In recent years, multifunctional nanoparticles have received considerable attention in many applications such as biosensors, diagnostic nanoprobes and targeted drug delivery. These efforts have been driven to a large extent by the need to improve biological specificity in diagnosis and therapy through the precise, spatio-temporal control of agent delivery. In order to achieve this goal, continuous efforts have been dedicated to develop stimuli-responsive nanoplatforms. Environmental stimuli that were exploited include pH, temperature, enzymatic expression, redox reaction and light induction. Among these activating signals, pH trigger is one of the most extensively studied stimuli based on two types of pH differences: (a) pathological (e.g. tumor) vs. normal tissues and (b) acidic intracellular compartments. For example, due to the unusual acidity of the tumor extracellular microenvironment (pHe ≈ 6.5), several pHe-responsive nanosystems were reported to increase the sensitivity of tumor imaging or the efficacy of therapy.
To target the acidic endo-/lysosomal compartments, nanovectors with pH-cleavable linkers were reported to improve payload bioavailability. Furthermore, several smart nanovectors with pH-induced charge conversion were designed to increase drug efficacy. Despite these remarkable advances, specific transport and activation of nanoparticles in different endocytic organelles during endocytosis in living cells is not well documented. The endocytic system is comprised of a series of compartments that have distinctive roles in the sorting, processing and degradation of internalized cargo. Selective targeting of different endocytic compartments by pH-sensitive nanoparticles is challenging due to the short nanoparticle residence times (~mins) and small pH differences in these compartments (e.g. <1 pH unit between early endosomes and lysosomes).
Herein we report a set of tunable, pH-activatable micellar (pHAM) nanoparticles based on the supramolecular self-assembly of ionizable block copolymer micelles (Figure 1). Micelle formation and its thermodynamic stability are driven by the delicate balance between the hydrophobic and hydrophilic segments. Ionizable groups can act as tunable hydrophobic groups at different pH values. Amino groups have been incorporated into polymers as ionizable groups to render pH sensitivity. In this study, we introduced tertiary amines with precisely controlled hydrophobic substituents as ionizable hydrophobic blocks. Micellization dramatically sharpens the ionization transition of tertiary amines in the hydrophobic block, rendering fast and ultra-sensitive pH response. Nanoparticles with different transition pH can be selectively activated in specific endocytic compartments such as early endosomes or lysosomes in human cells.
For proof of concept, we synthesized two series of block copolymers (PEO-b-PR, Figure 1) with tertiary amine-containing (PR) and poly(ethylene oxide) (PEO) segments by atom transfer radical polymerization (Supplementary Table S1). In the linear di-alkyl series, we varied the chain length from methyl to butyl groups; in the cyclic series, we varied the ring size from 5- to 7-membered rings. The two series were systematically used to adjust the pKa values of ammonium groups (Supplementary Table S2) and PR hydrophobicity. A pH-insensitive dye, tetramethyl rhodamine (TMR), was used as a model fluorophore and conjugated in the PR block as an imaging beacon to investigate the pH responsive properties. At higher pH, neutral PR segments self-assemble into the hydrophobic cores of micelles, resulting in the aggregation of fluorophores and quenching of fluorescent signals through mechanisms of Förster resonance energy transfer between TMR molecules (homo-FRET) and photoinduced electron transfer (PeT) from tertiary amines to TMR. At lower pH, PR segments become protonated and positively charged, leading to micelle disassembly and dramatic increase in fluorescence emission due to the increase in TMR distance and decrease in PeT (Figure 1).
