Pallarès et al. Journal of Hematology & Oncology (2020) 13:36 https://doi.org/10.1186/s13045-020-00863-9 RESEARCH Open Access An Auristatin nanoconjugate targeting CXCR4+ leukemic cells blocks acute myeloid leukemia dissemination Victor Pallarès1,2,3†, Ugutz Unzueta1,3,4†, Aïda Falgàs1,2,3, Laura Sánchez-García3,4,5, Naroa Serna3,4,5, Alberto Gallardo1,6, Gordon A. Morris7, Lorena Alba-Castellón1, Patricia Álamo1,3, Jorge Sierra2,8, Antonio Villaverde3,4,5, Esther Vázquez3,4,5*, Isolda Casanova1,2,3* and Ramon Mangues1,2,3* Abstract Background: Current acute myeloid leukemia (AML) therapy fails to eliminate quiescent leukemic blasts in the bone marrow, leading to about 50% of patient relapse by increasing AML burden in the bone marrow, blood, and extramedullar sites. We developed a protein-based nanoparticle conjugated to the potent antimitotic agent Auristatin E that selectively targets AML blasts because of their CXCR4 receptor overexpression (CXCR4+) as compared to normal cells. The therapeutic rationale is based on the involvement of CXCR4 overexpression in leukemic blast homing and quiescence in the bone marrow, and the association of these leukemic stem cells with minimal residual disease, dissemination, chemotherapy resistance, and lower patient survival. Methods: Monomethyl Auristatin E (MMAE) was conjugated with the CXCR4 targeted protein nanoparticle T22- GFP-H6 produced in E. coli. Nanoconjugate internalization and in vitro cell viability assays were performed in CXCR4+ AML cell lines to analyze the specific antineoplastic activity through the CXCR4 receptor. In addition, a disseminated AML animal model was used to evaluate the anticancer effect of T22-GFP-H6-Auristatin in immunosuppressed NSG mice (n = 10/group). U of Mann-Whitney test was used to consider if differences were significant between groups. Results: T22-GFP-H6-Auristatin was capable to internalize and exert antineoplastic effects through the CXCR4 receptor in THP-1 and SKM-1 CXCR4+ AML cell lines. In addition, repeated administration of the T22-GFP-H6- Auristatin nanoconjugate (9 doses daily) achieves a potent antineoplastic activity by internalizing specifically in the leukemic cells (luminescent THP-1) to selectively eliminate them. This leads to reduced involvement of leukemic cells in the bone marrow, peripheral blood, liver, and spleen, while avoiding toxicity in normal tissues in a luminescent disseminated AML mouse model. (Continued on next page) * Correspondence: esther.vazquez@uab.cat; icasanova@santpau.cat; rmangues@santpau.cat † Victor Pallarès and Ugutz Unzueta contributed equally to this work. 3 CIBER en Bioingeniería, Biomateriales y Nanomedicina (CIBER-BBN), 08041 Barcelona, Spain 1 Biomedical Research Institute Sant Pau (IIB-Sant Pau), Hospital de la Santa Creu i Sant Pau, 08041 Barcelona, Spain Full list of author information is available at the end of the article © The Author(s). 2020 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Pallarès et al. Journal of Hematology & Oncology (2020) 13:36 Page 2 of 19 (Continued from previous page) Conclusions: A novel nanoconjugate for targeted drug delivery of Auristatin reduces significantly the acute myeloid leukemic cell burden in the bone marrow and blood and blocks its dissemination to extramedullar organs in a CXCR4+ AML model. This selective drug delivery approach validates CXCR4+ AML cells as a target for clinical therapy, not only promising to improve the control of leukemic dissemination but also dramatically reducing the severe toxicity of classical AML therapy. Keywords: Acute myeloid leukemia, CXCR4, Targeted nanoparticle, Auristatin nanoconjugate, Leukemic stem cells, Disseminated AML mouse model Background inhibition of AML dissemination in the absence of C-X-C chemokine receptor type 4 (CXCR4) receptor toxicity, because of its low uptake in normal tissues, overexpression (CXCR4+) associates with dissemin- ation, relapse, minimal residual disease, and poor sur- as we reported for a NC loaded with a CRC sensitive vival in acute myeloid leukemia (AML) patients [1–4]. drug [22]. To that aim, we generated a biolumines- Moreover, CXCR4+ leukemic blasts become anchored cent CXCR4+ disseminated AML model, which allows and quiescent in the bone marrow (BM) through a non-invasive follow-up of the response in BM, liver, CXCR4-CXCL12 (C-X-C motif chemokine 12 or stro- and spleen, the clinically relevant sites. We describe mal cell-derived factor 1; SDF1) interaction [5], dis- placing normal hematopoietic stem cells (HSCs) to CXCR4-dependent internalization and killing of cause patient death [6]. CXCR4+ AML cells also cir- CXCR4+ AML cells for this novel NC in vitro. We culate in the blood [4] to reach extramedullar also report in this in vivo model, a potent antineo- CXCL12 producing sites [4, 7], mainly the liver and plastic effect on BM and blockade of dissemination to spleen [5, 7–11]. All these functions support CXCR4+ extramedullar sites. AML cells as Leukemic stem cells (LSCs), despite the lack of a formal proof. Similarly, CXCR4+ cancer cells are metastatic stem cells in several solid tumors [12– Methods 17]. Consequently, different groups are developing Drugs, nanoparticle, and nanoconjugate small drug CXCR4 inhibitors for AML therapy [5, 7, Monomethyl Auristatin E (MMAE) is a potent tubulin 18] aimed at eliminating CXCR4+ LSCs, responsible polymerization inhibitor that was acquired by custom for AML aggressiveness and patient death. systhesis from Levena Biopharma (Levena Biopharma, Nanotechnological approaches can increase selectiv- San Diego, CA, USA) as a maleimide conjugated ity regarding target cancer cell killing, while reducing MMAE (MC-MMAE). T22-GFP-H6 is a CXCR4 tar- systemic toxicity [19]. Here, we specifically tested if a geted protein nanoparticle produced in E. coli as pre- novel T22-GFP-H6-Auristatin nanoconjugate (NC) se- viously described [21]. T22-GFP-H6-Auristatin lectively eliminates CXCR4+ AML cells, and therefore nanoconjugates were synthesized by covalent binding LSCs, in an AML disseminated model, to improve of the targeting vector (T22-GFP-H6) with the thera- chemotherapy performance, which reduces the differ- peutic moiety (MC-MMAE). For that, an excess of entiated tumor bulk but does not eliminate LSCs MC-MMAE was incubated with T22-GFP-H6 nano- [20]. The NC incorporates a protein nanoparticle to particles and reacted with amino groups of external selectively internalize in CXCR4+ AML cells by lysines in a 1:50 ratio (protein to MC-MMAE) for 4 h exploiting high CXCR4 overexpression in LSCs as at room temperature. T22-GFP-H6-Auristatin nano- compared to HSCs [6, 20], similarly to a previous ap- conjugates were then again purified by IMAC affinity proach in colorectal cancer (CRC) [21, 22]. The NC chromatography using HiTrap Chelating HP 5 mL is loaded with Auristatin E, a microtubule- columns in an ÄKTA pure (GE Healthcare, Chicago, destabilizing toxin, introduced in clinical onco- IL, USA) in order to remove non-reacted free MC- hematology [23, 24]. This drug eliminates quiescent MMAE. Finally, re-purified nanoconjugates were dia- cancer cells [23, 24], overcoming the limited effective- lyzed against sodium carbonate buffer (166 mM ness of current AML chemotherapy, unable to kill NaCO3H, 333 mM NaCl pH = 8) and conjugation ef- LSCs in the G0 phase [6, 20]. ficiency and presence of free MMAE checked by MALDI-TOF mass spectrometry. We expected the novel T22-GFP-H6-Auristatin NC The volume size distribution of T22-GFP-H6 nano- to selectively deliver Auristatin E to cycling and qui- particles and resulting nanoconjugates (T22-GFP-H6- escent CXCR4+ LSCs target cells achieving a potent Auristatin) was determined by dynamic light Pallarès et al. Journal of Hematology & Oncology (2020) 13:36 Page 3 of 19 scattering at 633 nm in a Zetasizer Nano (Malvern In- Cell lines struments, Malvern, UK). Measurements were per- The AML cell lines used in this study were purchased formed in triplicate. In addition, ultrastructural from DSMZ (Leibniz Institute DSMZ-German Collec- morphometry of T22-GFP-H6-Auristatin nanoconju- tion of Microorganisms and Cell Cultures, Braun- gates (size and shape) was determined