Dysregulation of Autophagy Contributes to Anal Carcinogenesis

Abstract

Introduction: Autophagy is an intracellular catabolic process that removes and recycles unnecessary/dysfunctional cellular components, contributing to cellular health and survival. Autophagy is a highly regulated cellular process that responds to several intracellular signals, many of which are deregulated by human papillomavirus (HPV) infection through the expression of HPV-encoded oncoproteins. This adaptive inhibitory response helps prevent viral clearance. A strong correlation remains between HPV infection and the development of squamous cell carcinoma (SCC) of the anus, particularly in HIV positive and other immunosuppressed patients. We hypothesize that autophagy is inhibited by HPV-encoded oncoproteins thereby promoting anal carcinogenesis (Fig 1).

Materials and methods: HPV16 transgenic mice (K14E6/E7) and non-transgenic mice (FVB/N), both of which do not spontaneously develop anal tumors, were treated topically with the chemical carcinogen, 7,12-Dimethylbenz[a]anthracene (DMBA), to induce anal cancer. The anuses at different time points of treatment (5, 10, 15 and 20 weeks) were analyzed using immunofluorescence (IF) for two key autophagy marker proteins (LC3β and p62) in addition to histological grading. The anuses from the K14E6/E7 mice were also analyzed for visual evidence of autophagic activity by electron microscopy (EM). To see if there was a correlation to humans, archival anal specimens were assessed histologically for grade of dysplasia and then analyzed for LC3β and p62 protein content. To more directly examine the effect of autophagic inhibition on anal carcinogenesis, nontransgenic mice that do not develop anal cancer with DMBA treatment were treated with a known pharmacologic inhibitor of autophagy, chloroquine, and examined for tumor development and analyzed by IF for autophagic proteins.

Results: Histologically, we observed the progression of normal anoderm to invasive SCC with DMBA treatment in K14E6/E7 mice but not in nontransgenic, syngeneic FVB/N background control mice. With the development of low-grade dysplasia in the K14E6/E7 mice, there was an increase in both punctate LC3β and p62 expression while EM revealed increased autophagosomes without evidence of autophagolysosomes. These observations are consistent with autophagy being inhibited at a later stage in the autophagic process. In contrast, in high-grade dysplasia and SCC in the DMBA-treated K14E6/E7 mice, there were decreased levels of p62 with a continued increase in punctate LC3β expression by IF, while autophagolysosomes were seen on EM, consistent with the process of autophagy proceeded to completion. Similar findings, including histological grade dependent changes in LC3β and p62 expression, were noted with human samples upon analysis of IF. Finally, with pharmacologic inhibition of autophagy in DMBA-treated, nontrangenic FVB/N mice, there was a significant increase in anal cancer development similar to that observed in DMBA- treated K14E6/E7 mice.

Conclusion: Autophagic dysregulation is noted early on in HPV-associated anal carcinogenesis (low-grade dysplasia), with normalization of the autophagic process arising in late stages of HPV-associated anal carcinogenesis (high-grade dysplasia and invasive carcinoma).

Fig 1. Representative diagram of proposed role of autophagy in anal carcinogenesis.

This figure demonstrates inhibition of autophagy via HPV. With normal autophagic function p62 levels do not accumulate as it is continuously being degraded via the autophagic pathway. With late autophagic inhibition, there is blockage of the fusion of the autophagosome with the lysosome. This results in the accumulation of punctate LC3β and p62. There is also an accumulation of damaged organelles such as mitochondria which result in subsequent cellular and genetic damage and creates an environment that promotes carcinogenesis.

Fig 2. Histological analysis of the anal transition zone at various time points with and without DMBA in FVB/N mice.

Anal histology identified by a trained pathologist for each time point (25 mice/group) following treatment with and without DMBA (0.12μmole topically to anus weekly). There is evidence of increased number of animals with high-grade dysplasia, however overt carcinoma is not seen with increasing treatment times with DMBA in mice without HPV oncogenes.

Fig 3. Histological analysis of the anal transition zone at various time points of DMBA treatment in K14E6/E7 mice.

Anal histology identified by a trained pathologist for each time point (25 mice/group) following treatment with and without DMBA (0.12μmole topically to anus weekly). There is a statiscally significant increase in anal dysplasia over the time course of DMBA treatment, with the majority of the mice at 5 and 10 weeks of DMBA treatment having low-grade dysplasia, while 75% of mice at 15 weeks. By 20 weeks of DMBA treatment 100% of K14E6/E7 mice have overt carcinoma. During this same time course none of the K14E6/E7 mice not treated with DMBA developed anal cancer.

