There are several major observations from our current study: First, UVC selectively increased the green AF of mice’s skin. Second, the UVC-induced increases in the green AF of mice’s skin were dose dependent, which were highly associated with the UVC dosages. Third, UVC-induced increases in the green AF of B16-F10 cells were dose dependent, which were also highly associated with UVC dosages. Fourth, UVB exposures produced dose-dependent increases in skin’s green AF of mice, which were significantly associated with the UVB dosages. Fifth, UVC-induced skin damage was highly associated with both UVC dosages and UVC-induced green AF intensity of mice’s skin. Sixth, the capacity of UVC, UVB, UVA and X-ray to produce increases in the skin’s green AF is distinctly different, with the following order: UVC > UVB > UVA > X-ray.

So far the dosages of UV exposures to a person can be determined only by UV dosimeters during the course of UV exposures. Our current study has indicated significant associations between the skin’s green AF intensity and UVB/UVC dosages, indicating that the skin’s green AF may become the first biomarker for non-invasive evaluations of the dosages of UVB/UVC exposures to a person during post-UV exposure period. This finding has provided a critical basis for us to determine the biomarkers for predicting UV-induced skin damage.

Evaluations of UV-induced skin damage has been mainly based on physicians’ visual observation of the alterations of the skin’s appearance, including erythema and swollen state of the skin, which has two major shortcomings for diagnosis and treatment of UV-induced skin damage: First, since there has been no non-invasive and practical approach for predictions of UV-induced skin damage, early intervention of the UV-induced skin damage is exceedingly difficult; and second, the current approaches are incapable of providing quantitative evaluations of UV-induced skin damage. Therefore, discovery of biomarkers for non-invasive prediction of UV-induced skin damage is highly valuable for early treatment of UV-induced skin damage. Our study has provided first evidence indicating that skin’s green AF is the first biomarker for predicting UVB/UVC-induced skin damage: Both our cell culture study and our animal study have indicated that UVB/UVC dosages are significantly associated with skin’s green AF intensity. Since UVB/UVC dosages are also significantly correlated with the UVB/UVC-induced skin damage, the increased skin’s green AF induced by UVB/UVC could become the first biomarker for non-invasive prediction of UVB/UVC-induced skin damage.

Discovery of the skin’s green AF as a biomarker for predicting UVB/UVC-induced skin damage is of great clinical significance: Since there has been no practical method that can be used for non-invasive prediction of UV-induced skin damage, treatment of the damage can be conducted only when UV-induced symptoms become obvious under the visual evaluations of physicians. This delayed treatment leads to markedly decreased efficacy of the treatment. Based on our skin’s green AF-based approach, early treatment of the UV-induced skin damage may be achieved before the occurrence of clinical symptoms. The early treatment is expected to significantly enhance the therapeutic efficacy of the treatment, which may even fully prevent UVB/UVC-induced skin damage.

In addition to its capacity for early diagnosis and treatment of UV-induced skin damage, the skin’s green AF-based approach for evaluating UV-induced skin damage also has the following three distinct advantages compared to other clinical or experimental approaches: First, our approach can provide quantitative evaluations of UVB/UVC-induced skin damage. In contrast, evaluations by physicians’ visual observations can provide only qualitative evaluations; and second, our approach is non-invasive and label-free. In contrast, the reported experimental approach for predicting UV-induced cellular damage—the cyclobutane pyrimidine dimer (CPD)-based approach—requires cell lysis [11].

Our findings indicating that the skin’s green AF is a novel biomarker for predicting UV-induced skin damage may have highly valuable and extensive applications: First, based on our findings, the devices that can predict non-invasively UV-induced skin damage may be developed. The devices may be used in clinical settings to predict UV-induced skin damage, so that early interventions may be conducted to minimize skin damage. Second, the devices may also be used by natural populations, particularly the populations under high risk of UV-induced skin damage, to monitor the amount of their UV exposures as well as the risk of sun exposure-induced skin damage. Our study has raised the possibility that we may develop devices for home and personal use for evaluating people’s risk of developing major UV-induced skin damage. However, it is certainly a challenging task to produce this type of devices that should be not only portable and economical, but also have significant sensitivity and precision. Third, our novel method can provide non-invasive and label-free prediction of UV-induced skin damage, which may revolutionize animal studies on UV-induced skin damage: Unlike currently used approaches, our novel method does not require animal killing for skin damage assays, which could both decrease dramatic experimental cost and enhance the ethical levels of the animal studies. Fourth, since the tests using our method require only a short period of time, our novel method may lead to remarkably increased efficacy for screening and discovering new effective ingredients for decreasing UV-induced skin damage. Moreover, our finding has established a basis for future, potential development of equipment for non-invasive and economic prediction of UVR-induced skin damage.

