Colostrum from healthy cows was purchased from a ranch named Ha-yan after routine milking. Milk exosomes from colostrum (Milk-exo) and exosome free fraction of milk (Exo-free milk) were isolated by ultracentrifugation and filtration. All centrifugations were performed at 4°C. Milk was centrifuged at 5,000 g for 30 min, 12,000 g for 1 h to remove milk fat globules, somatic cells, and cell debris. This defatted milk was stored at -80°C until use. After the milk was thawed, milk was then subjected to serial centrifugation steps at 35,000 g for 1 h, 70,000 g for 3 h using an Optima XE-100 Ultracentrifuge (Beckman Instruments Inc., Palo Alto, CA, USA) to remove residual milk fat globules, casein. The supernatant was filtered through 0.80, 0.45, and 0.2 μm filters (Sartorius, Gottingen, Germany). The filtered supernatant was ultracentrifuged at 100,000 g for 1 h. The obtained exosomal pellet was resuspended in cold PBS. The supernatant was further depleted from exosomes by ultracentrifugation at 200,000 g for 3 h. After the last ultracentrifugation, the remaining Exo-free milk supernatant was used for comparison purposes. Pierce™ BCA protein assay kit (Thermo Fisher, Waltham, MA, USA) was used for quantifying the total protein concentration of Milk-exo and Exo-free milk. The size distribution of Milk-exo and Exo-free milk was obtained by dynamic light scattering (DLS) using Zetasizer Nano ZS (Malvern Instruments, United Kingdom). For DLS measurements, Milk-exo and Exo-free milk (100 μg) were diluted in 1 ml of PBS and then transferred to a disposable cuvette. The mean values of particle sizes were obtained from more than three independent preparations. The structures of Milk-exo were detected by transmission electron microscope (TEM) (Tecnai F20 G2). 10 μg of Milk-exo dissolved in PBS was transferred to an electron microscope grid and 1% uranyl acetate solution was used for negative staining.

Milk-exo and Exo-free milk were quantitated by bicinchoninic acid (BCA) protein assays, and the proteins were separated with 8–12% sodium dodecyl sulfate polyacrylamide gel and transferred to nitrocellulose membrane using a Trans-Blot Turbo (Bio-Rad, Hercules, CA, USA). The membrane was blocked in 5% skim milk for 1 h at room temperature followed by incubating with primary antibodies at 4°C overnight: anti-Tsg101 (1:1000, Abcam, Cambridge, MA, USA), anti-Alix (1:500, Novus Biologicals, Littleton, CO, USA), anti-MFG-E8 (1:1000, R&D Systems, Wiesbaden-Nordenstedt, Germany), anti-GM130 (1:1000, Invitrogen™ Thermo Fisher Scientific, Waltham, MA, USA), anti-Lactoferrin (1:1000, Santa Cruz Biotechnology, Dallas, TX, USA), anti-Wnt3a (1:1000, Abcam), anti-β-catenin (1:500, Abcam), GAPDH (1:1000, R&D Systems). After 3 times wash with TBST (Tris-buffered saline, 0.1% Tween 20), the membranes were incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies for 1 h at room temperature. The protein bands were visualized with iBright™ CL750 Imaging System (Invitrogen™ Thermo Fisher Scientific, Waltham, MA, USA).

Human follicle dermal papilla cells (PromoCell, Heidelberg, Germany) were cultured in follicle dermal papilla cell medium (PromoCell, Heidelberg, Germany) with fetal calf serum (0.04 ml/ml), bovine pituitary extract (0.004 ml/ml), basic fibroblast growth factor (1 ng/ml), insulin (5 μg/ml), and 1% antibiotic antimycotic (Gibco, Darmstadt, Germany). The passage 4–6 cells were maintained in 5% CO₂ at 37°C and subcultured when the cells reached 70–80% confluency. The medium was replaced with fresh medium every 3 days.

DP cells (5 × 10⁴) were seeded on a 35 mm glass-bottom confocal dish (SPL Life Science). After 12 h incubation, the medium was removed and washed with DPBS (Welgene, Seoul, Korea). For the fluorescent labeling of exosomes, Milk-exo was incubated in a 65 μM Cyanine 5.5-NHS ester (BioActs, Incheon, Korea) staining solution for overnight at 4°C, and free dye was removed using air-fuge (Beckman Instruments Inc., Palo Alto, CA, USA). After centrifugation for 1 h, washed with PBS and centrifugation again. The labeled exosomal pellet was resuspended in PBS. The labeled exosomes (100 μg) were then added into the dish and incubated in 5% CO₂ at 37°C for 1, 12, and 24 h. After incubation, cells were washed with DPBS and fixed in 4% paraformaldehyde for 10 min. The nucleus of cells was stained with 4’, 6-Diamino-2-Phenylindole (DAPI) (Sigma Aldrich, St. Louis, MO, USA). Images were analyzed with a confocal microscope Leica TCS SP6 (Leica Microsystems, Wetzlar, Germany).

