Reversible Intracellular Gelation of MCF10A Cells Enables Programmable Control Over 3D Spheroid Growth

Abstract

In nature, some organisms survive extreme environments by inducing a biostatic state wherein cellular contents are effectively vitrified. Recently, a synthetic biostatic state in mammalian cells was achieved via intracellular network formation using bio-orthogonal strain-promoted azide-alkyne cycloaddition (SPAAC) reactions between functionalized poly(ethylene glycol) (PEG) macromers. In this work, the effects of intracellular network formation on a 3D epithelial MCF10A spheroid model are explored. Macromer-transfected cells are encapsulated in Matrigel, and spheroid area is reduced by ~50% compared to controls. The intracellular hydrogel network increases the quiescent cell population, as indicated by increased p21 expression. Additionally, bioenergetics (ATP/ADP ratio) and functional metabolic rates are reduced. To enable reversibility of the biostasis effect, a photosensitive nitrobenzyl-containing macromer is incorporated into the PEG network, allowing for light-induced degradation. Following light exposure, cell state and proliferation return to control levels, while SPAAC-treated spheroids without light exposure (i.e., containing intact intracellular networks) remain smaller and less proliferative through this same period. These results demonstrate that photodegradable intracellular hydrogels can induce a reversible slow-growing state in 3D spheroid culture.

Graphical Abstract

Intracellular crosslinking in MCF10A spheroids is achieved via transfection of SPAAC-functionalized PEG macromers, inducing a biostatic state. The intracellular network increases the proportion of quiescent cells and decreases the growth of Matrigel-encapsulated spheroids, which display reduced bioenergetics. These effects can be reversed using photodegradation, enabling a user-directed synthetic method to control proliferation and cell state in 3D spheroid culture.

1. Introduction

Cryptobiosis is an adaptive metabolic state that allows certain organisms to survive drastic environmental changes and return to a functioning state.^[1] Tardigrades are one of the best known examples of a cryptobiotic organism,^[2] with their ability to reversibly vitrify their cytosol through the production of tardigrade disordered proteins (TDPs).^[3] These proteins stabilize water binding sites and preserve protein structures, enabling survival under extreme conditions.^[4] In general, mammalian cells do not undergo cryptobiosis, and consequently require specific culture and storage processes, with narrow temperature ranges and environmental conditions. To reduce the reliance on these storage techniques and develop strategies for inducing synthetic cryptobiotic states in mammalian cells, an emerging field of science is focused on the role of macromolecular crowding via intracellular polymerization for stabilizing cell membranes and protein structures.^[5]

Macromolecular crowding dictates diffusion-dependent reaction rates that many metabolic processes, cellular functions, and associated protein folding rates rely on.^[6] The crowded cytosol, filled with biomacromolecules and organelles, reduces the available intracellular volume in addition to nonspecific steric repulsions of macromolecules.^[7] These interactions therefore dictate the equilibrium and kinetics of many macromolecular reactions.^[8] Macromolecular crowding occurs across the kingdoms of life, in particular with studies exploring the role of macromolecular crowding in bacteria^[9] and yeast.^[10] In live mammalian cells, macromolecular crowding has been shown to affect essential intracellular processes, and Boersma et al. developed a fluorescence resonance energy transfer (FRET) sensor to directly measure macromolecular crowding.^[11] The sensor was expressed in HEK293 cells and successfully detected the change in crowding induced by osmotic shock. Protein stability can been tuned by various crowding agents, including synthetic polymers.^[12] Overall, it has been shown that macromolecular crowding usually serves to optimize intracellular reaction rates and preserve protein structures and functionality, even in harsh conditions.^[13]

Previous work by our groups introduced a method to reversibly increase macromolecular crowding in mammalian cells (MCF10A, C2C12, and others) through the polymerization of intracellular photodegradable poly(ethylene glycol) (PEG) macromers.^[14] The resultant hydrogels increased the viscosity of the cytosol and resulted in decreased DNA replication, protein synthesis, cellular motility, and other biological effects suggesting an induced quiescent state. These effects were reversed through exposure to UV light, degrading the synthetic network and reducing the artificial crowding of the cytosol. Here, to expand on these methods and study the effects of synthetic intracellular hydrogels on a more complex multicellular model, we investigate the effects of intracellular gelation on MCF10A spheroids. MCF10A cells are a non-tumorigenic human mammary epithelial cell line that forms spheroids when cultured in a 3D matrix.^[15] In this study, a 3D transfection model was developed to achieve a biostatic state within these multicellular constructs using intracellular PEG networks. The biostasis effects were reversed on-demand by exploiting photodegradable linkers in the network, thus returning the spheroids to an active state.