at nearly native schweig, Germany). THP-1 and SKM-1 were cultured state with field emission scanning electron microscopy in RPMI-1640 medium supplemented with 10% FBS, (FESEM). Samples were directly deposited on silicon 10 mmol/L L-glutamine, 100 U/mL penicillin, 10 mg/ wafers (Ted Pella Inc., Redding, CA, USA) for 30 s, mL streptomycin and 0.45 μg/mL fungizone (Gibco, excess of liquid blotted, air dried, and immediately Thermo Fisher Scientific (TFS), Waltham, MA, USA). observed without coating with a FESEM Zeiss Merlin Cells were kept at 37 °C in a humidified atmosphere (Zeiss, Oberkochen, Germany) operating at 1 kV and of 5% CO2. THP-1 cells were transfected with a plas- equipped with a high resolution in-lens secondary mid encoding the luciferase gene that confers bio- electron detector. Representative images of a general luminescence (BLI) to the cells using Lipofectamine field were captured at two high magnifications (× LTX and PLUS reagents (A12621, Invitrogen, TFS) 100,000 and × 120,000). In a quantitative approach, according to the manufacturer’s instructions. nanoconjugates average size from FESEM images were analyzed by Image J software (1.8.0.172, National In- Quantitation of CXCR4 membrane levels in AML cell lines stitutes of Health, USA) [25]. To detect cell surface expression of CXCR4 in AML cell The average molar mass of T22-GFP-H6 nanopar- lines, fluorescence-activated cell sorting (FACS) analysis ticles and T22-GFP-H6-Auristatin nanoconjugates was performed as described before [26]. Two technical was measured by a size exclusion chromatography and three biological replicates were performed. Data ac- coupled to a multi-angle light scattering (SEC- quisition was analyzed by Cell Quest Pro software (BD MALS). Samples were injected in a Superdex 200 in- Biosciences, San Jose, CA, USA) and results were crease 10/300 GL column (GE Healthcare, Chicago, expressed as mean fluorescence intensity (MFI) ± stand- IL, USA) and run in a degassed sodium carbonate ard error (SE). buffer with Nickel (166 mM NaCO3H, 333 mM NaCl, 0.1 mM NiCl2 pH = 8). The eluent was monitored Antineoplastic effect and internalization analyses of T22- by an in-line UV-Vis detector, a Dawn Heleos MALS GFP-H6-Auristatin in CXCR4+ AML cells in vitro detector and an Optilab rEX RI detector (Wyatt To assess the antineoplastic effect of the nanoparticle Technology Corporation, Santa Barbara, CA, USA). T22-GFP-H6 and the nanoconjugate T22-GFP-H6- All data were analyzed using Astra 6.0.2.9 software Auristatin in AML cell lines, XTT cell viability assays (Wyatt Technology Corporation). Molecular weights were performed. Cells were cultured at 20 × 104 were double-checked from MALS with UV and RI cells/mL for 24 h in 96-well plates. After that, T22- signals using ASTRA software and dn/dc (mL/g) GFP-H6-Auristatin was added in a range of 30–300 values of 0.185 and UV extinction Coefficient (mL/ nM and T22-GFP-H6 at 500 nM. Forty-eight hours (mg.cm)) values of 1.099 for all proteins. Conjugated later, cells were incubated for 4 h with the XTT re- MMAE ratio was calculated by subtracting the aver- agents (Cell Proliferation Kit II, Roche Diagnostics, age molecular weight of T22-GFP-H6 nanoparticles Basel, Switzerland) and then the absorbance was mea- to the average molecular weight of T22-GFP-H6- sured at 492 nm in a FLUOstar OPTIMA spectropho- Auristatin nanoconjugates and dividing by the mo- tometer (BMG Labtech, Ortenberg, Germany). Five lecular weight of MMAE molecule. technical and three biological replicates were per- Fluorescence of T22-GFP-H6 nanoparticles and formed for each condition. The growth inhibitory ac- T22-GFP-H6-Auristatin nanoconjugates was deter- tivity was obtained by subtracting the absorbance of mined in a Cary Eclipse fluorescence spectrophotom- the blanks and expressed as the percentage of cell eter (Varian Inc, Palo Alto, CA, USA) at 510 nm growth inhibition (± SE), as compared with untreated upon excitation at 488 nm. controls. T22-GFP-H6-Auristatin nanoconjugates proteolytic To find out the mechanism of cell death mediating stability in serum was determined by western blot T22-GFP-H6-Auristatin antineoplastic activity in AML immunostaining using a monoclonal anti-His anti- cells, nuclear staining with DAPI dye in THP-1 and body (Santa Cruz Biotechnology, Dallas, TX, USA) SKM-1 cells exposed to 240 nM T22-GFP-H6-Aurista- upon incubation in front of human serum (Sigma- tin or buffer for 24 or 48 h, in 12-well plates (50 × Aldrich, St-Louis, MO, USA) at different time points 104 cells/mL), was performed. After incubation, the (0, 2, 5, and 24 h) at 37 °C and a final concentration media was collected and centrifuged to obtain the of 1 mg/mL. suspended cells. Cells were rinsed once with PBS. Pallarès et al. Journal of Hematology & Oncology (2020) 13:36 Page 4 of 19 Fig. 1 (See legend on next page.) Pallarès et al. Journal of Hematology & Oncology (2020) 13:36 Page 5 of 19 (See figure on previous page.) Fig. 1 T22-GFP-H6-Auristatin nanoconjugates characterization and stability. a T22-GFP-H6 nanoparticle conjugation with maleimide functionalized Monomethyl Auristatin E (MMAE). b MALDI-TOF mass spectrometry of wild-type T22-GFP-H6 nanoparticles and T22-GFP-H6-Auristatin nanoconjugates (top) and free MMAE before and after nanoconjugates re-purification (bottom). Each peak corresponds to the covalent addition (+ 911 Da) of a single MMAE molecule (top). Red boxes indicate free Auristatin mass spectrometry spectra (bottom). c Hydrodynamic volume size distribution of wild-type T22-GFP-H6 nanoparticles (red) and T22-GFP-H6-Auristatin nanoconjugates (blue) determined by dynamic light scattering. Samples were analyzed in triplicate and data represented as mean +/− SE. PDI indicates polydispersion index. d Average molar mass distribution of wild-type T22-GFP-H6 nanoparticles (red) and T22-GFP-H6-Auristatin nanoconjugates (blue) determined by size exclusion chromatography coupled to multi-angle light scattering (SEC-MALS). e Representative electron microscopy (FESEM) images of T22-GFP-H6- Auristatin nanoconjugates presented in two magnifications (see inset). Scale bars indicate 100 nm. In the bottom, the quantitative average size of nanoconstructs determined by image analysis and shown as mean ± SE. f Proteolytic stability of T22-GFP-H6-Auristatin nanoconjugates in human serum at different incubation times up to 24 h analyzed by western blot immunodetection with a monoclonal anti-His antibody. “S” indicates human serum control. MW, molecular weight; SE, standard error Afterwards, they were centrifuged and fixed (3.7% p- AML mouse model, 19 mice were intravenously (IV) formaldehyde in PBS, pH= 7.4) for 10 min at − 20 °C, injected with luciferase-transfected THP-1 cells (THP-1- washed with PBS, and resuspended in 10 μL of PBS. Luci; 1 × 106 cells/200 μL). These mice were divided ran- Finally, cells were mounted on a slide with DAPI dye domly into two different experimental groups (day 0) (ProLong™ Gold Antifade Mountant with DAPI, (Fig. 5). One group (VEH; n = 10) was IV injected (Day P36935, Invitrogen, Carlsbad, CA, USA) and observed 2) with the vehicle of the nanoconjugate (Buffer for the appearance of the apoptotic bodies or mitotic NaCO3H + NaCl pH = 8) and another group (T22-AUR; catastrophe (MC), as induced events, under a fluores- n = 9) with 100 μg of T22-GFP-H6-Auristatin daily for a cence microscope (Olympus BX53, Olympus, Tokyo, total of 9 doses. The evolution of AML dissemination in Japan). Cell death events such as MC or apoptosis mice was monitored using the IVIS Spectrum Imaging were quantified using the point tool in Image J soft- System (PerkinElmer, Waltham, MA, USA) three times ware (1.8.0.172) and 4 randomly chosen fields for per week until the day of the euthanasia. Animal weight each condition. was measured the same day as that of BLI analysis. All To evaluate the internalization capacity of the nanocon- mice were euthanized the day that the first animal pre- jugate T22-GFP-H6-Auristatin in AML cell lines, FACS sented relevant signs of disease such as lack of mobility analyses were performed as described before [26] exposing or 10% weight loss. Animals were intraperitoneally