The Cell-Counting Kit 8 (Dojindo Laboratories, Kumamoto, Japan) was used for cell proliferation assay. DP cells (7.5 × 10³/well) were seeded on a 96–well plate and cultured in the complete medium. After 12 h incubation, the medium was removed from each well. Cells were washed in DPBS and treated with various concentrations of Milk-exo (50, 100, 200, 300, 400, 500, and 600 μg/ml) or Exo-free milk (400 μg/ml) for 24 h in serum-free medium. CCK-8 solution was added, and absorbance at 450 nm was measured with a SpectraMax 34 microplate reader (Molecular Devices, Sunnyvale, CA, USA). To measured DP cell proliferation against DHT, DP cells (7.5 × 10³) were seeded on a 96-well plate and cultured in the complete medium. After 12 h incubation, the medium was removed from each well. Cells were washed in DPBS and treat with 30 μM DHT solution for 24 h in a serum-free medium. The cells were then added with Milk-exo or Exo-free milk (400 μg/ml). After 24 h, CCK-8 solution was added, and the optical density was measured at 450 nm.

All animal experiments were conducted in accordance with the International Guide for the Care and Use of Laboratory Animals and approved by the Korea Institute of Science and Technology. C57BL/6 mice (male, 7 weeks old) were purchased from Orient Bio Animal Center (Seongnam, Korea) and maintained standard habitat conditions for 1 week.

To determine the optimal time interval, the biodistribution study was performed using labeled Milk-exo as described above. Animals were intradermally injected into five spots (20 μL per site) with saline or 100 μg Milk-exo in 100 μL saline was intradermally injected into five spots (20 μL per site) on the dorsal skin of C57BL/6 mice (n = 3). The fluorescence intensity was analyzed using an IVIS Lumina Series Ⅲ in vivo imaging system (PerkinElmer, MA, USA). 24 h after injection, mice were sacrificed, and the skin and other major organs were collected. The IVIS Lumina series Ⅲ imaging system was used to measure fluorescence intensities of skin and organs. Quantification of fluorescence intensities was analyzed using Living Image software (PerkinElmer, MA, USA).

To determine whether Milk-exo could induce hair growth, we examined the effect of Milk-exo in C57BL/6 mice model. After acclimatizing, mice (n = 12) were randomly divided into four groups. The dorsal hair was shaved using animal clippers and applied with hair removal cream to observe the pink skin. Animals were intradermally injected into five spots (20 μL per site) with saline, 200 μg Milk-exo in 100 μL saline, and 200 μg Exo-free milk in 100 μL saline. 2.5% minoxidil was topically applied as a positive control. Each treatment was applied every other day for 19 days. Hair regeneration was assessed for each animal following the assignment of a hair regeneration scoring, which also assessed skin color and hair area. Mice were imaged following treatment and quantification of hair regeneration area was assessed using ImageJ based on the criteria below.

After sacrificing the mice to collect the dorsal skin, the skin tissue was fixed with 4% paraformaldehyde and washed with PBS. Fixed skin samples were dehydrated through serial graded ethanol, cleared in xylene, and embedded in paraffin to obtain longitudinal sections. The paraffin blocks were cut longitudinally into 7 µm thick sections. To observe the histological change, sections were stained with hematoxylin and eosin (H&E). Briefly, sections were deparaffinized in xylene for 30 min, hydrated in serial graded ethanol (100%, 90%, 80% and 70%) and then stained with hematoxylin for 1 min, followed by washes for 15 min and eosin staining for 15 s. They also dehydrated three times for 2 min in graded ethanol (70, 80, 90 and 100%) stepwise. They mounted the cover glass after cleaning with xylene. The number of hair follicles was counted in the deep subcutis (n = 6). The immunohistochemistry (IHC) staining was performed using anti-β-catenin (1:200, Abcam) and was visualized by the Mouse and Rabbit Specific HRP/DAB (ABC) Detection IHC Kit (Abcam), according to the manufacturer’s instructions. The digital photomicrographs were taken from representative areas using a digital camera. Deparaffinized and hydrated sections were performed immunofluorescence (IF) staining. After heat-mediated antigen retrieval for 2 min in citrate buffer, sections were blocked with 1% BSA buffer for 20 min and incubated with anti-β-catenin (1:200, Abcam), anti-COX2 (1:200, Abcam), anti-Ki67 (1:200, Abcam) and anti-CD31 (1:100, Invitrogen, Carlsbad, CA, USA) overnight at 4°C. After PBST washing, sections were incubated with Alexa Fluor™ 647 goat anti-rabbit IgG (1:500, Invitrogen), Alexa Fluor™ 488 goat anti-rabbit IgG (1:500, Invitrogen) for 1 h at room temperature. Following PBST rinsing, the stained slides were mounted with Fluoromount-G^® (SouthernBiotech, Birmingham, AL, USA). Stained sections were observed by a confocal microscope Leica TCS SP6.