2. Results and Discussion

2.1. 2D Biostasis Transfection of MCF10A Cells Reduces Spheroid Growth in Matrigel

To induce biostasis in 3D cellular constructs, intracellular polymer networks were formed using 8-arm PEG macromers (M_n = 20 kg mol⁻¹) functionalized with SPAAC moieties (Figure 1a). 8-arm PEG was functionalized with either azide or dibenzocyclooctyne (DBCO) moieties to allow for rapid, spontaneous network formation (Figure 1b). Macromer transfection followed two approaches. In the first, adherent MCF10A cells were incubated with Lipofectamine and PEG-azide (5 wt%) in reduced serum media for 2 h, followed by washing with reduced serum media prior to a similar PEG-DBCO (5 wt%) treatment (Figure 1c). Alternatively, we directly transfected pre-formed MCF10A spheroids with hydrogel-forming macromer in 3D. In contrast to the 2D approach, spheroids were dissociated from Matrigel with cold media, washed, and collected by centrifugation before re-encapsulation after each transfection step to prevent extracellular hydrogel formation (Figure 1d). Our previous study focused on subcellular effects of biostasis induction via macromer transfection and established that homogenous uptake of macromer throughout the cytoplasm is not required for achieving the desired effects, including decreased proliferation. Based on these findings, we primarily use spheroid area as a functional metric to describe how treated cells are affected by treatment with and without macromer, but do not attempt to quantify where or how macromers are distributed throughout the cells.

Figure 1.

a) Schematic of 8-arm PEG functionalized with alkyne (green) and azide (pink) reactive groups. b) The bio-orthogonal SPAAC click reaction forms crosslinks between PEG macromers to generate a polymer network. c) Through Lipofectamine-mediated transfection, macromers are sequentially introduced into the cytoplasm of adherent MCF10A cells to induce intracellular hydrogelation (green and pink together) and a state of reduced biological activity which we refer to as biostasis. d) In this work, we demonstrate induction of biostasis in multicellular MCF10A spheroids cultured in 3D. Arrows in panels c) and d) indicate washing steps to remove macromer that is not internalized by cells to avoid extracellular hydrogel formation.

Twenty-four hours after 2D transfection, MCF10As were encapsulated into Matrigel. Matrigel is a 3D matrix formed from a basement membrane protein solution made up of primarily collagen IV, laminin, and perlecan, isolated from a Engelbreth–Holm–Swarm mouse tumor.^[16] A benefit of using Matrigel during the 3D transfection of MCF10A spheroids with PEG macromers is that excess macromer on the exterior of the cell is easily removed by the dissociation and washing steps, which would be challenging in covalently crosslinked 3D matrices. Since Matrigel is a liquid at 4 °C and forms a gel at 37 °C, addition of cold media to the spheroid-laden gels allows for collection of the transfected spheroids and removal of excess macromer via centrifugation. After MCF10As transfected in 2D were encapsulated into Matrigel, their growth (measured by projected spheroid area) was followed over the course of several days. Uptake of fluorescently labeled PEG macromers was not homogenous (Figure S1), corresponding to our previous results;^[14] we interpret this to mean that only a small amount of polymer network is required to enter the cell to observe a biological effect, as the cytosol is at a near-optimal state of crowding at baseline, before gel-forming macromers are introduced.^[7] Spheroid area was analyzed over time by tracking with an endogenously expressed H2B-mCherry nuclear marker (Figure 2a). These results demonstrate that the 3D spheroids experienced slower growth rates when treated with SPAAC macromers compared to control samples, which were treated with Lipofectamine in reduced serum media for the same incubation period with and without non-network forming macromer (Figure 2b). Nine days after transfection (eight days after encapsulation), a statistical difference in spheroid area was measured between controls and SPAAC conditions (Figure 2c). This suggests that the formation of an intracellular network influences proliferation and leads to a reduction in the spheroid growth rates. Although our results indicate successful internalization of macromer, other transfection methods, including thiol- and dithiolane-mediated uptake, may enable more efficient delivery of macromolecular cargo to cells, especially in dense 3D structures.^[17]

Figure 2.

Transfection of adherent cells with SPAAC macromers, followed by encapsulation in Matrigel. a) Day 6 and day 14 (post-transfection, encapsulation on day 2) representative images showing size difference in control, PEG-azide only, and SPAAC-treated spheroids, nuclear H2B-mCherry marker (blue), scale bars = 100 μm. b) Quantification of growth of individual spheroids at days 6 and 14, n ≥ 100 spheroids per condition, pooled across three technical replicates. c) Spheroid area normalized to day 6 average (across all conditions), demonstrating progressive growth retardation in SPAAC-treated spheroids. Data shown as mean of three replicate wells ± SD. Data were analyzed using a one-way ANOVA, * = p < 0.05.

