Rational hydrogel design to improve brain modulus matching for implantation

Abstract

Treating the brain is challenging due to the restrictive blood–brain barrier, and modulus-mismatched implants often cause problems. Herein, we have fabricated copolymer hydrogels from thermoresponsive polymer, poly(N-isopropylacrylamide) (PNIPAAm), -r-hydrophilic polymer, poly(acrylic acid) (PAA), which are injectable and transform into soft implants above their lower critical solution temperature (LCST). PAA concentration can be leveraged to tune the LCST and viscosity of the PNIPAAm–r–PAA hydrogel in solution. Furthermore, the Young’s moduli of these materials, ranging from 1–4 kPa, are close to rat and human brain tissue, potentially leading to less inflammation and rejection due to significant modulus mismatch.

1. Introduction

Due to biologic barriers [1], most treatable central nervous system diseases require invasive surgeries that lead to complications with limited therapy. To improve outcomes, invasive surgical procedures can be coupled with direct implantation of biomaterials to release adjuvant drugs. The Gliadel wafer, releasing the anticancer drug carmustine, is an FDA-approved polymeric biomaterial for surgical application post-resection of brain tumors. Once maximum tumor resection is completed and bleeding stops, these wafers are packed into the wound [2]. However, its effectiveness was questioned with a recently completed clinical trial (NCT00525590). Remarkably, 67.8 % of patients had severe adverse events and 15.25 % of patients died. Wound healing and brain regeneration are commonly interrupted by the stiff (Young’s modulus of 10⁵ kPa [3]) wafer placement [4], leading to a significant translation gap in the field of neurologic regenerative biomaterials.

Recently, the field has suggested a tissue-focused approach to biomaterials design [5], recognizing the importance of biophysical and biochemical matching to target tissues. We hypothesize that a more “brain-matched” biomaterial would lead to milder side effects for patients with implants. Therefore, we developed a soft, injectable copolymer hydrogel using the thermoresponsive, injectable polymer poly (N-isopropylacrylamide) (PNIPAAm) and the hydrophilic polymer poly acrylic acid (PAA), (PNIPAAm–r–PAA) to eliminate invasive surgical placements. Both PNIPAAM and PAA are deemed neurocompatible [6,7]. PNIPAAm is attractive due to its thermoresponsive nature, along with its ease of cross-linking. PNIPAAm undergoes a phase transition at a lower critical solution temperature (LCST) of 32 °C, below which the polymer remains in solution and above which PNIPAAm chains precipitate out [6] with a Young’s modulus on the order of 10¹ kPa [8]. The transition is a result of an entropic increase, with the release of hydrophobically bound water molecules, or the hydrophobic effect [9]. Altering the hydrophilic fraction with the addition of hydrophilic PAA can raise the LCST [7]. This increase in the hydrophilic fraction provides an increased capacity for bonding to water molecules, thus increasing the degree of swelling, forming a softer material [9]. Here, we demonstrate that an increase in PAA content results in higher LCST and a decrease in Young’s modulus of our PNIPAAm–r–PAA hydrogels, more adequately mimicking neurologic tissue.

2. Experimental procedure

2.1. Hydrogel Synthesis

PNIPAAm, N-acryloxy-succinimide, and PAA were dried under nitrogen. Dichloromethane was added to dissolve the polymers, prior to incubation in a 50 °C oil bath. Azobisisobutyronitrile was added for copolymerization. The hydrogel was extracted, washed, and dried.

2.2. Characterization

Details of characterization were presented in supplemental information (SI).

3. Results and discussion

Previous work on PNIPAAm–r–PAA demonstrated that greater than 20 mol% PAA resulted in hydrogels with LCSTs above body temperature,[10] rendering them ineffective for implantation. Therefore, we chose to explore 6 and 10 mol% PAA. Copolymer fabrication was proven via attenuated total reflectance Fourier-Transform infrared spectroscopy on freeze-dried hydrogels (Fig. S1). Our hydrogels dissolve into polymer chains in water, thus soluble below their LCST. Physical crosslinking of polymers occurs as the solution heats to and above LCST, resulting in the insoluble hydrogel formation. A change in opacity of the fluid (Fig. S2), confirms this physical change [11] and was used to evaluate the average time for hydrogel formation after incubation at 37 °C. Notably, all hydrogels formed in under 5 min (Table 1), irrespective of PAA content. This is a significant reduction in gelation time compared to neat PNIPAAm (~25 min [12]), likely due to the increased hydrophilicity. The change in opacity is caused by structural changes with the large hydrophobic environment produced by the collapse of the PNIPAAm network via intermolecular aggregation between the hydrophobic backbone and isopropyl side chains [13]. The increase in PAA leads to structural differences (Fig. 1), with increased PAA resulting in more pores within a tighter network, expected as the intermolecularly bound water within the polymer chains is expelled above the LCST.

