HIV nucleoside reverse transcriptase inhibitors efavirenz and tenofovir change the growth and differentiation of primary gingival epithelium




An increasing number of HIV-infected patients are combating HIV infection through the use of antiretroviral drugs, including reverse transcriptase inhibitors. Oral complications associated with these drugs are becoming a mounting cause for concern. In our previous studies, both protease inhibitors and reverse transcriptase inhibitors have been shown to change the proliferation and differentiation state of oral tissues. This study examined the effect of a nonnucleoside and a nucleoside reverse transcriptase inhibitor on the growth and differentiation of gingival epithelium.


Organotypic (raft) cultures of gingival keratinocytes were treated with a range of efavirenz and tenofovir concentrations. Raft cultures were immunohistochemically analysed to determine the effect of these drugs on the expression of key differentiation and proliferation markers, including cytokeratins and proliferating cell nuclear antigen (PCNA).


These drugs dramatically changed the proliferation and differentiation state of gingival tissues when they were present throughout the growth period of the raft tissue as well as when drugs were added to established tissue on day 8. Treatment with the drugs increased the expression of cytokeratin 10 and PCNA and, conversely, decreased expression of cytokeratin 5, involucrin and cytokeratin 6. Gingival tissue exhibited increased proliferation in the suprabasal layers, increased fragility, and an inability to heal itself.


Our results suggest that efavirenz and tenofovir treatments, even when applied in low concentrations for short periods of time, deregulated the cell cycle/proliferation and differentiation pathways, resulting in abnormal epithelial repair and proliferation. Our system could be developed as a potential model for studying the effects of HIV and highly active antiretroviral therapy (HAART) in vitro.


The advent of antiretroviral drugs has greatly decreased the mortality and improved the life expectancy of HIV-infected patients. With the use of these drugs, HIV-infected patients can live long lives with suppressed immune systems. However, oral complications are the most common manifestations in HIV-infected patients on highly active antiretroviral therapy (HAART) drug regimens. The usual HAART regimen combines three or more different drugs, such as two nucleoside reverse transcriptase inhibitors (NRTIs) and a protease inhibitor (PI), two NRTIs and a nonnucleoside reverse transcriptase inhibitor (NNRTI) or other such combinations. These HAART regimens have been shown to reduce the viral load in patients. By contrast, HAART has been shown to have several oral complications, such as oral warts [1, 2], erythema multiforme [3, 4], xerostomia [3, 4], toxic epidermal necrolysis, lichenoid reactions [4, 5], exfoliative cheilitis [3], oral ulceration and paraesthesia [3, 6]. These complications affect both the quality of life of patients and noncompliance with drug regimens.

Efavirenz (trade name Sustiva, Bristol-Myers Squibb, New York, NY, USA) is an NNRTI which is often used in HAART treatment. Tenofovir (trade name Viread, Gilead Sciences, Foster City, California, USA), also an important component of HAART, is a nucleoside reverse transcriptase inhibitor (NRTI). Previous studies have examined the effect of both protease inhibitors and an NRTI on gingival tissue [7, 8]. This study sought to expand on those results by comparing the effects of an NRTI and an NNRTI on gingival tissue.

Materials and methods

Primary gingival keratinocytes and organotypic raft cultures

Primary gingival keratinocytes were isolated from a mixed pool of tissues obtained from patients undergoing dental surgery [9] in accordance with Penn State University College of Medicine Institutional Review Board (IRB# 25284) procedures. The tissue was processed as described previously [9].

Raft cultures were grown as previously described [10, 11]. Rafts were fed with E-media supplemented with either efavirenz or tenofovir. Efavirenz capsules (Bristol-Myers Squibb Company, New York, NY, USA) and tenofovir capsules (Gilead Sciences, Foster City, California, USA) were purchased from the Pharmacy at The Milton S. Hershey Medical Center, Penn State University. The contents of either a 200 mg or 300 mg capsule were removed and resuspended in sterile phosphate-buffered saline (PBS). Serial dilutions were made directly in E-media to obtain the correct concentration. The maximum concentration (Cmax) of efavirenz is 4 μg/mL [12-14]. In the first set of experiments, the rafts were treated with efavirenz at concentrations of 1, 2, 4, 5 and 7 μg/mL from day 0. Control rafts were fed with E-media only. Additional experiments were performed using the same concentrations of efavirenz but beginning treatment of the raft cultures on day 8. All rafts were fed every other day and harvested at the indicated time-points. The Cmax of tenofovir is 362 ng/mL [12-14]. The rafts were treated with tenofovir at concentrations of 200, 250, 362, 400 and 600 ng/mL from day 0 or day 8 as described above.