2.2. 3D Biostasis Transfection Reduces Growth of Pre-Formed MCF10A Spheroids

The formation of intracellular networks in single cells resulted in a reduction in growth of 3D multicellular structures after encapsulation in Matrigel. Next, to determine whether this transfection approach could be translated to multicellular 3D structures, expanding our biostasis strategy, MCF10A cells were encapsulated into Matrigel at a density of 125,000 cells/mL and incubated for 6 days to allow for the formation of spheroids. The 3D transfection protocol was carried out over 2 days to reduce cellular stress and spheroid shearing during the Matrigel dissociation steps. After the first transfection, the spheroids were encapsulated into fresh Matrigel and sequentially transfected 24 h later. The transfection conditions were the same as with adherent cells (5 wt% PEG macromer, 2 h incubation). At early time points (48 h after final transfection step, 9 days after encapsulation), spheroids were visualized with intracellular fluorescence signal from labeled PEG macromers, suggesting that successful 3D transfection of macromers had occurred (Figure S2). At later time points, there was a decrease in spheroid growth in the SPAAC condition compared to controls (Figure 3b, c). Introducing only the PEG-azide macromer (i.e., no intracellular network formation) in 3D did not result in a biostasis effect, consistent with the results observed with cells transfected in 2D. Of note, MCF10A spheroids cultured in 3D before transfection were significantly larger than any condition (controls included) transfected in 2D before encapsulation. We attribute this to pre-conditioning from the 2D state, where MCF10As were grown to confluence and were thus contact-inhibited before transfection. In contrast, cells directly encapsulated in Matrigel and cultured for 6 days before transfection were subconfluent before encapsulation (see Experimental Section), resulting in more rapid spheroid growth at early timepoints. This also accounts for the differences seen between the number of spheroids and normalized spheroid area over time, where 3D-transfected spheroids grow less relative to the average day 6 (pre-transfection) spheroid area compared to 2D-transfected spheroids over a similar timecourse. Essentially, due to a 5 to 10-fold difference in day 6 baseline area between these conditions, 3D-transfected spheroids grow “less” because they are already quite large (as a result of both proliferation and fusion between neighboring spheroids), even before the transfection steps on days 6 and 7 post-encapsulation.

Figure 3.

3D transfection of MCF10A spheroids encapsulated in Matrigel. a) Representative images of pre-transfection (day 6 post-encapsulation) and 9 days post-transfection (day 15 post-encapsulation) spheroids showing size differences in control spheroids and SPAAC-treated spheroids, visualized by nuclear H2B-mCherry marker (blue), scale bar = 200 μm. b) Quantification of growth of individual spheroids over time, n ≥ 20 spheroids per condition, pooled across three technical replicates. c) Projected area of 3D-transfected spheroids normalized to day 6 average (across all conditions), showing reduced growth in SPAAC-treated spheroids. Dashed line on day 7 indicates final step of transfection. Data shown as mean of three replicate wells ± SD. Data were analyzed using a one-way ANOVA, * = p < 0.05.

2.3. Intracellular Gelation Increases Quiescent Population and Alters Bioenergetics in MCF10A Spheroids

To better understand how spheroid growth was decreased by intracellular gelation via transfection of SPAAC macromers, a more detailed characterization was conducted of the change in cell state induced by the intracellular network. First, an MCF10A cell line stably expressing fluorescently-tagged p21 enabled quantification of cells entering a quiescent state after intracellular gelation.^[28] The p21 reporter MCF10A cells were encapsulated in Matrigel and cultured for 6 days before sequential macromer transfections on days 6 and 7. Quantification of the p21 marker 72 hours after transfection (9 days post-encapsulation) revealed an increased number of p21⁺ cells, suggesting that more cells entered a quiescent state after network formation compared to the controls (Figure 4a). To complement the p21 quiescence marker, the spheroids were also stained for phospho-Rb (pRb), as cells hyper-phosphorylate Rb upon commitment to the cell cycle.^[18] SPAAC-treated spheroids had fewer pRb⁺ cells 24 h after transfection, suggesting that cells exited the cell cycle after the formation of an intracellular network (Figure S3). This is in accordance with previous work^[14] and indicates that intracellular gelation has similar effects in spheroids as in single cells. It is worth noting that only a relatively small fraction of cells transfected in 3D enter into a quiescent state, which may relate to earlier observations of heterogenous macromer uptake. However, even this relatively small increase in the proportion of quiescent cells as indicated by p21 expression (~10–20% greater than controls) appears to influence the growth dynamics of the spheroids (Figures 2 and 3). Additionally, p21/pRb expression may only partially account for cells affected by the presence of intracellular macromer.

Figure 4.

Cellular effects of SPAAC-treated spheroids in Matrigel. a) MCF10A spheroids expressing p21 (magenta) 72 h after 3D transfection, nuclei (blue), scale bars = 50 μm. Data shown as mean of spheroids in triplicate wells ± SD (large symbols), small symbols indicate individual spheroids (n = 15 spheroids per condition). Data were analyzed using an unpaired t-test, * = p < 0.05. b) Bioluminescence ATP/ADP assay for control vs. SPAAC-treated spheroids 24 h after transfection: i) intracellular ATP levels, ii) intracellular ADP levels, iii) ATP/ADP ratio. Data shown as mean (n = 3 wells) ± SD, circles with black borders represent the mean of each well. Data were analyzed using an unpaired t-test, ** = p < 0.01, *** = p < 0.001. c) Change in OCR of control vs. SPAAC-treated spheroids following drug treatments, data shown as mean (n = 3 wells) ± SD, circles with black borders represent the mean of each well. Data were analyzed using an unpaired t-test, ** = p < 0.01. d) Fluorescence lifetime imaging microscopy (FLIM) images for control vs. SPAAC-treated spheroids 12 day after transfection, i) Representative FLIM images showing relative fluorescence lifetime (NADH channel, 2P excitation at 800 nm, 80 MHz) in control and SPAAC-treated spheroids, scale bars = 250 μm. ii) Quantification of relative fluorescence lifetimes. Data shown as mean (n ≥ 10 spheroids) ± SD. Data were analyzed using a two-way ANOVA, # = p < 0.001 for control vs. SPAAC for long fluorescence lifetimes, $ = p < 0.0001 for control vs. SPAAC for short fluorescence lifetimes.

