• Log in with Facebook Log in with Twitter Log In with Google      Sign In    
  • Create Account
  LongeCity
              Advocacy & Research for Unlimited Lifespans

Photo

A versatile drug delivery system targeting senescent cells

chemotherapy fibrosis nanomedicine palbociclib senescence

  • Please log in to reply
No replies to this topic

#1 Engadin

  • Guest
  • 198 posts
  • 580
  • Location:Madrid
  • NO

Posted 30 June 2019 - 01:33 PM


Published at Rejunevation Science News (RSN)

 

 

F U L L   T E X T :   NCBI _ PMC 

 

 

ABSTRACT

 

Senescent cells accumulate in multiple aging‐associated diseases, and eliminating these cells has recently emerged as a promising therapeutic approach. Here, we take advantage of the high lysosomal β‐galactosidase activity of senescent cells to design a drug delivery system based on the encapsulation of drugs with galacto‐oligosaccharides. We show that gal‐encapsulated fluorophores are preferentially released within senescent cells in mice. In a model of chemotherapy‐induced senescence, gal‐encapsulated cytotoxic drugs target senescent tumor cells and improve tumor xenograft regression in combination with palbociclib. Moreover, in a model of pulmonary fibrosis in mice, gal‐encapsulated cytotoxics target senescent cells, reducing collagen deposition and restoring pulmonary function. Finally, gal‐encapsulation reduces the toxic side effects of the cytotoxic drugs. Drug delivery into senescent cells opens new diagnostic and therapeutic applications for senescence‐associated disorders.

 

 

INTRODUCTION

 

Severe or unrepairable cellular damage often triggers a stereotypic cellular response known as senescence. This cellular program is controlled by relatively well‐understood pathways that stably block cell proliferation and are conserved across vertebrates (Muñoz‐Espín & Serrano, 2014). The main purpose of cellular senescence is to prevent the proliferation of damaged cells and, at the same time, to trigger tissue repair through the secretion of a complex mixture of extracellular factors, known as senescence‐associated secretory phenotype or SASP. Specifically, senescence‐initiated tissue repair involves the recruitment of inflammatory cells, the dismissal of senescent cells by phagocytic and immune cells, and the activation of stem/progenitor features in non‐damaged surrounding cells (Krizhanovsky et al2008; Demaria et al2014; Yun et al2015; Mosteiro et al2016; Chiche et al2017; Ritschka et al2017). Even during vertebrate development, embryos use senescence in a damage‐independent manner to initiate specific tissue remodeling processes (Muñoz‐Espín et al2013; Storer et al2013).

 

Upon persistent damage or during aging, senescent cells accumulate, probably due to an inefficient clearance by immune cells, and this accumulation may lead to chronic inflammation and fibrosis (Muñoz‐Espín & Serrano, 2014). Indeed, evidence in mice indicates that the accumulation of senescent cells actively contributes to multiple diseases and aging (Muñoz‐Espín & Serrano, 2014). In this regard, genetic ablation of senescent cells delays and ameliorates some aging‐associated diseases, reverts long‐term degenerative processes associated with chemotherapy, and extends longevity (Baker et al2016; Childs et al2016; Demaria et al2017). Importantly, senescent cells present vulnerabilities to particular small molecule inhibitors, known as “senolytics”, that trigger apoptosis preferentially in senescent cells (Zhu et al2015). For example, the combination of dasatinib (a tyrosine kinase inhibitor of broad specificity) and quercetin (a flavonol with antioxidant and estrogenic activity) has preferential toxicity over senescent cells compared to control cells (Zhu et al2015). Also, the survival of senescent cells is highly dependent on elevated levels of the BCL‐2 family of anti‐apoptotic factors, and accordingly, senescent cells are hyper‐sensitive to apoptosis induced by navitoclax, a BCL‐2 family inhibitor (Chang et al2016; Yosef et al2016; Zhu et al2016; Pan et al2017). Similarly, inactivation of the transcription factor FOXA4 with a peptide derivative preferentially eliminates senescent cells over non‐senescent ones (Baar et al2017). These pharmacological treatments reduce the number of senescent cells in vivo and show therapeutic activity against senescence‐associated diseases and aging (Zhu et al20152016; Chang et al2016; Roos et al2016; Yosef et al2016; Baar et al2017; Pan et al2017; Schafer et al2017).

 

Senescent cells in vitro are characterized by high levels of lysosomal β‐galactosidase activity, known as senescence‐associated β‐galactosidase (SAβGal; Dimri et al1995; Muñoz‐Espín & Serrano, 2014; Kurz et al2000; Lee et al2006). In addition to β‐galactosidase, senescent cells present high levels of most tested lysosomal hydrolases (Knaś et al2012). Indeed, senescent cells show a remarkable accumulation of lysosomes, together with abnormal endosomal traffic and autophagy (Cho et al2011; Narita et al2011; Ivanov et al2013; Udono et al2015; Tai et al2017). Interestingly, damaged or diseased tissues generally contain cells that are positive for SAβGal, while normal healthy tissues are negative for this marker (Sharpless & Sherr, 2015).