Histochemical analyses

Raft cultures were fixed in 10% buffered formalin, and embedded in paraffin. Four-micrometre sections were cut and stained with haematoxylin and eosin as described previously [10].

Immunostaining was performed with the Vectastain Elite ABC kit (Vector Laboratories, Burlingame, CA) [10]. The slides were stained as described previously, using the following primary antibodies: mouse monoclonal keratin 5 (clone XM26; dilution 200 mg/mL), keratin 10 (clone DEK10; dilution 200 mg/mL), involucrin (200 mg/mL), keratin 6 (clone LHK6B; dilution 10 ng/mL) (all from Lab Vision, Fremont, CA) and rabbit polyclonal proliferating cell nuclear antigen (PCNA) (cloneFL-261; dilution 2 mg/mL; Santa Cruz Biotechnology Inc., Santa Cruz, CA) [9].


Effect of efavirenz and tenofovir on morphological differentiation and stratification of gingival keratinocytes in raft cultures

The raft culture system has been shown to accurately mimic the in vivo physiology of the gingival epidermis [7-9, 15, 16]. Haematoxylin and eosin staining was performed to examine the effect of these drugs on gingival epithelial morphology and stratification. Figure 1 shows the results for both drugs after 10 days of treatment (panels E1−E6 and T1−T6). There was a dramatic change in morphology and stratification, as was seen previously with the NTRI zidovudine [9]. Normally, nuclei are only present in the basal layer of cells, as was the case with our untreated rafts; however, both abnormal nuclei and keratin pearls were visible in treated tissues.

Figure 1.

Effects of efavirenz and tenofovir on gingival epithelium morphology, stratification and expression patterns of differentiation markers. Primary gingival keratinocytes were grown in organotypic (raft) cultures and treated with different concentrations of drug. Drug treatment began at day 0 and continued for 10 days. Panels E1−E6 and T1−T6 show haematoxylin and eosin staining at the indicated drug concentrations. Panels E7−E12 and T7−T12 show keratin 5 staining at the indicated drug concentrations. Panels E13−E18 and T13−T18 show involucrin staining, and panels E19−E24 and T19−T24 show keratin 10 staining. Panels E25−E30 and T25−T30 show keratin 6 staining, and panels E31−E36 and T31−T36 show proliferating cell nuclear antigen (PCNA) staining. Images are at 17.6× original magnification.

Efavirenz and tenofovir treatment changes the expression pattern of differentiation markers in gingival epithelium

Involucrin and the cytokeratins 5 and 10 are associated with the terminal differentiation of gingival epithelium [17]. Immunohistochemistry was used to assess the expression pattern of biochemical markers of differentiation in treated and untreated samples.

Cytokeratin 5 and its partner cytokeratin 14 form dimers that help give tissue its integrity. Both drug treatments decreased and changed the expression pattern of cytokeratin 5 at all drug concentrations, beginning with tissue harvested at day 8 (Fig. 1, panels E7−E12 and T7−T12, and data not shown). Tissues that were grown to day 8 and then exposed to drugs were also affected even if they were only exposed to drugs for 6 h (Fig. 2, panels E6−E9 and T9−T12). It is apparent that reverse transcriptase inhibitors reduce the amount of this cytokeratin in gingival tissue.

Figure 2.

Effects of efavirenz and tenofovir on gingival epithelium morphology, stratification and cytokeratin expression pattern in established gingival raft cultures. Primary gingival keratinocytes were grown in organotypic (raft) cultures to day 8 without drug treatment. On day 8, the indicated amount of drug was added. The tissue was then harvested 6, 12, 24 or 48 hours later. Panels 1, 5, 10, 15, 20 and 25 are untreated rafts. Panels E2−E5 and T2−T5 show haematoxylin and eosin (H&E) staining at the indicated drug concentrations. Panels E6−E9 and T6−T9 show keratin 5 staining at the indicated drug concentrations. Panels E10−E14 and T10−T14 show involucrin staining, and panels E16−E19 and T16−T19 show keratin 10 staining. Panels E20−E24 and T20−T24 show keratin 6 staining, and panels E26−E29 and T26−T29 show proliferating cell nuclear antigen (PCNA) staining. Images are at 20× original magnification.

Involucrin is expressed in response to the same pathway as cytokeratin 5 and is present in keratinocytes in the epidermis and other stratified squamous epithelia. Efavirenz and tenofovir both decreased the expression pattern of involucrin at all drug concentrations and at all time-points (Fig. 1, panels E13−E18 and T13−T18, and Fig. 2, panels E11−E14 and T11−T14).