As intracellular gelation induces a state of increased quiescence in 3D spheroids, we next sought to characterize the bioenergetic state of biostatic MCF10A spheroids.^[19] SPAAC-treated spheroids had decreased levels of ATP and ADP 24 h after transfection compared to control spheroids, leading to a reduced ATP/ ADP ratio (Figure 4b, panels i-iii), suggesting that the intracellular network likely also suppresses cellular metabolism. Given that glycolysis and oxidative phosphorylation constitute key metabolic pathways involved in cellular bioenergetics, we next evaluated oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) using a Seahorse assay; OCR and ECAR are functional measurements of oxidative phosphorylation and glycolytic flux, respectively.^{[19b, 20]} Consistent with the trends observed in ATP and ADP, basal OCR and ECAR were reduced after SPAAC network formation compared to the control spheroids (Figure S4). Additionally, real-time perturbation of the electron transport chain (ETC) produced more dramatic changes in control spheroids compared to SPAAC-treated spheroids. Oligomycin, carbonyl cyanide-4 (trifluoromethoxy) phenylhydrazone (FCCP), and rotenone plus antimycin A (Rnt/AA) were sequentially injected to target different parts of the ETC. Oligomycin inhibits complex V, FCCP disrupts the membrane potential and Rnt/AA inhibit complex I and III of the ETC.^[21] Changes in OCR after injections of these treatments were small after intracellular network formation in comparison to controls (Figure 4c, S4), further enforcing that SPAAC-treated spheroids are relatively unresponsive to extrinsic metabolic perturbation.

In addition, the effect of intracellular gelation on 3D spheroids were measured using fluorescence lifetime imaging microscopy (FLIM). FLIM is a technique that can be used to assess live cell metabolism via autofluorescence of nicotinamide adenine dinucleotide+hydrogen (NADH). Free NADH has a short fluorescence lifetime (<1 ns), which increases when it is bound to protein.^[22] A semi-quantitative ratio of free to bound NADH was used to monitor changes in individual MCF10A spheroids.^[23] Two factors make this data difficult to precisely interpret in terms of specific metabolic effect. First, the DBCO macromer has an intermediate fluorescence lifetime (~2–3 ns) that is convoluted with the lifetime of intracellular NADH. Second, the lifetime of free NADH is affected by viscosity,^[24] which we previously identified as being increased in SPAAC-transfected cells.^[14] Together, both increased viscosity and convolution with the fluorescence lifetime of intracellular macromer could result in a prolonged lifetime in the emission spectrum of NADH (Figure S5). However, as these effects are both related to successful macromer transfection, we take these FLIM measurements to be representative of a global change in cell state that affects bioenergetics and proliferation rates, as is indicated by our other data. Frequency-domain FLIM was used to quantify the relative lifetime of NADH autofluorescence within spheroids, indicating changes in cell state for control and SPAAC spheroids after intracellular gelation. Single MCF10A cells were encapsulated in Matrigel and then underwent the sequential transfection procedure with the SPAAC PEG macromers as previously described. On day 12 (5 days post-transfection), NADH autofluorescence was measured with FLIM analysis (Figure 4d). In general, SPAAC spheroids showed longer lifetimes, indicating presence of intracellular macromer compared to control spheroids. This result suggests that network formation between reactive macromers may influence cell metabolism by altering NADH mobility through viscosity modification of the cytosol, although future studies are required to fully parse the effects of specific macromer chemistries (with unique fluorescence lifetimes) on this measurement.

2.4. Biostasis in 3D Spheroids can be Reversed via Photodegradation of Intracellular Hydrogels

By introducing functionalized PEG macromers into single MCF10A cells and pre-formed spheroids, an intracellular network forms which in turn decreases spheroid growth and bioenergetics compared to controls. However, for intracellular gelation to provide true control over cell state in 3D, these effects must be reversible such that cellular functions can be restored at a later time. To achieve reversibility with the SPAAC material system, a photocleavable nitrobenzyl moiety was incorporated into the PEG-azide macromer to produce PEG-oNB-azide (Figure 5a). When exposed to low doses of near-UV light (i.e., 5 mW/cm² of 365 nm light for 3–5 min), the network can be degraded on demand, reversing the biostasis effect and enabling spheroids to proliferate. We confirmed photodegradation of this hydrogel formulation with both in situ oscillatory shear rheology and fluorescence recovery after photobleaching (FRAP), which validated that bulk hydrogels thicker than cells (> 10 μm) degrade completely in ~3 min of irradiation and that suppressed diffusion of protein cargo encapsulated within these gels is reversed on the order of minutes (Figure S6).

