Long Evans (LE) Rats with nursing pups were purchased from Envigo Laboratories Inc. (8520 Allison Pointe Blvd Suite 400 Indianapolis, IN). Nursing LE rat pups were randomly assigned to receive subcutaneous injections in the nape of the neck with either vehicle of PBS (P5119-100ML, Sigma-Aldrich, St.Louis), cumene hydroperoxide (247504, Sigma-Aldrich, St. Louis, MO), or hydrogen peroxide (H1009, Sigma-Aldrich, St. Louis, MO). The cumene hydroperoxide solution was prepared at a concentration of 20 µL/mL by volume in PBS (molar concentration of 0.134 M), of which 100 µL was used per injection. Each hydrogen peroxide injection included 40 µL of the 30% stock solution and 60 µL of PBS for a total injected volume of 100 µL at a concentration of 12% by volume (molar concentration of 3.92 M). Daily injections were administered to the LE pups from day 5 until day 10 after birth, following which chemotherapy administration was started.

Nursing LE rat pups were randomly assigned to receive subcutaneous injections in the nape of the neck with either PBS, low-dose cumene hydroperoxide, the P53 activator RITA (Reactivation of p53 and Induction of Tumor cell Apoptosis) (S2781, Selleck Chemicals, Houston, TX), or a combination of low-dose cumene hydroperoxide and RITA. Low-dose cumene hydroperoxide was prepared at a concentration of 4 µL/mL in PBS, of which 50 µL was administered with each injection for a 10-fold decrease in cumene hydroperoxide used. RITA was initially solubilized in DMSO in accordance with manufacturer suggestions and diluted to a final concentration of 40 µg/mL in PBS, of which 50 µL was administered with each injection. The rats receiving the combination of low-dose cumene hydroperoxide and RITA received both the 50 µL injection of cumene hydroperoxide and the 50 µL injection of RITA for a final total volume of 100 µL. Daily injections were administered to the LE pups from day 5 until day 10 after birth along the same timeline as previously discussed, following which chemotherapy administration was started.

Following pre-treatment with either PBS, cumene hydroperoxide, or hydrogen peroxide, the LE rat pups were then given chemotherapy intraperitoneally starting at 11 days of age. Stock solutions of VP-16 (Etoposide, 1268808, Sigma-Aldrich, St. Louis MO, USA) was prepared at a concentration of 20 mg/ml in 30.5% ethanol in H₂O supplemented with 2 mg/ml citric acid, 30 mg/ml benzyl alcohol, 80 mg/ml Tween-80, and 150 mg/ml PEG-300. VP-16 stock solution was diluted in PBS and was given intraperitoneally at a concentration of 1.5 mg/kg for three days. Cytoxan (cyclophosphamide, PHR1404, Sigma-Aldrich, St. Louis MO, USA) was prepared in H₂O supplemented with 5% dextran and was also diluted appropriately in PBS. Cytoxan and was given at a concentration of 35 mg/kg for one day only. The rat pups that received the combination of Cytoxan and Adriamycin (Doxorubicin, 1225703, Sigma-Aldrich, St. Louis MO, USA) were first administered with Cytoxan at a concentration of 25 mg/kg for one day followed by three days of Adriamycin that was prepared in H₂O and given at a concentration of 1.5 mg/kg. Alopecia was recorded seven days after the last injection of chemotherapy. Since all rats pre-treated with either cumene hydroperoxide or hydrogen peroxide exhibited 100% or near 100% protection from hair loss, data were displayed in the figures by either (+) or (–) for whether any hair was protected at the site of injection.

Skin at the site of vehicle and cumene hydroperoxide injections were isolated from the nape of rats on day 20 after birth (day 7 after treatment with VP-16, day 9 after treatment with Cytoxan, day 6 after treatment with Cytoxan-Adriamycin). Histological analyses were performed following our previously-described protocol (19). Briefly, Paraffin sections were stained with hematoxylin and eosin, and the stained slides were subsequently evaluated using an Axio Observer D1 microscope and AxioVision software (Carl Zeiss Microimaging, Thornwood, NY). Images were captured using a Keyence Biozero BZ–800 K Microscope and analyzed using Image J image analysis software.