cells to the vehicle or with T22-GFP-H6-Auristatin be- injected with luciferin; after 5 min, blood was obtained tween 30 and 300 nM. Two technical and three biological by intracardiac extraction, and mice were killed by cer- replicates were performed for each condition. Data acqui- vical dislocation. Blood samples were collected for BLI sition was analyzed by Cell Quest Pro software (BD Biosci- analysis in IVIS and CD45-positive cells detection by ences) and results were expressed as MFI ± SE. flow cytometry. Tissues were excised to analyze the BLI Receptor competition assays were performed with levels ex vivo, and preserved in paraffin for further histo- the antagonist of CXCR4, AMD3100. To that aim, logical and immunohistochemical analyses. BLI mea- internalization analyses or XTT cell viability assays surements were expressed as total flux of BLI (photons/ were performed as described above, treating cells second; radiance photons) ± SE in both in vivo and with 240 nM of T22-GFP-H6-Auristatin with or ex vivo studies. without a pre-treatment for 1 h of AMD3100 (2000 nM). Detection of AML cells in circulating blood Evaluation of the antineoplastic effect of T22-GFP-H6- To detect leukemic cells in the blood, CD45 leukocyte Auristatin in a disseminated CXCR4+ AML mouse model marker detection by FACS was performed. Red blood NSG (NOD-scid IL2Rgammanull) female mice (4 weeks cells were lysed from the blood samples according to old) were obtained from Charles River Laboratories the instructions of the lysis buffer (eBioscience, TFS). (Wilmington, MA, USA). Mice were housed in microiso- After that, cells were incubated with 10 μL of PE-Cy5 lator units with sterile food and water ad libitum. All Mouse Anti-Human CD45 (ref. 555484, BD Biosci- procedures were conducted in accordance with the ences) for 20 min in agitation and dark at 4 °C. Cells guidelines approved by the institutional animal Ethics were washed and re-suspended with PBS-BSA. CD45- Committee of Hospital Sant Pau. Three healthy (not positive cells were detected by flow cytometry using having cancer) and untreated mice (not administered FACS Calibur cytometer (BD Biosciences). Data ac- with vehicle or nanoconjugate) were assigned to the quisition was analyzed by Cell Quest Pro software Normal mouse group (normal, n = 3). To generate the (BD Biosciences) and results were expressed as the Pallarès et al. Journal of Hematology & Oncology (2020) 13:36 Page 6 of 19 Fig. 2 (See legend on next page.) Pallarès et al. Journal of Hematology & Oncology (2020) 13:36 Page 7 of 19 (See figure on previous page.) Fig. 2 Internalization capacity and antineoplastic activity of T22-GFP-H6 and T22-GFP-H6-Auristatin in AML cell lines. a CXCR4 membrane expression of THP-1 and SKM-1 AML cell lines by flow cytometry. b Evaluation of T22-GFP-H6-Auristatin internalization, by flow cytometry, in CXCR4+ leukemic cells after 1 h of incubation at different concentrations. c Competition assay with AMD3100 (240 nM T22-GFP-H6-Auristatin and 2000 nM AMD3100) to determine the specificity of internalization through CXCR4 receptor. d Evaluation of antineoplastic activity of T22-GFP-H6 after 48 h of treatment in AML cell lines performed by XTT assay. e Anticancer activity of T22-GFP-H6-Auristatin in AML cell lines after 48 h of incubation by XTT assay. f Competition of antineoplastic activity of T22-GFP-H6-Auristatin at 240 nM with AMD3100 at 2000 nM after 48 h exposure by XTT assay. Results are presented as mean ± SE MFI in a, b, and c, and mean ± SE percent of cell viability in d, e, and f. U of Mann- Whitney test was used to test differences between groups. In b and e, T22-GFP-H6-Auristatin treated cells were compared with the buffer treated cells (Ctrl). Statistical significant differences are indicated by * when p value < 0.05 and ** when p value < 0.001. n.s. means no significance. AML, acute myeloid leukemia; Ctrl, control; MFI, mean fluorescence intensity; SE, standard error number of CD45-positive events of a total of 10,000 ALTL ref. 20764957 322; CREJ2 ref. 04810716 190, events ± SE. respectively. Roche Diagnostics), adapted for a COBAS 6000 autoanalyzer (Roche Diagnostics), to Immunohistochemical staining of leukemic cells in AML evaluate toxicity in tissues such as the liver, kidneys, affected organs or heart after T22-GFP-H6-Auristatin treatment in Immunohistochemical analysis was performed using mice. paraffin-embedded tissue samples as described before [26] using anti-human CD45 antibody (ref. IR75161-2, Statistical analysis Dako, Agilent, Santa Clara, CA, USA) to detect infil- Statistical analyses were performed in the IBM SPSS tration of AML cells and anti-human CXCR4 anti- Statistics (Release 22.0.0.0, New York, NY, USA). body (ref. AB124824, Abcam, Cambridge, UK) to Mann-Whitney U test was used in both in vitro and evaluate CXCR4 expression in tissues. AML infiltra- in vivo assays, and any differences were considered tion by CD45 detection was evaluated with cellSens statistically significant when the p value was < 0.05. Dimension 1.9 software (Olympus, Tokyo, Japan) IC50 were calculated with SigmaPlot (Release using the Counter and measurement tool. Results 12.0.0.182; Systat Software, Inc., San Jose, CA, USA) were expressed as the percentage of surface stained ± using non-linear regression test with Hill-3 parameter SE. Stain intensity was calculated using Image J soft- adjustment. ware (1.8.0.172) and the Colour Deconvolution Plugin with the H DAB vector to split the brown staining adjusting the threshold to 150. After that, the Results “Analyze particles” plugin was used to detect all T22-GFP-H6-Auristatin nanoconjugate synthesis and stained areas and the mean gray value was obtained characterization combining all selected black areas and measuring T22-GFP-H6-Auristatin nanoconjugates (NCs) were using the ROI Manager. The intensity value of each generated by covalent reaction of maleimide func- analyzed slide was calculated subtracting 255 to the tionalized MMAE with amine groups of exposed ly- mean gray value obtained by the Image J analysis. Six sines in T22-GFP-H6 nanoparticles (Fig. 1a). slides for each treatment type and tissue were ana- Reaction efficacy was then checked by MALDI-TOF lyzed and results were expressed as mean ± SE inten- mass spectrometry. In this context, MALDI-TOF sity values. Two independent observers evaluated all spectra of T22-GFP-H6-Auristatin NCs showed up samples to made consensus decisions. to 14 peaks with sequential molecular weight in- crease of around 911 Da over the wild-type T22- Toxicity analyses in tissues and blood GFP-H6 molecular weight (30.6 kDa) corresponding Hematoxylin and eosin staining was performed in the to the addition of differences number of MC-MMAE organs infiltrated by leukemia cells (bone marrow, molecules (911 Da) (Fig. 1b top). Moreover, low mo- liver, and spleen), and also in non-affected organs to lecular weight spectra showed efficient elimination of evaluate possible off-target toxicities of the nanocon- non-conjugated free MC-MMAE after nanoconju- jugate (lung, brain, pancreas, heart, and kidney). In gates re-purification by IMAC affinity chromatog- addition, plasma glutamic oxaloacetic transaminase raphy (Fig. 1b bottom). (GOT) and glutamic pyruvic transaminase (GPT) en- Auristatin conjugation resulted in no significant spe- zyme activities, and creatinine levels were determined cific fluorescence alteration (data not shown) but in a using commercial kits (ASTL ref. 20764949 322; significant increase of NCs size (17.6 nm) over the wild- Pallarès et al. Journal of Hematology & Oncology (2020) 13:36 Page 8 of 19 Fig. 3 (See legend on next page.) Pallarès et al. Journal of Hematology & Oncology (2020) 13:36 Page 9 of 19 (See figure on previous page.) Fig. 3 Study of cell death in AML cell lines caused by T22-GFP-H6-Auristatin treatment. Study of the mechanism of T22-GFP-H6-Auristatin- induced cell death. a Induction of apoptosis and mitotic catastrophe by T22-GFP-H6-Auristatin at 240 nM for 24 and 48 h of exposure detected by DAPI staining in CXCR4-positive THP-1 and SKM-1 leukemic cell lines. Yellow and red arrows point to MC figures and apoptotic bodies, respectively. b