Fig 4. Histological analysis of the anal transition zone at various time points in FVB/N mice during DMBA treatment time course.

Anal histology identified by a trained pathologist for each time point (25 mice/group) following treatment with and without DMBA (0.12μmole topically to anus weekly). A) H&E staining of animals not treated with DMBA at the 5 week timepoint reveals normal epithelium. (B) Following 5 weeks of DMBA treatment, there is inflammation seen on H&E staining. (C) At 10 weeks of DMBA treatment there is histological evidence of low-grade dysplasia. (D) Low-grade dysplasia is present at 15 weeks of DMBA treatment and (E) high-grade dysplasia at 20 weeks of DMBA treatment. All images are acquired at 20x magnification.

Fig 5. Histological examination of anal transition zone (ATZ) of K14E6/E7 double transgenic mice throughout a time course of no treatment versus DMBA treatment.

(A) H&E staining of animals not treated with DMBA at the 5 week timepoint reveals normal epithelium. (B) Following 5 weeks of DMBA treatment, there is inflammation on H&E staining. (C) At 10 weeks of DMBA treatment there is histological evidence of low-grade dysplasia. (D) High-grade dysplasia is present at 15 weeks of DMBA treatment and (E) invasive squamous cell carcinoma at 20 weeks of DMBA treatment. All images are acquired at 20x magnification.

Fig 6. LC3β and p62 immunofluorescence levels for each FVB/N mouse over the DMBA treatment time course as determined by FIJI analysis of 20x images.

The above graph shows that in general FVB/N treatment groups are very similar to each other without profound differences in autophagic function. ANOVA analysis demonstrated a statistically significant increase in only in LC3β at 10 weeks of DMBA treatment compared to no DMBA treated mice (p-value = 0.001), indicating increase in autophagic function at this one time point compared to others. There were no significant differences with regards to p62 levels.

Fig 7. LC3β and p62 immunofluorescence levels in K14E6/E7 mice over DMBA treatment time course as determined by FIJI analysis.

ANOVA analysis demonstrated a statistically significant increase in p62 levels at 10 weeks of DMBA treatment (yellow triangles) compared to all other treatment groups, indicating a blockage of autophagic degradation function. There was also a statistically significant increase in LC3β in the 15 and 20 week DMBA treated mice compared to no DMBA treated mice, demonstrating a significant increase in autophagic induction at these two time points. 10 weeks of DMBA treatment also showed an increase in LC3β levels compared to no treatment controls, but it did not reach statistical significance. The above graph shows that mice at the 10 week DMBA treatment time point (yellow triangles) are very different from the other treatment groups in terms of evidence of autophagic dysfunction.

Fig 8. LC3β immunofluorescence staining of K14E6/E7 mice over the DMBA treatment time course with evidence of increased punctate LC3β noted at all treatment time points.

(A). No DMBA treated K14E6/E7 mouse with low levels of autophagy can be seen by low levels of punctate LC3β expression. (B) With 5 weeks of DMBA treatment where again low levels of autophagy can be seen by low levels of punctate LC3β, (C) 10 weeks of DMBA treatment with increasing levels of punctate LC3β, (D) 15 weeks of DMBA treatment with continued increase in punctate LC3β, and (E) and 20 weeks of DMBA treatment there is also a continued increase in the extent of punctate of LC3β expression, indicating autophagic induction is intact in K14E6/E7 mice treated with DMBA.

Fig 9. Immunofluorescence for autophagic proteins following DMBA treatment.

Immunofluorescence for LC3β (cytoplasmic, green), p62 (cytoplasmic, red) and DAPI (nuclear, blue) at various time points of treatment. (A) At 0 weeks of treatment, there is a low level of punctate LC3β expression and minimal p62 noted, indicating low levels of autophagy in this untreated specimen. (B) After 5 weeks of DMBA treatment, there is evidence of autophagic induction with the increase in punctate, granular LC3β compared to 0 week treatment mice (A). (C) An overlay image with co-expression (LC3β and p62) expressed as orange at the 5 week time point compared to LC3β alone in panel B indicates a mild increase in p62 in addition to LC3β at this time point. (D) With 10 weeks of DMBA treatment, there is significant accumulation of p62 expression and an increase in punctate LC3β indicating autophagic induction without degradation of autophagy-specific substrate p62. (E) At 15 weeks of DMBA treatment, the time of development of high-grade dysplasia, p62 levels begin to decrease with continued evidence of punctate LC3β indicating autophagic induction with the ability to degrade p62. (F) At 20 weeks of treatment, where carcinoma is present, there is again a significant increase in LC3β punctate expression with low levels of p62. All images are acquired at 20x magnification.