 

Here, we have explored the possibility of using lysosomal β‐galactosidase as a vulnerable trait of senescent cells that can be exploited to deliver tracers or drugs preferentially to diseased tissues with high content of senescent cells. Our approach is based on the encapsulation of diagnostic or therapeutic agents with β(1,4)‐galacto‐oligosaccharides and their delivery to lysosomes via endocytosis. We show that this delivery strategy is effective in vivo, and it is therapeutic in the context of cancer chemotherapy and also against pulmonary fibrosis. Moreover, our encapsulation system has the added value of reducing the systemic toxicities of cytotoxic drugs.

 

 

RESULTS

 

Validation of gal‐encapsulation to target senescent cells in vitro

 

Sugar‐coated beads (~100 nm diameter), based on a silica porous scaffold known as MCM‐41 (Kresge et al1992), are efficiently internalized into cells by endocytosis, targeted to the lysosomes, and eventually released by exocytosis (Slowing et al20062011; Tao et al2009; Bernardos et al2010; Hocine et al2010; Yanes et al2013; Aznar et al2016). In a previous work (Agostini et al2012), we designed beads pre‐loaded with a fluorophore and then coated with a layer of galacto‐oligosaccharides of mixed lengths (referred to as GosNP). We observed that β‐galactosidase can digest the sugar coating of the beads, thus allowing the diffusion of the cargo out of the silica scaffold. Interestingly, fluorophore‐loaded GosNP efficiently release their content within senescent cells, in agreement with the high levels of β‐galactosidase activity of these cells (Agostini et al2012). Here, we have improved this system by using a homogeneous coating mostly consisting of a 6‐mer galacto‐oligosaccharide (referred to as GalNP) (Fig 1A and Appendix Fig S1A). Using this strategy, we have encapsulated rhodamine B [GalNP(rho)], doxorubicin [GalNP(dox)], and navitoclax [GalNP(nav)], and enzymatic digestion of the three types of beads with fungal β‐galactosidase demonstrated efficient release of their respective cargos (Appendix Fig S1B–E). In general, 100 mg of drug is encapsulated per gram of beads and, upon digestion in vitro, ~30 mg of drug is released per gram of beads (Appendix Tables S1–S3). Targeting of senescent cells with GalNP(rho) was validated in three human cancer cell lines treated with palbociclib (a selective CDK4/CDK6 inhibitor approved in combination with letrozole or fulvestrant for hormone receptor+/HER2 metastatic breast cancer; Fig 1B and Appendix Fig S1F and G). Of note, the three cell lines used, SK‐MEL‐103, NCI‐H226, and UT‐SCC‐42B, have an active retinoblastoma pathway (RB1‐proficient; Ikediobi et al2006; Giefing et al2011; Massaro et al2017) and undergo senescence upon treatment with palbociclib (Fig 1B and Appendix Fig S1F). In these cellular models, the fluorophore was released more efficiently in senescent cells compared to control cells, demonstrating the functionality of the GalNP encapsulation method. Co‐staining of cells with a lysosomal marker indicated that a substantial fraction of rhodamine was present in lysosomes (Appendix Fig S1H).

 

 

EMMM-10-e9355-g002.jpg
Release of gal‐encapsulated fluorophores in xenografts

 

  • A

    GalNP beads are based on a mesoporous silica scaffold (MCM‐41) that can be loaded with different cargoes encapsulated by a coat of 6‐mer β(1,4)‐galacto‐oligosaccharides. Cellular uptake of the GalNP beads occurs via endocytosis and, after fusion with lysosomal vesicles, the beads are released by exocytosis. The high lysosomal β‐galactosidase activity of senescent cells allows a preferential release of the cargo by a β‐galactosidase‐mediated hydrolysis of the cap.

  • B

    SK‐MEL‐103 melanoma cells were treated with palbociclib (1 μM) for 1 week, and senescence induction was assessed by SAβgal staining. Next, cultures were exposed to GalNP(rho) (50 μg/ml, for 16 h). Pictures show representative images illustrating rhodamine release by confocal microscopy. Cells were co‐stained with Calcein, and nuclei were stained with Hoechst. Graphs to the right show the rhodamine intensity relative to cell surface in senescent cells and non‐senescent (control) cells. Each assay was repeated at least three times with similar results. Scale bar: 50 μm.

  • C

    Subcutaneous tumor xenografts of SK‐MEL‐103 melanoma cells in athymic female nude mice. Upon tumor formation, mice were treated daily with palbociclib (oral gavage, 100 mg/kg) during 7 days. The left panel picture shows representative whole tissue portions of tumors after SAβGal staining. The right panel shows sections of control and palbociclib‐treated tumors processed for SAβGal staining, and Ki67 and phosphorylated Rb (p‐Rb) immunohistochemistry. This experiment has been repeated at least two times with similar results. Scale bar: 50 μm.