The statistical analysis was performed using Prism 8.0 (Graphpad). Statistical significance was determined by ANOVA or t-test. A p-value less than 0.05 was considered statistically significant.

Prior to investigating the hair regeneration ability of colostrum-derived exosomes, milk exosome (Milk-exo) was isolated from colostrum as previously described method with minor modification (Figure 2A) (Vaswani et al., 2017). We also prepared an exosome-free fraction (Exo-free milk) as a control group in which exosomes were completely removed from colostrum through an additional centrifugation step. The size of the extracted Milk-exo and Exo-free milk was analyzed through dynamic light scattering (DLS) analysis (Figure 2B). Milk-exo showed a size distribution of about 50 ∼ 100 nm as previously reported (Sedykh et al., 2020), whereas >10 nm nano-sized vesicles were not observed in Exo-free milk (data not shown). The shape of Milk-exo observed with a transmission electron microscope (TEM) was a round sphere. Furthermore, western blot analysis showed that the exosome markers TSG101 and Alix were identified in Milk-exo samples, and GM130, a Golgi marker commonly used as a negative control for exosome protein markers, was detected only in Exo-free milk (Figure 2C). In addition, MFG-E8, also known as lactadherin, and lactoferrin, which are known to be abundant in milk, were highly loaded in Milk-exo (Ripolles et al., 2018; Superti, 2020). In particular, interestingly, it was found that lactoferrin, which is known to prevent hair loss and promote the growth of dermal papilla (DP) cells through the Erk/Akt and Wnt signaling pathway, is overexpressed only in Milk-exo (Li H. et al., 2019; Huang et al., 2019).

Next, we investigated the influence of Milk-exo on the proliferation of DP cells. First, the cellular uptake efficiency of exosomes in DP cells was examined using confocal microscopy with Milk-exo labeled with Cy-5.5 (Supplementary Figure S1). The confocal image showed that Milk-exo was gradually absorbed by DP cells over time, and almost all exosomes were uptaken within 12 h. Further, we examined the change in the proliferation rate of DP cells according to the concentration of Milk-exo (Figure 3A). As the concentration of Milk-exo increased, the proliferation rate of DP cells gradually increased and saturated at a concentration of about 400 μg/ml. However, there was no change in the cell proliferation rate in the DP cell group treated with the same concentration of Exo-free milk compared to the control group (Figure 3B). In addition, we explored whether Milk-exo could rescue the decrease in DP cell growth caused by the DHT treatment, a male hormone known as a representative cause of hair loss, with its ability to enhance the proliferation rate of DP cells (Figure 3C) (Ustuner, 2013). While Exo-free milk-treated group had no effect, surprisingly, the Milk-exo-treated group showed improved DP cell proliferation compared to the control group. Ultimately, the expression level of β-catenin, which promotes the regulation and differentiation of keratinocytes as well as the induction of DP cell to the anagen phase, was compared by Western blot analysis (Figure 3D). Similar to the results of proliferation of DP cells, the expression of β-catenin in the group treated with Milk-exo was increased by about 1.5 compared to the control group despite the DHT treatment. Collectively, these results suggest that Milk-exo may affect hair growth by enhancing the proliferation of DP cells and the expression of β-catenin.

Prior to examining in vivo hair growth ability of Milk-exo, the biodistribution of exosomes was confirmed by intradermal injection of Milk-exo on the dorsal side of C57BL/6 mice (Figures 4A,B). Milk-exo injected into the back of the mouse was remained localized for about 2 days, then gradually dissipated, and all was excreted on the 11th day. The level of Milk-exo was the highest in skin tissues compared to other tissues of mice that were sacrificed 2 days after the intradermal injection. Therefore, in vivo experiment was scheduled with Milk-exo injection every 2 days (Figure 4C) (Zhang et al., 2013; Son et al., 2018). Minoxidil, a commercially available drug for hair loss, was used as a positive control. Minoxidil, as the first-line pharmacologic recommendation for most patients with androgenetic alopecia, has been proven to be efficient in slowing hair loss by a number of double-blind, randomized, and case-control studies. However, minoxidil treatment is known to induce several adverse effects including contact dermatitis, pruritus, dryness, and facial hypertrichosis even when applied topically. Indeed, the intradermal injection of minoxidil caused severe skin irritation, and immunostaining of skin tissues confirmed that the expression of COX2, which is related to the inflammatory response, was very high (Figure 4D). However, the expression of COX2 was hardly observed in mice skin tissues treated with Milk-exo even after the three consecutive injections.

