 with function of pyroptotic
   killing specific to cancers rather than normal tissues is pivotal.

2.3. Chemical Design of Activatable PyAs

   In light of the flexible phenol caging/uncaging chemistry,^[ [121]^26 ,
   [122]^48 , [123]^49 ^] PyPS‐1 allows one to tailor multiple
   conditionally activatable PyAs by facilely installing PyPSs with
   various CG that sensitive to biospecies of interest (Figure [124]3a).
   Based on PyPS‐1 scaffold, we prepared a series of activatable PyPSs
   (PyPS‐CG) in response to different small molecules and enzymes,
   including hydrogen peroxide‐activatable PyPS‐PB, superoxide‐activatable
   PyPS‐SAR, GSH‐activatable PyPS‐DBS, and NQO1‐activatable PyPS‐BQ
   (Figure S24a, Supporting Information). These synthesized PyPS‐CG
   molecules all exhibited excellent stimuli‐triggered turn‐on response in
   photoactivity, including ICT process (Figure S24b–e, Supporting
   Information) and photosensitivity (Figure S25, Supporting Information).
   Considering the ubiquitous hypoxic feature of solid tumors,
   hypoxia‐activatable PyPS‐NF was selected as a proof‐of‐concept example
   for further in vitro and in vivo antitumor evaluation.

Figure 3.

   Figure 3
   [125]Open in a new tab

   Chemical design of activatable PyAs. a) Schematic illustration of the
   design and activation mechanism of a panel of activatable PyPS
   molecules. b) UV–vis absorption spectra of PyPS‐NF (5.0 μm) before and
   after NTR/NADH treatment in aqueous media. c) Time‐dependent changes in
   absorbance at 603 nm recorded for PyPS‐NF (5.0 μm) upon incubation
   without (blue) or with (red) NTR/NADH at 37 °C (n = 3). d) Binding
   conformation of PyPS‐NF at the hydrophobic interspace of NTR. e) HPLC
   chromatograms of PyPS‐NF without treatment (top); with NTR/NADH
   treatment for different time intervals (middle); and PyPS‐1 only
   (bottom).

   Since tumor hypoxia is always accompanied by elevated level of
   nitroreductase (NTR),^[ [126]^50 , [127]^51 , [128]^52 ^] an
   NTR‐responsive 5‐nitrofuran (NF) group was installed to provide
   hypoxia‐activatable PyPS‐NF. As expected, the ICT process of PyPS‐NF
   was inhibited after etherification, as demonstrated by the vanishment
   of the ICT band (Figure [129]3b) and fluorescence quenching (Figure
   S26b,c, Supporting Information). Meanwhile, negligible change was
   observed no matter in QDPBF absorption (Figure S27a,b, Supporting
   Information) or DHE emission (Figure S28a,b, Supporting Information) of
   PyPS‐NF following light irradiation, suggesting the successful
   photoactivity suppression of the caged PyPS.

   Then, we set out to investigate its NTR‐induced activation under
   hypoxic conditions. As sketched in Figure S26a, Supporting Information,
   NTR induces the reduction of the nitro group of the NF moiety to an
   amine group,^[ [130]^50 , [131]^52 ^] resulting in self‐immolation^[
   [132]^53 ^] and release of the phenolic group. Of note, NTR‐induced
   reduction, and self‐immolation occurred only under hypoxic
   conditions,^[ [133]^54 , [134]^55 ^] which could further ensure that
   PyPS‐NF was specifically activated in hypoxic tumor cells. Upon adding
   NTR (5 μg mL^−1) to the aqueous solution of PyPS‐NF (containing 0.5 mm
   NADH) at 37 °C, the characteristic absorption band at 603 nm of PyPS‐1
   emerged and was restored within 40 min (Figure [135]3b,c).
   Simultaneously, a fluorogenic response of PyPS‐NF toward NTR was
   observed (Figures S26b,c and S29a, Supporting Information), allowing
   fluorescence monitoring of the activation process and imaging‐guided
   PDT. The response of PyPS‐NF to NTR was shown to be highly selective
   over other potentially interfering biomolecules (Figure S29b,c,
   Supporting Information). Docking calculations revealed a high affinity
   between PyPS‐NF and NTR with a docking energy of −9.6 kcal mol^−1 and
   also indicated that PyPS‐NF tends to approach the cleft of NTR via
   hydrogen bonding and hydrophobic interactions (Figure [136]3d and
   S30a,b, Supporting Information). The binding affinity between PyPS‐NF
   and NTR was further assessed by microscale thermophoresis, which gave a
   dissociation constant (K [d]) of ≈2 μm for PyPS‐NF/NTR (Figure S30c,d,
   Supporting Information). High‐performance liquid chromatography (HPLC)
   analysis and high‐resolution mass spectrometry (HRMS) further
   demonstrated the NTR‐triggered full conversion from PyPS‐NF into PyPS‐1
   (Figure [137]3e). Significantly, after exposing the NTR‐treated PyPS‐NF
   solution to light irradiation, remarkable ROS‐induced QDPBF consumption
   (Figure S27c,d, Supporting Information) and drastic O[2] ^•−‐induced
   DHE fluorescence enhancement (Figure S28c,d, Supporting Information)
   were observed, indicating restored type‐I photoactivity.