Samples of rat skin were cleared of hair and debris prior to submission to Ocean Ridge Biosciences (array@oceanridgebio.com) for mRNA-Sequencing. According to the vendor provided protocol, total RNA was isolated using the TRI Reagent (Molecular Research Center; Part # TR118) method. After isolation, RNA was quantified by O.D. measurement and assessed for quality on a 1% agarose – 2% formaldehyde RNA Quality Control (QC) gel. The RNA was then digested with RNase free DNase I (Epicentre; Part # D9905K) and re-purified using Agencourt RNAClean XP beads (Beckman Coulter; Part # A63987). The newly digested RNA samples were then quantified by O.D. measurement. Amplified cDNA libraries suitable for sequencing were prepared from 250 nanograms (ng) of DNA-free total RNA using the TruSeq Stranded mRNA Library Prep (Illumina Inc.; Part # 20020595). The quality and size distribution of the amplified libraries were determined by chip-based capillary electrophoresis (Bioanalyzer 2100, Agilent Technologies). Libraries were quantified using the KAPA Library Quantification Kit (Kapa Biosystems, Boston, MA). Libraries were pooled at equimolar concentrations and sequenced on the Illumina NextSeq 500 sequencer using one Mid Output v2 150 cycle kit (part# FC-404-2001). In each case the libraries were sequenced with 76 nt paired-end reads plus 8 nt dual-index reads on the instrument running NextSeq Control Software version 2.2.0.4. Reads were generated into FASTQ files using the default setting on Illumina’s bcl2fastq program v2.17.1.14 and were filtered and trimmed. The reads were aligned to the hg38 genome using HISAT2 version 2.0.5 along the following parameters (N = 1, L = 20, i = S,1,0.5). Read count was performed using the featureCounts program version 1.5.1 for gene-level counting (http://subread.sourceforge.net). Normalized RPKM values were calculated from the raw featureCounts read counts using the formula (Gr x 1,000,000)/(Gl x mRNArt) where Gr is raw read count for a gene, Gl is length of the exon model for the gene, and mRNArt is the total read count for exons from nuclear-derived protein-coding mRNAs. The RPKM values were further filtered to retain a list of genes with a minimum of 50 mapped reads in all 3 samples to ensure maximum technical reproducibility. Differential expression analysis was performed in R studio (R version 3.6.3) using the DESeq2 package with a q-value < 0.05 and a fold change > 1.5. Heatmap analysis was also performed in R studio using the pheatmap package (version 1.0.12) along parameters that specify two distinct clusters. Gene ontology and pathway analyses were performed using EnrichR (20–22). Raw RNA-seq data is provided in Table S1 , and processed and filtered RNA-seq data is provided in Table S2 .

Cumene hydroperoxide is known to exert oxidative stress by inducing intracellular alkyl radicals in murine keratinocytes, and cumene hydroperoxide has been widely used to model oxidative stress in vivo (23–26). Fortunately, studies have shown that short term (< 14 days) subcutaneous administration of cumene hydroperoxide produces minimal cutaneous irritation in rats (26), making cumene hydroperoxide an ideal choice for exogenously introducing cutaneous oxidative stress. Using the previously described and widely used rat pup model of chemotherapy-induced alopecia (27), we administered daily subcutaneous injections of either cumene hydroperoxide or PBS (phosphate-buffered saline) into the nape of Long Evans rat pups for six days, followed by daily intraperitoneal injections of the chemotherapeutic agent VP-16 (etoposide) for three days. Upon seven days after the completion of VP-16 injections, all rats pre-treated with cumene hydroperoxide maintained a coat of hair at the site of cumene hydroperoxide injection whereas all rats pre-treated with vehicle did not ( Figures 1A, B ). Cumene hydroperoxide provided similar hair loss protection against monotherapy with CTX (cyclophosphamide) and combination therapy with CTX-ADM (Cyclophosphamide-Doxorubicin) ( Figure 1B ). This demonstrates that pre-treatment with subcutaneous cumene hydroperoxide prior to the start of chemotherapy prevents alopecia induced by multiple classes of chemotherapeutic agents. Histological analyses of skin biopsies taken from the site of cumene hydroperoxide and vehicle injections show that VP-16 induced extensive hair follicle damage and dispersion of melanin in the skin of rats pre-treated with vehicle only ( Figure 1C ). However, the skin of rats pre-treated with cumene hydroperoxide showed minimal to no hair follicle damage nor melanin dispersion following VP-16 treatment ( Figure 1C ). Additionally, the skins of cumene hydroperoxide pre-treated rats appeared to be thicker than that of vehicle pre-treated rats, although objective measurements were not obtained. This lends further evidence to suggest that pre-treatment with subcutaneous cumene hydroperoxide protects against the cytotoxic effects of chemotherapies such as VP-16.