  • D

    Mice bearing SK‐MEL‐103 xenografts, control or treated with palbociclib for 7 days, as in ©, were tail vein injected with 200 μl of a solution containing GalNP(rho) (4 mg/ml). At 6 h post‐injection, mice were sacrificed, tumors were collected, and fluorescence was analyzed by an IVIS spectrum imaging system. The graph indicates the average difference in tumor radiance between GalNP‐injected control and palbociclib‐treated groups. The inset shows the absolute values of radiance (p/s/cm2/sr × 106) for each group. The corresponding differences are highlighted in black or red. Values are expressed as mean ± SD, and statistical significance was assessed by the two‐tailed Student's t‐test.

 

 

 

Release of gal‐encapsulated fluorophores in xenografts

 

To evaluate the release of gal‐encapsulated fluorophores in vivo, we employed tumor xenografts treated with senescence‐inducing chemotherapy. Subcutaneous xenografts were generated using SK‐MEL‐103 melanoma cells and NCI‐H226 lung squamous carcinoma cells. Upon tumor formation, mice were treated daily with palbociclib for 7 days, and this resulted in high levels of intratumoral senescence, as inferred from elevated SAβGal activity, absence of the proliferative marker Ki67, and reduction in phosphorylated Rb (Fig 1C). In a first approach, palbociclib‐treated mice carrying SK‐MEL‐103 xenografts were given a single intravenous injection of GalNP(rho) and fluorescence was analyzed 6 h later. Palbociclib‐treated tumors were strongly autofluorescent (Fig 1D). Importantly, however, rhodamine was detectable above background in mice treated with GalNP(rho), and the signal attributed to rhodamine was higher in tumors treated with palbociclib compared to non‐treated tumors (Fig 1D). Fluorescence was not detected in other organs at 6 h post‐injection, including liver, spleen, and lungs (Appendix Fig S1I and see below Fig 2B). To avoid the detection of autofluorescence, we used beads loaded with indocyanine green (see Appendix Fig S1E), a fluorophore that emits in the far‐red spectrum and therefore is minimally affected by the autofluorescence of palbociclib‐senescent cells. Confirming the rhodamine release data, SK‐MEL‐103 and NCI‐H226 xenografts treated with palbociclib and gal‐encapsulated indocyanine green showed much higher fluorescence compared to tumors treated with a single agent alone (Appendix Fig S1J and K).

 

 

EMMM-10-e9355-g003.jpg
Release of gal‐encapsulated fluorophores in fibrotic lungs

 

  • A

    C57BL/6 male mice were subjected to a single intratracheal administration bleomycin (1.5 U/kg) and analyzed 2 weeks later. Pictures at the left correspond to representative lungs after whole tissue SAβGal staining. The right panel shows sections of control and bleomycin‐treated lungs processed for SAβGal and Masson's trichrome staining to detect collagen fibers (stained in blue). Scale bar: 50 μm.

  • B

    Control and bleomycin‐treated mice, as in (A), were tail vein injected with 200 μl of a solution containing GalNP(rho) (4 mg/ml). At 6 h post‐injection, mice were sacrificed and the lungs were analyzed by an IVIS spectrum imaging system, as in Fig 1D. The graph indicates the average difference in lung radiance between GalNP‐injected control and bleomycin‐treated groups. The inset shows the absolute values of radiance (p/s/cm2/sr × 106) for each group. The corresponding differences are highlighted in black or red. Values are expressed as mean ± SD, and statistical significance was assessed by the two‐tailed Student's t‐test.

  • C

    Control and bleomycin‐treated mice were injected with GalNP(rho), as in (B), and lung sections were analyzed 6 h post‐injection by confocal microscopy. Pictures correspond to representative sections of control and bleomycin‐treated lungs. The graph shows % of rhodamine+ cells in fibrotic areas from bleomycin‐treated mice compared to normal lung tissue from control mice. Values are expressed as mean ± SD, and statistical significance was assessed by the two‐tailed Student's t‐test. Scale bar: 25 μm.

  • D

    Lung cell suspensions from control and bleomycin‐treated mice, as in (B), were analyzed by flow cytometry. The upper panels show representative dot plots of rhodamine staining in CD45CD31 cells. The gating strategy is shown in detail in Appendix Fig S2C using as example the panel corresponding to the bleomycin‐treated lung shown here. Values in boxes correspond to the mean ± SEM, and statistical significance was assessed by the two‐tailed Student's t‐test. The lower panels show representative dot plots of rhodamine+ cells (after exclusion of CD45+ and CD31+ cells) separated in EpCAM(fibroblasts) and EpCAM+ (epithelial cells) subpopulations. Values in boxes correspond to the mean ± SEM of these two populations. Statistical significance between the fibroblast:epithelial ratio of positivity was assessed by the two‐sided Fisher exact test (P = 0.008).