2.4. In Vitro Studies of Hypoxia‐Activatable PyPS‐NF

   Encouraged by the earlier findings, we proceeded to assess the
   performance of PyPS‐NF in specifically activating and inducing
   pyroptosis in hypoxic cells (Figure [138]4a). We quantified the NTR
   levels in both normoxic and hypoxic MCF‐7 cells via ELISA assay,
   indicating that the NTR level in hypoxic cells is approximately twofold
   to that in normoxic cells (Figure S31c, Supporting Information). MCF‐7
   cells were incubated with PyPS‐NF under normoxia or hypoxia for 6 h and
   then washed. Upon imaging, the fluorescence in hypoxic cells was
   intense, in contrast to the slight fluorescence observed in normoxic
   cells (Figure [139]4b, S32, Supporting Information). Moreover, hypoxic
   cells pretreated with a known NTR inhibitor, dicumarin (DIC),^[
   [140]^50 ^] showed a negligible fluorescence (Figure S31a,b, Supporting
   Information). The earlier results confirmed the specific hypoxia
   activation of PyPS‐NF in living cells. As expected, the converted
   PyPS‐1 was well anchored on ER (Figure [141]4c). Activatable
   photosensitivity of PyPS‐NF in generating ROS within MCF‐7 cells was
   demonstrated via flow cytometry (Figure [142]4d, S33 and S34,
   Supporting Information). MTT assay showed negligible dark cytotoxicity
   of PyPS‐NF in both normoxia and hypoxia (Figure [143]4e), indicating
   its good biocompatibility.

Figure 4.

   Figure 4
   [144]Open in a new tab

   In vitro studies of hypoxia‐activatable PyPS‐NF. a) Mimic of a hypoxic
   environment (2% O[2]) with a hypoxia incubator chamber. b) Confocal
   fluorescence imaging of PyPS‐NF in response to hypoxia. Scale bar:
   50 μm. c) Colocalization studies of PyPS‐NF toward ER in MCF‐7 cells.
   Scale bar: 10 μm. d) Flow cytometry analysis of the ROS levels in MCF‐7
   cells after different treatments using DCF‐DA as ROS fluorescence
   probe. N and H represent normoxia and hypoxia, respectively. e) Cell
   viability of MCF‐7 cells incubated with PyPS‐NF in dark or upon light
   irradiation determined by MTT assay (n = 4). f) Representative phase
   contrast images of MCF‐7 cells incubated with PyPS‐NF in dark or upon
   light irradiation under normoxic or hypoxic conditions; the red arrows
   indicate the characteristic large bubbles accompanying pyroptosis.
   Scale bar: 20 μm. g) Western blotting of the cleavage of GSDME in MCF‐7
   cells after treatment of PyPS‐NF and light under hypoxic conditions.
   h,i) Investigation of cell surface exposure of CRT and release of
   HMGB‐1 (h), and extracellular secretion of ATP (n = 4) (i). N and H in
   (b), (d), and (e) represent normoxia and hypoxia, respectively. Light
   source: Xe lamp (490–700 nm, 10 mW cm^−2). ***p < 0.001, ****p < 0.0001
   determined by Student's t‐test. Scale bar: 75 μm.