Table 1.

Characteristics of 6 and 10 mol% PNIPAAm–r–PAA hydrogels.

	6 mol%	10 mol%
Gelation Time [m:s.ms]	3:44.2 ± 0.01	3:45.4 ± 0.01
Injectability [%]	96 ± 1	95 ± 5
Swelling Ratio [%]	113 ± 17	112 ± 48
LCST [°C]	25.4 ± 0.005	32.7 ± 0.005
Viscosity (RT) [mPa·s]	2104 ± 350	1232 ± 133
Young’s Modulus [kPa]	1.62 ± 0.309*	2.88 ± 0.573*

^*p < 0.05.

Fig. 1.

Scanning electron microscopy of PNIPAAm-r-PAA hydrogels at 37°C.

Both hydrogel formulations maintained > 90 % injectability at RT, showing the aqueous form is injectable (Table 1). Increasing PAA content allows the hydrogel to hold more water, leading to increased swelling and decreased stiffness [7]. However, at 37 °C, no measurable difference in swelling ratios was observed, though more deviation occurred with higher PAA content. Rheological testing confirmed findings (Fig. 2). As the temperature increases, viscosity increases, indicating gelation as crosslinking occurs. As shear stress plateaus, the physical-crosslinked network between the PNIPAAm–r–PAA strands has formed, and further stress cannot change the phase. This transition occurs where the slope changes from positive to near zero as the polymer becomes more solid-like. The transition occurs at LCST values of 25.4 and 32.7 °C for 6 and 10 mol% PAA, respectively (Table 1). The large variation in shear stress and viscosity after the LCST with 10 mol% PAA could be due to increased variability in overall swelling ratio. Both hydrogel solutions exhibited viscosities between 1200–2500 mPa’s, deemed injectable. As we increase the PAA content, viscosity decreases due to the increased water content within the system.

Fig. 2.

(A) Shear stress and (B) viscosity as a function of temperature for PNIPAAm-r-PAA hydrogels. Dashed lines indicate transition to near zero slope.

Finally, we measured the Young’s modulus of hydrogels via mechanical indentation (Table 1). For comparison, we also characterized the Young’s modulus of brains from adult male Sprague Dawley rats (AUP 2023–0105), calculated as 6.47 ± 0.65 kPa, agreeing with literature [14]. If averaged across all regions, human brain tissue is reported to have a Young’s modulus around 2 kPa [15]. Notably, our PNIPAAm–r–PAA have Young’s moduli close to human brain tissue (Table 1). Fig. 3 highlights the Young’s modulus of our hydrogels compared to various FDA-approved materials used to interface with the brain, showing major progress toward decreasing the modulus mismatch gap [15–22]. Our PNIPAAm–r–PAA hydrogels are close to the Young’s modulus of brain tissue, demonstrating an improvement over FDA-approved materials [23].

Fig. 3.

Young’s Moduli for FDA-approved materials for brain implants, PNIPAAm-r-PAA hydrogels, and healthy brains.

4. Summary

In conclusion, our research highlights the potential of PNIPAAm–r–PAA hydrogels as effective, injectable implants within the brain. This could enable them to conform to the brain’s irregular contours following surgery. Importantly, we have demonstrated that these hydrogels can be engineered with an LCST close to body temperature and to better match brain tissue stiffness better than FDA-approved implants. Future considerations of our NIPAAm-r-PAA hydrogel should evaluate problems associated with high brain water content [6], which could cause excess hydrogel swelling and intracranial pressure. Simultaneously, diffusion of necessary molecules within the gel should be explored to ensure that nutrient transfer is not limited upon gel placement; particularly, oxygen and glucose diffusion is necessary due to the high metabolic needs of the brain [23].

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.matlet.2025.138187.

Data availability

Data will be made available on request.