Fig 10. Electron microscopy following DMBA treatment.

(A) Control K14E6/E7 animals with various organelles in normal quantities noted within the cell. (B) Following 5 weeks of DMBA treatment, there is an accumulation of mitochondria in the cell cytoplasm. (C) At 10 weeks, an accumulation of lysosomes is noted and is depicted with the red arrow. By 15 (D) and 20 (E) weeks of DMBA the accumulation of lysosomes is no longer present and there is evidence of autophagolysosomes (white arrow), which is formed after the autophagosome has fused with the lysosome.

Fig 11. LC3β immunofluorescence staining of human samples to identify autophagic induction at various histological stages of carcinogenesis.

Panels A-C are at 40x magnfication while Panels A-2 to C-2 are acquired at 63x magnification. Panel A contains a representative normal human anal specimen with evidence of punctate LC3β. Panel B demonstrates continued LC3β punctate formation with low-grade squamous intraepithelial lesion (LSIL) which equates to low-grade dysplasia in the mouse samples. Panel C depicts a human anal sample with high-grade squamous intraepithelial lesion (HSIL) which equates to high-grade dysplasia in the mouse samples, and continues to show evidence of punctate LC3β. These images demonstrate, as in the K14E6/E7 mouse samples, that autophagic induction is intact throughout anal carcinogenesis.

Fig 12. Immunofluorescence for autophagic proteins for human anal samples.

Immunofluorescence for LC3β (cytoplasmic, green), p62 (cytoplasmic, red) and DAPI (nuclear, blue) for human anal samples with evidence of normal, LSIL, or HSIL. There are three distinct samples per histological subtype represented in this figure. There is low expression of LC3β and no to low expression of p62 in the normal anal samples. With the development of LSIL, there is significant accumulation of p62 expression in addition to punctate LC3β expression. In cases of of HSIL, there is again a significant increase in punctate LC3β protein expression, but with normalization of p62 levels. All images are acquired at 20x magnification. These images show evidence of autophagic dysfunction, similar to that seen in K14E6/E7 mouse anal samples.

Fig 13. LC3β and p62 immunofluorescence intensity of human anal samples based on histological classification demonstrating autophagic dysfunction, similar to K14E6/E7 mice, with the development of low-grade dysplasia that is not evident in normal or high-grade dysplasia samples.

There is statistical difference in p62 expression levels between normal and LSIL and between HSIL and LSIL (p-value = 0.001). The expression levels of LC3β and p62 show similar changes as noted in K14E6/E7 mice over the time course of DMBA treatment (Fig 7) with evidence of autophagic dysfunction in samples with low-grade dysplasia.

Fig 14. Percent Tumor Free Survival curves for FVB/N mice with and without DMBA and chloroquine treatment.

The curves in black represent FVB/N mice with and without DMBA that did not receive chloroquine. The curves in red are FVB/N mice receiving chloroquine with and without DMBA. There is a statistically significant increase in the number of FVB/N mice that developed anal tumors with DMBA treatment when the mice were also treated with the late autophagic inhibitor, chloroquine. Both FVB/N mice without DMBA and with chloroquine alone did not develop anal tumors over the time course, resulting in the lines overlapping.

Fig 15. LC3β and p62 immunofluorescence intensity in FVB/N mice treated with and without DMBA and with and without a late autophagic inhibitor demonstrating the importance of autophagy in anal carcinogenesis.

FVB/N mice treated with DMBA alone do not develop anal cancer (red triangles). However, with autophagic inhibition 100% of FVB/N mice developed anal tumors with DMBA treatment (orange triangles). ANOVA analysis demonstrated as statistically significant increase in p62 levels in mice received chloroquine with and without DMBA (yellow circles and orange triangles respectively). There was also a statisically significant increase in LC3β in FVB/N mice treated with DMBA alone (red triangles) and mice treated with chloroquine and DMBA (orange triangles) compared to those not treated with DMBA with and without chloroquine (yellow and blue circles, respectively).