   We further assessed the effect of PyPS‐NF on the precise induction of
   cell pyroptosis under hypoxic conditions. MCF‐7 cells were incubated
   with PyPS‐NF under normoxia or hypoxia, followed by exposure to light
   irradiation (10 mW cm^−2) for 10 min. Both MTT and live/dead staining
   assays indicated the selective phototoxicity of PyPS‐NF (Figure
   [145]4e, S35a, Supporting Information). Remarkably, the MTT assay gave
   PyPS‐NF an IC[50] value of only 0.38 μM under hypoxia, indicating a
   superb hypoxia‐specific anticancer performance of PyPS‐NF in vitro. Not
   surprisingly, dying MCF‐7 cells evoked by precise PDT with PyPS‐NF
   exhibited typical features of pyroptosis (Figure [146]4f), which was
   demonstrated to be GSDME mediated by Western blotting (Figure [147]4g).
   Collectively, the earlier results demonstrated that applying flexible
   hydroxyl protection/deprotection chemistry to PyPS‐1 scaffold is a
   simple yet robust strategy for preparing activatable PyAs.

   Notably, accompanying cell pyroptosis, release of damage‐associated
   molecular patterns,^[ [148]^56 , [149]^57 , [150]^58 ^] including cell
   surface exposure of calreticulin (ecto‐CRT), release of HMGB‐1, and
   extracellular secretion of ATP were observed (Figure [151]4h,i, S35b,
   Supporting Information). These findings suggested that PyPS‐NF‐induced
   pyroptosis is highly immunogenic. To further study the immune responses
   induced by PyPS‐NF/light treatment, we evaluated gene expression in
   MCF‐7 cells after treatments via RNA‐seq. Compared with the control
   group, 2215 differential expressed genes (DEGs) were detected in the
   PyPS‐NF/light‐treated group, while only 10 DEGs were detected in the
   group of PyPS‐NF without light (Figure [152]5a). In the
   PyPS‐NF/light‐treated group, genes associated with ER‐stress and immune
   system were strikingly upregulated (Figure [153]5b), and these genes
   were closely associated with each other (Figure [154]5c). Kyoto
   Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis
   showed that after PyPS‐NF/light treatment, numerous DEGs were enriched
   in inflammatory related as well as immune response‐associated signaling
   pathways, such as tumor necrosis factor (TNF) signaling pathway, IL‐17
   signaling pathway, MAPK signaling pathway, and cytokine‐cytokine
   receptor interaction signaling pathway (Figure [155]5d). These results
   suggested that PyPS‐NF‐induced pyroptosis can enhance tumor
   immunogenicity, and thus has great potential for antitumor
   immunotherapy.

Figure 5.

   Figure 5
   [156]Open in a new tab

   RNA sequencing analysis. a) Volcano plots displayed the differentially
   expressed genes (DEGs). b) Expression profiles of representative DEGs
   in different groups. c) Interaction network of proteins encoded by
   representative DEGs. d) Kyoto Encyclopedia of Genes and Genomes (KEGG)
   pathway enrichment analysis based on the DEGs.

2.5. In Vivo Studies of Hypoxia‐Activated Antitumor Efficacy of PyPS‐NF

   Motivated by the fascinating in vitro performance, we subsequently
   evaluated the specific immunotherapy of hypoxic solid tumors with
   PyPS‐NF in vivo. As murine breast cancer cells, after PyPS‐1 mediated
   PDT, 4T1 cells underwent marked pyroptosis similar to MCF‐7 cells (a
   human breast cancer cell line) under both normoxic and hypoxic
   conditions (Figure S36, Supporting Information). We then assessed the
   hypoxia‐triggered activation of PyPS‐NF using 4T1 tumor‐bearing BALB/c
   mice as the model. After intratumoral injection of PyPS‐NF, a
   significant fluorescence increase at the tumor site was observed within
   0.5 h, reaching maximum fluorescence at 4 h after injection (Figure
   [157]6a,b). In contrast, tumor regions in mice pretreated with DIC
   showed weaker fluorescence, suggesting that PyPS‐NF was catalyzed by
   tumor‐overexpressed NTR to PyPS‐1. In addition, the weak fluorescence
   of the control group injected directly into healthy hindlimb muscles
   further demonstrated the specific hypoxic tumor activation of PyPS‐NF.

Figure 6.