Similar to cumene hydroperoxide, H₂O₂ (hydrogen peroxide) is a ROS that is known to generate oxidative stress and has been clinically used in numerous cutaneous applications (28). However, in contrast to cumene hydroperoxide, H₂O₂ is produced endogenously in most living organisms, and low-cost commercial H₂O₂ solutions are widely available. We therefore performed the same set of experiments using subcutaneous H₂O₂ injections. Like cumene hydroperoxide, pre-treatment with daily H₂O₂ injections for six days also provided full protection against alopecia induced by VP-16 as well as by CTX monotherapy and CTX-ADM combination therapy at the site of injection ( Figures 1D, E ). Of note, we initially intended to score the alopecia on a scale of 0 to 4 based on the degree of hair loss. However, since all rats pre-treated with either cumene hydroperoxide or H₂O₂ exhibited 100% or near 100% protection from hair loss, data were displayed in the figures by either (+) or (–) for whether any hair was protected at the site of injection. These results therefore show that pre-treatment with ROS-producing compounds can protect hair follicles against the cytotoxic effects of chemotherapy and prevent chemotherapy-induced alopecia.

To further investigate the molecular mechanisms underlying how cumene hydroperoxide protects against chemotherapy-induced alopecia, we performed total RNA-seq using skin harvested from the injection sites of vehicle as well as cumene hydroperoxide-treated rats upon completing the six days of cumene hydroperoxide injections. We found that 1,307 genes and 1,465 genes were significantly down- and upregulated, respectively, in the skin of cumene hydroperoxide-treated rats when compared to that of the vehicle-treated rats ( Figure 2A ). Gene enrichment analysis showed that E2F Target genes as well as genes that regulate the G2-M checkpoint were significantly downregulated ( Figure 2B – top). Additionally, gene ontology analysis shows downregulation of processes such as mitotic cell cycle phase transition and G2/M transition of mitotic cell cycle ( Figure 2B – bottom). Specific genes important for cell cycle progression and cellular division such as PCNA (proliferating cell nuclear antigen), CCND1 (cyclin D1), and CCND3 (cyclin D3) were also significantly downregulated by cumene hydroperoxide treatment ( Figure 2C ). This therefore shows that treatment with cumene hydroperoxide causes cell cycle arrest in the skin and hair follicles of cumene hydroperoxide-treated rats. On the other hand, gene enrichment analysis of the 1,465 upregulated genes showed significant upregulation in genes implicated in the P53 pathway as well as P53-regulated processes such as Apoptosis ( Figure 2D ). Genes involved in inflammation-related pathways such as cholesterol homeostasis and IL-2/STAT5 signaling were also significantly upregulated ( Figure 2D ). Specific target genes of the P53 pathway (i.e. BTG1, BTG2, FOXO3) and the apoptotic pathway (SAT1, RHOB) were significantly upregulated by cumene hydroperoxide treatment ( Figure 2E ). Interestingly, genes involved in glutathione-mediated detoxification pathway such as GSTZ1, GSTK1, and MGST1 were also significantly upregulated ( Figure 2E ). Since glutathione is one of the main mechanisms of protection against ROS in mammalian cells (29), its upregulation indicates a coordinated protective response against cumene hydroperoxide-induced inflammation and oxidative stress in the rat skin and hair follicles. This therefore suggests that cumene hydroperoxide-induced oxidative stress produces inflammation that leads to P53-mediated cell cycle arrest in rat skin and hair follicles.

To further investigate the link between cumene hydroperoxide, P53 pathway activation, and cell cycle arrest, we decided to test whether treatment with the P53 activator RITA (Reactivation of p53 and Induction of Tumor cell Apoptosis) can potentiate the protective effects of cumene hydroperoxide (30, 31). We therefore pre-treated four groups of six rats with daily subcutaneous injections of either vehicle, low-dose cumene hydroperoxide, RITA (140 µM), or a combination of low-dose cumene hydroperoxide and RITA for six days prior to chemotherapy with VP-16. We found that VP-16 induced total alopecia in all rats pre-treated with either vehicle, low-dose cumene hydroperoxide, or RITA ( Figure 2F ). However, only the combination of RITA and low-dose cumene hydroperoxide preserved a patch of hair at the site of injection in all six treated rats ( Figure 2F ). Although the protected area of hair is smaller than that of therapeutic levels of cumene hydroperoxide ( Figure 1A ), this result nevertheless shows that co-treatment with the P53-activating RITA enhances the protective effects of low-dose cumene hydroperoxide that is ineffective at preventing alopecia when given alone. This result thereby provides further evidence to suggest that cumene hydroperoxide protects against chemotherapy-induced alopecia by inducing P53-mediated cell cycle arrest in hair follicles.