Quantitation of the number of cells undergoing mitotic catastrophe triggered by treatment with T22-GFP-H6-Auristatin at 240 nM in THP-1 and SKM-1 for 24 (orange bars) or 48 h (yellow bars). c Quantitation of the number of apoptotic bodies in THP-1 and SKM-1 induced by exposure to 240 nM T22-GFP-H6-Auristatin for 24 (orange bars) or 48 h (yellow bars). Results in b and c are presented as mean ± SE cell number for each cell death event (MC or apoptosis) in all fields analyzed. * indicates significant differences between groups with a p value < 0.05. CTRL, control; MC, mitotic catastrophe; SE, standard error type T22-GFP-H6 nanoparticles size (12.4 nm) (Fig. 1c). unconjugated T22-GFP-H6 nanoparticle at 500 nM. SEC-MALS analysis indicated an average molar mass in- XTT assays showed that T22-GFP-H6 had no anti- crease of 186.3 kDa in T22-GFP-H6-Auristatin NCs cancer activity in the two tested cell lines (Fig. 2d). In (522 kDa) over the wild-type T22-GFP-H6 nanoparticles contrast, nanoconjugate had a dose-dependent anti- (335.7 kDa). Considering that each T22-GFP-H6 nano- neoplastic effect after cell exposure (Fig. 2e), with a particle is generated by the assembling of an average of IC50 of 156.0 and 105.8 nM in THP-1 and SKM-1, re- 11 protein monomers (30.6 kDa each) and the molecular spectively. Finally, the competition assays showed that weight of MC-MMAE represents 911 Da, this total pre-treating the cells with AMD3100 significantly de- molar mass increase over the whole nanoparticle indi- creased the anticancer activity of T22-GFP-H6-Auris- cates the addition of an average of 18.5 MMAE mole- tatin in both AML cell lines (THP-1: 54.5 ± 6.3 vs cules per protein monomer or 204.5 MMAE molecules 28.3 ± 5.6% viability, p < 0.001; SKM-1: 42.1 ± 6.9 vs per nanoparticle (Fig. 1d). Moreover, we also performed 15.5 ± 1.9% viability, p < 0.001) (Fig. 2f). a morphometric characterization of the NC using field Additionally, the mechanism of cell death induced emission scanning electron microscopy (FESEM) which by T22-GFP-H6-Auristatin was analyzed in both AML allows to visualize the size and shape of our NCs with- cell lines using nuclear staining with DAPI dye (Fig. out any staining. The FESEM images (Fig. 1e) showed 3). Cell treatment with 240 nM T22-GFP-H6-Aurista- circular NCs compatible in size with dynamic light scat- tin resulted in the induction of apoptosis and mitotic tering data. Finally, the stability of the NC in serum was catastrophe (MC) at 24 and 48 h after cell exposure determined by western blot, showing the absence of deg- for both AML cell lines (Fig. 3a). More precisely, MC radation for at least 24 h in the blood (Fig. 1f). events were significantly increased after 24 h of incu- bation with the NC compared to the number of MCs CXCR4-dependent NC internalization and killing of CXCR4 of the CTRL cells in THP-1 (51.8 ± 2.9 vs 7.8 ± 1.0 + AML cells in vitro MCs, p = 0.020) and SKM-1 (42.0 ± 3.5 vs 6.8 ± 1.8 First, CXCR4 levels of THP-1 and SKM-1 cell lines MCs, p = 0.020) (Fig. 3b). Similarly, the number of were evaluated by flow cytometry. In this context, apoptosis found in NC-treated cells was increased at high membrane expression of CXCR4 was detected in 24 h in THP-1 (50.5 ± 5.3 vs 4.3 ± 0.9 apoptotic both tested AML cell lines (THP-1: 70.1 ± 6.7 MFI; bodies, p = 0.020) and SKM-1 (22.3 ± 4.3 vs 5.0 ± SKM-1: 66.6 ± 6.3 MFI) (Fig. 2a). Afterwards, the in- 1.1 apoptotic bodies, p = 0.021) compared to the ternalization capacity of the T22-GFP-H6-Auristatin apoptosis observed in buffer-treated cells (CTRL) (Fig. NC was assessed by flow cytometry in these AML cell 3c). After 48 h, the number of cell death events were lines. T22-GFP-H6-Auristatin internalized in a dose- similar to those found at 24 h (Fig. 3b, c). However, dependent manner after 1 h exposure reaching similar in SKM-1, apoptotic bodies increased to reach 39.3 ± levels in THP-1 and SKM-1 cells, in the 30–300 nM 2.2 at 48 h (p = 0.020) (Fig. 3c). range (Fig. 2b). Finally, a competition assay using AMD3100, a CXCR4 antagonist, showed that the pre- Generation of a CXCR4+ AML model with bone marrow treatment of cells with AMD3100 decreased signifi- and extramedullar involvement cantly the internalization capacity of T22-GFP-H6- A disseminated AML mouse model overexpressing Auristatin after 1 h exposure (THP-1: 13.0 ± 0.5 vs CXCR4 and expressing luciferase that allows the follow- 5.4 ± 0.2 MFI, p = 0.004; SKM-1: 14.8 ± 0.6 vs 5.5 ± up of the fate of the target leukemic cells under therapy 0.1 MFI, p = 0.004) (Fig. 2c). was generated. This model was afterwards used to evalu- The antitumor activity of T22-GFP-H6-Auristatin ate the anticancer effect of T22-GFP-H6-Auristatin. was also evaluated in vitro in THP-1 and SKM-1 Model generation consisted in the IV injection of THP- AML cell lines using XTT cell viability assays. First, 1-Luci cells in immunosuppressed NSG mice. After the we analyzed the antineoplastic activity of the injection of 1 × 106 cells, the dissemination of AML can Pallarès et al. Journal of Hematology & Oncology (2020) 13:36 Page 10 of 19 Fig. 4 Follow-up of bioluminescence, detection of CD45 and CXCR4 expression in the disseminated AML mouse model. a Evolution of bioluminescence along 14 days period after the intravenous (i.v.) injection of CXCR4+ luciferase-transfected THP-1 cells (THP-1-Luci) in NSG mice measuring radiance intensity using the IVIS Spectrum. b Detection of CD45 and CXCR4 levels, by IHC, in the bone marrow, liver, and spleen of the leukemic mice 14 days after i.v. injection of THP-1-Luci cells be non-invasively monitored with the IVIS Spectrum of cancer growth started giving clear signs of ad- equipment (Fig. 4a). In this model, bioluminescence was vanced disease or a significant weight loss in mice be- started to be detected 6 days after injection of cells and tween days 14 and 15 after tumor cell injection, the until the day of the euthanasia. The increasing burden time point at which the mice had to be killed to Pallarès et al. Journal of Hematology & Oncology (2020) 13:36 Page 11 of 19 avoid suffering. At necropsy, after mouse euthanasia, Nanoconjugate reduction of leukemia dissemination in tissues were preserved in paraffin and an exhaustive the bone marrow and extramedullar sites analysis was performed to evaluate CD45 (presence of The antineoplastic activity of the T22-GFP-H6-Auris- leukemic cells) and CXCR4 levels by IHC. Expression tatin NC was assessed in vivo, in the disseminated of both proteins was detected in the bone marrow, THP-1 AML model described above. The experimen- liver, and spleen of the mice, therefore generating a tal design is depicted in Fig. 5a. On day 12 after disseminated THP-1 AML model displaying high AML cell injection, the VEH group presented a sig- CXCR4 overexpression in leukemia-involved tissues nificant decrease in body weight in comparison to the (Fig. 4b). CXCR4 expression was detected not only in T22-AUR-treated group (19.2 ± 0.3 vs 20.5 ± 0.4 gr. the cell surface of the THP-1-Luci cells, but also in- resp. p = 0.018) (Fig. 5b). In contrast, the mouse ternalized in endosomes within cells. weight in the Normal group, not bearing cancer and Fig. 5 Antineoplastic activity of T22-GFP-H6-Auristatin in a disseminated AML mouse model. a Experimental design of the in vivo assay used to assess the antineoplastic activity of T22-GFP-H6-Auristatin. b Measurement of total body weight of mice during the experimental time according to the treatment group. Results are presented as mean ± SE body weight in grams. c Comparison of evolution of bioluminescence emission in mice treated with vehicle (VEH) or T22-GFP-H6-Auristatin (T22-AUR) and untreated healthy mice (Normal group) the days 4, 8, 11, and 13 after injection of THP-1-Luci cells, measured by the IVIS Spectrum. d Follow-up of total body bioluminescence emission of mice treated with T22-GFP- H6-Auristatin (T22-AUR) or vehicle (VEH) and untreated healthy mice (Normal) during all the experiment, analyzed in IVIS Spectrum. Results are presented as mean ± SE luminescence values in photons per second (total flux [p/s]). U of Mann-Whitney test was used to assess significant differences