   Figure 6
   [158]Open in a new tab

   In vivo studies of hypoxia activatability of PyPS‐NF. a) Timelapse
   in vivo fluorescence imaging of BALB/c mice after injection of PyPS‐NF
   subcutaneously into tumors (top), DIC pretreated tumors (middle), or
   healthy hindlimb muscles (bottom). b) MFI of the tumor or muscle
   regions quantified from (a); n = 3. c) Time‐lapse in vivo fluorescence
   imaging of BALB/c mice after intravenous injection of PyPS‐NF and
   ex vivo imaging of major organs and tumors collected from mice after
   48 h of injection. d) MFI of the ex vivo major organs and tumors
   (n = 3). e) Schematic illustration of the caged and activated forms of
   the two PyPS‐CG molecules (inert PyPS‐Et and activatable PyPS‐NF). f)
   In vivo biodistribution of PyPS‐CG in both caged and activated forms
   (n = 5). g) Conversion rate (R [c]) of PyPS‐CG from caged to activated
   form in different main organs and tumor tissues (n = 5).

   Encouraged by the earlier results, we continued to evaluate the
   performance of PyPS‐NF through systemic administration. As shown in
   Figure [159]6c and S37, Supporting Information, after intravenous
   injection of PyPS‐NF, the fluorescence of the converted PyPS‐1 in the
   xenograft tumors gradually increased over 24 h and remained at a high
   level for at least 48 h. Of note, the activated PyPS‐1 was mainly
   distributed in the solid tumor, as evidenced by the ex vivo
   fluorescence imaging data (Figure [160]6c,d). This remarkable
   tumor‐specific targeting of PyPS‐1 is mainly due to
   hypoxia‐responsiveness of PyPS‐NF, bringing great promise for
   overcoming the stubborn dilemmas of off‐target damage. In addition, the
   turn‐on fluorescence concomitant with PyPS‐NF activation allows for
   imaging‐guided tumor treatment, further improving therapeutic accuracy.

   To further determine the on‐target activation of PyPS‐NF in tumor site,
   we synthesized an ethyl‐caged inert PyPS‐CG molecule (PyPS‐Et) as a
   control (Figure [161]6e). The biodistribution of PyPS‐CG (PyPS‐Et or
   PyPS‐NF) in main organs and tumor tissues of 4T1 tumor‐bearing BALB/c
   mice (n = 5) was then studied. Tissue homogenates after 24 h of
   intravenous injection of PyPS‐CG were collected and analyzed by HPLC to
   determine the levels of PyPS‐CG in both caged and activated forms in
   different tissues (Figure [162]6e). As shown in Figure [163]6f, the
   levels of PyPS‐Et and PyPS‐NF in tumor tissues were comparable. The
   difference is that PyPS‐Et exists almost exclusively in the caged form
   (Figure S38a, Supporting Information), whereas PyPS‐NF exists mainly in
   the activated form (Figure S38b, Supporting Information). Furthermore,
   the conversion rate (R [c]) of the two molecules from caged PyPS‐CG to
   activated PyPS‐1 in tumor tissues was calculated. PyPS‐Et was barely
   activated with an R [c] value of 0.7%, while about 66.2% of PyPS‐NF was
   successfully activated (Figure [164]6g). The above results combined
   with the in vivo fluorescence data (Figure [165]6a–d) demonstrated the
   on‐target activation of PyPS‐NF in solid tumors.

   Subsequently, PyPS‐NF‐mediated photoimmunotherapy was investigated in
   mice inoculated with 4T1 cancer cells in the right flank as the primary
   tumors (Figure [166]7a).

Figure 7.

   Figure 7
   [167]Open in a new tab

   In vivo studies of antitumor efficacy of PyPS‐NF. a,b) Schematic
   illustrations of the specific photoimmunotherapy of hypoxic solid
   tumors with PyPS‐NF in vivo (a) and the schedule for tumor implantation
   and PyPS‐NF‐mediated photoimmunotherapy (b). c) Size of the primary
   tumors in different treatment groups (n = 5). d) Photographs recorded
   for the primary tumors collected after 2 weeks of observation. e)
   Representative IHC images of Ki67, cleaved caspases‐3 (c‐Caspase‐3),
   and GSDME‐N in the primary tumors 2 weeks after different treatments.
   Scale bar: 50 μm.