  • E

    GSEA plots of published signatures of SASP and SIR (senescence‐inflammatory response) (Lasry & Ben‐Neriah, 2015) against the ranked list of differential expression between Rho+ and Rho cells (all CD45CD31) from bleomycin‐treated mice (n = 3), at 2 weeks post‐bleomycin, as in (D).

 

 

 

Release of gal‐encapsulated fluorophores in pulmonary fibrosis

 

Cellular senescence is abundant in pulmonary fibrosis, both in humans and in mice, and actively contributes to the pathological manifestations of this disease (Aoshiba et al20032013; Hecker et al2014; Pan et al2017; Schafer et al2017). We wondered whether our senescence delivery system would also work in a mouse model of pulmonary fibrosis. Intratracheal instillation of bleomycin in mice produced full‐blown lung fibrosis in a period of 2 weeks, accompanied by focal areas of SAβGal activity and strong collagen deposition as indicated by Masson's trichrome staining (Fig 2A). Two weeks post‐bleomycin administration, mice were intravenously injected with GalNP(rho) and 6 h later fluorescence was measured in the lungs. In this in vivo senescence model, autofluorescence was less prominent than in the case of palbociclib‐treated tumors. Importantly, rhodamine release occurred preferentially in fibrotic lungs compared to healthy lungs (Fig 2B). Moreover, confocal microscopy indicated that Rho+ cells were more abundant in fibrotic lung lesions compared to non‐fibrotic lungs (Fig 2C).

 

The differential fluorescence observed between fibrotic and healthy lungs could conceivably reflect, at least in part, a different accessibility and accumulation of the GalNP beads. To evaluate this, we measured the levels of silicon in the lungs and other organs, 6 h after i.v. injection of GalNP beads, by ICP‐MS (inductively coupled plasma mass spectroscopy). Interestingly, the levels of silicon in the lungs and in other tissues were similar between control and bleomycin‐treated mice (Appendix Fig S2A). Therefore, the silica beads reach equally well both healthy and fibrotic lungs (Appendix Fig S2A); however, the release of the fluorophore preferentially occurs within fibrotic lungs (Fig 2B and C). We also wondered if the GalNP beads would retain their activity when administered intratracheally rather than intravenously. Indeed, as in the case of i.v. injection, intratracheal administration of the beads also produced preferential cargo release in fibrotic lungs compared to healthy lungs (Appendix Fig S2B).

 

Next, we set to characterize in detail the cells targeted by GalNP(rho) in fibrotic lungs using flow cytometry. After excluding endothelial (CD31+) and hematopoietic (CD45+) cells (Appendix Fig S2C), we quantified the relative number of Rho+ cells in double‐negative CD45CD31 cells, which are mostly comprised by lung epithelial cells and fibroblasts. Importantly, bleomycin‐treated lungs showed higher levels of Rho+CD45CD31 cells than control lungs (Fig 2D). Further analyses using the epithelial marker EpCAM suggested that the large majority of Rho+CD45CD31 cells corresponded to fibroblasts (EpCAM) (Fig 2D). To directly test whether Rho+CD45CD31 cells are indeed senescent, CD45CD31 cells from bleomycin‐treated lungs were sorted into Rho+ and Rho subpopulations and subjected to RNAseq. Gene set enrichment analyses (GSEA) using published signatures of senescence (Lasry & Ben‐Neriah, 2015) indicated that Rho+CD45CD31 cells present a significant upregulation of senescence signatures (Fig 2E and Appendix Fig S2D and Dataset EV1). We also examined the levels of Rho+ cells in endothelial, total hematopoietic cells, lymphocytes, macrophages, and granulocytes. The majority of Rho+ cells, both in healthy and fibrotic lungs, were macrophages. However, the relative levels of Rho+ macrophages were reduced in bleomycin‐treated lungs, and the same trend was observed in the other cell types (Appendix Fig S2E–G). Although the significance of this reduction in Rho+ non‐fibroblastic cells remains to be explored, it could be due to competition by the Rho+ fibroblasts present in the bleomycin‐treated lungs. These results demonstrate in vivo that GalNP beads release their cargoes within senescent fibroblasts and can be used as a tool to detect and isolate senescent fibroblasts from fibrotic tissues.

 

 

Therapeutic activity of gal‐encapsulated cytotoxic drugs on tumor xenografts

 