Figure 5:

UV-degradable intracellular hydrogels allow for reversal of biostasis effects a) Schematic of nitrobenzyl moiety incorporated into linear PEG-bis(azide) backbone and the degradation product of the photodegraded crosslink. b) Representative images of control and SPAAC-treated spheroids (with and without photodegradation) on day 11 (top) and day 21 (bottom) visualized with the H2B nuclear tag (blue), scalebar = 200 μm. c) Quantification of individual spheroid area on days 11 and 21 (irradiation on day 11 in the + light condition). Data shown as mean (n = 3 wells) ± SD, large symbols with black borders represent the mean of each well. Data were analyzed using a one-way ANOVA, * = p < 0.05. d) Projected area of spheroids normalized to day 4 average (across all conditions), showing reduced growth in SPAAC-treated spheroids which can be reversed on demand following light exposure on Day 11 (indicated by dashed line). Data shown as mean of three replicate wells ± SD. Data were analyzed using a one-way ANOVA, * = p < 0.05.

To assess reversibility of biostasis in 3D culture, adherent single MCF10As were transfected with the photodegradable macromer system and encapsulated into Matrigel 24 h later. The cell-laden matrices were cultured for 11 days, at which point we observed a significant decrease in spheroid size between the SPAAC-treated condition and controls (SPAAC spheroids were 23±7% smaller than control spheroids by area, Figure 5b). Controls and half of the SPAAC-treated samples were then exposed to 365 nm light at 5 mW/cm² for 3 min to induce degradation of the intracellular hydrogels. After exposure to light and subsequent de-gelation, cell proliferation over 10 days resulted in spheroids that were comparable in size to controls. Spheroids not exposed to light (i.e., containing intact intracellular polymer networks) were significantly smaller than their photodegraded counterparts (SPAAC-treated spheroids were 42±8% smaller than spheroids exposed to light 21 days after encapsulation, Figure 5c and d). Translating this approach to the 3D transfection method, the oNB-azide system produced the same trends regarding spheroid growth as the adherent transfection method (Figure S7). As a degradation cue, light can be depth-limited; however, these samples are optically thin in the near-UV range (i.e., minimal light attenuation up to several hundred microns). As a result, the biological activity of SPAAC-treated spheroids is recovered shortly (within days) after light exposure, suggesting that light is delivered at a sufficient intensity throughout the spheroid volume to induce network degradation. Looking to the future, many alternative chemical approaches are focused on producing redshifted photocleavable and photocrosslinkable units, so that visible or near-IR light might be used to induce similar photocleavage reactions at greater depths (e.g., cells embedded in tissues).^[25]

Next, we utilized frequency-domain FLIM to assess spheroid state before photodegradation and 2, 6, and 8 days after light exposure. As before, we assessed the potential role of macromer or photocleavage product lifetime in affecting our measurements in spheroids, and found that the lifetime of the nitrobenzyl moiety was typically shorter than that of intracellular NAD. Additionally, this lifetime measurement was relatively unaffected by photodegradation of the macromer. When gels were formed from solutions of PEG-oNB-azide and PEG-DBCO containing NADH, the collective lifetimes for these species were generally shorter than those observed in non-photodegradable PEG-DBCO/PEG-azide gels (Figure S5), suggesting that prolonged fluorescence lifetime in biostasis-transfected spheroids can be attributed to metabolic effects, alterations in viscosity/NADH mobility, or successful macromer uptake, all of which relate to and help explain the biostatic cell state.

Initially, spheroids were imaged 8 days after transfection, but before photodegradation, to ensure that gel crosslinking with PEG-oNB-azide produced the same effects on spheroid fluorescence lifetime as the non-degradable SPAAC system (Figure 6a.i.). As expected, the photodegradable SPAAC-treated spheroids had significantly longer lifetime compared to control spheroids (Figure 6a.ii.). Two days later (i.e., 10 days after transfection), the control spheroids and half of the SPAAC-treated spheroids were exposed to 365 nm light at 5 mW/cm² for 3 minutes. When visualized 2 days later (day 12 post-transfection), both the control and SPAAC spheroids not exposed to light were comparable to their state on day 8 (before photodegradation). In contrast, the spheroids in which the intracellular network was photodegraded (SPAAC + light) showed an increase in the intermediate fluorescence lifetime, suggesting that cell state was beginning to recover from induced biostasis (Figure 6b.i.). Interestingly, shorter fluorescence lifetimes in the SPAAC + light condition were not immediately restored to control levels, suggesting that although the network is degraded, the timescale for a return to control metabolic state, intracellular viscosity, or complete exocytosis of macromer may occur over the course of multiple days (Figure 6b.ii.). FLIM analysis 4 days later (6 days after light exposure, 16 days after transfection) revealed a continued increase in shorter fluorescence lifetimes in SPAAC + light spheroids, which were by this point comparable to matched FLIM measurements in control spheroids (Figure 6c.i.). SPAAC-treated spheroids with intact intracellular hydrogels remained smaller and had significantly longer fluorescence lifetimes (Figure 6c.ii.).