Fluorescent images of a series of nanoprobe solutions (Figure 2a) at different pH values illustrate a sharp fluorescence transition for each nanoprobe. In this study, poly(ethylene oxide)-b-poly((dimethyl-amino)ethyl methacrylate) (PEO-b-PDMA, 1) was used as an “always ON” control where no micelles or fluorescence quenching was observed in the tested pH range (4.5–8.0) due to the strong hydrophilicity of the PDMA block (see discussion below). Normalized fluorescence intensity (NFI) vs. pH curves (Figure 2b) allowed for quantitative assessment of the ultra-pH responsive properties. NFI is calculated as the ratio of [F-Fmin]/[Fmax-Fmin], where F is the fluorescence intensity of the nanoprobe at any given pH, and Fmax and Fmin are the maximal and minimal fluorescence intensities at the ON/OFF states, respectively. To quantify the sharpness in pH response, we evaluated ΔpH10–90%, the pH range in which the NFI value varies from 10% to 90%, for all the pHAM nanoprobes. The sharpness values were 0.21, 0.23, 0.24, and 0.20 pH unit for nanoprobes 4, 6, 3 and 7, respectively. The small values indicate a remarkable pH sensitivity as it represents a <2-fold change in proton concentration. In comparison, for small molecular dyes, the sharpness value is about 2 pH unit (100-fold in [H+]) for the same degree of emission change, consistent with Henderson-Hasselbalch equation. In addition to the pH sharpness, we also measured the ratio of Fmax and Fmin (RF=Fmax/Fmin) to quantify the fluorescence response between the ON/OFF states. The values of RF range from 10 to 55 fold (Table S2, Figure S1). Consistent with the decreased emission intensity in the micelles, data show that excited state of TMR had a much shorter life time (e.g. 0.44 ns for nanoprobe 3, Figure S2) in the micelles (pH = 7.4) than the free dye (1.97 ns) at pH 7.4 or the disassembled unimers at pH 5.5 (1.84 ns).
For the pH temporal reponse, stopped-flow experiments showed that fluorescence activation was very fast, with most nanoprobes fully activated within 5 ms at lower pH (e.g. τ1/2 = 3.7 ms for 4, Figure 2c). The ultra-sensitive pH response was only observed with 4, 3, 7 and 6. The fluorescence transition pH values (pHt, the pH at which F = 0.5×(Fmax+Fmin)) were 5.4, 6.3, 6.8 and 7.2 for nanoprobes 4, 3, 7 and 6, respectively (Figure 2b). The other copolymers either did not show any pH response (e.g., 1 in Figure 2a) or only broad pH responses (e.g. 2, 5, data not shown). We hypothesize that hydrophobic micellization is the driving force of the ultra-pH responsive properties of pHAM, and a critical threshold of hydrophobicity in the PR segment is necessary to achieve the co-operative response. To test this hypothesis, we used copolymers 5 and 7 as examples and compared their pH titration curves as well as the corresponding monomers (Figure 3a). Larger ring size (i.e. 7) resulted in higher hydrophobicity in the PR segment due to the extra methylene groups. Copolymer 5 showed a broad pH response, similar to both monomers over added volumes of NaOH. In contrast, copolymer 7 had a dramatically sharpened pH transition, demonstrating a greatly increased buffer capacity. Deuterated 1H NMR spectra of 5 and 7 at different ionization states of tertiary amines further support the hypothesis (Figure 3b). The PEO segment did not change its peak intensity and was used as an internal standard. Throughout the ionization states, the proton resonance peaks for the PR segment of 5 were easily visualized although the peak intensity decreased with broadened peak width at higher pH, reflecting the bulk aggregation behavior of the copolymer. For 7, neutral state of the copolymer (i.e. 0%) led to completely suppressed resonance peaks in the PR segment due to the formation of highly compact micelle cores. Transmission electron microscopy (TEM) of 7 in aqueous solution demonstrated the formation of micelles above its pKa (6.7) at pH 7.4 and complete micelle dissociation at pH 5.5 (Figure 3c). In comparison, no micelles were formed from 5 at either pH (data not shown).
To investigate the intracellular activations of pHAM, we examined nanoprobe 3 in human H2009 lung cancer cells by confocal laser scanning microscopy (Figure S3). It should be noted that nanoprobe 3 has an optimal pH transition at 6.3, which is ideally suited for the study of nanoparticle activation in early endosomes (pH = 5.9–6.2)[13a, 21]. Because pHAM nanoprobes are “silent” at neutral pH, we directly applied them in the culture medium and monitored the kinetics of their uptake and activation without the need to remove the medium. Right after the nanoprobe addition, neither the H2009 cells nor the medium showed observable fluorescence signal. At 15 min, punctuate fluorescent dots appeared inside the cells. The number of fluorescent dots increased over time. Signal to noise ratio of the H2009 cells (SNRCell, using fluorescence intensity at time 0 as the background noise) allowed further quantification of the increased nanoprobe uptake and activation over time. At 60 min, a 31-fold increase in SNRCell (2.14±0.17×103) was observed over the medium (SNRMed=69.3±9.1, P<0.001) where majority of the nanoprobes were still present. Then 0.1N HCl solution was added to acidify the medium pH to 5.0 and considerable increase in fluorescence intensity in the medium background was found. A reverse trend of fluorescence contrast was observed, where SNRCell was 74% of SNRMed (P<0.05). These data illustrate that pHAM nanoprobes can dramatically increase the contrast sensitivity of cancer cells compared to potentially always ON nanoprobes (as in the case after HCl was added).