After demonstrating that GalNP beads preferentially release fluorescent cargoes within senescent cells, we wondered whether gal‐encapsulated cytotoxics would also target senescent cells in vivo. We screened a panel of 80 anti‐cancer drugs with the aim of identifying drugs with killing activity against both senescent and non‐senescent cells. The killing activity of this panel of drugs was tested in three RB1‐proficient cancer cell lines (SK‐MEL‐103, NCI‐H226, and the hepatocarcinoma cell line Huh7), under normal proliferating conditions or rendered senescent with palbociclib (Appendix Fig S3A–C). We noted that the commonly used cytotoxic drug doxorubicin was among the most efficient cytotoxic drugs against both normal and senescent cells in the three cancer cell lines used. Based on this, we generated gal‐encapsulated beads loaded with doxorubicin [GalNP(dox)] with the aim of preferentially killing senescent cells (Appendix Fig S1E). We took advantage of the intrinsic fluorescence of doxorubicin to assess its preferential release within senescent cells. Addition of GalNP(dox) to senescent SK‐MEL‐103 cells (treated with palbociclib for 14 days) resulted in a strong fluorescent signal after 30 min (Fig 3A). In contrast, growing SK‐MEL‐103 cells showed much weaker fluorescence. As an additional control, we also used SK‐MEL‐103 cells treated with palbociclib for only 1 day, which was not sufficient to induce senescence, but efficiently reduced the levels of phosphorylated RB1 and FOXM1, indicative of CDK4/CDK6 inhibition (Fig 3A and Appendix Fig S3D). Treatment of these short‐term palbociclib‐treated cells with GalNP(dox) resulted in low levels of fluorescence, similar to those of control cells (Fig 3A). To further link drug release with senescence, we performed the same experiment with SAOS‐2 osteosarcoma cells. These cells are null for the RB1 gene (Li et al1995) and, therefore, are resistant to palbociclib. Interestingly, treatment of SAOS‐2 cells with GalNP(dox) did not result in doxorubicin release even after long‐term palbociclib treatment (Appendix Fig S3E). It is important to note that the subcellular localization of doxorubicin fluorescence was different depending on its formulation: Free doxorubicin localized in the nucleus, whereas encapsulated doxorubicin rendered perinuclear fluorescence (Fig 3A and Appendix Fig S3E). This reinforces the concept that GalNP beads enter cells through endocytosis and release their cargo in the lysosomal compartment.

 

 

EMMM-10-e9355-g004.jpg
Therapeutic activity of gal‐encapsulated cytotoxic drugs on tumor xenografts

 

  • A

    SK‐MEL‐103 melanoma cells were treated with palbociclib (1 μM) for 1 or 14 days, and senescence induction was assessed by SAβgal staining. Next, cultures were exposed to free doxorubicin (50 μM) or GalNP(dox) (1 mg/ml, filtered) for 30 min. Pictures show representative images illustrating doxorubicin fluorescence by confocal microscopy. Scale bar: 50 μm.

  • B

    SK‐MEL‐103 melanoma cells were treated with palbociclib (5 μM) for 14 days, cultures were exposed to GalNP(dox) (0.06 mg/ml, filtered), and annexin V signal was quantified over time. Representative pictures are shown in Appendix Fig S3F.

  • C

    Athymic nude female mice carrying subcutaneous SK‐MEL‐103 xenografts were treated daily with palbociclib (oral gavage, 50 mg/kg) and/or GalNP(dox) (tail vein injection, 200 μl of a solution with 4 mg/ml of GalNP containing a total of 1 mg/kg of deliverable doxorubicin), alone or in combination, as indicated. For each tumor, the relative tumor volume change was calculated relative to its baseline prior to treatment. Values are expressed as mean ± SEM. Individual tumor size measurements are shown in Appendix Fig S3G.

  • D

    Similar to (B) but using GalNP(nav) (tail vein injection, 200 μl of a solution with 4 mg/ml of GalNP containing a total of 1 mg/kg of deliverable navitoclax), as indicated. Individual tumor size measurements are shown in Appendix Fig S3F.

  • E

    Left, fold change of tumor size of SK‐MEL‐103 xenografts, after the indicated daily treatments. Data for palbociclib, and for palbociclib plus GalNP(dox), correspond to the same data in panel ©, at day 17. Data for free doxorubicin (daily tail vein injection, 1 mg/kg, for 17 days) were obtained in parallel. Right, mRNA levels of cardiotoxicity markers in hearts from the same mice. Actb and Gapdh were used for input normalization. Values are relative to control mice and are expressed as mean ± SD, and statistical significance was assessed by one‐way ANOVA and Dunnett's multiple comparisons test (versus palbociclib‐alone treated group).

  • F

    Left, fold change of tumor size, as in ©, after the indicated daily treatments. Data for palbociclib, and for palbociclib plus GalNP(nav), correspond to the same data in panel (D), at day 13. Data for free navitoclax (daily oral gavage, 25 mg/kg, for 13 days) were obtained in parallel. Right, platelet levels in the blood of the same mice. Values are expressed as mean ± SEM in the case of tumor size, and as mean ± SD in the case of platelet counting, and statistical significance was assessed by one‐way ANOVA and Dunnett's multiple comparisons test (versus palbociclib‐alone treated group).

 

 

 

It is known that modified forms of doxorubicin with lysosomal tropism efficiently induce apoptosis (Nair et al2015; Sheng et al2015). To test whether this was the case in senescent cells treated with GalNP(dox), we measured by time‐lapse fluorescent microscopy the appearance of annexin V‐positive cells. Senescent cells treated with gal‐encapsulated doxorubicin underwent cell death within 48 h, whereas control cells remained viable during this time (Fig 3B and Appendix Fig S3F).