Next, in vivo hair regeneration effect of Milk-exo was monitored over about 3 weeks period after shaving the dorsal hair of mice with animal clippers. The severe skin reactions from minoxidil injection inhibited the intradermal group, and topical administration of minoxidil (at a concentration of 2.5%) were used for the experiments Hair coverage level was used as a direct indicator of hair regeneration, and clinical visual scores were evaluated on a scale of 0–6 (Park et al., 2012). The rate of hair regeneration was not significantly different between groups until around the 7th day, but from the 9th day, the dorsal skin darkened and hair began to appear in the Milk-exo and minoxidil-treated groups (Figure 5A). On day 13, the Exo-free/saline group had an average of only 25% fur coverage, and the fur coverage of the Milk-exo group was approximately 50%. The highest hair coverage was seen on day 15 in the Milk-exo treatment group, at the level comparable to minoxidil treatment (Figure 5B). Visual scoring data also showed consistent results (Figure 5C).

To determine the effect of Milk-exo on hair follicles growth, hematoxylin and eosin (H&E) staining was performed on dorsal skin sections after 9 and 13 days of treatment (Figure 6A). In the stained section of the dorsal skin on day 9, typical anagen features such as thickened subcutis and large hair bulbs were found in the Milk-exo and minoxidil treatment groups. In addition, in the dorsal skin tissue section at day 13 of treatment, the Milk-exo treatment group significantly increased the number of hair follicles compared to the saline/Exo-free milk treatment group (Figure 6B). Furthermore, we performed immunofluorescence staining for Ki67, a proliferation marker of DP cells (Figure 6C). Ki67 signals in the nucleus of the hair matrix cells were found to be 50% and approximately 35% for minoxidil and Milk-exo, respectively. In association with angiogenesis, CD31, which is highly expressed in early anagen, was additionally identified (Mecklenburg et al., 2000; Li et al., 2019b), and the Milk-exo treatment group showed the highest expression rate (Supplementary Figure S2). This result may be due to recently reported angiogenesis-related proteins such as MFG-E8, which were highly expressed in Milk-exo identified in the characterization experiments (Uchiyama et al., 2014; Zhang et al., 2021). Taken together, these data indicated that Milk-exo in vivo could further enhance hair matrix proliferation and accelerate the telogen-to-anagen transition of hair follicles (Sugaya and Hirobe, 2014; Bai et al., 2015).

In particular, β-catenin, known as the primary initiator of the anagen phase, is a major regulator of hair follicle growth (Bejaoui et al., 2019). We investigated whether Milk-exo influences the expression of β-catenin using immunohistochemical analysis (Figures 7A,B, Supplementary Figure S3). Consistent with the previous data, β-catenin was expressed in the Milk-exo and minoxidil-treated groups on the 9th day after treatment. Specifically, quantification data of β-catenin–positive (β-catenin+) cells showed that β-catenin expression by Milk-exo treatment was gradually increased from day 9 (18%) to day 19 (30%).

Among a variety of inter-/intracellular signaling molecules, Wnt signaling acts a crucial role in hair follicle development via triggering the entry into the anagen phase. The secreted Wnt proteins bind with the Frizzled receptor, thereby stabilizing hypophosphorylated β-catenin by inhibiting glycogen synthase kinase-3β-mediated protein degradation as well as promote self-renewal of stem cells (Huelsken et al., 2001; Yang et al., 2012). As expected, the expression of Wnt3a was increased in the Milk-exo and minoxidil-treated groups (Figures 7C,D). As with the previous histological data, the expression level of β-catenin in each group also showed a similar trend to that of Wnt3a. Collectively, Milk-exo could activate Wnt/β-catenin signaling that enables to induce the transition to the growth phase.

Milk, which contains various nutrients, is an ecosystem-optimized diet that humans have been consuming for thousands of years. The bioactive components of milk pass through the digestive system where encounters multiple stress factors such as shifting pH and are well absorbed into the body. Here we investigated whether Milk-exo, which has undergone freeze-drying, maintains its physicochemical properties and hair regeneration activity (Figure 8). First, Milk-exo powder prepared by freeze-drying at -85°C and 58 mTorr for 2 days were again dissolved in PBS (FD Milk-exo), and their size, morphology, and the expression of exosome-specific markers were evaluated. FD Milk-exo also expressed exosome markers (TSG101 and Alix) and milk-specific proteins (MFG-E8 and lactoferrin), and their size and shape did not change even after lyophilization (Figures 8A,B). Furthermore, the hair regeneration ability of FD Milk-exo was evaluated in the C57BL/6 mouse model (Figures 8C–E). FD Milk-exo showed remarkable hair growth ability compared to saline from the 13th day, and this hair regeneration efficiency was similar to that of Milk-exo without freeze-drying.