Figure 6:

Frequency-domain FLIM (NADH channel) for UV-degradable intracellular hydrogels in MCF10A spheroids a.i) Day 8 representative images of control and SPAAC treated spheroids. Scale bars = 250 μm. a.ii) Quantification of relative distribution of fluorescence lifetimes. b.i) Day 12 representative FLIM images of control and SPAAC treated spheroids with and without light exposure (5 mW/cm², 365 nm, 3 minutes). Scale bars = 100 μm. b.ii) Quantification of relative distribution of fluorescence lifetimes at day 12. c.i) Day 16 representative FLIM images of control and SPAAC treated spheroids with and without light exposure (5 mW/cm², 365 nm, 3 minutes). Scale bars = 100 μm c.ii) Quantification of relative distribution of fluorescence lifetimes at day 16. d.i) Day 18 representative images of control and SPAAC treated spheroids with, without light, and with a higher UV light dose (10 mW/cm², 385 nm light, 5 min irradiation). Scale bars = 100 μm d.ii) Quantification of relative distribution of fluorescence lifetimes,. Data shown as mean (n ≥ 6 spheroids) ± SD. Data were analyzed using a two-way ANOVA. Significance markers are all based on comparison with control values and are as follows: # = p < 0.01 for long fluorescence lifetimes, & = p < 0.05 for intermediate fluorescence lifetimes, $ = p < 0.01 for short fluorescence lifetimes.

Final FLIM analysis was conducted 8 days after exposure to light (18 days after transfection). At this timepoint, the fluorescence lifetimes observed in the photodegraded SPAAC network spheroids were statistically indistinguishable from those of the control spheroids (Figure 6d.i). SPAAC spheroids not exposed to light still maintained long fluorescence lifetimes, indicating that the intracellular network likely remains intact over extended periods of time (at least 18 days post-transfection). Additionally, exposing spheroids to a higher light dose (10 mW/cm², 385 nm light, 5 min irradiation) 48 h before FLIM was conducted revealed that the shorter fluorescence lifetime increased rapidly, suggesting that the biostasis effect can be reversed at arbitrary timepoints by light exposure, with the rate of functional recovery following photodegradation of the network being potentially tunable with light dose (Figure 6d.ii.). Although it is outside the scope of this work, whether cells in synthetic crowding-induced biostasis are more resilient to altered culture conditions or harsh stimuli is a salient question and is an active field of study.^[26]

3. Conclusion

In this work, SPAAC PEG macromers were transfected into 3D MCF10A spheroids and a spontaneously crosslinked polymer network was formed intracellularly. The intracellular network resulted in reduced spheroid growth and an increased population of quiescent cells. Bioenergetic and intracellular fluorescence lifetime measurements further confirmed macromer internalization and biostasis effects. To reverse these effects, a photodegradable PEG macromer was introduced to the network and the reductions in spheroid area and intracellular state were restored to control levels following exposure to light. Overall, a strategy for reversible intracellular network formation has been developed for encapsulated spheroids resulting in user-directed control over biological activity in 3D.

4. Experimental Section

Materials

8-arm PEG-amine and 8-arm PEG-hydroxyl (both Mn = 20 kg mol⁻¹, tripentaerythritol core) and linear PEG (Mn = 3.4 kg mol⁻¹) were purchased from JenKem Technology, USA. DBCO acid was purchased from Click Chemistry Tools (Scottsdale, AZ, USA). Phenol red free, growth factor-reduced Matrigel was purchased from Corning (356231). All other chemicals were purchased from Sigma Aldrich (St. Louis, MO, USA) or Thermo Fisher Scientific (Waltham, MA, USA).

Cell Culture

All cells were cultured in humidified incubators at 37 °C and 5% CO₂. MCF10A cells were fed every 3–4 days and maintained in full growth media consisting of Dulbecco’s modified Eagle’s medium (DMEM)/F12 (Gibco, Waltham, MA, USA), supplemented with 5% horse serum (Gibco), 20 ng mL⁻¹ epidermal growth factor (EGF), 0.5 mg mL⁻¹ hydrocortisone, 100 ng mL⁻¹ cholera toxin, 10 μg mL⁻¹ insulin, and 1× penicillin/streptomycin. MCF10A cells encapsulated in Matrigel were maintained in the same media. Before cell feeding, media was supplemented with 4% Matrigel and 2% epidermal growth factor (EGF).

Stable Cell Lines

MCF10A cells expressing H2B-mCherry and mCitp21were generated through CRISPR-Cas9 editing, as described and validated by Moser et al.^[27]

Synthesis of PEG Macromers

8-arm PEG-DBCO (M_n = 20 kg mol^–1) and 8 arm PEG-azide (M_n = 20 kg mol^–1) and difunctional PEG nitrobenzyl azide were synthesized as previously described.^[14] 4-(4-(1-Hydroxyethyl)-2-methoxy-5-nitrophenoxy) butyric acid was purchased from ChemImpex.