 

Based on the above date, we decided to test this strategy in vivo using gal‐encapsulated doxorubicin and also gal‐encapsulated navitoclax (also known as ABT‐263), which is one of the most efficient senolytic compounds reported to date (Zhu et al2016; Chang et al2016; Yosef et al2016; Pan et al2017Appendix Fig S1E). The first model that we used to test senolysis in vivo using GalNP(dox) and GalNP(nav) consisted of tumor xenografts treated with palbociclib. Nude mice carrying subcutaneous SK‐MEL‐103 xenografts were treated with daily doses of palbociclib and GalNP(dox), alone or in combination (Fig 3C and Appendix Fig S3G); and a similar experiment was performed with GalNP(nav) (Fig 3D and Appendix Fig S3H). Interestingly, both GalNP(dox) and GalNP(nav) had a clear therapeutic benefit in combination with palbociclib. In contrast, GalNP(dox) and GalNP(nav) had no effect on tumor growth when administered in the absence of palbociclib, demonstrating that their therapeutic activities require the concomitant induction of senescence by palbociclib (Fig 3C and D; and Appendix Fig S3G and H). Similar results were obtained with NCI‐H226 xenografts treated with GalNP(dox) (Appendix Fig S3I). As an additional control, we used empty nanoparticles, GalNP(empty). These empty particles did not have senolytic activity on in vitro senescent cells (Appendix Fig S3J), and treatment of mice with GalNP(empty) did not affect the growth of xenografts and did not improve the growth‐inhibitory effect of palbociclib (Appendix Fig S3J).

 

A potential beneficial aspect of drug encapsulation is reduced drug toxicity. Cardiotoxicity is the most serious side effect of doxorubicin (Chatterjee et al2010), whereas thrombocytopenia is the main toxicity of navitoclax (Kile, 2014). To assess the effect of gal‐encapsulation on the toxicities of doxorubicin and navitoclax, we selected doses of free and encapsulated drug that were similarly effective in reducing palbociclib‐treated xenografts after daily treatments for ~2 weeks (Fig 3E and F; left panels). For doxorubicin‐treated mice, we measured the mRNA levels of Myh7 and Nppa, which reflect cardiac hypertrophy in response to doxorubicin‐induced cardiomyocyte death (Barry et al2008; Richard et al2011). In the case of navitoclax‐treated mice, we measured the levels of serum platelets. Interestingly, gal‐encapsulation completely prevented the cardiotoxicity of doxorubicin (Fig 3E; right panel). Also, gal‐encapsulation of navitoclax produced a mild reduction in platelet counts, which was in contrast to the profound and significant reduction produced by free navitoclax (Fig 3F; right panel). We conclude that gal‐encapsulation of chemotherapeutic drugs is effective in clearing senescent cells in vivo, with the added benefit of reducing the toxicities associated with the drugs.

 

 

Therapeutic activity of gal‐encapsulated cytotoxic drugs on pulmonary fibrosis

 

Based on the above data, we decided to test whether gal‐encapsulated doxorubicin could be therapeutic on bleomycin‐induced pulmonary fibrosis. In this disease model, it has been well established that treatment with senolytics reduces fibrosis and favors functional recovery (Pan et al2017; Schafer et al2017). We first confirmed that GalNP(dox) efficiently released doxorubicin in lungs from bleomycin‐treated mice, but not in healthy lungs, as determined by total fluorescence (Appendix Fig S4A). Based on this, we performed a longitudinal study evaluating the development of the disease by plethysmography and by computerized tomography (CT) according to the schedule shown in Fig 4A. Bleomycin was administered intratracheally and, when analyzed 10 days later by plethysmography, all the mice presented a significant increase in the LR/Cdyn ratio (LR: lung resistance; Cdyn: compliance dynamics), which is indicative of pulmonary fibrosis (Fig 4B). Bleomycin‐treated mice were subsequently treated with daily doses of free or gal‐encapsulated doxorubicin (1 mg/kg) for ~2 weeks. Importantly, at the end of the treatment, mice treated with GalNP(dox), but not with free doxorubicin, presented LR/Cdyn values similar to those of healthy controls (Fig 4B and Appendix Fig S4B). These data were confirmed by computerized tomography, which indicated a significant reduction in the lung volume affected by inflammation and fibrosis in GalNP(dox)‐treated mice (Fig 4C and D and Appendix Fig S4C). Histological analysis of the bleomycin‐treated lungs showed characteristic features of lung fibrosis, namely focal areas of high cellularity with interstitial collagen deposits (stained with Sirius red; Fig 4E). Similar lesions were observed in mice treated with free doxorubicin (Fig 4E). In contrast, mice treated with gal‐encapsulated doxorubicin for ~2 weeks presented a significant collagen reduction in the damaged areas of bleomycin‐treated lungs (Fig 4E). No evidence of hepatic or renal damage was observed in any of the groups, as evaluated by serum markers and histology (Appendix Fig S4D and E).