between groups in these studies (b, d), and these differences were considered statistically significant when the p value was lower than 0.05. * indicates differences between the VEH and T22-AUR groups, # between the Normal and VEH groups, and ‡ between the Normal and T22-AUR groups. BLI, bioluminescence; T22-AUR, T22-GFP-H6-Auristatin group; VEH, vehicle group.;SE, standard error Pallarès et al. Journal of Hematology & Oncology (2020) 13:36 Page 12 of 19 being untreated, increased significantly as compared to the data, T22-GFP-H6-Auristatin treatment significantly T22-AUR group (on day 13) or to the VEH group (days decreased the dissemination of leukemic cells to the 11–13) (Fig. 5b). From day 8 and until the end of the ex- bone marrow (7.8E+06 ± 8.7E+05 vs 1.3E+07 ± periment, T22-GFP-H6-Auristatin-treated group showed a 1.4E+06(p/s), p = 0.007), liver (1.5E+06 ± 1.2E+05 vs total luminescence intensity in the whole animal signifi- 8.9E+06 ± 4.5E+05(p/s), p = 0.0002), and spleen cantly lower than that observed in the VEH group (Fig. 5c, (3.3E+05 ± 6.8E+04 vs 2.2E+06 ± 2.3E+05(p/s), p = d). These observed differences between groups progres- 0.0002) as compared to the VEH group (Fig. 6a, b). sively increased from the day 8 until 13, even during the 3 BLI levels emitted by the Normal group were signifi- days lapse after the last dose of T22-GFP-H6-Auristatin, cantly lower than the other groups (VEH and T22- which was administered on day 10. The background BLI AUR) in the bone marrow, liver, and spleen (Fig. 6a, detected in the Normal group was significantly lower than b). In addition, blood samples were analyzed with two that registered in VEH or T22-AUR treated groups from different techniques to evaluate the anticancer activity day 6 after cell injection until the end of the experiment. of the nanoconjugate on circulating leukemic cells. In In addition, we measured ex vivo the biolumines- the T22-AUR group the luminescence emitted by cence in the bone marrow and extramedullar sites AML cells circulating in the blood was lower than in separately (liver and spleen). Luminescence emission the VEH group (2.7E+04 ± 3.5E+03 vs 6.7E+04 ± by hindlimbs measured leukemic cell load in the bone 1.4E+04(p/s), resp. p = 0.004) (Fig. 6c). In addition, marrow, whereas luminescence emission by abdominal CD45-positive cells (a surface marker of AML cells mouse sections, which included the liver and spleen, and therefore of THP-1 cells) were also detected by measured the leukemic burden at extramedullar sites. flow cytometry in the blood. Consistently with the A reduction in bioluminescence in T22-GFP-H6-Aur- observed luminescence results, the relative percentage istatin-treated, as compared to the VEH-treated group of CD45-positive cells was significantly lower in the was evident soon after treatment initiation, for full- T22-AUR group than in the VEH group (18.7 ± 6.4 body emission, bone marrow emission, and extrame- vs 100 ± 15.2 %, p = 0.001) (Fig. 6d, e). Therefore, in dullar site emission, all measured sites becoming the CXCR4+ disseminated THP-1 model, T22-GFP- highly significant at day 8 after treatment initiation, a H6-Auristatin treatment reduces the leukemic cell significance that was maintained until the end of the burden in the bone marrow and circulating blood, experiment (day 13) (Table 1). blocking also the leukemic spread to the liver and spleen. Nanoconjugate reduction of AML cell burden in the bone marrow, circulating blood, liver, and spleen NC anticancer effect measured as organ affectation by The anticancer effect of the NC was assessed ex vivo, CD45 cells and associated toxicity measuring luminescence emission by leukemic cells in We analyzed the capacity of the NC to block AML all AML affected organs as well as in circulating involvement in the spleen, liver, and bone marrow blood at the end of the experiment. Based on these (hindlimbs) using IHC anti-CD45 staining in paraffin- Table 1 Bioluminescence intensity levels emitted in different body sections of mice along the experimental time Day after i.v. injection of cells Treatment 1 4 6 8 11 12 13 Full body VEH 6.61E+05 7.91E+05 1.29E+06 3.76E+06 2.95E+07 4.37E+07 6.44E+07 T22-AUR 6.33E+05 8.43E+05 1.25E+06 1.65E+06 1.22E+07 1.92E+07 2.96E+07 p value 0.4624 0.5675 0.3904 0.0002 0.0003 0.0002 0.0002 Hindlimbs section (bone marrow) VEH 2.03E+05 2.80E+05 4.94E+05 1.57E+06 1.54E+07 2.21E+07 2.93E+07 T22-AUR 1.95E+05 3.01E+05 4.67E+05 7.20E+05 8.34E+06 1.25E+07 1.80E+07 p value 0.4375 0.4377 0.1777 0.0002 0.0043 0.0009 0.0033 Abdominal section (extramedullar) VEH 1.94E+05 2.51E+05 4.76E+05 1.51E+06 9.13E+06 1.52E+07 2.70E+07 T22-AUR 1.77E+05 2.88E+05 4.44E+05 5.39E+05 1.41E+06 2.45E+06 4.70E+06 p value 0.2699 0.1651 0.5401 0.0002 0.0002 0.0002 0.0002 Levels of luminescence detected using the IVIS Spectrum in mice treated with vehicle (VEH) or T22-GFP-H6-Auristatin in full body, hindlimbs section (bone marrow), and in abdominal section (including liver and spleen) of mice during the experiment. Results are presented by mean total flux [photons/second] of radiance detected in mice for each group. U of Mann-Whitney test was used to assess the significance of the differences between groups. These differences were considered statistically significant when the p value was lower than 0.05 (bold values) BLI bioluminescence, i.v. intravenously, T22-AUR T22-GFP-H6-Auristatin group, VEH vehicle group Pallarès et al. Journal of Hematology & Oncology (2020) 13:36 Page 13 of 19 Fig. 6 (See legend on next page.) Pallarès et al. Journal of Hematology & Oncology (2020) 13:36 Page 14 of 19 (See figure on previous page.) Fig. 6 Antineoplastic activity of T22-GFP-H6-Auristatin in the bone marrow, circulating blood, and affected extramedullar sites ex vivo. a Comparison of bioluminescence emission between the T22-GFP-H6-Auristatin-treated (T22-AUR, orange bars), vehicle-treated (VEH, blue bars), and untreated healthy mice (Normal, green bars) groups in the bone marrow, liver, and spleen. b Bioluminescence Images comparing leukemia dissemination in the bone marrow, liver, and spleen of mice treated with the nanoconjugate (T22-AUR) or vehicle (VEH), and untreated healthy mice (Normal). c Comparison of levels of bioluminescence emitted by circulating blood between groups (VEH vs T22-AUR) detected ex vivo at the end of the experiment. d Detection of CD45-positive cells in the peripheral blood by flow cytometry in mice treated with vehicle (VEH, blue) or T22-GFP-H6-Auristatin (T22-AUR, orange). Results are presented as percentage, normalizing the number of CD45-positive cells found in a sample by the mean of CD45-positive cells found in all vehicle-treated mice. e Histogram and dot plot representations of FL3-H fluorescence detection in a mouse blood sample treated with vehicle (VEH, blue) overlapped with another treated with T22-GFP-H6-Auristatin (T22-AUR, orange). M1 and R2 sections indicate CD45-positive cells, whereas R1 section indicates CD45-negative cells. Results in studies a and c are presented as mean ± SE luminescence values in photons per second (total flux [p/s]), and as mean ± SE of the percent of CD45-positive cells in d. U of Mann-Whitney test was used in studies a, c, and d to assess significant differences and are presented in the figures with an * to compare between the VEH and T22-AUR groups, # between the Normal and VEH groups, and ‡ between the Normal and T22-AUR groups. The differences between groups were considered significant when the p value was lower than 0.05. PB, peripheral blood; T22-AUR, T22-GFP-H6-Auristatin group; VEH, vehicle group; SE, standard error embedded tissues to detect THP-1 AML cells. The a potent blockade of leukemic cell spread to extrame- percent of positive CD45 within tissue sections was dullar organs. These results were obtained by measur- significantly lower in the T22-AUR group than the ing the bioluminescence emitted by the CXCR4+ VEH group in the spleen (8.3 ± 4.2 vs 35.2 ± 8.3%, luciferase+ AML cells at each affected site, findings resp. p = 0.002), liver (7.3 ± 2.4 vs 25.2 ± 3.7%, resp. that were validated observing a reduction of CD45+ p = 0.004), and bone marrow (82.6 ± 2.8 vs 95.3 ± AML