Functionalization of PEG Macromers with Fluorophores

8-arm PEG-DBCO fluorescein and 8-arm PEG-azide Cyanine 5 were prepared as previously described.^[14]

Intracellular PEG Uptake of SPAAC Macromers using Lipofectamine on Adherent cells and Encapsulation

MCF10A cells were seeded on to tissue culture treated plastic 6-well plate at a seeding density of 450,000 cells mL^–1 and incubated for 48 h which typically marked the onset of confluence. Cells were transfected with SPAAC PEG macromers as previously described.^[14] Dry PEG macromers were dissolved in sterile PBS at a final concentration of 50 wt%. Transfection media was prepared using Lipofectamine 3000 (10 μL mL^–1) and PEG macromer (5 wt%) in Opti-MEM reduced serum media. Control media contained Lipofectamine (10 μL mL^–1) and Opti-MEM only, with the macromer volume replaced by PBS. Growth media was removed from wells and transfection media containing either PBS or macromer dissolved in PBS was added before incubating for 2 h. The cells were washed twice with Opti-MEM followed by incubation with the second macromer transfection media. For single macromer conditions, control conditions (Lipofectamine and Opti-MEM only plus PBS) were used for the second transfection. After 2 h, the cells were washed with Opti-MEM twice and incubated with full growth media for 48 h. After 48 h, a layer of Matrigel (100 μL cm⁻²) was coated onto a multi-well plate and solidified in an incubator at 37 °C. Transfected cells were trypsinized, counted and diluted to the correct concentration with assay media (125,000 cells mL^–1). A stock of assay media was supplemented with 4% Matrigel and 2% EGF and combined with the suspended cells in a 1:1 ratio.^[15] The cell/assay media solution was added to the Matrigel layer and incubated. Media was changed every 4 days with assay media supplemented with 4% Matrigel and 2% EGF.

Encapsulation of MCF10A cells in Matrigel and 3D transfection of SPAAC PEG Macromers

Cells were seeded on a 10 cm dish at a seeding density of 10,000 cell mL⁻¹ and incubated for 72 hours, then encapsulated into Matrigel in the same manner as above. The MCF10A cells were allowed to grow for 6 days with 1 media change, then transfected in situ. Transfection media was prepared using Lipofectamine 3000 (10 μL mL^–1) and PEG macromer (5 wt%) in Opti-MEM. Media was removed from each well and replaced with transfection media. The plate was incubated for 2 h before dissociating the Matrigel with cold assay media. The media-Matrigel-spheroid mixture was spun down at 1000 rpm for 5 min. The supernatant was aspirated, and the spheroids were resuspended in media and reseeded into Matrigel with assay media supplemented with 4% Matrigel and 2% EGF. The spheroids were incubated for 24 h, then the same procedure was repeated for the second macromer.

Live-cell imaging and spheroid size analysis

Spheroids were imaged using a Nikon Eclipse Ti2 microscope with a 10 × 0.45 NA or 20 × 0.45 NA objective in a humidified, 37 °C chamber at 5% CO₂. All live cell images were converted from z-stacks to maximum intensity projection images. Views of either the top or bottom (or both) of spheroids were collected, meaning that hollow cavities within spheroids generally did not affect the projected area measurement; any cavities not filled by the maximum intensity projection account for an estimated <2% variation in area. The images were threshold using the nuclear H2B-mCherry marker and the projected spheroid area was quantified with the “Analyze Particles” function in FIJI with a minimum particle size of 1000 μm².

Fixing Spheroids in Matrigel and Immunostaining

Samples in Matrigel were fixed with 3.2% paraformaldehyde and 0.1% glutaraldehyde in PBS at room temperature for 20 min, then washed with PBS for 5 min. Samples were incubated with sodium borohydride (10 mM in PBS) at room temperature for 5 min and then washed 3 times with PBS. The samples were stored at 4 °C until imaging. Fixed spheroids were permeabilized with 0.2% Triton-X100 for 1 h and subsequently blocked with 2% Goat serum with 0.2% Triton-X100 (blocking buffer), all at at room temperature. Phospho-Rb (1:250, (Ser807/811) primary antibody staining was performed overnight at 4 °C in the presence of blocking buffer, followed by incubation with secondary antibody conjugated to Alexa Fluor 647 (#A21245, Thermo Fisher, 1:250). The secondary antibody was incubated overnight at 4 °C before samples were washed 3 times with PBS. Immunofluorescence pRb images were taken on a Nikon Ti Spinning Disc Confocal 20 × 0.75 NA air objective.