 

 

EMMM-10-e9355-g005.jpg
Therapeutic activity of gal‐encapsulated doxorubicin on pulmonary fibrosis

 

  • A

    C57BL/6 male mice were subjected to a single intratracheal administration of bleomycin at 1.5 U/kg BW. Beginning at day 10, mice were treated daily with free doxorubicin (tail vein injection, 1 mg/kg) or with GalNP(dox) (tail vein injection, 200 μl of a solution with 4 mg/ml of GalNP containing a total of 1 mg/kg of deliverable doxorubicin), for 18 days, that is, until day 28 post‐bleomycin. Plethysmography and CT were performed at the indicated days.

  • B

    Plethysmography was used to determine the ratio between lung resistance (LR) and compliance dynamics (Cdyn) in the indicated groups before and after the indicated treatments. Values are expressed as mean ± SEM, and statistical significance was assessed for each group by the two‐tailed Student's t‐test comparing the LR/Cdyn values at day 28.

  • C

    Representative CT images of the indicated treatments at days 11 and 29 post‐bleomycin injury. Each pair of images, at days 11 and 29, correspond to the same mouse. The graph represents the ratio between the volume of fibrosis at end of treatment (day 29) and the volume of fibrosis at the beginning of treatment (day 11). Values are expressed as mean ± SEM, and statistical significance was assessed by the two‐tailed Student's t‐test.

  • D

    3D isocontour‐based volume rendering of a representative lung before and after treatment with GalNP(dox). Healthy lung tissue is shown in red and fibrotic lesions are shown in gray. Ventral and lateral views are shown.

  • E

    Representative images of Sirius red staining of lungs from mice subjected to the indicated treatments 2 days after the last CT analysis, visualized by bright field or polarized light. The graph represents the fraction of collagen per damaged area. Red represents densely packed collagen, yellow represents intermediately packed collagen, and green represents loosely packed collagen. Values are expressed as mean ± SD, and statistical significance comparing the total fraction of collagen per damaged area was assessed by the two‐tailed Student's t‐test. Scale bar: 50 μm.

 

 

 

 

Discussion

 

In recent years, it has become evident that multiple diseases are associated with the presence of senescent cells and that the elimination of these cells has therapeutic benefits in mouse models (Childs et al2017; Soto‐Gamez & Demaria, 2017). Here, we report a versatile vehicle to deliver small therapeutic compounds to senescent lesions in vivo. We show that gal‐encapsulated cytotoxic drugs are therapeutic against tumors treated with senescence‐inducing chemotherapy and against bleomycin‐induced pulmonary fibrosis.

 

Gal‐encapsulation is based on the high levels of lysosomal β‐galactosidase activity present in many senescent cells (abbreviated as SAβGal). Although SAβGal is not a perfect marker of senescence, damaged or diseased tissues are in general positive for this marker (Sharpless & Sherr, 2015). Gal‐encapsulation consists on spherical particles (100 nm diameter) of porous silica loaded with the chosen therapeutic cargo and then coated with galacto‐oligosaccharides that prevent the diffusion of the cargo out of the silica matrix (Agostini et al2012). Using this approach, we previously generated gal‐encapsulated rhodamine and we showed preferential release of rhodamine in progeric fibroblasts (dyskeratosis congenita) compared to healthy fibroblasts (Agostini et al2012). We show here that this is also the case for human cancer cell lines (melanoma, lung squamous cell carcinoma, and head and neck squamous cell carcinoma) responsive to palbociclib. This compound is a CDK4/CDK6 inhibitor that efficiently induces senescence in responsive cancer cells. Importantly, we demonstrate that gal‐encapsulated rhodamine, abbreviated GalNP(rho), preferentially releases rhodamine within senescent lesions in vivo. In particular, we demonstrate this in palbociclib‐treated tumor xenografts and in lungs damaged with bleomycin. When we used a different fluorescent cargo, indocyanine green, we also observed preferential release in palbociclib‐treated tumors.

 

We have used bleomycin‐induced lung fibrosis to investigate which cell types are labeled with rhodamine after in vivo treatment with GalNP(rho). Flow cytometry analysis has revealed that lung epithelial cells and fibroblasts are preferentially labeled with rhodamine upon GalNP(rho) injection in bleomycin‐treated lungs compared to control lungs. Interestingly, the RNAseq profile of these cells was enriched in signatures of senescence. Cargo release by GalNP(rho) was also observed in lung macrophages, both in control and bleomycin‐treated lungs. In this regard, it is relevant to mention that macrophages are known to metabolize nanoparticles (Wilhelm et al2016) and can present SAβGal activity without actually being senescent (Hall et al2017). Future analyses should determine how gal‐encapsulated drugs affect macrophage populations.

 

To test the therapeutic potential of our encapsulation method, we first screened a panel of 80 chemotherapeutic agents and we found that doxorubicin was among the most potent agents killing both senescent and non‐senescent cells. In this regard, it is important to point out that doxorubicin kills cells through multiple mechanisms, many of which are not related to cell division (Gewirtz, 1999). In addition, doxorubicin has intrinsic fluorescence thus allowing to track its delivery. Based on this, we generated gal‐encapsulated doxorubicin, or GalNP(dox), and we determined its delivery in cultured cells. As it was the case for rhodamine, gal‐encapsulated doxorubicin was more efficiently released in senescent cells than in control cells.