   Compared to the bulky xenon lamp used, a portable 633 nm laser that can
   better penetrate animal tissue was used during in vivo experiments.^[
   [168]^59 , [169]^60 ^] We first evaluated cell pyroptosis upon
   PyPS‐NF/laser treatment. Gratifyingly, as shown in Figure S39,
   Supporting Information, both hypoxic 4T1 and MCF‐7 cells after
   PyPS‐NF/laser treatment underwent marked pyroptosis, similar to
   PyPS‐NF/Xe lamp treatment. Notably, due to the serious energy loss of
   laser when penetrating skin and tumor tissue, the laser intensity was
   increased to 0.1 W cm^−2 during in vivo experiments to obtain favorable
   antitumor efficacy.^[ [170]^61 , [171]^62 ^]

   These tumor‐bearing mice were then randomized into four groups (n = 5
   mice): i) PBS, ii) PBS + light (633 nm, 0.1 W cm^−2, 10 min), iii)
   PyPS‐NF (5 mg kg^−1), and iv) PyPS‐NF + light, and tumor volumes were
   monitored every other day to assess the treatment outcome (Figure
   [172]7b). As shown in Figure [173]7c,d, compared with the rapid growth
   of the primary tumors in the three control groups (i–iii), the tumor
   growth in the treatment group (group iv) was significantly inhibited
   during the treatment. Hematoxylin and eosin (H&E) staining of the
   collected tumors and main organs (heart, liver, spleen, lung, and
   kidney) was also performed for histological analysis. Obvious
   extermination of cells was observed in the treatment group, showing
   cell shrinkage and separation (Figure S40a, Supporting Information).
   However, in the other control groups, no appreciable cell damage was
   observed. On the other hand, we compared the antitumor efficacy of
   PyPS‐Et/light and PyPS‐NF/light treatments (Figure S41, Supporting
   Information). Gratifyingly, for the group of PyPS‐NF/light treatment, a
   clear regression of the tumors was observed. In contrast, inert
   PyPS‐Et/light failed to inhibit tumor growth in view of the continuous
   tumor growth similar to the untreated group. These findings
   demonstrated the intriguing on‐target antitumor activity of PyPS‐NF in
   solid tumors.

   Immunohistochemical (IHC) staining of the tumors showed a reduced
   expression of Ki67 in the PyPS‐NF/light‐treated group, indicating that
   the proliferation of tumor cells was significantly inhibited after
   treatment (Figure [174]7e). A significant increase of c‐caspase‐3 and
   GSDME‐N was observed in the tumors of the treatment group (Figure
   [175]7e), indicating pyroptosis of tumor cells after treatment.
   Notably, negligible cell necrosis or inflammation lesions were observed
   in the main organs of all mice (Figure S40a, Supporting Information),
   and their body weight increased steadily (Figure S40b, Supporting
   Information), suggesting the biocompatibility of our PyPS‐NF.

2.6. Systematic Immune Response after PyPS‐NF/Light Treatment

   Infiltrating immune cells were then analyzed to assess the immune
   response in vivo. The maturation of dendritic cells (DCs) was pivotal
   for initiating antitumor immunotherapy,^[ [176]^63 ^] which we
   investigated via flow cytometry. As shown in Figure [177]8a,c, and
   Figure S42a and S43a, Supporting Information, CD80^+CD86^+ DCs in both
   tumors and draining lymph nodes of the treatment group were
   significantly upregulated compared with other control groups (group i
   and iii), suggesting that PyPS‐NF‐mediated photoimmunotherapy
   successfully promoted DC maturation. Furthermore, the DC‐mediated
   immune response was further studied by profiling intratumoral
   infiltration of cytotoxic T cells (CD8^+) and helper T cells (CD4^+),
   which play a key role in cancer immunotherapy.^[ [178]^64 , [179]^65 ^]
   Compared with the control groups, the populations of CD3^+CD8^+ and
   CD3^+CD4^+ T cells in the tumors of the treatment group were obviously
   higher (Figure [180]8b,d,e, S42b and S43b, Supporting Information). The
   results of immunostaining further indicated that the infiltration of
   cytotoxic and helper T cells in the tumor tissues of the treatment
   group was increased remarkably (Figure [181]8f). These findings
   suggested that PyPS‐NF‐mediated PDT successfully recruited cytotoxic
   and helper T cells to activate the immune response. In addition,
   cytokine profiling in serum was investigated via ELISA. As shown in
   Figure [182]8g–i, TNF‐α, interferon γ (IFN‐γ), and interleukin 1β
   (IL‐1β) were significantly increased after PyPS‐NF/light treatment,
   suggesting a pronounced immune activation. Significant
   immunotherapeutic outcome induced by this pyroptosis‐mediated immune
   activation was further demonstrated by the efficient suppression of
   distant tumors (Figure [183]8j,k). Taken together, the above results
   adequately demonstrated the promise of PyPS‐NF in precise and efficient
   immunotherapy against solid tumors.