cells in the liver and spleen at the end of treat- 0.9%, resp. p = 0.003) (Fig. 7a, b). In addition, CD45 ment. This potent blockade of leukemic dissemination staining intensity exhibits a similar level between was achieved using a NC dosage that did not induce groups in the liver and bone marrow, whereas the any histological alteration or toxicity on normal or- spleen shows a significant reduction in the T22-AUR gans; therefore, this nanoconjugate displays a high group compared to the VEH-treated group (122 ± 1.8 therapeutic index. vs 131 ± 2.7 p = 0.025, respectively) (Fig. 7c). Finally, Our results show the importance of using a dissem- we performed hematoxylin and eosin staining of the inated model for preclinical AML drug development, three AML affected organs such as the spleen, liver, displaying bone marrow homing, circulating leukemic and bone marrow, and the off-target organs, including cells in the blood, and AML cell involvement at the lung, kidney, heart, pancreas, and brain. We did extramedullar sites. The luminescence emitted by lu- not observe any histological alteration in normal cells ciferase+ AML cells allows the non-invasive follow-up found in these organs, indicating a lack of toxicity in of the response of target CXCR4+ leukemic cells to these tissues (Fig. 8). In addition, GOT, GPT, and NC treatment in each clinically relevant site. This creatinine levels were determined in plasma to evalu- model outperforms most current models used to ate the toxicity, mainly in the liver, kidneys and heart. evaluate the effectiveness of novel drugs or thera- No significant differences were found in these bio- peutic approaches in AML therapy. Thus, many au- chemical parameters between mice treated with 9 thors evaluate only antitumor activity measuring doses of T22-GFP-H6-Auristatin compared to buffer subcutaneous (SC) tumor size reduction in mouse treated mice (VEH) (Additional Fig. 1). Thus, T22- models [27–30], which cannot evaluate the relevant GFP-H6-Auristatin treatment decreased the amount drug effect on AML cells anchored in the bone mar- of CD45 AML cells detected in the spleen, liver, and row [6]. Other groups complement the SC model as- bone marrow, validating the reduction of leukemic sessment with orthotopic or patient-derived PDX cell burden in these organs, without inducing toxicity models [31–36]; however, the lack of luciferase ex- on normal organs. pression in most models prevents the direct measure- ment of the drug effect on AML cells affecting the Discussion bone marrow, liver, or spleen [10]. We developed a novel targeted drug delivery strategy We explain the potent antineoplastic effect obtained for acute myeloid leukemia (AML) treatment. Re- in the bone marrow, circulating cells, and extramedul- peated administration of the T22-GFP-H6-Auristatin lar sites without associated toxicity, in our model, by nanoconjugate (NC) in a CXCR4+ THP-1 AML the achievement of a high selectivity by the T22-GFP- mouse model achieved potent anticancer activity, by H6-Auristatin NC in CXCR4+ AML cell killing, and significantly reducing the leukemic cell burden in the the consequent elimination of leukemic stem cells bone marrow and circulating blood, and by inducing (LSCs). Thus, CXCR4+ AML cells can be considered Pallarès et al. Journal of Hematology & Oncology (2020) 13:36 Page 15 of 19 Fig. 7 (See legend on next page.) Pallarès et al. Journal of Hematology & Oncology (2020) 13:36 Page 16 of 19 (See figure on previous page.) Fig. 7 Detection of CD45-positive cells in the leukemia-infiltrated tissues after treatment with T22-GFP-H6-Auristatin. a Detection of CD45-positive cells by IHC in the bone marrow and extramedullar organs (spleen and liver) of mice treated with vehicle (VEH) or T22-GFP-H6-Auristatin (T22- AUR) 14 days after injection of CXCR4+ THP-1-Luci cells, and comparison with CD45-positive cell detection in same tissues of a normal non- leukemic mouse. b Quantification of the CD45 detection by IHC in the spleen, liver, and bone marrow in mice treated with vehicle (VEH, blue bars) and T22-GFP-H6-Auristatin (T22-AUR, orange bars). This quantification is presented as mean ± SE percentage of stained surface with CD45 antibody in the tissue of all mice. c Quantitation of CD45 staining intensity by IHC in the liver, spleen, and bone marrow in mice treated with vehicle (VEH, blue bars) or T22-GFP-H6-Auristatin (T22-AUR, orange bars). Results are presented as the mean ± SE of the mean gray value obtained in Image J (see the “Methods” section). Significant differences were indicated by * when the p value was lower than 0.05, using U of Mann-Whitney test between groups. T22-AUR, T22-GFP-H6-Auristatin group; VEH, vehicle group; SE, standard error LSCs based on their association with bone marrow CXCR4+ AML cell uptake, and selective elimination by homing, circulation in the blood, and extramedullar the payload drug, differs from the molecularly targeted dissemination, all associated with poor prognosis [2, small drugs that inhibit proteins involved in maintaining 5, 20]. Two additional reasons also support the notion the cancer stem cell phenotype (e.g. Wnt, Hedgehog that this NC achieves selective killing of LSCs, with- NFkB). The passive diffusion of small drugs to all cells out having an impact on normal cells: implies a lack of selectivity in their biodistribution to cancer sites; therefore, the achievement of therapeutic 1. Selectivity in CXCR4+ AML cell killing exploits concentrations in cancer cells is limited by associated the high overexpression of CXCR4 in LSCs as toxicity in normal tissues [40, 41]. compared to normal HSCs, as previously The direct killing of CXCR4+ quiescent AML blasts described [6, 37], leading to the specific in the BM reached by our nanoconjugate differs also internalization and selective delivery of the from the effect of low molecular weight inhibitors of conjugated drug to CXCR4+ AML cells in all the CXCR4 receptor, which in AML are used in com- clinically relevant sites, since our THP-1 model bination with chemotherapy. CXCR4 inhibitors highly express the CXCR4 receptor in the AML mobilize AML blasts from BM to the bloodstream, cell membrane in BM and extramedullar sites. losing their quiescence, and becoming cycling cells This selective effect is also supported by the that chemotherapy targets. Nevertheless, AML cell in vitro THP-1 and SKM-1 AML cells results, in mobilization induces hyper-leucocytosis that could, which the NC showed a strict selectivity and unexpectedly, generate extramedullar dissemination CXCR4 dependence in internalization and cell [5]. killing of CXCR4+ AML cells. Regarding strategies similar to our targeted drug 2. Ability of the NC payload, Auristatin E, for delivery approach, no antibody-drug conjugates killing quiescent cancer cells, as a microtubule (ADCs) targeting CXCR4 have been introduced in destabilizing agent, through the induction of the clinic so far. We believe that our protein-based mitotic catastrophe and apoptosis as described nanoparticle could, however, improve the reported [23, 24], which particularly, in our model, biodistribution described for ADCs, reaching only eliminates the LSCs that enter quiescence after 0.01–0.1% of the injected dose in cancer cells [42]. homing in the BM [6, 20]. This effect may In contrast, we have reported that T22-GFP-H6 overcome the limited effectiveness of current nanoparticles and the derivate NCs achieve around chemotherapy that blocks AML cells in division 85% of the administered dose in tumor tissues, both but induces life-threatening adverse effects [38], in solid tumors [22, 43] and hematological neoplasias being unable to kill quiescent LSCs. This is likely [26]. the reason why Auristatin E is extensively used Finally, because of the reported overexpression of in antibody-drug conjugates under development, CXCR4 in AML cells which associates with chemo- one of which, Brentuximab vedotin, is currently therapy resistance, LSC development, relapse, and used in the clinic to treat CD30+ Hodgkin’s minimal residual disease in AML patients [1–4, 11, lymphoma and anaplastic large-cell lymphoma 44–47], we believe this nanoconjugate could be used patients [39]. to treat resistant AML. Thus, in the near future, we plan to develop disseminated AML models resistant We also think these arguments, and particularly their to therapy to evaluate this hypothesis. cancer-selective biodistribution, underlie our expectation that this nanoconjugate may outperform the efficacy of Conclusions drugs used in the clinic or being under development as In summary, we developed a novel nanoconjugate for we following explain. Thus, the NC selectivity in targeted drug delivery of Auristatin E that achieves a Pallarès et al. Journal of Hematology & Oncology (2020) 13:36 Page 17 of 19 Fig. 8 Hematoxylin and eosin staining in leukemia-infiltrated tissues after treatment with T22-GFP-H6-Auristatin. Hematoxylin and eosin staining of the liver, spleen, bone marrow, lung, pancreas, brain, heart, and kidneys of mice treated with vehicle (VEH) or T22-GFP-H6-Auristatin (T22-AUR) 14 days after the injection of CXCR4+ THP-1-Luci cells in NSG mice, compared with the stained tissues of a healthy mouse (Normal). T22-AUR, T22-GFP-H6-Auristatin group; VEH, vehicle group Pallarès et al. Journal of Hematology & Oncology (2020) 13:36 Page 18 of 19 significant reduction of the leukemic cell burden in We are also indebted for funding by the ISCIII Networking Research Center the bone marrow and circulating blood, and a potent on Bioengineering, Biomaterials and Nanomedicine (CIBER-BBN) and its ICTS Nanbiosis Platform. The bioluminescent follow-up of cancer cells and toxicity blockade of leukemic cell spread to extramedullar or- studies has been performed in the ICTS-141007 Nanbiosis Platform, using its gans in a CXCR4+ AML model. These results prove CIBER-BBN Nanotoxicology Unit (http://www.nanbiosis.es/portfolio/u18-nano- the high selectivity of the nanoconjugate in control- toxicology-unit/). Protein production has been partially performed by the ICTS “NANBIOSIS”, more specifically by the Protein Production Platform of ling AML dissemination without associated toxicity CIBER-BBN/ IBB (http://www.nanbiosis.es/unit/u1-protein-production-plat- and, importantly, that CXCR4+ AML cells are a rele- form-ppp/). vant target for clinical therapy. Availability of data and materials All data generated or analyzed during the current study are available from Supplementary information the corresponding author on reasonable request. Supplementary information accompanies this paper at https://doi.org/10. 1186/s13045-020-00863-9. Ethics approval and consent to participate All procedures performed on animals were conducted in accordance with Additional file 1: Figure S1. GOT, GPT and creatinine levels in plasma the guidelines approved by the institutional animal Ethics Committee of after treatment with T22-GFP-H6-Auristatin. Determination of GOT, GPT Hospital Sant Pau. (A) and Creatinine (B) in plasma of mice treated with 9 doses of vehicle (VEH) or T22-GFP-H6-Auristatin (100μg/dose) (T22-AUR) 14 days after the injection of THP-1-Luci cells in NSG mice. Results are presented as mean Consent for publication ± SE enzyme activity in U/L for GOT and GPT (A) and mean ± SE creatin- Not applicable. ine levels in mmol/L (B). GOT, oxaloacetic transaminase; GPT, glutamic pyruvic transaminase; T22-AUR, T22-GFP-H6-Auristatin group; VEH, Vehicle Competing interests group. SE, standard error. EV, RM, and AV are co-founders of Nanoligent, a company devoted to de- velop anticancer drugs based on proteins. UU, EV, AV, RM, and IC are cited as inventors in a patent application Abbreviations (EP11382005.4) covering the therapeutic use of T22. All other authors report ADC: Antibody-drug conjugate; AML: Acute myeloid leukemia; no conflicts of interest in this work. BLI: Bioluminescence; BM: Bone marrow; BSA: Bovine serum albumin; CRC: Colorectal cancer; CXCL12: C-X-C motif chemokine 12; CXCR4: C-X-C chemokine receptor type 4; FACS: Fluorescence-activated cell sorting; Author details 1 GFP: Green fluorescence protein; GOT: Oxaloacetic transaminase; Biomedical Research Institute Sant Pau (IIB-Sant Pau), Hospital de la Santa GPT: Glutamic pyruvic transaminase; HSC: Hematopoietic stem cell; Creu i Sant Pau, 08041 Barcelona, Spain. 2Josep Carreras Research Institute, IHC: Immunohistochemistry; IV: Intravenously; LSC: Leukemic stem cell; 08041 Barcelona, Spain. 3CIBER en Bioingeniería, Biomateriales y MALDI-TOF: Matrix-Assisted Laser Desorption/Ionization - Time-Of-Flight; Nanomedicina (CIBER-BBN), 08041 Barcelona, Spain. 4Departament de MC: Mitotic catastrophe; MC-MMAE: Maleimide conjugated monomethyl Genètica i de Microbiologia, Universitat Autònoma de Barcelona, 08193 Auristatin E; MFI: Mean fluorescence intensity; MMAE: Monomethyl Auristatin Bellaterra, Spain. 5Institut de Biotecnologia i de Biomedicina, Universitat E; NC: Nanoconjugate; NSG: NOD-scid IL2Rgammanull; PBS: Phosphate- Autònoma de Barcelona, 08193 Bellaterra, Spain. 6Department of Pathology, buffered saline; SC: Subcutaneous; SE: Standard error; SEC-MALS: Size Hospital de la Santa Creu i Sant Pau, Barcelona, Spain. 7Department of exclusion chromatography – multi-angle light scattering Chemical Sciences, School of Applied Sciences, University of Huddersfield, Huddersfield HD1 3DH, UK. 8Department of Hematology, Hospital de la Acknowledgements Santa Creu i Sant Pau, Barcelona, Spain. The authors would like to thank Annabel García-León, Yaiza Núñez, and Lídia Cedó (IIB-Sant Pau, Barcelona) for their technical support. Received: 4 November 2019 Accepted: 24 March 2020 Authors’ contributions Conception and design: VP, UU, AV, EV, IC, and RM; Development of References methodology: VP, UU, AF, LS-G, NS, and IC; Acquisition of data: VP, UU, AF, 1. Spoo AC, Lübbert M, Wierda WG, Burger JA. CXCR4 is a prognostic marker PA, LS-G, and NS. Histopathological analysis: VP, AF, PA, and AG; Analysis and in acute myelogenous leukemia. Blood. 2007;109:786–91. interpretation of data: VP, UU, AF, AG, GAM, JS, AV, EV, RM, and IC. Drafted 2. Burger JA, Bürkle A. The CXCR4 chemokine receptor in acute and chronic the manuscript: VP and UU; Review and revision of the manuscript: LA-C, JS, leukaemia: a marrow homing receptor and potential therapeutic target. Br J RM, IC, AV, and EV. Study supervision: GAM, RM, IC, AV, and EV. All authors Haematol. 2007;137:288–96. critically revised and approved the final manuscript. 3. Konoplev S, Lin P, Yin CC, Lin E, Nogueras González GM, Kantarjian HM, et al. CXC chemokine receptor 4 expression, CXC chemokine receptor 4 Funding activation, and wild-type nucleophosmin are independently associated with This work was supported by Instituto de Salud Carlos III (ISCIII, Co-funding unfavorable prognosis in patients with acute myeloid leukemia. Clin from FEDER) [PI18/00650, PIE15/00028, PI15/00378 and EU COST Action CA Lymphoma Myeloma Leuk. 2013;13:686–92. 17140 to R.M., FIS PI17/01246, RD16/0011/0028 and FIS PI14/00450 to J.S.; FIS 4. Mannelli F, Cutini I, Gianfaldoni G, Bencini S, Scappini B, Pancani F, et al. PI15/00272 to E.V.]; Agencia Estatal de Investigación (AEI) and Fondo Europeo CXCR4 expression accounts for clinical phenotype and outcome in acute de Desarrollo Regional (FEDER) (grant BIO2016-76063-R, AEI/FEDER, UE) to myeloid leukemia. Cytometry B Clin Cytom. 2014. https://doi.org/10.1002/ A.V.; CIBER-BBN [CB06/01/1031 and 4NanoMets to R.M., and VENOM4CANCER cytob.21156. to A.V.]; AGAUR [2017 FI_B 00680 to A.F.; 2017-SGR-865 to R.M., 2017-SGR- 5. Cho B-S, Kim H-J, Konopleva M. Targeting the CXCL12/CXCR4 axis in acute 1395 to J.S. and 2017SGR-229 to A.V.]; Josep Carreras Leukemia Research In- myeloid leukemia: from bench to bedside. Korean J Intern Med. 2017;32: stitute [P/AG to R.M.]; a grant from the Cellex Foundation, Barcelona [to J.S.]; 248–57. a grant from La Generalitat de Catalunya (PERIS) [SLT002/16/00433 to J.S.]; a 6. Schepers K, Campbell TB, Passegué E. Normal and leukemic stem cell grant from the Generalitat de Catalunya CERCA Programme. The work was niches: insights and therapeutic opportunities. Cell Stem Cell. 2015;16:254– also supported by PERIS program from the health department of the Gener- 67. alitat de Catalunya (SLT006/17/00093) [grated to U.U.] and Fundación Espa- 7. Zhang Y, Patel S, Abdelouahab H, Wittner M, Willekens C, Shen S, et al. ñola de Hematología y Hemoterapia (FEHH) [granted to V.P.]