Moreover, this preferential release of doxorubicin by GalNP(dox) resulted in higher levels of apoptosis in senescent cells. Interestingly, we noted that soon after addition of GalNP(dox) to senescent cells, doxorubicin was confined to a perinuclear compartment. This is consistent with the expected route of entry of the nanoparticles via endocytosis. Previous investigators working with conjugated forms of doxorubicin have observed that delivery of doxorubicin into lysosomes is very efficient in inducing apoptosis (Nair et al2015; Sheng et al2015). In contrast to the perinuclear localization of encapsulated doxorubicin, administration of the free form of the drug resulted in nuclear localization in all cells, senescent or non‐senescent.

 

These observations support the use of GalNPs as a vehicle to deliver doxorubicin to senescent cells and to induce their apoptotic death.

 

After validating the use of GalNP(dox) with senescent cultured cells, we focused on the two experimental models mentioned above, namely, bleomycin‐induced lung fibrosis and palbociclib‐induced tumor senescence. Upon administration of GalNP(dox), doxorubicin‐derived fluorescence was higher in bleomycin‐treated lungs compared to control lungs. Bleomycin‐damaged mice were daily treated with intravenous administration of GalNP(dox) or free doxorubucin, from day 10 to day 28 post‐bleomycin. Importantly, gal‐encapsulated doxorubicin significantly improved lung elasticity, as evaluated by plethysmography, while free doxorubicin did not restore lung function. These observations were further supported by the quantification of fibrosis by computerized tomography and by histology.

 

In addition, GalNP(dox) showed tumor‐regressing activity in combination with palbociclib when administered to mice carrying tumor xenografts (SK‐MEL‐103 melanoma, and H226 lung squamous cell carcinoma). This was also the case when we used encapsulated navitoclax, a well‐established senolytic drug. It is important to note that GalNP(dox) or GalNP(nav) had no effect in the absence of palbociclib, that is, were effective against senescent tumors but not on growing tumors. Also, empty nanoparticles, GalNP(empty), had no effect on tumor size, either alone or in combination with palbociclib. These observations reinforce the concept that gal‐encapsulation is a useful vehicle to deliver therapeutic drugs into senescent tumor cells, but not into non‐senescent tumor cells. The elimination of senescent tumor cells by GalNP‐mediated senolysis may also affect tumor growth by reducing the SASP and its potential pro‐tumorigenic effects. Moreover, the delivery of drugs into senescent cells may have a local bystander effect on neighboring cells and this may enhance its therapeutic effects.

 

As it is the case of many drugs, doxorubicin and navitoclax present toxicities, which limit their clinical use. The encapsulation of these agents in GalNPs may reduce the exposure of healthy tissues to the drugs and, thereby, diminish undesired effects. In support of this, we have observed that gal‐encapsulation reduced the carditoxicity of doxorubicin and the thrombocytopenia characteristic of navitoclax.

 

Besides the therapeutic potential of gal‐encapsulation for senescence‐related diseases, our vehicle can be of use for the detection of senescence by in vivo imaging. In our work, the encapsulation of fluorophores has allowed us to detect lung tissue regions with senescent cells and palbociclib‐responsive tumors. This opens up the possibility of encapsulating tracers and contrast agents for biomedical imaging of senescent areas that could be of use in multiple age‐related disorders. For example, diagnostic GalNPs could release gadolinium or positron‐emitting radioisotopes in senescent lesions to be detected by MRI or PET, respectively. Gal‐encapsulation could also serve to evaluate the response of solid tumors to the administration of senescence‐inducing chemotherapies or radiotherapies. In addition, cellular senescence is a defining feature of a wide variety of premalignant lesions both in humans and in mice (Collado & Serrano, 2010). Based on this, it is tempting to speculate that GalNPs could also be used for the diagnosis of precancerous tumors. Even more, GalNPs could be potentially employed as a novel theranostic tool, aimed to simultaneously detect and eradicate senescent lesions associated with numerous pathologies, or during aging.

 

In summary, we contribute a versatile strategy to deliver essentially any small compound to senescent lesions in vivo. This novel therapeutic tool can be used to eradicate senescent cells and could be employed in clinical imaging. As a proof of principle, we show that gal‐encapsulated cytotoxic drugs are therapeutically efficient against tumors treated with senescence‐inducing chemotherapy and against pulmonary fibrosis. Finally, and equally important, gal‐encapsulation reduces the systemic toxicities of chemotherapeutic drugs.

 

 

 

.... / ....

 

 

 

 

.


  • Cheerful x 1
  • like x 1





Also tagged with one or more of these keywords: chemotherapy, fibrosis, nanomedicine, palbociclib, senescence

0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users