Cytokeratin 10 expression indicates terminal differentiation of tissue and is usually present at low levels in the suprabasal layers of oral keratinocytes [18, 19]. Efavirenz and tenofovir treatments both induced the expression of cytokeratin 10 and changed its expression pattern, although the effect of tenofovir was more dramatic; small increases could be seen as early as 6 h, but more tissue-wide changes were evident at 48 h (Fig. 1, panels E16−E24 and T16−T24). These results are similar to those seen in zidovudine-treated rafts [9].

Effects of efavirenz and tenofovir treatment on the expression of keratin 6

Cytokeratin 6 expression is related to the wound-healing process and is present in the suprabasal layer. Epidermal injury results in induced cytokeratin 6 expression in keratinocytes undergoing activation at the wound edge [20, 21]. In our study, cytokeratin 6 expression was dramatically reduced at all concentrations of both drugs. A noticeable decrease in cytokeratin 6 was seen after just 6 hours when either drug was added to tissue after 8 days of growth (Fig. 1, panels E25−E30 and T25−T30, and Fig. 2, panels E21−E24 and T21−T26). Such an immediate decrease in expression of cytokeratin 6 days post treatment suggests an impaired wound-healing response of tissue to drug-induced injury.

Efavirenz and tenofovir treatments induce cell proliferation

We evaluated the effect of these drugs on the expression of a well-characterized cell proliferation marker, PCNA. PCNA is a nuclear protein associated with DNA polymerase delta, which is present throughout the cell cycle in the nuclei of proliferating cells. Typically, cells only proliferate in the basal layer of tissue. In this study, PCNA expression in untreated rafts was limited to the basal layer. In drug-treated rafts, however, PCNA was strongly expressed in both the basal and differentiating layers of the tissue (Fig. 1, panels E31−E36 and T31−T36). This effect on PCNA expression was seen as early as 6 to 12 h post treatment (Fig. 2, panels E26−E29 and T26−T29). In addition, tissues in this study also showed an increase in cyclin A staining, similar to that seen previously in zidovudine-treated tissues (data not shown) [9]. Proliferation was most increased in tenofovir-treated rafts, although proliferation in differentiating tissues was very marked in efavirenz-treated rafts. Around days 16–18, all drug-treated rafts showed a marked decline in PCNA expression (data not shown), suggesting that over time tissue becomes less proliferative as a result of the cytostatic effects of the drug. The changed expression pattern of PCNA indicates the activation of the wound-healing pathway and that drug-treated rafts have lost control of cell cycle events, which could play a role in the generation of oral complications in HIV-infected patients.


HIV-positive patients taking HAART experience many oral complications, which have a major impact on their overall health and quality of life. Protease inhibitors have been shown to affect both the growth and differentiation of oral epithelium; however, the effects of NNRTIs on the growth and differentiation of oral epithelium are currently unknown.

The growth of gingival epithelium was inhibited when either the drug efavirenz, an NNTRI, or tenofovir, an NRTI, was present throughout the growth period, even at concentrations below their Cmax. They also disrupted its proliferation and stratification status. Our observations suggest the possibility that oral epithelium in HIV-infected patients exposed to HAART that includes these drugs experience drug-induced abnormalities in the molecular and cellular biology of the tissue, leading to oral complications.

Normally, gingival stratified epithelia express the cytokeratin pair of cytokeratins 5 and 14 only in the proliferative basal layer [17, 22]. Application of tenofovir and efavirenz decreased the expression of cytokeratin 5 (Fig. 1, panels E7−E12 and T7−T12). The decrease in cytokeratin 5 was seen in all levels of the oral tissue. Cytokeratin 5 and its partner cytokeratin 14 form dimers that help give tissue its integrity. Without these cytokeratins present, tissue becomes very fragile and the smallest injury causes the tissue to fall apart and blisters to form [18]. The production of these cytokeratins has also been shown to be increased in hyperproliferative situations such as wound healing [7, 21, 23]. These data suggest that reverse transcriptase inhibitor treatment is impairing the ability of oral tissue to heal itself.