. Finally, AV CXCR4 inhibitors selectively eliminate CXCR4-expressing human acute received an ICREA ACADEMIA Award supported by the Catalan Government. myeloid leukemia cells in NOG mouse model. Cell Death Dis. 2012;3:e396. Pallarès et al. Journal of Hematology & Oncology (2020) 13:36 Page 19 of 19 8. Fiegl M, Samudio I, Clise-Dwyer K, Burks JK, Mnjoyan Z, Andreeff M. CXCR4 31. Fetterly GJ, Aras U, Lal D, Murphy M, Meholick PD, Wang ES. Development expression and biologic activity in acute myeloid leukemia are dependent of a preclinical PK/PD model to assess antitumor response of a sequential on oxygen partial pressure. Blood. 2009;113:1504–12. aflibercept and doxorubicin-dosing strategy in acute myeloid leukemia. 9. Fukuda S, Onishi C, Pelus LM. Trafficking of Acute Leukemia Cells – AAPS J. 2013;15:662–73. Chemokine Receptor Pathways that Modulate Leukemia Cell Dissemination. 32. Lehmann C, Friess T, Birzele F, Kiialainen A, Dangl M. Superior anti-tumor In: Antica M, editor. Acute Leukemia - The Scientist’s Perspective and activity of the MDM2 antagonist idasanutlin and the Bcl-2 inhibitor Challenge. InTech. 2011. p. 137–56. venetoclax in p53 wild-type acute myeloid leukemia models. J Hematol 10. Habringer S, Lapa C, Herhaus P, Schottelius M, Istvanffy R, Steiger K, et al. Oncol. 2016;9:50. Dual Targeting of Acute Leukemia and Supporting Niche by CXCR4- 33. Christian S, Merz C, Evans L, Gradl S, Seidel H, Friberg A, et al. The novel Directed Theranostics. Theranostics. 2018;8:369–83. dihydroorotate dehydrogenase (DHODH) inhibitor BAY 2402234 triggers 11. Peled A, Tavor S. Role of CXCR4 in the pathogenesis of acute myeloid differentiation and is effective in the treatment of myeloid malignancies. leukemia. Theranostics. 2013;3:34–9. Leukemia. 2019. https://doi.org/10.1038/s41375-019-0461-5. 12. Croker AK, Allan AL. Cancer stem cells: implications for the progression and 34. Li F, Sutherland MK, Yu C, Walter RB, Westendorf L, Valliere-Douglass J, et al. treatment of metastatic disease. J Cell Mol Med. 2008;12:374–90. Characterization of SGN-CD123A, A Potent CD123-Directed Antibody-Drug 13. Zeelenberg IS, Ruuls-Van Stalle L, Roos E. The chemokine receptor CXCR4 is Conjugate for Acute Myeloid Leukemia. Mol Cancer Ther. 2018;17:554–64. required for outgrowth of colon carcinoma micrometastases. Cancer Res. 35. Tron AE, Belmonte MA, Adam A, Aquila BM, Boise LH, Chiarparin E, et al. 2003;63:3833–9. Discovery of Mcl-1-specific inhibitor AZD5991 and preclinical activity in 14. Müller A, Homey B, Soto H, Ge N, Catron D, Buchanan ME, et al. multiple myeloma and acute myeloid leukemia. Nat Commun. 2018;9:5341. Involvement of chemokine receptors in breast cancer metastasis. Nature. 36. Mambet C, Chivu-Economescu M, Matei L, Necula LG, Dragu DL, Bleotu C, 2001;410:50–6. et al. Murine models based on acute myeloid leukemia-initiating stem cells 15. Hermann PC, Huber SL, Herrler T, Aicher A, Ellwart JW, Guba M, et al. xenografting. World J Stem Cells. 2018;10:57–65. Distinct populations of cancer stem cells determine tumor growth and 37. Konopleva MY, Jordan CT. Leukemia stem cells and microenvironment: metastatic activity in human pancreatic cancer. Cell Stem Cell. 2007;1:313– biology and therapeutic targeting. J Clin Oncol. 2011;29:591–9. 23. 38. Centelles JJ. Treatment of Acute Myeloid Leukemia. Curr Synthetic Sys Biol. 16. Schimanski CC, Bahre R, Gockel I, Müller A, Frerichs K, Hörner V, et al. 2014;02. Dissemination of hepatocellular carcinoma is mediated via chemokine 39. Garnock-Jones KP. Brentuximab vedotin: a review of its use in patients with receptor CXCR4. Br J Cancer. 2006;95:210–7. hodgkin lymphoma and systemic anaplastic large cell lymphoma following 17. Zhang S, Han Z, Jing Y, Tao S, Li T, Wang H, et al. CD133(+)CXCR4(+) colon previous treatment failure. Drugs. 2013;73:371–81. cancer cells exhibit metastatic potential and predict poor prognosis of 40. Aberger F, Hutterer E, Sternberg C, Del Burgo PJ, Hartmann TN. Acute patients. BMC Med. 2012;10:85. myeloid leukemia - strategies and challenges for targeting oncogenic 18. Zhang Y, Saavedra E, Tang R, Gu Y, Lappin P, Trajkovic D, et al. Targeting Hedgehog/GLI signaling. Cell Commun Signal. 2017;15:8. primary acute myeloid leukemia with a new CXCR4 antagonist IgG1 41. Ashihara E, Takada T, Maekawa T. Targeting the canonical Wnt/β-catenin antibody (PF-06747143). Sci Rep. 2017;7:7305. pathway in hematological malignancies. Cancer Sci. 2015;106:665–71. 19. Das M, Mohanty C, Sahoo SK. Ligand-based targeted therapy for cancer 42. Lambert JM. Antibody-Drug Conjugates (ADCs): Magic Bullets at Last! Mol tissue. Expert Opin Drug Deliv. 2009;6:285–304. Pharm. 2015;12:1701–2. 20. Pollyea DA, Gutman JA, Gore L, Smith CA, Jordan CT. Targeting acute 43. Céspedes MV, Unzueta U, Álamo P, Gallardo A, Sala R, Casanova I, et al. Cancer- myeloid leukemia stem cells: a review and principles for the development specific uptake of a liganded protein nanocarrier targeting aggressive CXCR4+ of clinical trials. Haematologica. 2014;99:1277–84. colorectal cancer models. Nanomedicine. 2016;12:1987–96. 44. Ahn JY, Seo K, Weinberg OK, Arber DA. The prognostic value of CXCR4 21. Unzueta U, Céspedes MV, Ferrer-Miralles N, Casanova I, Cedano J, Corchero in acute myeloid leukemia. Appl Immunohistochem Mol Morphol. 2013; JL, et al. Intracellular CXCR4+ cell targeting with T22-empowered protein- 21:79–84. only nanoparticles. Int J Nanomedicine. 2012;7:4533–44. 45. Konoplev S, Rassidakis GZ, Estey E, Kantarjian H, Liakou CI, Huang X, et al. 22. Céspedes MV, Unzueta U, Aviñó A, Gallardo A, Álamo P, Sala R, et al. Overexpression of CXCR4 predicts adverse overall and event-free survival in Selective depletion of metastatic stem cells as therapy for human colorectal patients with unmutated FLT3 acute myeloid leukemia with normal cancer. EMBO Mol Med. 2018;10. karyotype. Cancer. 2007;109:1152–6. 23. Field JJ, Kanakkanthara A, Miller JH. Microtubule-targeting agents are 46. Rombouts EJC, Pavic B, Löwenberg B, Ploemacher RE. Relation between clinically successful due to both mitotic and interphase impairment of CXCR-4 expression, Flt3 mutations, and unfavorable prognosis of adult microtubule function. Bioorg Med Chem. 2014;22:5050–9. acute myeloid leukemia. Blood. 2004;104:550–7. 24. Bates D, Eastman A. Microtubule destabilising agents: far more than just 47. Tavor S, Petit I, Porozov S, Avigdor A, Dar A, Leider-Trejo L, et al. CXCR4 antimitotic anticancer drugs. Br J Clin Pharmacol. 2017;83:255–68. regulates migration and development of human acute myelogenous 25. Rueden CT. ImageJ2: ImageJ for the next generation of scientific image leukemia stem cells in transplanted NOD/SCID mice. Cancer Res. 2004;64: data; 2017. 2817–24. 26. Falgàs A, Pallarès V, Unzueta U, Céspedes MV, Arroyo-Solera I, Moreno MJ, et al. A CXCR4-targeted nanocarrier achieves highly selective tumor uptake in diffuse large B-cell lymphoma mouse models. Haematologica. 2019. Publisher’s Note https://doi.org/10.3324/haematol.2018.211490. Springer Nature remains neutral with regard to jurisdictional claims in 27. Kuhne MR, Mulvey T, Belanger B, Chen S, Pan C, Chong C, et al. BMS- published maps and institutional affiliations. 936564/MDX-1338: a fully human anti-CXCR4 antibody induces apoptosis in vitro and shows antitumor activity in vivo in hematologic malignancies. Clin Cancer Res. 2013;19:357–66. 28. Xie C, He Y, Zhen M, Wang Y, Xu Y, Lou L. Puquitinib, a novel orally available PI3Kδ inhibitor, exhibits potent antitumor efficacy against acute myeloid leukemia. Cancer Sci. 2017;108:1476–84. 29. Srivastava G, Matta A, Fu G, Somasundaram RT, Datti A, Walfish PG, et al. Anticancer activity of pyrithione zinc in oral cancer cells identified in small molecule screens and xenograft model: Implications for oral cancer therapy. Mol Oncol. 2015;9:1720–35. 30. Lin W-H, Yeh T-K, Jiaang W-T, Yen K-J, Chen C-H, Huang C-T, et al. Evaluation of the antitumor effects of BPR1J-340, a potent and selective FLT3 inhibitor, alone or in combination with an HDAC inhibitor, vorinostat, in AML cancer. PLoS One. 2014;9:e83160.