Figure 8.

   Figure 8
   [184]Open in a new tab

   Systematic immune response. a,b) Flow cytometric analysis of maturated
   DCs (CD80^+CD86^+CD11c^+) (a) and cytotoxic T cells (CD3 + CD8+) (b) in
   the primary tumors after different treatments. c–e) Populations of
   maturated DCs (c), cytotoxic T cells (d), and helper T cells (e)
   (CD3^+CD4^+) in the primary tumors after different treatments (n = 3).
   f) Representative images of immunostaining of CD8 and CD4 in the
   primary tumors 2 weeks after different treatments. Scale bar: 50 μm.
   g–i) Secretion of TNF‐α (g), IFN‐γ (h), and IL‐1β (i) in serum measured
   by ELISA assay (n = 4). **p < 0.01 (one way ANOVA). j) Photographs and
   k) weights of the distant tumors (n = 5). Data is mean ± SD (n = 3 or
   5). **p < 0.01, ***p < 0.001 determined by Student's t‐test.

3. Conclusion

   We have designed and synthesized three photosensitizers (PyPS‐1,
   PyPS‐2, and PyPS‐3), which showed excellent photosensitivity in
   producing ROS under both normoxic and hypoxic conditions. In addition,
   they all demonstrated unexpected ER‐targeting capabilities without
   chemical conjugation of any ER‐targeting moieties. This specific
   photodamage of ER thereby initiated rapid and strong cell pyroptosis.
   We successfully revealed the mechanism of GSDME‐mediated pyroptosis by
   studying the best‐performing PyPS‐1. Finally, we established a general
   strategy to prepare activatable PyAs based on the caging/uncaging
   chemistry of the phenolic hydroxyl group of PyPS‐1, and thus
   successfully obtained a set of activatable PyPSs with specific
   responsiveness to biological species. As a proof‐of‐concept example,
   NTR‐activatable PyPS‐NF targeting tumor hypoxia was selected and
   demonstrated significant performances in precisely inducing immunogenic
   pyroptosis of hypoxic tumor cells in vitro and in vivo. This work
   establishes a unified design to prepare activatable PyAs, opening an
   avenue to tailor and synthesize activatable PyAs capable of inducing
   pyroptosis in a controlled manner, which is essential for advancing
   pyroptosis‐inducing cancer immunotherapy.

4. Experimental Section

4.1.

4.1.1.

Cell Culture

   Cell lines including MCF‐7, A549, HepG2, A375, HeLa, NIH‐3T3, 293 T,
   and 4T1 cells were purchased from American Type Culture Collection
   (ATCC).

Tumor Model

   All animal experiments were approved by the Animal Ethics Committee of
   Nankai University (2021‐SYDWLL‐000234). BALB/c female (6–8 weeks) were
   purchased from Beijing Vital River Laboratory Animal Technology Co.,
   Ltd. (Beijing, China). For the subcutaneous tumor model, 4T1 cells
   (≈1 × 10^6) suspended in 100 μL of PBS were subcutaneously injected
   into the right flank of each mouse. After about 10 days, the
   tumor‐bearing mice were then used subsequently.

Statistical Analysis

   The data herein were given as means ± standard deviation. Statistical
   significance was determined using Student's t‐test or analysis of
   variance (ANOVA) analysis, and p < 0.05 was considered to be
   statistically significant (*, **, and *** indicate p < 0.05, p < 0.01,
   and p < 0.001, respectively).

Conflict of Interest

   The authors declare no conflict of interest.

Supporting information

   Supplementary Material
   [185]SMSC-4-2300135-s001.pdf^ (14.2MB, pdf)

Acknowledgements