Sex-dependent development of Kras-induced anal squamous cell carcinoma in mice

Anal squamous cell carcinoma (SCC) will be diagnosed in an estimated 9,080 adults in the United States this year, and rates have been rising over the last several decades. Most people that develop anal SCC have associated human papillomavirus (HPV) infection (~85–95%), with approximately 5–15% of anal SCC cases occurring in HPV-negative patients from unknown etiology. This study identified and characterized the Kras-driven, female sex hormone-dependent development of anal squamous cell carcinoma (SCC) in the LSL-KrasG12D; Pdx1-Cre (KC) mouse model that is not dependent on papillomavirus infection. One hundred percent of female KC mice develop anal SCC, while no male KC mice develop tumors. Both male and female KC anal tissue express Pdx1 and Cre-recombinase mRNA, and the activated mutant KrasG12D gene. Although the driver gene mutation KrasG12D is present in anus of both sexes, only female KC mice develop Kras-mutant induced anal SCC. To understand the sex-dependent differences, KC male mice were castrated and KC female mice were ovariectomized. Castrated KC males displayed an unchanged phenotype with no anal tumor formation. In contrast, ovariectomized KC females demonstrated a marked reduction in anal SCC development, with only 15% developing anal SCC. Finally, exogenous administration of estrogen rescued the tumor development in ovariectomized KC female mice and induced tumor development in castrated KC males. These results confirm that the anal SCC is estrogen mediated. The delineation of the role of female sex hormones in mediating mutant Kras to drive anal SCC pathogenesis highlights a subtype of anal SCC that is independent of papillomavirus infection. These findings may have clinical applicability for the papillomavirus-negative subset of anal SCC patients that typically respond poorly to standard of care chemoradiation.

scrub. An incision was made on the caudal aspect of the back just to the right of midline.

263
Capsules were inserted parallel to the spine, and the incision was closed with wound clips.

Statistical analysis 266
Comparisons of tumor development between groups was accomplished using the fisher's 267 exact test. The qPCR data was analyzed using an unpaired, two-tailed t-test with Welch's 268 correction to evaluate possible expression differences of Pdx1 and Cre in the sample groups.

269
Data was considered significant with a p-value <0.05. In contrast, large perianal tumors were grossly evident in 9-month-old female KC mice with invasive anal SCC present on histologic examination. This is indicated by the bracket (D).

Development of anal tumors in KC mice
were clearly evident by 6 months, and of significant size by 9 months (Fig 2). Mice with anal 283 tumors displayed no increase in lethality, with normal mobility and weight gain up until the time 284 of euthanasia (age 9 months). All anal SCC tumors were located at or just distal to the anorectal   In female KC mice, anal tumors are visible as an area of congestion at 4 months of age, with mild erythema around the anal region. By age 5 months, an anal tumor is generally evident. By 9 months of age, the anal tumors are significant in size. Despite the large size, the mice are able to maintain weight, consume food, and defecate normally. No mice experienced obstructive symptoms.

303
In addition to the sex-dependency of anal SCC development, there was complete 304 penetrance (n = 14/14) of anal SCC in female KC mice. These findings were confirmed by 305 histopathologic analysis, where excised anal tissue from wild type males, wild type females, and 306 KC males demonstrated normal microscopic anal histology in comparison to the anal SCC seen 307 in KC females (Fig 1). These findings demonstrate anal SCC development is dependent on sex.  cytopathic effect from histopathology of KC mouse anus ( Figure 3A) Fig 3A), and no 321 MmuPV1 DNA was detected within the anal tumors of the KC mice ( Fig 3B). Together, these 322 data show that the anal SCC in this study was not driven by papillomavirus infection. activated Kras G12D mutation (genomic DNA) in both male and female mice (Fig 4-6). In the KC  Kras G12D mutation (Fig 6 and S2 Fig), which was absent in WT mice. When excising the anus 351 from female KC mice, the specimen was removed en bloc with the large anal tumor, and DNA 352 isolation revealed clear presence of the activated Kras G12D mutation (Fig 6, S2 Fig). The male 353 KC anal tissue appeared grossly and histologically normal, yet genomic DNA isolated from 354 whole anal tissue demonstrated the same activated Kras G12D mutation (Fig 6, S2 Fig).  Interestingly, we were unable to detect a mutant-Kras band (or only a faint band) in one male 360 KC mice (Fig 6), which was likely due to detection error from the excised anal samples. Thus, to 361 confirm, we used FFPE sections generated from the 9-month male KC mice (cohort used to

401
concordant with results seen in the untreated ovariectomized mice (33% vs 15%, p-value =0.56) 402 (Table 2). Meanwhile, in the beta-estradiol (E2) dosed ovariectomized KC female mice, 100% 403 (5/5) developed macroscopically visible anal tumors by 4 months of age (Table 3), 'rescuing' the 404 tumor phenotype and again demonstrating stark contrast to ovariectomized mice (100% vs 405 15%, p-value = 0.001). In the sham-dosed castrated KC male group, 0% (0/5) developed a 406 tumor (Table 3), consistent with the results seen in the untreated castrated KC male mice (0/5 407 vs 0/11, p-value = 1). In contrast, 75% (3/4) of E2-dosed castrated KC male mice developed 408 anal SCC that was macroscopically visible by 8 months of age and confirmed on histopathologic 409 analysis (Fig 8). This remarkable and significant increase in tumor formation in E2-dosed KC   anal tumors would display ER expression. To investigate, we performed fluorescent IHC using 421 an antibody specific for ER [15] in the male and female anus. In particular, we assessed ER 422 in intact KC females and found robust presence of ER in the anal tumor (Fig 9)

508
It is important to note that following reduction of estrogen (ovariectomy), 15% (2/13) of 509 female KC mice and 2/6 (33%) sham dosed female KC mice still developed tumors. To confirm 510 that our ovariectomized KC female mice did experienced a significant reduction in circulating 511 estrogen, and that the E2 dosed mice possessed sufficient levels of circulating estrogen, we 512 used a standardized approach of uterine weights. This methodology is more accurate than 513 'single time point' levels of estrogen in circulating blood, due to the substantial variation of 514 circulating estrogen in normal females.
[20] In contrast, uterine weights reflect the steady levels 515 of estrogen stimulation over an extended period. These techniques helped to confirm successful 516 reduction of (ovariectomy) and rescue of (E2 dosed) circulating estrogen.

529
STAT3 activation in female rat brain which results in neuroprotection against ischemic brain 530 injury.
[39] The association between estrogen and STAT3 activation along with the association 531 between STAT3 activity and mutant Kras G12D -induced cancer formation suggests a possible 532 mechanism behind the phenotype of sex-dependent anal SCC development in KC mice.

533
Subsequent analyses will aim to clarify these questions and study limitations, as well as focus 534 on elucidating the specific underlying mechanism by which estrogen enhances Kras-mutant 535 anal SCC development.

536
Our study clearly shows the sex-dependent development of anal SCC is tied to presence of