To further investigate whether different endocytic organelles can selectively activate pHAM, we transfected H2009 cells with green fluorescent protein (GFP)-fused Rab5a and Lamp1 biomarkers in early endosomes and late endosomes/lysosomes, respectively. Two pHAM nanoprobes (3 and 4 with pHt of 6.3 and 5.4, respectively) were incubated with H2009 cells and confocal imaging was used to examine the subcellular locations for pHAM activation (Figure 4 and Figure S4). H2009 cells (N=30–50) with 20 or more colocalized dots (i.e. activated pHAM within early endosomes or lysosomes) were identified as positive and the percentage was quantified (Figure 4c and 4d). For nanoprobe 3, 80% of cells were positive in colocalization with early endosomes at 30 min, whereas only 12% colocalized with late endosomes/lysosomes (Figure 4a and Figure S4). Over time, colocalization of activated 3 decreased with early endosomes but increased with late endosomes/lysosomes (Figure 4c). In contrast, nanoprobe 4 (pHt = 5.4) showed a different pattern of subcellular location for activation. At all times, less than 10% of positive cells were found with early endosome colocalization (top panel of Figure 4b and Figure S4). Instead, almost all the activated nanoprobe 4 colocalized with late endosomes/lysosomes (Figure 4b bottom panel, Figure 4d). Figure 4e and Figure 4f depict the different processes of intracellular uptake and activation of the two nanoprobes. Nanoprobe 3 can be quickly activated inside early endosomes with higher vesicular pH (5.9–6.2)[13a, 21] and the activation is sustained as the nanoprobes traffic into late endosomes/lysosomes. By contrast, nanoprobe 4 is almost exclusively activated inside the late endosomes/lysosomes with lower vesicular pH (5.0–5.5).[13a, 21] Similar results were also found with human SLK tumor endothelial cells (data not shown). These data demonstrate the feasibility of targeting small differences in the vesicular pH inside different endocytic organelles by the pHAM nanoparticles.
To verify the intracellular activation mechanism of pHAM, we incubated H2009 cells with bafilomycin A1 for 1 hour and then added nanoprobe 3. Bafilomycin is a specific inhibitor of vacuolar-type H+-ATPase (V-ATPase), which is responsible for the proton pumping across the plasma membranes and acidification of intracellular organelles (e.g. lysosomes). Data show that in the presence of bafilomycin A1, nanoprobe 3 was not activated as indicated by the absence of TMR fluorescence (Figure 5a). After removal of bafilomycin A1 and 3 in the culture medium, the activation of 3 emerged with colocalization of TMR fluorescence with GFP labeled lysosomes (Figure 5b). Similar results were also found with nanoprobe 4 in H2009 cells (Figure S5).
In summary, we report the design of a series of pH-activatable micellar nanoparticles with tunable and ultra-sensitive pH response in the physiological range (5.0–7.4). These nanopartilces have fast temporal response (<5 ms), large increase of emission intensity between ON/OFF states (up to 55 times), and only require <0.25 pH for activation. Confocal imaging studies demonstrate the nanoparticles are “silent” in the media at pH 7.4 but can be activated upon uptake in cells. Moreover, nanoparticles with pH transitions at 6.3 and 5.4 can be selectively activated in different endocytic compartments such as early endosomes (pH 5.9–6.2) and lysosomes (5.0–5.5). This nanoplatform offers many exciting opportunities in the development of nonlinear ON/OFF nanosystems for diagnostic imaging and drug delivery applications with minimal effect at physiological pH (e.g. 7.4), but can be activated in acidic tumor pH or specific intracellular organelles (e.g. endosomes/lysosomes) upon uptake in targeted cells.