ATP/ADP measurements and Seahorse assay

For ATP/ADP measurements and Seahorse assays, spheroids were transfected in Matrigel in a similar manner to the above protocol. ATP/ADP measurements were measured using an ADP/ ATP kit (MAK135, Sigma-Aldrich) according to manufacturer’s instruction. Bioluminescence signal was measured on a microplate reader (Synergy H1, BioTek) using Gen5 software (version 1.11.5). To assess functional metabolism, the Seahorse mitochondrial stress test was conducted using the XF24 Flux Analyzer (Agilent Technologies). Briefly, Seahorse sensor cartridges were incubated overnight in a non-CO₂ incubator using the Seahorse calibrant at 37 °C. Spheroids were cultured in a similar manner to the above procedure. The Seahorse assay was conducted 48 h after transfection. 24 h before the assay (24 h after transfection), a thin layer of Matrigel in DMEM was placed on the bottom of each well in a Seahorse 24-well plate and allowed to solidify (10–30mins). The spheroids were then dissociated out of Matrigel and resuspended in assay media with 2% EGF and 4% Matrigel. The spheroids in Matrigel were placed in an incubator overnight to allow for the spheroids to embed. On the day of the assay, embedded spheroids were washed twice in Seahorse XF DMEM Medium (pH 7.4) supplemented with 25 mM glucose and 4 mM Glutamine (Agilent Technologies), then incubated for an 1 h in a non-CO₂ incubator at 37 °C. Oligomycin (final well concentration = 0.5 μM), carbonyl cyanide-4 (trifluoromethoxy) phenylhydrazone (final well concentration = 2 μM), and rotenone plus antimycin A (final well concentration = 0.5 μM) were all sourced from Agilent Technologies. Wave software (version 2.4.2) was used to export Seahorse data as mean ± S.D. Data was normalized to the average spheroid count/condition. The ΔOCR values were calculated as followed; for oligomycin the fourth value was subtracted from the third normalized OCR datapoint. Similarly, ΔOCR for FCCP was obtained by subtracting the seventh from the sixth normalized OCR datapoint. Lastly, ΔOCR for Rnt/A.A. was obtained by subtracting the tenth from the ninth normalized OCR datapoint.

Fluorescence lifetime imaging microscopy (FLIM) and analysis

For FLIM measurements, spheroids were transfected in Matrigel in a similar manner to the above protocol. NADH lifetime m 35easurements were taken on a Zeiss LSM780 (Carl Zeiss, Jena, Germany) confocal microscope with ZEN and ISS FastFLIM software and a titanium:sapphire MaiTai HP (Spectra-Physics, Milpitas, CA) two-photon laser. MCF10A spheroids were exposed to 800 nm two-photon excitation at 80 MHz and observed through 10× and 20× air Zeiss C-Apochromat objectives (Carl Zeiss, Jena, Germany). Signal was isolated using a bandpass filter to collect fluorescence signal from the 450/50 nm spectral range (NAD(P)H emission) and detected with a Hamamatsu H7422p-40 photon-counting PMT connected to a ISS A320 FastFLIM box (ISS, Champaign, IL); lifetime calibration was performed with an aqueous fluorescein calibration solution of 1 μM with a 4 ns lifetime. ISS Vista was used for analysis and lifetime fitting of frequency-domain FLIM data. Phasor plots were analyzed using data pooled from all conditions at each timepoint and smoothed with a weighted Gaussian filter. Three circular cursors were placed equidistantly across the data contained in the phasor plots corresponding to each timepoint, marking inscribed pixels as having longer (pseudocolored dark gray, representing a greater fraction of bound NADH) and shorter (pseudocolored light gray, representing relatively increased free NADH) fluorescence lifetimes. Pseudocolored lifetime images were exported from the ISS software and analyzed using a custom Python script, which calculated the relative area of each thresholded fluorescence lifetime in every image. To produce gradient-colored lifetime images for visualization of differences between conditions, representative images for each condition were imported into SimFCS 4 software and phasor data was pooled for all images at each timepoint, then colored with a gradient spanning two cursors placed outside the extrema (shortest and longest lifetime) contained in the phasor plot.

Photorheology and Fluorescence Recovery After Photobleaching on Degradable SPAAC Gels

5 wt% SPAAC gels were prepared by mixing equal volumes of 5 wt% stock solutions of PEG-oNB-azide and PEG-DBCO. A Discovery HR-3 oscillatory shear rheometer (TA Instruments) equipped with a UV curing accessory was used to measure (1% strain, 1 rad/s) the evolution of the storage modulus (G’) over time during gelation and photodegradation with 365 nm light (Omnicure) set to an intensity of 5 mW/cm². FRAP was performed on a Nikon A1R microscope with a 15 second baseline, 5 second bleach (100% power, 488 laser), and 2 minute recovery. Photodegradation for these studies was performed with a 405 laser (100% power) in an ROI approximately 5 times the width of the bleach diameter during FRAP measurements. Curve fitting for FRAP data was performed as previously described.^[28]

Statistical analysis

Statistical testing was performed in GraphPad Prism. Data are presented as mean ± standard deviation (SD) unless otherwise indicated. Normalization of spheroid area was based on condition-specific area at the earliest measurement timepoint. Normalization of Seahorse data was based on number of spheroids. Sample size (n) for each statistical analysis was n = 3 unless otherwise indicated. For statistical differences, p values were calculated using either an unpaired t-test or a one or two-way analysis of variance (ANOVA) with Tukey test for pairwise comparisons. Significance levels were reported as p < 0.05 (*), 0.01 (**), 0.001 (***), and 0.0001 (****) with corresponding star notations; “ns” denotes no statistical significance.