Our study has indicated the significant promise of the skin’s green AF as a biomarker for predicting UV-induced skin damage. It has been strongly indicated that UV irradiation is a major causative factor for skin cancer [2, 24], implicating that our technology might also be used for predicting incidence of skin cancer. However, in addition to UV doses, multiple other factors also significantly affect the incidence of skin cancer, including age and childhood exposures of UV [2, 24]. Therefore, it appears that UV-induced development of skin cancer is a chronical process that can be influenced by multiple factors. Based on these pieces of information, extensive future studies are warranted to determine the capacity of our technology in predicting the incidence of skin cancer.

Relatively high doses of UVB/UVC doses have been used our study. However, it is noteworthy that in our study single irradiation of UV on the mice was conducted. In contrast, in a number of previous studies, multiple irradiations of UV were conducted, in which the accumulated dosages could be similar or even much greater than the doses used in our study. For examples, in one study the mice were irradiated six times a week for five weeks. The highest UV radiation (UVR) dosage was 0.65 J/cm2, and the accumulated dosage was 14.88 J/cm2 [25]; and in another study the mice were irradiated three times a week for six weeks. The highest UVR dosage was 0.4 J/cm2 [26].

The article of Kozlova et al. has also reported increased AF in dead cells [22]. Our findings have the following profound differences compared with their finding: (1) Our study has shown that UVC dosages are correlated with the AF intensity. In contrast, the article of Kozlova has indicated that only dead cells emit AF, which should not show any association between UVC dosage and the AF intensity as observed in our study. (2) The distribution of their AF signals in their cells, as described by the article of Kozlova et al., is significantly differently from distribution of the AF in the cells under our experimental conditions: Their AF signal were present everywhere in the dead cells. In contrast, in our study, there were AF signals only in certain areas of B16-F10 cells. (3) Our cell culture experiment has shown that UVC did not induce any apoptosis or necrosis 1 h after the UVC irradiation, while our experiment showed that UVC induced significant green AF increase 0.3 or 1 h after the UVC irradiation (Supplemental Fig. 3). Therefore, our study has strongly indicated that UVC did not induce the AF increase in the cell cultures by inducing cell death.

It is of significance to elucidate the biological properties of the UV-induced green AF. Our determinations of the spectrum of the skin’s AF of normal mice as well as the spectrum of the skin’s AF of the mice exposed to UVC, UVB, or UVA have provided the following information: The general patterns of the spectra of these four groups of mice were similar, while the AF intensity induced by the three types of UV was significantly higher than that of the control. This observation has suggested that the three types of UV may induce increased green AF by enhancing the AF intensity of certain autofluorescent molecules of the normal skin, which may be achieved by changing the amount or the structural properties of the autofluorescent molecules. The AF showed distinct polyhedral structure. The keratinocytes in the spinous layer of epidermis have polyhedral structure [27], implicating that the AF structure might be the structure the keratinocytes in the spinous layer of epidermis. However, it is necessary to conduct future studies to determine if the AF is originated from the spinous layer.

There are multiple autofluorescent molecules in the skin, including melanin, keratins, NAD(P)H, FAD, collagen, and elastin [18, 19]. Our study has shown that UVC could increase the AF of the ear’s skin of ICR mice that are melanin deficient [28], thus arguing against the possibility that melanin is the origin of the green AF. The excitation wavelength for NAD(P)H’s AF is far below the excitation wavelength used in our study [29], thus NAD(P)H is unlikely the origin of the AF. Since the excitation wavelength and emission wavelength for the AF of collagen was 270–370 nm and 305–450 nm, respectively [30], it is unlikely that collagen is the molecule that is responsible for the UV-induced green AF. Since the AF spectrum of elastin is similar to that of collagen [31], it is unlikely that elastin is the molecule that is responsible for the UV-induced green AF. The AF properties of the UV-induced green AF, as determined using the excitation wavelength of 488 nm and emission wavelength 500–530 nm, matched the AF spectra of keratins and FAD [32]. It is necessary to conduct future studies to further investigate the origins of the UV-induced skin’s AF.