When cells make the commitment to terminally differentiate, a switch in cytokeratin gene expression occurs. Expression of cytokeratins 5 and 14 is shut off and that of cytokeratins 1 and 10 is turned on [24]. In this study, drug treatments induced the expression of cytokeratin 10. Increased levels of cytokeratin 10 in drug-treated gingival epithelium may be an attempt by the tissue to protect itself against damage caused by the drugs [18, 25, 26]. Additionally, it has been shown that cytokeratin 10 is more strongly expressed in both oral lesions after treatment with efavirenz and tenofovir and hyperproliferative epidermis when compared with ordinary epidermis [27]. Thus, the elevated levels of cytokeratin 10 may be linked to the proliferative effect of the drugs.

The response of tissues to efavirenz and tenofovir was similar to that seen previously with the NRTI zidovudine (Table 1) [9]. In that case, we suggested that a lack of involucrin available for cross-linking in the cornified envelope and suprabasal layers of the epidermis may explain the lack of a vacuolated, cornified layer seen in treated tissues and may account for the fragility of oral tissues in patients on HAART [3]. Efavirenz and tenofovir probably act on cytokeratin 5 and involucrin through the Sp1 transcription regulatory pathway [28] in a similar manner to zidovudine (Table 1) [9].

Table 1. Summary of effects of HIV drugs on gingival raft cultures
Effect on expression of:EfavirenzTenofovirZidovudine [9]Amprenavir [7]Kaletra [8]
  1. The table summarizes the effects of various HIV treatments on morphology and cytokeratin expression patterns in gingival raft cultures.
  2. PCNA, proliferating cell nuclear antigen.
Keratin 5DecreasesDecreasesDecreasesIncreasesIncreases
InvolucrinDecreasesDecreasesDecreasesNot testedNot tested
Keratin 10IncreasesIncreasesIncreasesIncreasesIncreases
Keratin 6DecreasesDecreasesDecreasesIncreasesIncreases

Previous data suggest that HAART drugs cause damage to the gingival epithelium [7-9]. To examine this possibility, we looked at expression patterns of cytokeratin 6, which is activated in response to injury in the suprabasal layer of stratified epithelium. In the current study, cytokeratin 6 expression was reduced significantly by all concentrations of efavirenz and tenofovir, at all time-points examined, including just 6 h post treatment. The inability of oral tissue, which is constantly in a wounding environment, to repair itself through the cytokeratin 6 mechanism could explain some of the oral complications seen in patients on HAART.

The results of haematoxylin and eosin staining suggested a change in the proliferation status of drug-treated rafts. Therefore, we examined the effect of efavirenz and tenofovir on PCNA. This nuclear protein plays an important role in DNA synthesis and cell cycle progression, allowing cell proliferation. PCNA is generally found in cell nuclei between the G1 and M phases of the cell cycle [29, 30]. Normally, PCNA is expressed in only a few basal layer cells [29, 30]. Increased expression of PCNA in drug-treated rafts was seen in this study. Not only was there increased PCNA expression in the basal layer of the drug-treated tissue; PCNA expression also became apparent in the suprabasal layers of the drug-treated tissue. This increased expression of PCNA could indicate the activation of wound-healing pathways attempting to counteract drug-induced tissue damage, a conclusion supported by the elevated levels of cytokeratin 10 in treated rafts. However, it is more likely that exposure to these drugs deregulated cell proliferation and differentiation pathways, allowing abnormal proliferation independent of wound-healing pathways. This argument is supported by the decrease in cytokeratin 6 expression in drug-treated tissues. Overall, the treated tissue was highly and abnormally proliferative and had impaired epithelial repair mechanisms, thus making the tissue more vulnerable to the oral complications seen in HIV-infected patients taking these drugs. The increased levels of PCNA and cytokeratin 10 indicate that the tissue was in a hyperproliferative state that may make it more susceptible to the viral tumours common in HIV-positive patients. Decreased levels of cytokeratins 5 and 6 and involucrin suggest that the tissue was fragile and unable to repair itself sufficiently, contributing to a permissive environment for opportunistic infections and other complications.

In conclusion, we observed that efavirenz and tenofovir, like zidovudine, were able to inhibit and change the growth of gingival tissue when either drug was added at day 0 or day 8 of raft growth. These drugs increased the expression of PCNA and cytokeratin 10. Conversely, the expression of cytokeratins 5 and 6 and involucrin was decreased. Together, these results suggest that these reverse transcriptase inhibitors deregulated growth, differentiation and proliferation profiles in human gingival raft tissue. These results are consistent with the finding of oral complications in patients undergoing long-term HAART. Additional studies will be needed to determine the exact mechanism through which these drugs are exerting their effects.


We thank Lynn Budgeon for technical assistance in preparing histological slides. This work was supported by NIDCR grant DE018305 to CM.